macroeconomic analysis and parametric control of a national economy
TRANSCRIPT
Macroeconomic Analysis and Parametric Controlof a National Economy
Abdykappar A. Ashimov l Bahyt T. SultanovZheksenbek M. Adilov l Yuriy V. BorovskiyDmitriy A. Novikov l Rakhman A. AlshanovAskar A. Ashimov
Macroeconomic Analysisand Parametric Controlof a National Economy
Abdykappar A. AshimovKazakh National Technical UniversityNational Academy of Sciences of theRepublic of Kazakhstan
Almaty City, Kazakhstan
Bahyt T. SultanovKazakh National Technical UniversityState Scientific and Technical ProgramAlmaty City, Kazakhstan
Zheksenbek M. AdilovKazakh National Technical UniversityAlmaty City, Kazakhstan
Yuriy V. BorovskiyKazakh National Technical UniversityState Scientific and Technical ProgramAlmaty City, Kazakhstan
Dmitriy A. NovikovInstitute of Control Sciences RASMoscow, Russia
Rakhman A. AlshanovKazakh National Technical UniversityAlmaty City, Kazakhstan
Askar A. AshimovKazakh National Technical UniversityState Scientific and Technical ProgramAlmaty City, Kazakhstan
ISBN 978-1-4614-4459-6 ISBN 978-1-4614-4460-2 (eBook)DOI 10.1007/978-1-4614-4460-2Springer New York Heidelberg Dordrecht London
Library of Congress Control Number: 2012948194
# Springer Science+Business Media New York 2013This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part ofthe material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission orinformation storage and retrieval, electronic adaptation, computer software, or by similar or dissimilarmethodology now known or hereafter developed. Exempted from this legal reservation are brief excerptsin connection with reviews or scholarly analysis or material supplied specifically for the purpose of beingentered and executed on a computer system, for exclusive use by the purchaser of the work. Duplicationof this publication or parts thereof is permitted only under the provisions of the Copyright Law of thePublisher’s location, in its current version, and permission for use must always be obtained fromSpringer. Permissions for use may be obtained through RightsLink at the Copyright Clearance Center.Violations are liable to prosecution under the respective Copyright Law.The use of general descriptive names, registered names, trademarks, service marks, etc. in thispublication does not imply, even in the absence of a specific statement, that such names are exemptfrom the relevant protective laws and regulations and therefore free for general use.While the advice and information in this book are believed to be true and accurate at the date ofpublication, neither the authors nor the editors nor the publisher can accept any legal responsibility forany errors or omissions that may be made. The publisher makes no warranty, express or implied, withrespect to the material contained herein.
Printed on acid-free paper
Springer is part of Springer Science+Business Media (www.springer.com)
Contents
1 Elements of Parametric Control Theory of MarketEconomic Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Components of Parametric Control Theory
of Market Economic Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Methods of Analysis of the Stability and Structural
Stability of Mathematical Models of National
Economic Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1 Development of the Methods for Evaluating
Stability Indicators of Mathematical Models . . . . . . . . . . . . . . . . . . 3
1.2.2 Development of Methods for Evaluating Weak
Structural Stability of a Discrete-Time Dynamical
System (Semi-cascade) Based on the
Robinson Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Approach to Synthesis and Choice (in the Environment
of a Given Finite Set of Algorithms) of Optimal Laws
of Parametric Control of a National Economic System’s
Development. Existence Conditions for a Solution
to Respective Variational Calculus Problems. Conditions
of Influence of Uncontrolled Parameters to These Problems. . . . . . . . . . 9
1.3.1 Analysis of the Existence Conditions for a Solution
of the Variational Calculus Problem of Synthesis
and Choice (in the Environment of a Given Finite
Set of Algorithms) of Optimal Laws of Parametric
Control of a Continuous-Time Deterministic
Dynamical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.2 Analysis of the Existence Conditions for a Solution
of the Variational Calculus Problem of Synthesis
and Choice (in the Environment of a Given Finite
Set of Algorithms) of Optimal Laws of Parametric
Control of a Discrete-Time Deterministic
Dynamical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
v
1.3.3 Analysis of the Existence Conditions for a Solution
of the Variational Calculus Problem of Synthesis
and Choice (in the Environment of a Given Finite
Set of Algorithms) of Optimal Laws of Parametric
Control of a Discrete-Time Stochastic
Dynamical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.3.4 Analysis of the Influence of Uncontrolled
Parameter Variations (Parametric Disturbances)
on the Solution of the Variational Calculus Problem
of Synthesis and Choice of Optimal Parametric
Control Laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.4 Algorithm of Application of Parametric Control Theory
and Rules of Interaction Between Persons Making
Decisions on Elaboration and Realization of the Effective
State Economic Policy on the Basis of an Information
System for Decision-Making Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4.1 Algorithm of the Application of Parametric
Control Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
1.4.2 Rules of Interaction for Decision Makers on the
Formulation and Implementation of an Effective
Public Economic Policy Based on the Information
Decision Support System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
1.5 Examples for Application of Parametric Control Theory . . . . . . . . . . . . 36
1.5.1 Mathematical Model of the Neoclassical Theory
of Optimal Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.5.2 One-Sector Solow Model of Economic Growth. . . . . . . . . . . . . . 42
1.5.3 Richardson Model for the Estimation of Defense Costs . . . . . 46
1.5.4 Mathematical Model of a National Economic
System Subject to the Influence of the Share
of Public Expense and the Interest Rate
of Government Loans on Economic Growth . . . . . . . . . . . . . . . . . 50
1.5.5 Mathematical Model of the National Economic
System Subject to the Influence of International
Trade and Currency Exchange on Economic Growth. . . . . . . . 70
1.5.6 Forrester’s Mathematical Model of Global Economy. . . . . . . . 83
1.5.7 Turnovsky’s Monetary Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
1.5.8 Endogenous Jones’s Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
vi Contents
2 Methods of Macroeconomic Analysis and Parametric Controlof Equilibrium States in a National Economy . . . . . . . . . . . . . . . . . . . . . . . . . 117
2.1 Macroeconomic Analysis of a National Economic State
Based on IS, LM, and IS-LM Models, Keynesian
All-Economy Equilibrium. Analysis of the Influence
of Instruments on Equilibrium Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
2.1.1 Construction of the IS Model and Analysis
of the Influence of Economic Instruments. . . . . . . . . . . . . . . . . . 118
2.1.2 Macroeconomics of Equilibrium Conditions
in the Money Market. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
2.1.3 Macroestimation of the Mutual Equilibrium State
in Wealth and Money Markets. Analysis of the
Influence of Economic Instruments. . . . . . . . . . . . . . . . . . . . . . . . . 124
2.1.4 Macroestimation of the Equilibrium State
on the Basis of the Keynesian Model
of Common Economic Equilibrium. Analysis
of the Influence of Economic Instruments. . . . . . . . . . . . . . . . . . 125
2.1.5 Parametric Control of an Open Economic
State Based on the Keynesian Model. . . . . . . . . . . . . . . . . . . . . . . 127
2.2 Macroeconomic Analysis and Parametric Control
of the National Economic State Based on the Model
of a Small Open Country. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
2.2.1 Construction of the Model of an Open Economy
of a Small Country and Estimation of Its Equilibrium
Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
2.2.2 Influence of Economic Instruments on Equilibrium
Solutions and Payment Balance States . . . . . . . . . . . . . . . . . . . . . 136
2.2.3 Parametric Control of an Open Economy State
Based on a Small Country Model . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3 Parametric Control of Cyclic Dynamics of Economic Systems. . . . . . . 141
3.1 Mathematical Model of the Kondratiev Cycle. . . . . . . . . . . . . . . . . . . . . . 141
3.1.1 Model Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.1.2 Estimating the Robustness of the Kondratiev
Cycle Model Without Parametric Control . . . . . . . . . . . . . . . . . . 143
3.1.3 Parametric Control of the Evolution of Economic
Systems Based on the Kondratiev Cycle Model . . . . . . . . . . . 144
3.1.4 Estimating the Structural Stability of the Kondratiev
Cycle Mathematical Model with Parametric Control . . . . . . 147
3.1.5 Analysis of the Dependence of the Optimal Value
of Criterion K on the Parameter for the Variational
Calculus Problem Based on the Kondratiev
Cycle Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Contents vii
3.2 Goodwin Mathematical Model of Market Fluctuations
of a Growing Economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
3.2.1 Model Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
3.2.2 Analysis of the Structural Stability of the Goodwin
Mathematical Model Without Parametric Control . . . . . . . . . 149
3.2.3 Problem of Choosing Optimal Parametric Control
Laws on the Basis of the Goodwin
Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
3.2.4 Analysis of the Structural Stability of the Goodwin
Mathematical Model with Parametric Control . . . . . . . . . . . . . 153
3.2.5 Analysis of the Dependence of the Optimal
Parametric Control Law on Values of the
Uncontrolled Parameter of the Goodwin
Mathematical Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
4 Macroeconomic Analysis and Parametric Controlof Economic Growth of a National Economy Basedon Computable Models of General Equilibrium. . . . . . . . . . . . . . . . . . . . . . . 157
4.1 National Economic Evolution Control Based on a
Computable Model of General Equilibrium of
Economic Branches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.1.1 Model Description, Parametric Identification,
and Retrospective Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
4.1.2 Macroeconomic Analysis on the Basis of the
Computable Model of General Equilibrium
of Economic Branches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.1.3 Finding Optimal Parametric Control Laws
on the Basis of the CGE Model
of Economic Branches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
4.2 National Economic Evolution Control Based on the
Computable Model of General Equilibrium with the
Knowledge Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
4.2.1 Model Description, Parametric Identification,
and Retrospective Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
4.2.2 Estimation of the Macroeconomic Theory Provisions
on the Basis of the Computable Model of
General Equilibrium with the Knowledge Sector . . . . . . . . . . 236
4.2.3 Finding Optimal Parametric Control Laws Based
on the CGE Model with the Knowledge Sector . . . . . . . . . . . . 238
viii Contents
4.3 National Economic Evolution Control Based
on the Computable Model of General Equilibrium
with the Shady Sector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
4.3.1 Model Description, Parametric Identification,
and Retrospective Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
4.3.2 Finding the Optimal Values of the Adjusted
Parameters on the Basis of the CGE Model
in the Shady Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
About The Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
Contents ix
Chapter 1
Elements of Parametric Control Theoryof Market Economic Development
1.1 Components of Parametric Control Theoryof Market Economic Development
The application of mathematical models of a national economy is an important
subject area for the analysis of an effective public policy in the area of the economic
growth [73].
Many dynamical systems, including the national economic system [33, 30], after
some transformations, can be described by the following systems of nonlinear
ordinary differential equations:
x� ðtÞ ¼ f xðtÞ; uðtÞ; að Þ; (1.1)
with the initial condition
x t0ð Þ ¼ x0: (1.2)
Here t is the time, t 2 t0; t0 þ T½ �; T>0, is a fixed number;
x ¼ xðtÞ 2 Rm is the state of system (1.1), (1.2);
x0 2 Rm is the initial state of the system (deterministic vector);
u ¼ uðtÞ 2 Rq is the vector of controlled (regulated) parameters; the functions uðtÞand their derivatives are to be uniformly bounded;
a 2 A � Rs is the vector of uncontrolled parameters; and A is an open connected set.
For a solution to system (1.1), (1.2) to exist, let’s assume that the vector function
f satisfies the Lipschitz condition and the following linear constraints on its growth
rate:
jf ðx; u; aÞj � cð1þ jxjÞ;
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2_1,# Springer Science+Business Media New York 2013
1
where c is a positive constant.As is well known, the solution (evolution) to the considered system of ordinary
differential equations depends on both the vector of initial values x0 and the values
of vectors of controlled (u) and uncontrolled (a) parameters. Therefore, the result of
evolution (development) of the nonlinear dynamical system, with a given vector of
the initial values x0 , is defined by the values of vectors of both controllable and
uncontrollable parameters.
It is also known [3] that the process described by (1.1) may be judged by the
solutions of this system only if the qualitative image of the trajectories of this
system is invariable under small—in some sense—disturbances of the right-hand
side part of (1.1). In other words, system (1.1) must possess the property of
robustness or structural stability.
For the reason just mentioned, the theory of parametric control of the market
economic development is proposed in [7, 8, 53–55].
This theory consists of the following components:
1. The methods for forming the set (library) of macroeconomic mathematical
models. These methods are oriented toward the description of various specific
socioeconomic situations, taking environmental safety conditions into
consideration.
2. The methods for estimating the conditions for robustness (structural stability) of
the models of national economic systems from the library without parametric
control. Here, the conditions of belonging to the considered mathematical
models of the Morse–Smale class of systems, the class of O-robust systems,
the class of uniformly robust systems, the class of У-systems, or the class of
systems with weak structural stability are verified.
3. The methods of control or attenuation of nonrobustness (structural instability) of
mathematical models of economic systems. This category involves choosing
(synthesis) the algorithm for control or attenuation of structural instability for the
respective mathematical models of a national economic system.
4. The methods of choice and synthesis of the laws of parametric control of market
economic development based on the mathematical models of a national eco-
nomic system.
5. The methods of estimating the robustness (structural stability) of mathematical
models of a national economic system from the library with parametric control.
In this group, the conditions of belonging to the considered mathematical models
with parametric control of the Morse–Smale class of systems, the class of
O-robust systems, the class of uniformly robust systems, the class of Y-systems,
or the class of systems with weak structural stability are verified.
6. The methods of adjustment of constraints on parametric control of market
economic development in the case of structural instability of mathematical
models of the national economic system with parametric control and adjustment
of constraints on the parametric control of market economic development.
7. The methods of research and analysis of bifurcations of extremals of variational
calculus problems of choosing optimal laws of parametric control theory.
2 1 Elements of Parametric Control Theory of Market Economic Development
8. Development of recommendations on the elaboration and implementation of
efficient governmental economic policy on the basis of parametric control theory
of market economic development, taking into consideration specific socioeco-
nomic situations.
1.2 Methods of Analysis of the Stability and Structural Stabilityof Mathematical Models of National Economic Systems
1.2.1 Development of the Methods for Evaluating StabilityIndicators of Mathematical Models
By Orlov’s definition [31], the mathematical model of an economic system in
general form is some function y ¼ f ðpÞ
f : A ! B; (1.3)
transferring values of initial (exogenous) data p 2 A to solutions (values of
endogenous variables) y 2 B:After constructing a mathematical model of some real-life phenomenon or
process f and defining some actual values of the point p by known measured data
or solving the parametric identification problem, the question arises about the
adequacy of the analyzed model. The condition of model stability against admissi-
ble perturbations of the initial data [31] is one of the conditions of the model’s
adequacy. In case of such stability, small perturbations of the model’s initial data
result in small changes in its solution. In monograph [31], the definitions of the
basic stability indicators are introduced (these definitions are presented below). But
Orlov [31] does not propose any algorithm for computing the considered indicators
of the mathematical model’s stability.
In this chapter we present the developed algorithms for evaluating the mathe-
matical model’s stability indicators, which characterize the stability of solutions of
the mathematical model against initial data perturbations. All of the model
parameters and variables must be made dimensionless first.
Let X ¼ ðX1;X2; :::;XkÞ be some vector of values of the model exogenous
parameters for a time interval t 2 f0; :::; Tg. Let X0 ¼ ðX10;X
20; :::; X
k0Þ denote the
respective vector of base values for the same time interval. The vector that
incorporates the values of parameters and initial values of the variables of differen-
tial (or difference) equations is the vector X. The vector of measured statistical data
used for finding the model equation coefficients is denoted by vector X for econo-
metric models.
Let p ¼ ðp1; p2; :::; pkÞ be a vector of the normalized input data of the mathemat-
ical model, where pi ¼ Xi
Xi0
; i ¼ 1; :::; k. The vector p0 ¼ ð1; 1; :::; 1Þ.
1.2 Methods of Analysis of the Stability and Structural Stability. . . 3
Let А be a space of the normalized input data vectors that includes all admissi-
ble sets p. A � Rk is a metric space with the Euclidean metric defined by the space
Rk; po 2 A.Let Y ¼ YðpÞ ¼ ðY1; Y2; :::; YnÞ be a selected vector of the values of endogenous
variables for some chosen interval (or moment) of time obtained for the selected
values of p. The vector that incorporates the values of some selected set of the
model endogenous variables for the aforesaid interval (or moment) of time is
denoted by vector Y for the dynamical models. The vector of coefficients of the
model equations or vector of values of some selected set of the model endogenous
variables for the aforesaid interval (or moment) of time is considered to be vector Yfor econometric models.
In particular, with p ¼ p0, we introduce the notation Y0 ¼ Yðp0Þ ¼ ðY10 ; Y
20 ; :::;
Yn0Þ. The normalized vector of values of the endogenous variables for the moment
of time T1 is denoted by y ¼ yðpÞ ¼ Y1
Y10
; Y2
Y20
; :::; Yn
Yn0
� �; y0 ¼ yðp0Þ ¼ ð1; 1; :::; 1Þ:
Let B � Rn be a region that contains all possible output values y for p 2 A with
the Euclidean metric of space Rn; y0 2 B . The considered model defines the
mapping f of set A into set B.For the selected point p 2 A and number a > 0, letUaðpÞ denote the intersection
of a neighborhood of the point p with radius a with set A:
UaðpÞ ¼ fx1 2 A : pðp1; pÞ � ag:
Here and below, r denotes the Euclidean distance between two points of the
Euclidean space.
For some subset B1 � B, let dðB1Þ denote the diameter of set B1; i.e.,
dðB1Þ ¼ supðrðy1; y2Þ; y1; y2 2 B1Þ:
Definition 1.1 The number
bðp; aÞ ¼ dð f ðUaðpÞÞ (1.4)
is defined as the stability indicator of the econometric model at the point X 2 Afor a > 0.
Algorithm 1.1 for evaluating the model stability indicator bðp; aÞ by the Monte
Carlo method is as follows:
1. Choose sets of input parameters (X) and output variables (Y), and compute their
normalized values.
2. Define the vector of normalized input data p ¼ ðp1; p2; :::; pkÞ, number a > 0,
and set UaðpÞ:3. Generate a set of sufficiently large number M of pseudo-random points (p1,
p2, . . ., pM) uniformly distributed in UaðpÞ:
4 1 Elements of Parametric Control Theory of Market Economic Development
For this purpose, consecutively generate the coordinates pijði ¼ 1; :::; k; j ¼ 1;:::;MÞof the point pj in numerical intervals ½pi � a; pi þ a�coveringUaðpÞusinga sensor of pseudo-random numbers distributed uniformly. If the inequality
Xk
i¼1ðpij � piÞ2 � a2
holds (i.e., xj 2 UaðpÞÞ, this point is added to the created set.
4. For each point pj of the set, define point yj ¼ f pj� �
, j ¼ 1; ::: ;M, by simulation.
5. Evaluate
b ¼ maxðrðyi; yjÞ : i; j ¼ 1; :::;MÞ:
6. Stop.
With a ¼ 0.01, the obtained number b/2 characterizes the (maximum) percent-
age change of values of the model output variables under the perturbed input
data by 1%.
Definition 1.2 The number
bðxÞ ¼ inf0<a�a0
bðp; aÞ (1.5)
is called the absolute stability indicator of the econometricmodel at point x 2 A. Here,a0 is the maximal admissible relative deviation of values of the model input data.
Algorithm 1.2 for evaluating the absolute stability indicator bðpÞ of the econo-
metric model is as follows:
For the selected value a0 and numbers j ¼ 0; 1; 2; . . . ; consecutively find
numbers bj ¼ b p; a0=2jð Þ;and then evaluate the number
bðpÞ ¼ infj¼0;1;2;:::
bj
by Algorithm 1.2. If bðrÞ turns out to be less than some a priori given small number
e [i.e., bðrÞ is considered to be approximately zero], then the mapping f defined by
the analyzed model is evaluated at point p continuously depending on the input
values.
Definition 1.3 The number
g ¼ supp2AbðpÞ (1.6)
is called the maximal absolute stability indicator of the model for region A.
1.2 Methods of Analysis of the Stability and Structural Stability. . . 5
Algorithm 1.3 for evaluating the maximal absolute stability indicator g of the
model by the Monte Carlo method is as follows:
1. Generate the set of sufficiently large numberM of pseudo-random points (p1, p2,. . ., pM) uniformly distributed in A.
2. For each point pj in the set and chosena0>0, find the numbers bðpÞ; j ¼ 1; :::;M;by Algorithm 1.2.
3. Determine the number
g ¼ maxj¼1;:::;M
bðpjÞ:4. Stop.
If the number g turns out to be less than some a priori given small number
e (i.e., g is considered to be approximately zero), then the mapping f definedby the analyzed model is evaluated in set A continuously depending on the
input values.
The developed algorithms were applied for evaluating the econometric model of
correlation of macroeconomic indicators and the CGEmodel of economic branches.
For the CGE model of the economic branches, we consider the set of possible
values of initial statistical data used for parametric identification of coefficients and
initial conditions from the difference equations of that model as set A. The resultsfrom the model’s simulation for some definite moments of time following the
period of model parametric identification are denoted as set B.
1.2.2 Development of Methods for Evaluating Weak StructuralStability of a Discrete-Time Dynamical System(Semi-cascade) Based on the Robinson Approach
The methods for analyzing the robustness (structural stability) of mathematical
models of national economic systems are based on
– Fundamental results on dynamical systems in the plane
– Methods of verification of mathematical models belonging to certain classes of
structurally stable systems (classes of Morse–Smale systems, O-robust systems,
Y-systems, systems with weak structural stability)
At present, the theory of parametric control of market economic development
has a number of theorems about the structural stability of specific mathematical
models (the model of the neoclassical theory of optimal growth; models of national
economic systems, taking into consideration the influence of the share of public
expenses and of the interest rate of governmental loans on economic growth;
models of national economic systems, taking into consideration the influence of
international trade and exchange rates on economic growth; and others formulated
and proved on the basis of the aforementioned fundamental results).
6 1 Elements of Parametric Control Theory of Market Economic Development
Along with analysis of the structural stability of specific mathematical models
(both with and without parametric control), based on results of the theory of
dynamical systems, one can consider approaches to the analysis of the structural
stability of mathematical models of national economic systems using computer
simulations.
We shall consider below the construction of a computational algorithm
for estimating the structural stability of mathematical models of national economic
systems on the basis of Robinson’s theorem (Theorem A) [69] on weak structural
stability.
LetN0 be some manifold and N a compact subset inN0 such that the closure of theinterior of N is N. Let some vector field be given in a neighborhood of the set N in
N0: This field defines the C1 -flux f in this neighborhood. Let Rðf ;NÞ denote the
chain-recurrent set of the flux f on N.Let Rðf ;NÞ be contained in the interior of N. Let it have a hyperbolic structure.
Moreover, let the flux f upon Rðf ;NÞ also satisfy the transversability conditions of
stable and unstable manifolds. Then the flux f on N is weakly structurally stable. In
particular, if Rðf ;NÞ is an empty set, then the flux f is weakly structurally stable on
N. A similar result is also correct for the discrete-time dynamical system (cascade)
specified by the homeomorphism (with image) f : N ! N0.Therefore, one can estimate the weak structural stability of the flux (or cascade) f
via numerical algorithms based on Theorem A via the numerical estimation of the
chain-recurrent set Rðf ;NÞ for some compact region N of the phase state of the
considered dynamical system.
Let’s further propose an algorithm of localization of the chain-recurrent set for a
compact subset of the phase space of the dynamical system described by a system of
ordinary differential (or difference) equations and algebraic system. The proposed
algorithm is based on the algorithm of construction of the symbolic image [33].
A directed graph (symbolic image), being a discretization of the shift mapping
along the trajectories defined by this dynamical system, is used for computer
simulation of the chain-recurrent subset.
Suppose an estimate of the chain-recurrent setRðf ;NÞof some dynamical system
in the compact set N of its phase space has been found. For a specific mathematical
model of the economic system, one can consider, for instance, some parallelepiped
of its phase space including all possible trajectories of the economic system’s
evolution for the considered time interval to be the compact set N.The localization algorithm for the chain-recurrent set consists of the following:
1. Define the mapping f defined on N and given by the shift along the trajectories of
the dynamical system for the fixed time interval.
2. Construct the partition C of the compact set N into cells Ni. Assign the directed
graph G with graph nodes corresponding to the cells and branches between the
cells Ni and Nj corresponding to the conditions of the intersection of the image of
one cell f(Ni) with another cell Nj.
3. Find all recurrent nodes (nodes belonging to cycles) of the graph G. If the set ofsuch nodes is empty, then Rðf ;NÞ is empty, and the process of its localization
1.2 Methods of Analysis of the Stability and Structural Stability. . . 7
ceases. One can draw a conclusion about the weak structural stability of the
dynamical system.
4. The cells corresponding to the recurrent nodes of the graphG are partitioned into
cells of lower dimension, from which a new directed graph G is constructed (see
item 2 of the algorithm).
5. Go to item 3.
Items 3–5 must be repeated until the diameters of the partition cells become less
than some given number e.The last set of cells is the estimate of the chain-recurrent set Rð f ;NÞ.The method of estimating the chain-recurrent set for a compact subset of the
phase space of a dynamical system developed here allows us to draw a conclusion
about the weak structural stability of the dynamical system when the obtained
chain-recurrent set Rð f ;NÞ is empty.
In the case when the considered discrete-time dynamical system is a priori the
semi-cascade f, one should verify the invertibility of the mapping f defined on N(since, in this case, the semi-cascade defined by f is the cascade) before applying
Robinson’s theorem for estimating its weak structural stability.
Let’s give a numerical algorithm for estimating the invertibility of the differ-
entiable mapping f : N ! N0 , where some closed neighborhood of the discrete-
time trajectory f f tðx0Þ; t ¼ 0; . . . ; Tg in the phase space of the dynamical system
is used as N. Suppose that N contains a continuous curve L connecting the points
f f tðx0Þ; t ¼ 0; . . . ; Tg. We can choose as such curve a piecewise linear curve with
nodes at the points of the above-mentioned discrete-time trajectory of the
semi-cascade.
An invertibility test for the mapping f : N ! N0 can be implemented in the
following two stages:
1. An invertibility test for the restriction of the mapping f : N ! N0 to the curve L,namely, f : L ! f ðLÞ. This test reduces to ascertaining the fact that the curve f ðLÞdoes not have points of self-crossing; that is, ðx1 6¼ x2Þ ) ð f ðx1Þ 6¼ f ðx2ÞÞ; x1;x2 2 L. For instance, one can determine the absence of self-crossing points by
testing the monotonicity of the limitation of the mapping f onto L along any
coordinate of the phase space of the semi-cascade f.2. An invertibility test for the mapping f in neighborhoods of the points of curve L
(local invertibility). Based on the inverse function theorem, such a test can be
carried out as follows: For a sufficiently large number of chosen points x 2 L;one can estimate the Jacobians of the mapping f using the difference
derivations: JðxÞ ¼ det@f i
@xjðxÞ
� �; i; j ¼ 1; . . . ; n: Here i, j are the coordinates
of the vectors, and n is the dimension of the phase space of the dynamical
system. If all the obtained estimates of Jacobians are nonzero and have the
same sign, one can conclude that JðxÞ 6¼ 0 for all x 2 L and, hence, that the
mapping f is invertible in some neighborhood of each point x 2 L.
8 1 Elements of Parametric Control Theory of Market Economic Development
An aggregate algorithm for estimating the weak structural stability of the
discrete-time dynamical system (semi-cascade defined by the mapping f ) with
phase space N0 2 Rn defined by the continuously differentiable mapping f can be
formulated as follows:
1. Find the discrete-time trajectory f f tðx0Þ; t ¼ 0; . . . ; Tg and curve L in a closed
neighborhood N that are required to estimate the weak structural stability of the
dynamical system.
2. Test the invertibility of the mapping f in a neighborhood of the curve L using the
algorithm described above.
3. Estimate (localize) the chain-recurrent set Rð f ;NÞ . By virtue of the evident
inclusion Rð f ;N1Þ � Rðf ;N2Þ for N1 � N2 � N0, one can use any parallelepipedbelonging to N0 and containing L as the compact set N.
4. IfRð f ;NÞ ¼ Ø, draw a conclusion about the weak structural stability of the
considered dynamical system in N.
This aggregate algorithm can also be applied to estimate the weak structural
stability of a continuous-time dynamical system (the flux f ) if the trajectory L ¼f f tðx0Þ; 0 � t � Tgof the dynamical system is considered to be the curve L. In thiscase, item 2 of the aggregate algorithm is omitted. The mapping f t for some fixed
t (t > 0) can be accepted as the mapping f in item 3.
1.3 Approach to Synthesis and Choice (in the Environmentof a Given Finite Set of Algorithms) of Optimal Lawsof Parametric Control of a National Economic System’sDevelopment. Existence Conditions for a Solutionto Respective Variational Calculus Problems. Conditionsof Influence of Uncontrolled Parameters to These Problems
1.3.1 Analysis of the Existence Conditions for a Solution ofthe Variational Calculus Problem of Synthesis and Choice(in the Environment of a Given Finite Set of Algorithms)of Optimal Laws of Parametric Control of aContinuous-Time Deterministic Dynamical System
1.3.1.1 Statement of the Variational Calculus Problems of Synthesisand Choice of Parametric Control Laws for a Continuous-TimeDeterministic Dynamical System
Let’s consider a nonlinear dynamical system (1.1) given by
x: ðtÞ ¼ f ðxðtÞ; uðtÞ; lÞ
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 9
with the initial condition (1.2)
xðt0Þ ¼ x0:
We’ll introduce the optimality criterion to be maximized as
K ¼Z t0þT
t0
F½xðtÞ�dt; (1.7)
where the function FðxÞ satisfies the Lipschitz condition.The following state-space constraints are on the system:
xðtÞ 2 XðtÞ; t 2 ðt0; t0 þ T�: (1.8)
The following constraints on control can also be imposed in problems consid-
ered ahead:
uðtÞ 2 UðtÞ; t 2 ½t0; t0 þ T�: (1.9)
Here XðtÞ � Rm;UðtÞ � Rq are some compact sets with nonempty interiors, and the
sets X ¼ [t2ðt0;t0þTÞ
XðtÞ and U ¼ [t2½t0;t0þTÞ
UðtÞ are bounded.Let’s formulate the following variational calculus problem called the variational
calculus problem of synthesis of optimal parametric control law for a continuous-time deterministic dynamical system.
Problem 1.1 From a known vector of uncontrolled parameters, find □ parametriccontrol law u satisfying condition (1.9) such that the respective solution of thedynamical system (1.1), (1.2) satisfies condition (1.8) and maximizes function (1.7).
In the second problem, we again consider a continuous-time dynamical control
system described by (1.1), (1.2) in the presence of state-space constraints (1.8).
However, unlike the previous case, the control is to be selected from this set of
given control laws:
ujðtÞ ¼ Gjðv; xðtÞÞ; t 2 ðt0; t0 þ T�; j ¼ 1; :::; r; (1.10)
where n ¼ ðn1; :::; nlÞ is the vector of coefficients to be adjusted (control parameters)
of the control law. It is assumed that vector function Gjðv; xÞ satisfies the Lipschitzcondition and linear constraints on growth rate:
jGjðn; xÞj � cð1þ jxjÞ; (1.11)
where c is some positive constant.
These adjusted coefficients are imposed by the constraints
v 2 V; (1.12)
10 1 Elements of Parametric Control Theory of Market Economic Development
where V is some compact subset of the space Rl. Moreover, it is assumed that the
control parameters must be such that the respective control law (1.11) satisfies
condition (1.9); i.e., the inclusion
Gjðv; xjðtÞÞ 2 UðtÞ; t 2 ðt0; t0 þ T� (1.13)
holds true, where xj(t) is the solution to problem (1.1), (1.2) for the selected values
of v,l and jth parametric control law.
The following optimality criteria are under consideration:
Kj ¼ Kjðv; lÞ ¼ðt0þT
t0
F½xjðtÞ�dt (1.14)
Let’s state the following extremal problem, called the variational calculus prob-lem of the choice (in the environment of a given finite set of algorithms) of optimalparametric control law for a continuous-time deterministic dynamical system.
Problem 1.2 From a known vector of uncontrolled parametersl 2 L, for each of rcontrol laws (1.10), find a vector of adjusted coefficients v such that the respectivesolution x ¼ xj of problem (1.1), (1.2) with control law u ¼ uj defined by formula(1.10) satisfies conditions (1.8), (1.12), (1.13) and maximizes function (1.14) withthe subsequent choice of the best obtained control law, which results in themaximum value of the optimum criterion.
1.3.1.2 Solvability Conditions for the Variational Calculus Problemof Synthesis of the Optimal Control Law for a Continuous-TimeDynamical System
Let’s now proceed to considering Problem 1.1.
We’ll consider the parametric control laws u ¼ uðtÞ ¼ ðu1ðtÞ; :::; uqðtÞÞ in subsetW of the Sobolev space ½H1ð0; TÞ�q of vector functions satisfying the relations
uðtÞ 2 UðtÞ; j _uðtÞj � c; t 2 ð0; TÞ: (1.15)
Here U(t) is a compact set in Rq defined above, and c is some positive constant.
We will refer to the set of such control laws u 2 W that yield the existence of the
solution to system (1.1), (1.2) and satisfy inclusion (1.8) as a set of admissible
control Wad.
Problem 1.1 consists of finding such an admissible control law u(t) that
maximizes function (1.7) on the set Wad.
The following theorem is valid.
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 11
Theorem 1.1 Under the above assumptions, Problem 1.1 is solvable if the set ofadmissible control Wad is a nonempty set.
The proof is presented in the appendix.
1.3.1.3 Solvability Conditions for the Variational Calculus Problemof Choice (in the Environment of a Given Finite Set of Algorithms)of the Optimal Control Law for a Continuous-Time DeterministicDynamical System
Consider Problem 1.2 of the variational calculus problem of choice (in the environ-
ment of a given finite set of algorithms) of the optimal control law on the basis of a
continuous-time dynamical system (1.1) with initial conditions (1.2).
For jth control law (1.10), system (1.1) is described by the equation
x: ðtÞ ¼ f ðxðtÞ;Gjðv; xðtÞÞ; lÞ (1.16)
with initial condition
xðt0Þ ¼ x0: (1.17)
Here, function f satisfies the constraints assigned in the beginning of Section 2.3,
and functionGj grows in absolute value not faster than a linear function and satisfies
the Lipschitz condition. The vector of adjusted coefficients is imposed by constraint
(1.12): u 2 V.Since, after the redesignation, (1.16) can be reduced to (1.1) with the same
functional properties, we can conclude that the solution to the considered system
above exists and is unique. As it is determined by concrete values of the control law
number j and adjusted coefficient v, hereinafter we shall use the notation x ¼ xvj ðtÞ.For any j ¼ 1; :::; r; we’ll define the set Vj
ad of admissible values of the adjusted
coefficient as a set consisting of such values of v satisfying condition (1.12),
yielding the respective solution to problem (1.16), (1.17) to satisfy the inclusions
Gjðv; xvj ðtÞÞ 2 UðtÞ; t 2 ðt0; t0 þ TÞ (1.18)
xvj ðtÞ 2 XðtÞ; t 2 ðt0; t0 þ TÞ: (1.19)
Problem 1.2 consists of choosing the best control law, namely, such a number jthat yields the maximum of function (1.14) KJ on the set Vj
ad to be the largest.
Theorem 1.2 Under the above assumptions and in the case when the sets Vjad are
nonempty, Problem 1.2 is solvable.This proof is presented in the appendix.
12 1 Elements of Parametric Control Theory of Market Economic Development
1.3.2 Analysis of the Existence Conditions for a Solutionof the Variational Calculus Problem of Synthesisand Choice (in the Environment of a Given Finite Setof Algorithms) of Optimal Laws of Parametric Controlof a Discrete-Time Deterministic Dynamical System
1.3.2.1 Statement of the Variational Calculus Problems of Synthesisand Choice of Parametric Control Laws for a Discrete-TimeDeterministic Dynamical System
Let’s consider the discrete-time stochastic control system
xðtþ 1Þ ¼ f ðxðtÞ; uðtÞ; lÞ; t ¼ 0; 1; :::; n� 1; (1.20)
xð0Þ ¼ x0; (1.21)
where the time t is a nonnegative integer value, and n is a fixed natural number;
x ¼ xðtÞ 2 Rm is the state vector of system (1.20), (1.21), a vector function of the
discrete-time argument;
x0 2 Rm is the initial state of the system and a deterministic vector;
u ¼ uðtÞ 2 Rq is the control parameters’ vector, a vector function of the discrete-
time argument;
l 2 L � Rs is the vector of the uncontrolled parameters, L is the open connected
set;
f is the defined vector function of its arguments.
The optimality criterion to be maximized is given by
K ¼Xnt¼1
FtðxðtÞÞ: (1.22)
Here Ft(x) are the known functions.
The state-space constraints on the system are as follows:
xðtÞ 2 XðtÞ; t ¼ 1; :::; n: (1.23)
The following constraints on the control can also be imposed in the problems
considered below:
uðtÞ 2 UðtÞ; t ¼ 0; :::; n� 1 (1.24)
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 13
Here XðtÞ � Rm; UðtÞ � Rq are some compact sets with nonempty interiors.
Let’s formulate the following variational calculus problem, called the varia-tional calculus problem of the synthesis of the optimal parametric control law fora discrete-time deterministic dynamical system.
Problem 1.3 From the known vector of uncontrolled parameters l, find aparametric control law u satisfying condition (1.24) such that the respectivesolution of dynamical system (1.20), (1.21) satisfies condition (1.23) and maximizesfunction (1.22).
In the second problem, we again consider the continuous-time dynamical control
system described by (1.20), (1.21) in the presence of state-space constraints (1.23).
Here, the control is to be selected from the set of given control laws:
ujðtÞ ¼ Gjðv; xðtÞÞ; t ¼ 0; :::; n� 1; j ¼ 1; :::; r: (1.25)
Here Gj is the vector function of its arguments satisfying the Lipschitz condition,
and v ¼ ðv1; :::; vlÞ is the vector of the coefficients of control law Gj to be adjusted.
These adjusted coefficients are imposed by the constraints
v 2 V; (1.26)
where V is some compact subset of space Rl. Moreover, it is assumed that the
control parameters must be such that the inclusion
Gjðv; xvj ðtÞÞ 2 UðtÞ; t ¼ 0; :::; n� 1; (1.27)
holds true. Here xvj ðtÞ is the solution to problem (1.20), (1.21) for the selected values
of v, a, and jth parametric control law.
We consider the following optimality criteria:
Kj ¼ Kjðv; lÞ ¼Xn
t¼1Ftðxvj ðtÞÞ: (1.28)
Let’s state the following extremal problem, called the variational calculusproblem of choice (in the environment of a given finite set of algorithms) of theoptimal parametric control law for a discrete-time deterministic dynamical system.
Problem 1.4 From a known vector of uncontrolled parameters l 2 L, for each of rcontrol laws from set (1.25), find a vector of the adjusted coefficients v such that therespective solution x ¼ xj of problem (1.20), (1.21) with this control law satisfiesconditions (1.23), (1.26), (1.27) and maximizes function (1.28).
14 1 Elements of Parametric Control Theory of Market Economic Development
1.3.2.2 Solvability Conditions for the Variational Calculus Problemof Synthesis of the Optimal Control Law for a Discrete-TimeDeterministic Dynamical System
We will refer to the set of such the control laws u satisfying (1.24) such that the
solution to system (1.20), (1.21) satisfying inclusion (1.23) is the set of admissible
controls Wad.
Problem 1.4 consists of finding such an admissible control law u that maximizes
function (1.22) on the set Wad.
The following statement is a natural development of the classical Weierstrass
theorem on the existence of the extremum of a continuous functionwithin an interval.
Theorem 1.3 Let the vector function f be continuous with respect to the aggregateof arguments, let the sets X(t), U(t) be closed and bounded for all definite t, and letthe function Ft(x) be continuous. Then, if the set Wad of the admissible control lawsis nonempty, Problem 1.4 is solvable.
The proof is presented in the appendix.
1.3.2.3 Solvability Conditions for the Variational Calculus Problemof Choice (in the Environment of a Given Finite Set of Algorithms)of the Optimal Control Law for a Discrete-Time DeterministicDynamical System
Let’s now consider Problem 1.4 formulated above.
Denote by xvj the solution of system (1.20), (1.21) for the selected jth parametric
control law (1.25), its adjusted coefficient v, and parameter l:
xvj ðtþ 1Þ ¼ f ðxvj ðtÞ;Gjðv; xvj ðtÞÞ; lÞ; t ¼ 0; :::; n� 1; (1.29)
xvj ðt0Þ ¼ x0: (1.30)
Similarly to Theorem 1.2, we’ll define the set of admissible valuesVjad consisting
of such values of v satisfying condition (1.26), which yields the respective solution
of problem (1.29), (1.30) to satisfy the inclusions
Gjðv; xvj ðtÞÞ 2 UðtÞ; t ¼ 0; :::; n� 1; (1.31)
xvj ðtÞ 2 XðtÞ; t ¼ 1; :::; n: (1.32)
Theorem 1.4 Assume that in Problem 1.3.4 the functions f, Gj, and Ft are continu-ous with respect to the aggregate of the arguments. Then, if sets Vj
ad are nonempty,Problem 1.3.4 is solvable.
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 15
1.3.3 Analysis of the Existence Conditions for a Solutionof the Variational Calculus Problem of Synthesisand Choice (in the Environment of a Given Finite Setof Algorithms) of Optimal Laws of Parametric Controlof a Discrete-Time Stochastic Dynamical System
1.3.3.1 Statement of the Variational Calculus Problems of Synthesisand Choice of Parametric Control Laws for a Discrete -TimeStochastic Dynamical System
Let’s consider the discrete-time stochastic control system
xðtþ 1Þ ¼ f ðxðtÞ; uðtÞ; lÞ þ xðtÞ; t ¼ 0; :::; n� 1; (1.33)
xð0Þ ¼ x0 (1.34)
Here
x ¼ xðtÞ 2 Rm is the state vector of system (1.33), (1.34), a random vector function
of the discrete-time argument (a vector random sequence);
x0 is the initial state of the system and a deterministic vector;
u ¼ uðtÞ 2 Rq is the control parameters’ vector, a vector function of the discrete-
time argument;
l 2 Rs is the vector of the uncontrolled parameters; l 2 L; L � Rsis a given set;
x ¼ xðtÞ ¼ ðx1ðtÞ; :::; xmðtÞÞ is the known vector random sequence corresponding to
the disturbances (e.g., Gaussian white noise);
f is the defined vector function of its arguments.
The optimality criterion to be maximized is given by
K ¼ EXn
t¼1FtðxðtÞÞ
n o(1.35)
Here Ft are known functions, and E is the expectation.
The following state-space constraints are on the system:
E½xðtÞ� 2 XðtÞ; t ¼ 1; :::; n; (1.36)
where X(t) is a given set.
The problems considered below retain the constraints on the control defined
above:
uðtÞ 2 UðtÞ t ¼ 0; :::; n� 1; (1.37)
where U(t) � Rq is the given set.
16 1 Elements of Parametric Control Theory of Market Economic Development
We’ll formulate the following problem, called the variational calculus problemof the synthesis of the optimal parametric control law for a discrete-time stochasticdynamical system:
Problem 1.5 For the known vector of uncontrolled parameters a, find a parametriccontrol law u satisfying condition (1.37) such that the respective solution ofdynamical system (1.33), (1.34) satisfies condition (1.36) and maximizes function(1.35).
In the second problem of parametric control of a discrete-time dynamical
system, we again consider the discrete-time dynamical control system described
by (1.20), (1.21) in the presence of state-space constraints (1.23). In this problem, as
in the previous Problems 1.2 and 1.3.4, the control is to be selected from the set of
given control laws:
ujðtÞ ¼ Gjðv; xðtÞÞ; t ¼ 0; :::; n� 1; j ¼ 1; :::; r; (1.38)
where Gj is the known vector function of its arguments, and v ¼ ðv1; :::vlÞ is the
vector of parameters of the control law Gj to be adjusted.
These adjusted coefficients are imposed by the constraints
v 2 V; (1.39)
where V is some compact subset of the space Rl. Moreover, it is assumed that the
control parameters must be such that the respective control law (1.43) satisfies
condition (1.42); i.e., the inclusion
E[Gjðv; xvj ðtÞÞ� 2 UðtÞ; t ¼ 0; :::; n� 1; (1.40)
holds true. Herexvj ðtÞ is a solution to problem (1.38), (1.39) for the selected values of
v, l, and jth parametric control law.
We consider the following optimality criteria:
Kj ¼ Kjðv; lÞ ¼ EXnt¼1
Ftðxvj ðtÞÞ( )
: (1.41)
Let’s state the following extremal problem, called the variational calculusproblem of choice of the optimal parametric control law for a discrete-timestochastic dynamical system.
Problem 1.6 From a known vector of uncontrolled parameters l 2 L, for each of rcontrol laws, find a vector of adjusted coefficients v such that the respective solutionx ¼ xj of problem (1.33), (1.34) with control law u ¼ uj defined by formula (1.38)
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 17
satisfies conditions (1.39), (1.40) and maximizes function (1.41) with the subsequentchoice of the best obtained optimal control laws, i.e., corresponding to the greatestvalue of the optimality criterion.
1.3.3.2 Solvability Conditions for the Variational Calculus Problemof Synthesis of the Optimal Control Law for a Discrete-TimeStochastic Dynamical System
Let’s now consider the solvability of Problem 1.5. Define set Uad of the admissible
control laws for the considered system as the aggregate of control laws u(t)satisfying constraint (1.37) such that the expectation E xðtÞ½ � of the respective
solution of the stochastic system satisfies inclusion (1.36).
Theorem 1.5 In Problem 1.5 with l 2 L and for any t ¼ 1; . . . ; n; let the randomvariables xðtÞ be absolutely continuous and have zero expectations. Let the setsXðtÞ;UðtÞðt0Þ � xvj ðt0Þ ¼ 0 be closed and bounded for all t, let function f satisfy theLipschitz condition, and let function Ft be continuous. Let the absolute values offunction f (foru 2 Uandl 2 L) and Ft not exceed some linear functions in |x|. Then,if set Uad of the admissible control laws is nonempty, Problem 1.3.5 is solvable.
The proof is presented in the appendix.
1.3.3.3 Solvability Conditions for the Variational Calculus Problemof Choice (in the Environment of a Given Finite Set of Algorithms)of the Optimal Control Law for a Discrete-Time StochasticDynamical System
Let’s now consider Problem 1.6 formulated above.
Denote by xvj , the solution to system (1.33), (1.33) for the selected jth parametric
control law (1.25), its adjusted coefficient v, and parameter a:
xvj ðtþ 1Þ ¼ f ðxvj ðtÞ;Gjðv; xvj ðtÞÞ; lÞ þ xðtÞ; t ¼ 0; :::; n� 1; (1.42)
xvj ð0Þ ¼ x0: (1.43)
For the considered problem, we’ll define the set of admissible values Vjad
consisting of such values of v 2 V satisfying condition (1.39), which yields the
respective solution of problem (1.33), (1.34) to satisfy the inclusions
E[Gjðv; xvj ðtÞÞ� 2 UðtÞ; t ¼ 0; :::; n� 1; (1.44)
18 1 Elements of Parametric Control Theory of Market Economic Development
and
E[xvj ðtÞ� 2 XðtÞ; t ¼ 1; :::; n: (1.45)
Problem 1.6 is called nontrivial if the respective setVjad is nonempty and contains
some open set for any j ¼ 1; :::; r.
Theorem 1.6 In Problem 1.6, let l 2 L, let the sets U(t), X(t), and V be compact,and let the functions f, Gj, and Ft satisfy the Lipschitz condition. Let these functionsalso satisfy the following constraints on growth: The functions jf ðx;Gjðv; xÞ; lj andjFtðxÞjdo not exceed some functions linear in jxjuniformly with respect to v 2 V. Letthe random variable xðtÞ be absolute and continuous and have zero expectation.Then, if set Vj
ad is nonempty, Problem 1.6 is solvable.The proof is presented in the Appendix.
1.3.4 Analysis of the Influence of Uncontrolled ParameterVariations (Parametric Disturbances) on the Solutionof the Variational Calculus Problem of Synthesisand Choice of Optimal Parametric Control Laws
Below we present the results of the analysis of the influence of uncontrolled
parameter variations and the bifurcation point changes under parametric
disturbances in the variational calculus problem of choosing optimal parametric
control laws in the environment of given finite sets of the algorithms with phase
constraints and constraints in the allowed form.
The functional or phase constraints, as well the constraints in the allowed form
of the considered problems, often depend on the values of the vector parameter. The
analysis of similar problems requires first finding sufficient conditions for stability
of optimal values of the criteria considered as a function of the uncontrolled
parameters.
With the application for solving the problems of choosing parametric control
laws in the environment of a given finite set of algorithms, we are required to define
the bifurcation point and conditions of its existence and to analyze the bifurcation
value of this parameter. With the application of parametric control of the mecha-
nisms of market economies, finding the extremal solution of the respective problem
and its type can depend on the values of some uncontrollable parameters, and the
task of defining the bifurcation value becomes practical.
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 19
1.3.4.1 Analysis of Continuous Dependence of Optimal Values of theCriteria of Variational Calculus Problems of Synthesis and Choiceof Optimal Parametric Control Laws on the Values of UncontrolledParameters
The following auxiliary definition and lemma are useful for proving the continuous
dependence of optimal values of the criterion of the above variational calculus
problem of synthesis of optimal parametric control laws on the values of uncon-
trolled parameters.
Definition Let the family of functions KaðuÞ; a 2 A and u 2 Ua be defined under afamily of subsets fUag of some set U in the Banach space with parameter a in somesubset A of a Euclidean space. The familyfUag is called K-continuous in set A if, forany e>0; there exists a number such that if inequalitiesja� bj � d and a; b 2 Ahold true, then for any point ub 2 Ub, there exists some point ub 2 Ub such thatinequality jKaðuaÞ � KaðubÞj<e holds true.
By this definition, the family of sets fUag is K-continuous if, in the case of
sufficient proximity of parameter b to a for any element from set Ua, there exists an
element of set Ub that is arbitrary close to it (by the value of the functions).
Lemma 1.1 Let A and U be some subsets of Euclidean and Banach spaces,respectively. Let U be compact; all elements in the family of closed subsets fUaglie in U. Let the mapping ða; uÞ ! KaðuÞbe continuous on the productA� U. Let thefamily of subsets fUag be continuous in all points of some neighborhood at pointa0 2 A. Then the mapping a ! maxu2Ua
KaðuÞ is continuous at point a0 2 A.The proof is presented in the appendix.
We’ll now formulate the theorem on sufficient conditions of the continuous
dependence of optimal values of criterion of Problems 1.1, 1.3, and 1.3.5 on the
uncontrolled parameter.
For all aforesaid problems, we will assume that sets X(t) andU(t) are compact, and
sets [t2ðt0;t0þTÞ
XðtÞ and [t2ðt0;t0þTÞ
UðtÞ for Problems 1.1, 1.3, and 1.3.5 are bounded; sets
L are connected and open. It is also assumed that fulfillment of the conditions of
Theorems 1.1, 1.3, and 1.5, which guarantee the continuous dependence of the
considered system, states xu;lðtÞ (or E[xu;l(t)] for the stochastic problems) and the
optimality criterion of Problems 1.1, 1.3, and 1.3.5; dependence of K ¼ Kðu; lÞ onthe respective values of control parameters or control laws (u) and uncontrolled
parameters (l) in the sets determined by theorems stated and with the metrics
indicated above.
Theorem 1.7 Let the conditions of Theorems 1.1, 1.3, and 1.5 hold true for anyl 2 L. Then the optimal values for the criteria of respective Problems 1.1, 1.3, and1.3.5 are continuous functions of parameter l 2 L:
The proof is presented in the appendix.
20 1 Elements of Parametric Control Theory of Market Economic Development
Let’s now analyze the conditions of continuous dependence of optimal values of
the criterion of the variational calculus problems of choice of the parametric control
laws on uncontrolled parameters.
Theorem 1.8 Let the conditions of Theorems 1.2, 1.4, and 1.6 hold true for anyl 2 L . Then for a selected law number j, the optimal values of criteria Kj
corresponding to Problems 1.2, 1.3.4, and 1.6 are continuous functions of parame-ter l 2 L.
The proof is presented in the appendix.
Corollary 1.1 If the conditions of Theorem 1.8 hold, the optimal values of thecriteria of Problems 1.2, 1.3.4, and 1.6 are continuous functions of parameterl 2 L:
The proof is given in the appendix.
1.3.4.2 Analysis of Bifurcation Points of the Variational CalculusProblems of Choice of Parametric Control Laws in theEnvironment of a Given Finite Set of Algorithms
Let’s introduce a definition characterizing such values of uncontrolled parameter
l 2 L; which admits a change of one optimal control law to another.
Definition The parameter l 2 L is said to be the bifurcation point for Problem1.2, 1.3.4, or 1.6 if there exist two different numbers of control laws i; j 2 f1; :::; rgsuch that the relation
maxv2Vi
ad;l
Kiðv; lÞ ¼ maxv2Vj
ad;l
Kjðv; lÞ ¼ maxj¼1;:::;r
maxv2Vj
ad;l
Kjðv; lÞ
holds true. An additional requirement is the existence of such a point l1 2 L in anyneighborhood around point l that yields the sole law number j corresponding tomaxj¼1;:::;r
maxv2Vj
ad;l1
Kjðv; l1Þа.
The following theorem establishes sufficient conditions of existence for the
bifurcation point of extremals of the considered variational calculus problems for
choosing a parametric control law in a given finite set of algorithms for the cases of
a continuous-time deterministic or discrete-time deterministic or stochastic
dynamical system.
Theorem 1.9 Let the conditions in Theorem 1.2, 1.4, or 1.6 be fulfilled for any l2 L; where L is an open compact set. Suppose there exist two different values of
1.3 Approach to Synthesis and Choice (in the Environment of a Given Finite Set. . . 21
parameter l0 and l1 l0 6¼ l1; l0; l1 2 Lð Þ that yield the solutions of the respectiveproblems for choice of the optimal parametric control laws to be attained for twodifferent laws Gj0and Gj1ðj0 6¼ j1Þ in the given finite set; i.e.,
maxj¼1;::;r
j6¼j0
maxv2Vj
ad;l0
Kjðv; l0Þ < maxv2Vj0
ad;l0
Kj0ðv; l0Þ;
maxj¼1;::;r
j 6¼j1
maxv2Vj
ad;l1
Kjðv; l1Þ < maxv2Vj1
ad;l1
Kj1ðv; l1Þ:
Then there exists at least one bifurcation point of extremals in Problem 1.2,
1.3.4, or 1.6.The proof can be found in the appendix.
The following statement immediately follows from Theorem 1.9.
Corollary 1.2 Let the conditions of Theorem 1.2, 1.4, or 1.6 be fulfilled for anyl 2 L; where 0:0012<pt � 0:0480; t ¼ 10; :::; 14; is an open compact set. For l¼ l0; let the control law Gj0 from a given finite set of the parametric control lawsyield the solution of Problem 1.2, 1.3.4, or 1.6. For l ¼ l1; ðl0 6¼ l1; l0; l1 2 LÞ;the control law Gj0 from this set does not result in the solution of the consideredProblem 1.2, 1.3.4, or 1.6. Then there exists at least one bifurcation point of theextremals of the considered problems.
The proof is presented in the appendix.
In conclusion, we’ll introduce a numerical algorithm for finding the bifurcation
value of parameter l of one of Problems 1.2, 1.3.4, or 1.6 regarding choice of the
parametric control laws (in the environment of a given finite set of algorithms)
under fulfillment of Theorem 1.9’s conditions.
Connect the points l0 and l1 by a smooth line s � L. Partition this line into nequal parts with sufficiently small steps. For the obtained values bi 2 s; i ¼ 0; :::;n; b0 ¼ l0; bn ¼ l1 define the optimal numbers ji of the control laws solving
Problem 1.2, 1.3.4, or 1.6 with l ¼ bi . Then find the first value of i at which the
respective number of the law differs from j0. In this case, the bifurcation point of
parameter l lies on the arc bi�1; bi of line S.For the obtained section of the line, the algorithm of defining the bifurcation
point with given accuracy e implies the application of the bisection method. As a
result, we find the point g 2 ðbi�1; biÞ, from one side of which, within the limits of
deviation e from value g, the set of lawsGj0 is optimal, but from another side within
the limits of deviation e from the value g, this set is not optimal. From Corollary 1.2,
it follows that there exists a bifurcation point of the solved problem on the
mentioned arc.
22 1 Elements of Parametric Control Theory of Market Economic Development
1.4 Algorithm of Application of Parametric Control Theoryand Rules of Interaction Between Persons MakingDecisions on Elaboration and Realization of the EffectiveState Economic Policy on the Basis of an InformationSystem for Decision-Making Support
1.4.1 Algorithm of the Application of ParametricControl Theory
The application of the theory of parametric control of market economic evolution
for the definition and implementation of efficient public economic policy developed
here seems to be as follows [7, 11, 53, 54]:
1. The choice of direction (strategy) for economic development of a country on the
basis of estimating its economic state in the context of phases of the economic cycle.
2. The choice of one or several mathematical models addressing the problems of
economic development from the library of mathematical models of economic
systems.
3. The estimate of the adequacy of mathematical models of the problems. The
calibration of mathematical models (parametric identification and retrospective
prediction by the current indices of evolution of the economic system) and
additional verification of the chosen mathematical models via econometric
analysis and political-economic interpretation of sensitivity matrices.
4. The evaluation (if necessary) of indicators of stability and/or structural stability
(robustness) of mathematical models without parametric control in accordance
with the aforesaid methods of estimation of stability indicators and robustness
conditions. If the mathematical model is stable and/or structurally stable, it may
be used after the econometric analysis and political-economic interpretation of
the results of robustness analysis, and for solving the problem of choosing the
optimal control law for economic parameters and the prediction of macroeco-
nomic indicators.
5. If the mathematical model is unstable or nonrobust (structurally unstable), then it
is moved to a reserve section of the model library.
6. The choice of optimal laws of control for economic parameters.
7. The estimation (if necessary) of indicators of stability and/or structural stability
(robustness) of mathematical models with the chosen laws of parametric control
according to the aforesaid methods of estimation of robustness conditions. If the
mathematical model with the chosen laws of parametric control is stable and/or
structurally stable, then from the obtained results, after carrying out the respective
econometric analysis, the completing the political-economic interpretation, and
garnering approval of the decision makers, the obtained results can be put into
practice. If the mathematical model with the chosen laws of parametric control is
unstable and/or structurally unstable, then the choice of parametric control laws
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 23
must be refined. The corrected decisions on choosing the parametric control laws
are also to be considered according to the above-mentioned scheme.
8. Analysis of the dependence of the chosen optimal laws of parametric control on
the variation of uncontrolled parameters in an economic system. In this regard,
replacing one optimal parametric control law by another is possible.
This aggregate scheme for making decisions on the development and implemen-
tation of an efficient public economic policy via the choice of optimal values of
economic parameters must be maintained by methods of analysis and computer
simulations. This aggregate scheme for making decisions is presented in Fig. 1.1.
As is known, decision making is a process [25, pp. 126, 146] that can be divided
into the following three stages (G. Simon, [23]): information search, searching for
and finding of alternatives, and choosing the best alternative.
The information decision support system (IDSS) for applications based on the
parametric control system is intended for the support decision makers in the field of
economic policy under conditions of small open national economy on the basis of
parametric control theory.
The subjects of interaction of the decision makers (DM) with the IDSS are as
follows:
1. Evaluation of national economic conditions and choice of direction of economic
policy;
2. Choice of one or several econometric models and(or) dynamical models (in the
form of dynamical systems) from the library of mathematical models that are
consistent with economic policy problems;
3. Statement and problem solving of estimation values (variation laws) of eco-
nomic instruments for the economic policy implementation in the chosen
direction;
4. Analysis of the influence of values of exogenous social and economic indicators
on the results of the problem’s solution and estimation of values (variation laws)
of economic instruments of the economic policy implementation in the chosen
direction;
5. Formulating recommendations on the choice of values (variation laws) of
economic instruments of implementation of an economic policy in the chosen
direction.
1.4.2 Rules of Interaction for Decision Makerson the Formulation and Implementationof an Effective Public Economic Policy Basedon the Information Decision Support System
Let’s consider the rules of interaction between the DM and IDSS at the levels of
problems listed above.
24 1 Elements of Parametric Control Theory of Market Economic Development
Yes No
1. Evaluation of the national economic conditions and choice of direction of economic developmentin coordination with the decision maker
2. Selection of a set consisting of one or several econometric models and (or) dynamic models(models in form of dynamic system) from the library which are consistent with problems of developmentdirection in coordination of the results with the preferences of the decision maker. Parametricidentification of the selected models (if necessary)
3. Evaluation (if necessary) of the stability indicators of each econometric model in the set and/orweak structural stability (robustness) of each dynamic model in the set.
4. Are all models in the setstable or weakly structurally
stable (robust)?
6. Moving of unstable or non-robust models to the reservesection of the model library.
Go to block 2.
5. Statement and solution to problems of evaluating values (law of variation) of economic instruments foran economic policy implementation in the chosen direction based on the set of considered modelsin coordination with the decision maker. Econometric analysis, political-economic interpretation ofthe choice of the optimal economic policy
7. Evaluation (if necessary) of stability indicators of each econometric model in the set with selectedoptimal values of economic instruments. Evaluation (if necessary) of robustness of each dynamical modelin the set with selected optimal values of economic instruments. Econometric analysis, political-economicinterpretation of the results of the analysis of stability and robustness of the mathematical models withthe selected optimal values of economic instruments
No Yes8. Are all models in the
set with selected optimal values ofeconomic instruments stable
or robust?
9. Correction of the constraints in the problems of choice of optimal economicpolicy based on unstable (non-robust) mathematical models. Econometricanalysis, political-economic interpretation of the results of correction ofconstraints. Go to block 7.
10. Analysis of influence of exogenous social and economic indicators on results of solving the problem ofchoice of optimal economic policy in coordination with the decision maker. Econometric analysis,political-economic interpretation of the results of the extremals and adjustment in accordance with thedecision maker
11. Formulation of recommendations on choosing values (variation laws) of economic instrumentsof implementation of economic policy in the chosen direction in coordination with the decision maker
Fig. 1.1 Aggregate scheme of the algorithm for decision making and the implementation
of efficient public economic policy via the choice of optimal values of economic instruments:
(a) Sheet 1; (b) Sheet 2; (c) Sheet 3
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 25
1.4.2.1 Rules of Interaction Between the DM and IDSS at the Levelof the Problem “Evaluation of the National Economic Conditionsand Choice of an Economic Policy Direction”
According to G. Simon, a part of the considered problem “evaluation of the national
economic conditions” is connected with the stage of information search for struc-
turing the problem of choosing an economic policy’s direction. It results in forming
the necessary initial information based on the IDSS for the mentioned problem in
the operating model of man–machine procedures of interaction between the DM
and IDSS.
The rules of interaction between the DM and IDSS at the considered stage are
formulated as the following stages:
1. Computations. Administrator(s) of the IDSS presents to the DM the initial
information obtained from the IDSS in the form of, e.g., equilibrium values of
credit rate i, gross national income y in welfare and monetary markets by years
characterizing points of effective demand, and indicators of equilibrium based
on the Keynesian model of yearly all-economic equilibrium; indicators of the
national economic equilibrium based on the model of an open economy of a
small country by years and data on actual annual values of national economy
indicators within the limits of their equilibrium values in their respective mac-
roeconomic markets; causes and characteristics of market cycles in economics;
indicators of activity of economic agents; national economic functioning in the
past period and middle-time forecasting period based on various computable
models of general equilibrium and information obtained on the basis of other
mathematical models of the national economy.
2. DM analysis. Evaluating the presented information, the DM determines if the
information is admissible. If it is, then the procedure of interaction between the
DM and IDSS is executed. Otherwise, the DM requests new information.
3. Computations. If the request for new information can be provided, the adminis-
trator(s) of the IDSS presents the requested information based on working with
the IDSS (possibly, this is made in coordination with the DM).
The process of interaction between the DM and IDSS within the framework of
steps 1–3 continues until the DM receives admissible information.
The other part of the considered problem, “choice of economic policy’s direc-
tion,” is solved in two stages. In the first stage, based on information obtained from
the IDSS, information on the results of monitoring the socioeconomic area
conditions, and political stability and national security protection, the problem of
choosing an economic policy direction is structured in the form of a tree open graph
for multiobjective choice in coordination with the DM in the following steps:
Step 1. Based on the information obtained from the IDSS and the results from
monitoring national economic conditions and properties of a tree open graph
(tree of objectives), the administrator(s) of the IDSS presents a version of
decomposition of the main goal to a subgoal within the framework of hierarchy,
26 1 Elements of Parametric Control Theory of Market Economic Development
and relative significance [23] to the DM. The formulation of such goals that
cannot be divided and give the final results defined by the main goal is the
complete flag for decomposition of the main goal.
Step 2. Evaluating the presented version of the tree of objectives, the DM
determines whether or not it is admissible. If not, the DM formulates the remarks
and requests adjustment of the presented version of the tree of objectives.
Step 3. The administrator(s) of the IDSS adjusts the initial version of the tree of
objectives in accordance with the remarks and requests of the DM and presents
the renewed version of the tree of objectives for the problem of choosing the
economic policy’s direction.
Steps 2 and 3 must be repeated until the version of the tree of objectives for the
problem of choosing the economic policy’s direction is agreed upon and accepted
by the DM.
The second stage of solving the problem “choice of the economic policy’s
direction” reduces to making the decision based on the tree of objectives via
interaction of the DM with the administrator(s) of the IDSS within the framework
of the following steps, which are realized by a hierarchal analysis method or
paired comparison method (such as ELECTRE I, ELECTRE II, ELECTRE III,
and others), with corresponding means for evaluating the results of coordination
with the DM.
Step 1. Rank the goals of each level by the relative significance of attaining a
higher-level goal.
Step 2. Evaluate the importance of each goal with respect to higher-level goals and
the main goal. The sum of the significance coefficients of goals of each level
must be equal to 1 or 100.
Step 3. Evaluate the significance coefficients of this goal with respect to the main
goal.
Step 4. Choose an economic policy direction based on the results of steps 1–3.
1.4.2.2 Rules of Interaction Between the DM and IDSS at the Levelof Problem “Choice of One or Several Econometric Modelsand (or) Dynamical Models (in Form of Dynamical Systems)from the Library of Mathematical Models That Are Consistentwith Economic Policy Problems”
Solving the stated problem reduces to making a decision via interaction of the DM
with administrator(s) of the IDSS within the framework of the following steps:
Step 1. Based on results of choosing the economic policy’s direction, the adminis-
trator(s) of the IDSS proposes a set of mathematical models (or one model) that
are consistent with the economic policy’s direction from the library of mathe-
matical models.
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 27
Step 2. Evaluating the proposed set of mathematical models (one model), the DM
determines whether or not it is admissible. If yes, the interaction between the
DM and IDSS is complete. Otherwise, the DM puts new requests within the
framework of choice from the set of mathematical models (one model).
Step 3. If the DM’s request can be implemented, the IDSS administrator(s) presents
the DM with a new version of the set of mathematical models (one model) from
the library.
The process of interaction between the DM and IDSS continues within the
framework of steps 1–3 until the DM receives the admissible set of mathematical
models (one model).
1.4.2.3 Rules of Interaction Between the DM and IDSS at the Levelof Problem “Statement and Solution of the Problem(s)of Estimation(s) of Values (Variation Laws) of EconomicInstruments for Implementation of the Economic Policyin the Selected Direction”
The considered problem “statement and solution to the problem(s) of estimation(s)
of values (variation laws) of economic instruments for implementation of the
economic policy in the selected direction” is concerned with its agreement at the
level of criterion (criteria) and constraints on the range of variation of economic
instruments. It reduces to making the agreed-upon decisions via the interaction
between the IDSS administrator(s) and the DM at the level of the following steps:
Step 1. Based on the selected direction of the economic policy and agreed-upon
version of the set of mathematical models (one model), the IDSS administrator
(s) proposes the statement of problem(s) of estimation(s) of values (variation
laws) of economic instruments for implementation of the economic policy in the
selected direction to the DM.
Step 2. Evaluating the proposed statement for the problem of estimation(s) of values
(variation laws) of economic instruments for implementation of the economic
policy in the selected direction, the DM determines whether or not it is admissi-
ble with respect to criteria and constraints. If yes, the interaction between the
DM and IDSS is complete. Otherwise, the DM formulates the remarks and
requests with respect to the criterion (criteria) and constraints of the stated
problem.
Step 3. The IDSS administrator(s) presents a new statement for the problem of
estimating values (variation laws) of economic instruments for implementation
of the economic policy in the selected direction, taking into account initial
values and requests of the DM by the criterion (criteria) and constraints of the
initial statement of problem of estimation(s) of values (variation laws) of
economic instruments for implementation of the economic policy in the selected
direction.
28 1 Elements of Parametric Control Theory of Market Economic Development
The interaction between the DM and IDSS continues until the statement for the
problem of estimation(s) of values (variation laws) of the economic instruments for
implementation of the economic policy in the selected direction is not accepted by
the DM.
Another part of the aforesaid problem, “statement and solution of the problem(s) of
estimation(s) of values (variation laws) of the economic instruments for implementa-
tion of the economic policy in the selected direction,” is solved using two stages.
Based on the interaction between the DM and IDSS, in the first stage, the
methods for solving the formulated problem (if this problem is multiobjective) of
estimation(s) of values (variation laws) of economic instruments for implementa-
tion of the economic policy in the selected direction occur within the framework of
the following procedure:
Step 1. Based on the formulated problem “statement and solution of the problem(s)
of estimation(s) of values (variation laws) of the economic instruments for
implementation of the economic policy in the selected direction” and prelimi-
nary evaluations of characteristics of the methods of decision-making support in
the following classes:
• Methods of solution search with no participation of the DM;
• Methods using the DM’s preferences for constructing the rules of choice of a
unique solution or a small number of Pareto-effective ones;
• Interactive procedures to solving the problem with participation of the DM;
• Methods based on the approximation of the Pareto bound and informing the
DM about this in one or another form. Furthermore, the DM indicates the
most preferable criterion point on the Pareto bound. The preferable solution is
derived for this criterion point.
The methods from the first two classes are based on the construction of a solving
rule, i.e., a rule for finding one or several solutions from an admissible set. The main
difference between the first and second classes is that for the first class, the solving
rule is constructed with no participation from the DM. For the second class,
information about the DM’s preferences is used.
Step 2. Evaluating proposed methods for solving the problem “statement and
solution of the problem(s) of estimation(s) of values (variation laws) of eco-
nomic instruments for implementation of economic policy in the selected direc-
tion,” the DM determines whether or not it is admissible. In it is admissible, the
interaction between the DM and IDSS is complete. Otherwise, the DM suggests
choosing a method for solving the formulated problem “statement and solution
of the problem(s) of estimation(s) of values (variation laws) of economic
instruments for implementation of the economic policy in the selected
direction.”
Step 3. The IDSS administrator(s) illustrates the application of methods proposed
by the DM for solving the formulated problem and evaluating its characteristics.
The interaction between the DM and IDSS continues within the framework of
steps 1–3 until the method for solving the formulated problem “statement and
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 29
solution of the problem(s) of estimation(s) of values (variation laws) of economic
instruments for implementation of the economic policy in the selected direction” is
not accepted by the DM.
The second stage for solution of the considered problem is concerned with
solving the formulated problem “solution of the problem(s) of estimation(s) of
values (variation laws) of economic instruments for implementation of the eco-
nomic policy in the selected direction” based on the selected method with partici-
pation of the DM. We’ll illustrate it with the following example.
Example: In the first stage of solving the problem(s) of estimation(s) of values
(variation laws) of economic instruments for implementation of the economic
policy in the selected direction, based on the interaction between the DM and
IDSS, let the methods be selected that use the DM’s preferences for construction
of the rule for choosing a unique solution or a small number of the Pareto-effective
solutions of the multiobjective problem defined here. Let the hierarchal analysis
method [25] be selected as one of these methods.
The interaction between the IDSS administrator(s) and the DM while solving the
“problem(s) of estimation(s) of values (variation laws) of economic instruments for
implementation of the economic policy in the selected direction” according to the
hierarchal analysis method is implemented by the following scheme:
Step 1. For certain values of the exogenous parameters, the IDSS administrator(s)
finds the estimation of the Pareto set P of the considered multiobjective problem
using the method stated in the appendix.
Step 2. For the set of criteria for the considered multiobjective problem, the DM
defines the indicators of relative significance aij:ði; j ¼ 1; :::;m; j>iÞ for all pairsof criteria ðKi;KjÞ of the considered multiobjective problem.
Step 3. The IDSS administrator(s) computes the weight Wi of the linear utility
function UðK1;K2; :::;KmÞ ¼Pmi¼1
WiKi by the method of hierarchal analysis.
Step 4. In the constructed set P, the IDSS administrator(s) determines a point (set of
points) P+ maximizing function U.Step 5. Evaluating the presented point (set of points) P+ and the respective
estimations of values (variation laws) of economic instruments for implementa-
tion of the economic policy in the selected direction, the DM determines whether
it is admissible (Pareto-effective solutions). If it is not, then the DM formulates a
new set of indicators of relative significance of the criterion pair aij.Steps 2–5 are repeated until the version agreed upon with the DM’s Pareto-
effective solutions is accepted.
Step 6.When the set P+ contains more than one point, the DM selects a unique point
in P+ and respective estimations of values (variation laws) of economic instru-
ments for implementation of the economic policy in the selected direction.
We shall illustrate these steps on the basis of the following computational
algorithm for estimation of the Pareto-effective solution of a multiobjective
30 1 Elements of Parametric Control Theory of Market Economic Development
optimization problem based on parametric control (considering an example of a
computable model with the knowledge sector and a model of a small open
economy).
The combined model, including the computable model of general equilibrium
with the knowledge sector and an econometric model of a small open economy, is
constructed in the following way.
The following two combined models are considered:
Model 1 is a variant of the computable model of general equilibrium with the
knowledge sector. The parameters in this model are determined from known statis-
tical data from the Republic of Kazakhstan for the years 2000–2010 and solution of
the respective parametric identification problem. This model incorporates the
constraints on possible admissible values of its endogenous parameters.
Model 2 is a variant of the model of a small open economy with the coefficients
evaluated on the basis of known statistical data from the Republic of Kazakhstan for
the years 2000–2010 and values of the exogenous economic indicatorsPZ, iZ, and ее andeconomic instruments M and G forecasted in 2011. This model incorporates the
constraints on the possible equilibrium values of its endogenous parameters Y, P, i,and N.
As for the adjusted parameters’ vector, we consider vector p ¼ ðp1; p2; p3Þ wherepi; i ¼ 1; 2; 3, are values of additional investments into budget of 2011 in sectors
1, 2, and 3 of model 1, respectively. Also, the respective valueDG ¼ p1 þ p2 þ p3 isadded to the real volume of public expenses G of model 2.
Let’s define the following three criteria for the two- and three-criteria problems
of parametric control.
1. The calculated normalized value for the GDP of the country from 2011 to 2015
for model 1 is
K1 ¼ YðpÞY0
! max:
2. The calculated normalized value of welfare export in 2011 for model 2 is
K2 ¼ QexðpÞ=Qex0 ! max:
3. The calculated normalized value of the consumption of imported welfare in
2011 (with the negative sign) for model 2 is
K3 ¼ �QimðpÞ=Qim0 ! max:
Here Y0;Qex0; andQim0 are some base values of the respective variables.
Let’s formulate two multiobjective problems of parametric control on the basis of a
combined model.
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 31
Problem 1 Find the estimate of the Pareto set on the plane ðK1;K2Þ on the basis ofmodels 1 and 2, criteria K1 and K2, under the following constraints on adjusted
parameters:
p1 0; p2 0; p3 0; p1 þ p2 þ p3 � pm;DG � pm: (1.46)
Here, pm is the maximal value of additional investments from the public budget to
the producing sectors of the economy.
Problem 2 Find the estimate of the Pareto set in space ðK1;K2;K3Þ on the basis ofmodels 1 and 2, criteria K1, K2, and K3, under constraints (1.46) on the adjusted
parameters.
The algorithm to find the estimate of the Pareto sets indicated in the above
problems with the use of methods of concessions is as follows [25]:
Algorithm of Solution to Problem 1 Let the calculated ideal point of values of the
criteria of the stated problem have coordinates ðK1 ;K
2Þ; where K*
1, K*2 are the
maximal values of the criteria for models 1 and 2 for solutions of the respective
monoobjective problems under constraints (1.46). With the attainment of value K1
of the first model criterion, let the value of the second model criterion be equal to
K02 ; whereK
02<K
2 (in caseK02 ¼ K
2 the two-criteria problem is solved). Let’s divide
the numerical interval ½K02;K
2 � into m equal parts with step h ¼ ðK
2 � K02Þ=m .
We obtain the set of points Ki2 ¼ K0
2 þ ih; i ¼ 0; :::;m; where Km2 ¼ K
2 .
For each i ¼ 0; :::;m� 1 , we’ll solve the following auxiliary optimization
problem. Find the maximum of the criterion K1 ¼ K1ðpÞ under constraints (1.46)and K2ðpÞ Ki
2. As a result, we obtain optimum values of the first model criterion
Ki1; i ¼ 0; :::;m; and the set of mþ 1 points
Q ¼ ðKi1; K
i2Þ; i ¼ 0; :::;mg�
, which is
an estimate for the unknown Pareto set.
Algorithm of Solution of Problem 2 Let the calculated ideal point of values of the
criteria of the said problem have the coordinates ðK1 ;K
2;K
3Þ; where K
1 ;K2 ;K
3 are
the maximal values of the criteria for models 1 ðK1Þ and 2 ðK
2;K3Þ for solutions of
the respective monoobjective problems under constraints (1.46). With the attain-
ment of the criterion valueK1 ¼ K1, let the value of the criterionK3 ¼ K01
3 . With the
attainment of the criterion value K2 ¼ K2 , let the value of criterion K3 ¼ K02
3 ;
K03 ¼ minðK01
3 ;K023 Þ andK0
3<K3 (whenK
03 ¼ K
3, the three-criteria problem reduces
to a two-criteria problem).
Let’s divide the numerical interval ½K03;K
3 � into m equal parts with step
h ¼ ðK3 � K0
3Þ=m. We obtain the set of points Ki3 ¼ K0
3 þ ih; i ¼ 0; :::;m; where
Km3 ¼ K
3.
For each i ¼ 0; :::;m� 1, let’s solve the following auxiliary two-criteria optimi-
zation problems. Find an estimate of the Pareto set based on the combined model
32 1 Elements of Parametric Control Theory of Market Economic Development
for the problem with maximization of criteria K1 and K2 under constraints
(1.46) and
K3ðpÞ Ki3:
The solution to this problem is found by the aforesaid algorithm for estimating
the Pareto set for the two-criteria problem. As a result, for each fixed I, we obtain
the set of spatial points Kij1 ;K
ij2 ; ;K
i3g
n, where j ¼ 0; :::;m:
Finally, the estimation of the unknown Pareto set will consist of all pointsQ¼ Kij1 ;K
ij2 ;K
i3; i; j ¼ 0; :::;m
n odetermined before.
Let the DM’s preferences of the m-criteria problem be defined by the linear
utility function as
UðK1;K2; :::;KmÞ ¼Xmi¼1
WiKi; (1.47)
where Ki are the criteria and Wi>0 are the weights whose values are not known
beforehand.
It is assumed thatPmi¼1
Wi ¼ 1.
According to themethod of hierarchal analysis [25, p. 176], within the framework
of determining the values of the weightsWi, the DM asks to present the indicators ofrelative significance aijði; j ¼ 1; :::;m; j>iÞ for all pairs of criteria ðKi;KjÞ instead ofthe immediate values of the weights. The values of aij are selected from some scale,
e.g., {1,2,. . .,10}, and the number 10 means that the ith criterion is more significant
than the jth criterion. The number 1 means that these criteria are approximately
equivalent with respect to significance assessment.
With absolute logical answers, the DM must receive values
aij ¼ Wi
Wjði; j ¼ 1; :::;m; j>iÞ: (1.48)
It should be noted that for the two-criteria problem (m¼ 2), the values of weights
Wi are uniquely defined by values of the relative significance indicators aij. In this
case, the DM determines the valuea12 ¼ W1
W2. By adding conditionW1 þ W2 ¼ 1, we
obtain the system of two equations with two unknowns, which yields
W2 ¼ 11þa12
;W2 ¼ 1�W1:
According to the hierarchal analysis method, for m > 2, we define a matrix A(m � m) composed of indicators aij. Instead of the values of indicators determined
by the DM, in this matrix we suppose aii ¼ 1 and aij ¼ 1aji
for j<i. Let l be the
maximal eigenvalue of matrix A. Let W be the respective positive identity eigen-
vector. Then the coordinates of vector W ¼ ðW1;W2; :::;WmÞ are considered to be
the desired set of weights.
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 33
The unknown Pareto-effective set is estimated as a point (set of points) from
the estimation of the Pareto set which corresponds to the maximal value of the
utility function U.
1.4.2.4 Rules of Interaction Between the DM and IDSS at the Levelof Problem “Analysis of the Influence of Values of ExogenousSocial and Economic Indicators on the Results of the Problemof Solution and Estimation of Values (Variation Laws) ofEconomic Instruments of the Economic Policy Implementationin the Chosen Direction”
The considered problem reduces to decision making based on the interaction
between the DM and IDSS within the framework of the following procedures:
Step 1. Based on the analysis of “statement of the problem(s) of estimation(s) of
values (variation laws) of economic instruments for implementation of the
economic policy in the selected direction,” the IDSS administrator(s) informs
the DM about the planned analysis of the selected set of exogenous socioeco-
nomic indicators (one indicator) for the analysis of their influence on results
of the problem of solution and estimation of values (variation laws) of
economic instruments of the economic policy implementation in the chosen
direction.
Step 2. Evaluating the proposed analysis of the selected set of exogenous socio-
economic indicators (one indicator) for the analysis of their influence on results
of the solution of problem “analysis of influences of the values of exogenous
social and economic indicators on the results of the problem solution and
estimation of values (variation laws) of economic instruments of the economic
policy implementation in the chosen direction,” the DM determines whether the
selected set is admissible. If yes, the procedure is complete. Otherwise, the DM
formulates a suggestion on the set of the exogenous socioeconomic indicators
for the analysis of their influence on the results to the solution of problem
“analysis of influence of the values of exogenous social and economic
indicators on the results of the problem solution and estimation of values
(variation laws) of the economic instruments of the economic policy implemen-
tation in the chosen direction.”
Step 3. Based on the DM’s requests and analysis of the mathematical model of
“statement of the problem(s) of estimation(s) of values (variation laws) of
economic instruments for implementation of the economic policy in the selected
direction,” the IDSS administrator(s) proposes the new set of exogenous socio-
economic indicators for analysis of their influence on the results of solution to
problem “analysis of influences of the values of exogenous social and economic
indicators on the results of the problem of solution and estimation of values
(variation laws) of economic instruments of the economic policy implementa-
tion in the chosen direction.”
34 1 Elements of Parametric Control Theory of Market Economic Development
The interaction between the DM and IDSS continues in steps 1–3 until the set
of exogenous socioeconomic indicators for the analysis of their influence on the
results to the solution of problem “analysis of influences of the values
of exogenous social and economic indicators on the results of the problem
solution and estimation of values (variation laws) of the economic instruments
of the economic policy implementation in the chosen direction” is not accepted
by the DM.
1.4.2.5 Rules of Interaction Between the DM and IDSS at the Levelof Problem “Formulation of Recommendations on the Choiceof Values (Variation Laws) of Economic Instruments ofImplementation of the Economic Policy in the Chosen Direction”
The considered problem reduces to making a decision based on the interaction
between the DM and IDSS using the following procedure:
Step 1. Based on the estimation of values of the exogenous socioeconomic
indicators and results of the analysis of dependencies of the solutions to the
problems of estimation of values (variation laws) of economic instruments for
implementation of the economic policy in the selected direction, the IDSS
administrator(s) proposes a recommended version of values (variation laws) of
economic instruments for implementation of the economic policy in the selected
direction.
Step 2. Evaluating the recommended version of values (variation laws) of economic
instruments for implementation of the economic policy in the selected direction,
the DM determines whether or not the proposed version of values of economic
instruments for implementation of the economic policy in the selected direction
is admissible. If yes, the interaction between the DM and IDSS is complete.
Otherwise, the DM requests new information about possible values of exoge-
nous socioeconomic indicators and respective versions of the values (variation
laws) of economic instruments for implementation of the economic policy in the
selected direction.
Step 3. By request, the IDSS administrator(s) proposes new information about
possible values of exogenous socioeconomic indicators and respective versions
of the values (variation laws) of economic instruments for implementation of the
economic policy in the selected direction to the DM.
The interaction between the DM and IDSS continues within the framework of steps
1–3 until the set of values is accepted by the DM.
1.4 Algorithm of Application of Parametric Control Theory and Rules. . . 35
1.5 Examples for Application of Parametric Control Theory
1.5.1 Mathematical Model of the Neoclassical Theoryof Optimal Growth
1.5.1.1 Model Description
A mathematical model for economic growth [46] is given by the following system
of two ordinary differential equations containing the time (t) derivative:
dk
dt¼ Aka � c� ðnþ dÞk;
dc
dt¼ c
1� bðaAka�1 � ðdþ pÞÞ:
8>><>>: (1.49)
k is the ratio of capital (K) to labor (L). In this model, the country’s population and
labor force (labor) are not distinguished;
c is the mean consumption per capita;
n is the level of growth (or decrease) of population: LðtÞ ¼ L0ent;
d is the level of capital depreciation, d>0;
p is the discounting level;
e�pt is the discounting function (p>n);
A and a are parameters of the production function y ¼ ’ðkÞ ¼ Aka, where y is theratio of the gross domestic product to the labor force; that is, mean labor productiv-
ity (0<a<1; A>0);
b is a parameter of the social utility function characterizing the mean welfare of
the population:
UðcÞ ¼ Bcb ð0<b<1; B>0Þ:
The first equation in system (1.49) is the fundamental Solow equation taken from
the theory of economic growth. The second equation of this system is derived from
the maximum condition of the objective function
Z10
UðcÞLðtÞe�ptdt ¼ BL0
Z10
eb ln c�ðp�nÞtdt;
characterizing the total welfare of the entire population within time interval 0 � t<1. This function is maximized under the constraints
kð0Þ ¼ k0; k0 ¼ Aka � c� ðnþ dÞk; 0 � cðtÞ � ’ðkðtÞÞ
and constant values of parameters d, n, p, A, B, a, and b.
36 1 Elements of Parametric Control Theory of Market Economic Development
The solution to system (1.49) will be considered in some closed region Owhose
frontier is a simple closed curve belonging to the first quadrant of the phase plane
R2þ ¼ fk>0; c>0g. kð0Þ ¼ k0; cð0Þ ¼ c0; ðk0; c0Þ 2 O.
1.5.1.2 Analysis of the Structural Stability from a Mathematical Modelof the Neoclassic Theory of Optimal Growth Without ParametricControl
Estimation of robustness (structural stability) of the considered model without
parametric control in a closed region O whose boundary is a simple closed
curve belonging to the first quadrant of the phase plane R2þ ¼ fk>0; c>0g, kð0Þ
¼ k0; cð0Þ ¼ c0; ðk0; c0Þ 2 O relies on the theorem of necessary and sufficient
conditions of robustness [12]. First, let’s prove the following assertion:
Lemma 1.2 System (1.49) has the unique singular point
k ¼ aAdþ p
� � 11�a
;
c ¼ kðnþ dÞð1� aÞ þ p� n
a
� �8>>>><>>>>:
(1.50)
in R2þ. This point is a saddle point of system (1.49).
Proof Setting the right-hand side of the equations of system (1.49) to zero, we
obtain expressions (1.50). Obviously, k>0; c>0. Consider the determinant of the
Jacobian matrix for the right-hand side of (1.50) at the point (k; c):
D ¼ a� 1
ð1� bÞa ðpþ dÞððnþ dÞð1� aÞ þ p� nÞ:
Since for all stated values of the parameters A; a; b; p; n; d of the mathemati-
cal model we haveD<0, it follows that the singular pointk; c is the saddle point ofsystem (1.49).
Theorem 1.10 Let the right-hand sides of the system
x0 ¼ f1ðx; yÞ;y0 ¼ f2ðx; yÞ
((1.51)
be smooth functions in some region O1 � R2, and suppose that system (1.51) has aunique saddle singular point (x; y) in this region. Then system (1.51) is robust inthe closed region O (O � O1) containing the point (x; y).
1.5 Examples for Application of Parametric Control Theory 37
ProofLet’s make sure that system (1.51) does not have cyclic trajectories. Assume the
contrary. Let the region O1 have a cyclic trajectory. Then in its interior there exists
at least one singular point, and the sum of the Poincare indices of the singular points
within this cycle must be 1 [12, p. 117]. But in the regionO1 there is a unique saddle
point with index equal to –1. Thus, we have arrived at a contradiction.
Let’s make sure that the stable and unstable separatrices of the saddle point ðx;yÞ do not form the same trajectory in the region O1. Assume the contrary. Let the
stable and unstable separatrices of the saddle point ( x; y ) constitute the same
singular trajectory g lying in the region O1. Then this trajectory (or, if it exists, the
second trajectory composed of other stable and unstable separatrices), together with
the singular point, is the boundary of the closed cellO2 lying in the regionO1. Let’s
consider the semitrajectoryLþ coming from some point (x1; y1), where (x1; y1) is theinterior point of O2 . Then, by virtue of the absence of cyclic trajectories and the
uniqueness of the equilibrium point, the limit points of Lþ must be the boundary of
the cellO2 (the point (x1; y1) cannot be a unique limit point ofLþ since this point is a
saddle [9, p. 49]). Now, let’s consider the semitrajectory L� coming from the point
ðx1; y1) in the direction opposite Lþ. It is obvious that the boundary ofO2 cannot be
the limit points of L� . Since there are no other singular points and singular
trajectories in the region O2, we have a contradiction.
In accordance with [12, p. 146, Theorem 12], the assertion is proved.
Corollary 1.3 System (1.49) is robust in the closed region O (O � R2þ) contained
inside the point (k; c) for all fixed values of the parametersn; L0; d; p; A; a; B; bfrom the respective ranges of their definition.
In particular, it follows that there are no bifurcations of the phase-plane portrait
of system (1.49) in the regionO under variation of the given parameters within their
range of definition.
1.5.1.3 Choosing Optimal Laws of Parametric Control of MarketEconomic Development Based on a Mathematical Modelof the Neoclassical Theory of Optimal Growth
Consider now the feasibility of the realization of an efficient public policy on the
basis of model (1.49) by choosing the optimal control laws using the capital
depreciation level (d) as an example of the economic parameter.
Choosing the optimal parametric control laws is carried out in the environment
of the set of the following equations:
1:U1ðtÞ ¼ l1DkðtÞkð0Þ þ d; 2:U2ðtÞ ¼ �l2
DkðtÞkð0Þ þ d;
3:U3ðtÞ ¼ l3DcðtÞcð0Þ þ d; 4:U4ðtÞ ¼ �l4
DcðtÞcð0Þ þ d:
(1.52)
38 1 Elements of Parametric Control Theory of Market Economic Development
Here Ui is the ith law of the control of the parameter d (i ¼ 1; :::; 4);li is the adjustedcoefficient of the ith control law;li 0;d is the constant equal to the basic value ofthe parameter d ; DkðtÞ ¼ kiðtÞ � kð0Þ;DcðtÞ ¼ ciðtÞ � cð0Þ; (kiðtÞ , ciðtÞ ) is the
solution of system (1.49) with the initial conditions kið0Þ ¼ k0; cið0Þ ¼ c0 with useof the control lawUi. Use of the control lawUimeans the substitution of the function
from the right-hand side of (1.52) into system (1.49) instead of the parameter d; t ¼0 is the time of control commencement; t 2 ½0; T�.
The problem of choosing an optimal parametric control law at the level of one
economic parameter d can be formulated as follows: On the basis of mathematical
model (1.49), find the optimal parametric control law at the level of the economic
parameter d in the environment of the set of algorithms (1.52); that is, find the
optimal law from the set {Ui} that maximizes the criterion
K ¼ BL0
ZT0
eb ln ciðtÞ�ðp�nÞtdt ! maxfUi; lig
(1.53)
under the constraints
kiðtÞ � kðtÞj j � 0:09kðtÞ; ðkiðtÞ; ciðtÞÞ 2 O; where t 2 ½0; T�: (1.54)
Here ðkðtÞ; cðtÞÞ is the solution of system (1.49) without the parametric control.
The stated problem is solved in two stages:
– In the first stage, the optimal values of the coefficients li are determined for each
law Ui by the enumeration of their values on the respective intervals (quantized
with a small step) maximizing K under constraints (1.54).
– In the second stage, the law of optimal control of the parameter d is chosen on
the basis of the results of the first stage by the maximum value of the criterion K.
The considered problem was solved under the following conditions:
given parameter values a ¼ 0:5, b ¼ 0:5, A ¼ 1, B ¼ 1, k0 ¼ 4, c0 ¼ 0:8, T ¼ 3,
L0 ¼ 1;
for the following fixed values of the uncontrolled parameters: n ¼ 0:05, p ¼ 0:1;for the basic value of the controlled parameter d ¼ 0:2.The results of the numerical solution of the problem of choosing the optimal
parametric control law at the level of one economic parameter of the economic
system show that the best result K ¼ 1.95569 can be obtained with the use of the
following law:
d ¼ 0; 19DkðtÞ4
þ 0:2: (1.55)
Note that the criterion value without use of the parametric control is equal to K¼1.901038.
1.5 Examples for Application of Parametric Control Theory 39
1.5.1.4 Analysis of the Structural Stability of the Mathematical Modelof the Neoclassical Theory of Optimal Growth with ParametricControl
Let’s analyze the robustness of system (1.49), where the parameter d is given in
accordance with the solution to the parametric control problem, taking into account
the influence of variations of the uncontrolled parameters n and p by the expression
d ¼ l1k � k0k0
þ d0 (1.56)
with any fixed value of the adjusted coefficient l1>0. Here k0>0 and d0>0 are some
fixed numbers. Substitute (1.56) into the right-hand sides of the system (1.49) and
set them equal to zero. We obtain the following system with respect to the unknown
variables ðk; cÞ (other admissible values of variables and constants are fixed):
Aka � c� nþ l1k � k0k0
þ d0
� �k ¼ 0;
c
1� bðaAka�1 � l1
k � k0k0
þ d0Þ � p
� �¼ 0:
8>>><>>>:
(1.57)
Since the function from the right-hand side of the second equation of system
(1.49) is strictly decreasing as a function of variable k and takes on all values with
k > 0, then it follows that the second equation has a unique solution k . For thissolution, there exists a unique solution c of the first equation in (1.57); that is,
system (1.57) has the unique solution ðk; cÞ . If ðk; cÞ=2R2þ , then, obviously,
system (1.49) with the control law U1 is structurally stable in any closed region
O � R2þ.
Now, let ðk; cÞ 2 R2þ . Let’s find the determinant of the Jacobian of the
functions f1, f2; which are the left-hand sides of the respective equations of system
(1.57) at this point. Since
@f1@c
ðk; cÞ ¼ �1;@f2@k
ðk; cÞ ¼ c
1� bðaða� 1ÞAðkÞa�2 � l1Þ<0;
@f2@c
ðk; cÞ ¼ 0;
the determinant of the matrix isD<0. Therefore, in this case the point ðk; cÞ is thesaddle point of system (1.49) with control law U1. From Theorem 1.10, it follows
that the system is structurally stable in the closed region O � R2þ containing the
point ðk; cÞ inside.In particular, with the use of law (1.55), system (1.49) remains structurally
stable.
The methods above allow the analysis of the robustness conditions for system
(1.49) with the use of the optimal control law d ¼ �l1 c�c0c0
þ d0, when the values of
the parameters ðn; pÞ are within the closed region in R2þ.
40 1 Elements of Parametric Control Theory of Market Economic Development
1.5.1.5 Finding Bifurcation Points for the Extremals of the VariationalCalculus Problem Based on the Mathematical Modelof the Neoclassical Theory of Optimal Growth with ParametricControl
Let’s analyze the dependence of the results of choosing the parametric control law
at the level of parameter d on the uncontrolled parameters (n, p) with values in some
region (rectangle) L in the plane. In other words, let’s find possible bifurcation
points for the variational calculus problem of choosing the optimal parametric
control law of a given model of economic growth.
As a result of computational experiments, plots of dependencies of the optimal
value ofK in criterion (1.53) on the values of the parameters (n, p) were obtained foreach of four possible laws Ui. Figure 1.2 presents the plots for the laws U1 and U4,
which give the maximum values of the criterion in the region L, the intersection
curve for these surfaces, and the projection of the intersection curve onto the region
of the values of the parameters (n, p) consisting of the bifurcation points of these
parameters. This projection divides the rectangle L into two parts: In one, the
control law U1 is optimal, while in the other one, the law U4 is optimal. Along the
projection itself, both of these laws are optimal.
As a result of this analysis of the dependence of the results of the solution of the
considered variational calculus problem on the values of the uncontrolled parameters
(n, p), one can approach choosing optimal parametric control laws in the following
way: If the values of the parameters (n, p) lie to the left of the bifurcation curve in therectangleL (Fig. 1.2), then the lawU1 is recommended as the optimal law. If the values
of the parameters (n, p) lie to the right of the bifurcation curve in the rectangleL, thenthe lawU4 is recommended as the optimal law. If the values of the parameters (n, p) lieon the bifurcation curve in the rectangle L, then any of the laws U1, U4 can be
recommended as the optimal law.
Opt
imal
val
ues
of c
rite
rion
K
Fig. 1.2 Plots of optimal values of criterion K
1.5 Examples for Application of Parametric Control Theory 41
1.5.2 One-Sector Solow Model of Economic Growth
1.5.2.1 Model Description
The one-sector Solow model of economic growth is presented in the book [19].
The model is described by the system of equations (1.58), which includes one
differential equation and two algebraic equations:
LðtÞ ¼ Lð0Þent;dK
dt¼ �mKðtÞ þ rXðtÞ;
XðtÞ ¼ AKðtÞaLðtÞ1�a:
8>>><>>>:
(1.58)
Here t is the time (in months), L(t) is the number of people engaged in the economy,
K(t) is capital assets, X(t) is the gross domestic product, v is the monthly rate of
increase of the population engaged in the economy, m is the share of basic produc-
tion assets retired for a month, r is the ratio of gross investments to the gross
domestic product, A is the coefficient of neutral process improvement, and a is the
elasticity coefficient of the funds.
1.5.2.2 Estimation of the Model Parameters
In the context of the solution of the problem of preliminary estimation of the
parameters, we are required to estimate the values of the exogenous parameters n,m, r, А, and a by searching for the sense of the minimum of the criterion (sum of
squares of the discrepancies of the endogenous variables).
The parametric identification criterion is as follows:
K ¼ 1
9
Xð0Þ � Xð0ÞXð0Þ
� �2
þ Xð12Þ � Xð12ÞXð12Þ
� �2
þ Xð24Þ � Xð24ÞXð24Þ
� �2
þ
þ Xð36Þ � Xð36ÞXð36Þ
� �2
þ Xð48Þ � Xð48ÞXð48Þ
� �2
þ Kð12Þ � Kð12ÞKð12Þ
� �2
þ
þ Kð24Þ � Kð24ÞKð24Þ
� �2
þ Kð36Þ � Kð36ÞKð36Þ
� �2
þ Kð48Þ � Kð48ÞKð48Þ
� �2!
! min
(1.59)
Here XðtÞ represent data about the gross domestic product of the Republic of
Kazakhstan for the period 2001–2005, KðtÞ are the capital assets of the Republicof Kazakhstan for the period 2002–2005, andXðtÞ andKðtÞ are the calculated valuesof the variables of system (1.59).
42 1 Elements of Parametric Control Theory of Market Economic Development
In computations, we use a value of L(0) equal to 6.698 and a value of K(0) equalto 4004 (which corresponds to 2001), as well as the mean value of the exogenous
parameter v equal to 0.0017.
The relative value of the mean square deviation of the calculated values of the
endogenous variables from the respective observable values (statistical data) is
equal to 100ffiffiffiffiK
p ¼ 3.8%.
1.5.2.3 Analysis of the Structural Stability of the One-Sector Solow Modelof Economic Growth Without Parametric Control
By applying a numerical algorithm of the estimation of weak structural stability of
the discrete-time dynamical system for the chosen compact set N, defined by the
inequalities 3; 000 � K � 10; 000, 5 � L � 10 in the phase space of the variables
(K, L), we discover that the chain-recurrent setRðf ;NÞ is empty. This means that the
one-sector Solow model of economic growth describing the interaction between the
benefit market and the money market is estimated as weakly structurally stable in
the compact set N.
1.5.2.4 Choosing Optimal Laws of Parametric Control of MarketEconomic Development Based on the Solow Mathematical Model
Let’s consider now the feasibility of the realization of an efficient public policy on
the basis of model (1.58) by choosing the optimal control laws using the ratio of
gross investments to gross domestic product (r) as an example of an economic
parameter.
The choice of optimal parametric control laws is made in the environment of the
following relations:
1Þ rðtÞ ¼ r þ k1KðtÞ � Kð0Þ
Kð0Þ ; 2Þ rðtÞ ¼ r � k2KðtÞ � Kð0Þ
Kð0Þ ;
3Þ rðtÞ ¼ r þ k5XðtÞ � Xð0Þ
Xð0Þ ; 4Þ rðtÞ ¼ r � k6XðtÞ � Xð0Þ
Xð0Þ :
(1.60)
Here ki is the adjusted coefficient of the ith control law, and ki 0; r* is the value ofthe exogenous parameter r obtained as a result of the parametric identification of
the model.
The problem of choosing the optimal parametric control law at the level of one
of the economic parameters d can be formulated as follows: On the basis of
mathematical model (1.58), find the optimal parametric control law at the level of
the economic parameter r in the environment of the set of algorithms (1.60)
maximizing the performance criterion (mean value of the gross domestic product
on the considered time interval)
1.5 Examples for Application of Parametric Control Theory 43
K ¼ 1
49
X48t¼0
XðtÞ
under the constraints K > 0. The base value of the criterion (without application of
scenarios) is equal to 409.97.
The numerical solution of the problem of choosing the optimal parametric control
law at the level of one economic parameter of the economic system shows that the
best result, K ¼ 511.34, can be obtained with use of the following law:
rðtÞ ¼ r þ 0:268XðtÞ � Xð0Þ
Xð0Þ : (1.61)
The values of the endogenous variables of the model without using scenarios, as
well as with use of the optimal law, are presented in Figs. 1.3 and 1.4.
Years
Scenario 3 is used Without scenario
Fig. 1.3 Capital assets without parametric control and with use of law 3 optimal in the sense of
criterion K
Years
Scenario 3 is used Without scenario
Fig. 1.4 Gross domestic product without parametric control and with use of law 3 optimal in the
sense of criterion K
44 1 Elements of Parametric Control Theory of Market Economic Development
1.5.2.5 Analysis of the Structural Stability of the One-Sector Solow Modelof Economic Growth with Parametric Control
For carrying out this analysis, the expression for optimal parametric control law
(1.61) is substituted into the right-hand side of the second equation of system
(1.58) instead of parameter r. Then, by applying the numerical algorithm of
estimation of weak structural stability of the discrete-time dynamical system for
the chosen compact set N, defined by the inequalities 3,000 � K � 10,000, 5 � L�10 in the phase space of the variables (K, L), we obtain that the chain-recurrent
set R(f, N) is empty. This means that the one-sector Solow model with the
optimal parametric control law is estimated as weakly structurally stable in
the compact set N.
1.5.2.6 Analysis of the Dependence of the Optimal Value of CriterionK on the Parameter for the Variational Calculus Problem Basedon the Solow Mathematical Model
Let’s analyze the dependence of the optimal value of criterion K on the exogenous
parameter m, the share of the basic production assets retired for a month for
parametric control laws (1.60) with the found optimal values of the adjusted
coefficients ki . Plots of the dependencies of the optimal value of criterion K were
obtained from computational experiments (see Fig. 1.5). Analysis of the presented
plots shows that there are no bifurcation points of the extremals for the given
problem for the analyzed interval of values of the exogenous parameter m.
Opt
imal
val
ue o
f cr
iter
ion
Scenario 1 is used Scenario 3 is used Without parametric control
Fig. 1.5 Plots of the dependencies of the optimal value of criterion K on the exogenous
parameter m
1.5 Examples for Application of Parametric Control Theory 45
1.5.3 Richardson Model for the Estimation of Defense Costs
1.5.3.1 Model Description
The model is described by a system of two linear differential equations with
constant coefficients [20]
dx=dt ¼ ay� mxþ r;
dy=dt ¼ bx� nyþ s:
((1.62)
Here t is the time (in months), x(t) is the defense costs of the first country (group of
countries), y(t) is the defense costs of the second country (group of countries), a is thescale of threat for the first country (group of countries), b is the scale of threat for thesecond country (group of countries),m is the armament costs of the first country (group
of countries), n is the armament costs of the second country (group of countries), r isthe scale of the past damage suffered by the first country (group of countries), and s isthe scale of the past damage suffered by the second country(group of countries).
1.5.3.2 Estimation of Model Parameters
In the context of the solution of the problem of the preliminary estimation of the
parameters, we are required to estimate the values of the exogenous parameters a, b,m, n, r, and s by the searching method in a sense of the minimum of the criterion
(sum of the squares of the discrepancies of the endogenous variables).
The parametric identification criterion is as follows:
K ¼ 1
8
xð1Þ � xð1Þxð1Þ
� �2
þ xð2Þ � xð2Þxð2Þ
� �2
þ xð3Þ � xð3Þxð3Þ
� �2
þ
þ xð4Þ � xð4Þxð4Þ
� �2
þ yð1Þ � yð1Þyð1Þ
� �2
þ yð2Þ � yð2Þyð2Þ
� �2
þ
þ yð3Þ � yð3Þyð3Þ
� �2
þ yð4Þ � yð4Þyð4Þ
� �2!
! min :
(1.63)
Here x*(t) represents statistical data on the armament costs of France and Russia for
the years 1910–1913; y*(t) is statistical data about the armament costs of Germany
and the Dual Monarchy (Austria–Hungary) for the same years; and x(t), y(t) are therespective calculated values of the endogenous variables of system (1.62). The
statistical data (in millions of pounds sterling) are presented in Table 1.1.
The problem of preliminary estimation is solved by the Gauss–Seidel method
with the discrete divisor of the estimation range equal to 100,000. The number of
iterations of the algorithm is 50. To improve the result of parameter estimation, a
series of 1,000 experiments on random settings of the initial values of the estimated
exogenous parameters from the ranges of their estimation was conducted.
46 1 Elements of Parametric Control Theory of Market Economic Development
As a result of solving the problem of the preliminary estimation of the
parameters, the following values were obtained: a ¼ 0.4846, b ¼ 0.3498, m ¼0.2526, n ¼ 0.4390, r ¼ 0.3387, s ¼ –0.3386.
The relative value of the mean square deviation of the calculated values of the
endogenous variables from the corresponding observable values (100ffiffiffiffiK
p) is 3.2819%.
1.5.3.3 Analysis of the Structural Stability of the RichardsonMathematical Model Without Parametric Control
For obtained values of the parameters of system (1.62), its stationary point has the
coordinates x0 ¼ 0:2625; y0 ¼ � 0:5273ð Þ and does not lie in the first quadrant of
the phase plane R2þ ¼ fx>0; y>0g . Therefore, system (1.62) is robust for any
closed region O � R2þ.
1.5.3.4 Choosing Optimal Laws of Parametric Control of MarketEconomies on the Basis of the Richardson Mathematical Model
Let’s consider now the feasibility of the realization of an efficient public policy on
the basis of model (1.62) by choosing the optimal control laws using the threat level
for the second group of countries, b, as an example of the parameter.
Choosing the optimal parametric control laws is carried out in the environment
of the following relations:
0Þ bðtÞ ¼ b þ k1XðtÞ � Xð0Þ
Xð0Þ ;
1Þ bðtÞ ¼ b � k2XðtÞ � Xð0Þ
Xð0Þ ;
2Þ bðtÞ ¼ b þ k3YðtÞ � Yð0Þ
Yð0Þ ;
3Þ bðtÞ ¼ b � k4YðtÞ � Yð0Þ
Yð0Þ :
(1.64)
Here ki is the coefficient of the scenario, and b* is the value of the exogenous
parameter b obtained as a result of the preliminary estimation of the parameters.
The problem of choosing the optimal parametric control law at the level of one
of the economic parameters can be formulated as follows. On the basis of mathe-
matical model (1.62), find the optimal parametric control law at the level of the
economic parameter b in the environment of the set of algorithms (1.64)
maximizing the performance criterion
Table 1.1 Statistical data on
endogenous variables of the
Richardson model
Year 1909 1910 1911 1912 1913
t 0 1 2 3 4
x* 115.3 123.4 132.8 144.4 167.4
y* 83.9 85.4 90.4 97.7 112.3
1.5 Examples for Application of Parametric Control Theory 47
K ¼ 1
T
ZT0
yðtÞdt; (1.65)
under the constraints
yðtÞ � 1:1� xðtÞ (1.66)
Here the interval of control [0, T] corresponds to the years 1909–1913.
Numerical solution of the problem of choosing the optimal parametric control
law at the level of one economic parameter of the economic system shows that the
best result, K ¼ 111.51, can be obtained with use of the following law:
bðtÞ ¼ 0:3498þ 0:3208XðtÞ � Xð0Þ
Xð0Þ (1.67)
Note that the basic value of the criterion (without control) is equal toK¼ 96.8722.
The values of the endogenous variables of the model without the parametric
control, as well as with use of the parametric control, are presented in Figs. 1.6 and 1.7.
1.5.3.5 Analysis of Structural Stability of the Richardson MathematicalModel with Parametric Control
For carrying out this analysis, the expression for the optimal parametric control law
(1.67) is substituted into the right-hand side of the second equation of system (1.62)
instead of the parameter b. Then, by applying the numerical algorithm of the estima-
tion of the weak structural stability of the discrete-time dynamical system for the
chosen compact set N defined by the inequalities 100 � X � 150, 80 � Y � 120 in
the phase space of the variables (K, L), we find that the chain-recurrent set Rðf ;NÞ isempty. This means that the Richardson mathematical model with the optimal
parametric control law is estimated as weakly structurally stable in the compact setN.
Years
Fig. 1.6 Armament costs of the first group of countries without parametric control and with use of
the optimal law of parametric control. without parametric control, – law 0 is used
48 1 Elements of Parametric Control Theory of Market Economic Development
1.5.3.6 Analysis of the Dependence of the Optimal Value of CriterionK on the Parameter for the Variational Calculus ProblemBased on the Richardson Mathematical Model
Let’s analyze the dependence of the optimal value of the criterion K on the
exogenous parameter a, the threat level for the first group of countries for
parametric control laws (1.64) with the obtained optimal values of the adjusted
coefficientski . From computational experiments, the plots of dependencies of the
optimal value of the criterion K were obtained (see Fig. 1.8). Analysis of these plots
Law 1 is used
Opt
imal
val
ue o
f cr
iter
ion
Law 2 is usedLaw 3 is usedLaw 4 is used
Fig. 1.8 Plots of dependencies of the optimal value of criterion K on the exogenous parameter a
Years
Fig. 1.7 Armament costs of the second group of countries without parametric control and with use
of optimal law of parametric control. without parametric control, – law 0 is used
1.5 Examples for Application of Parametric Control Theory 49
shows that there are no bifurcation points of the extremals of the problem for the
analyzed interval of the values of the exogenous parameter a. There are bifurcationpoints of the extremals in this case for the values a ¼ 0.315 and a ¼ 0.345.
1.5.4 Mathematical Model of a National Economic SystemSubject to the Influence of the Share of Public Expenseand the Interest Rate of Government Loans on EconomicGrowth
1.5.4.1 Model Description
The mathematical model of a national economic system for analysis of the influ-
ence of the ratio of public expense to the gross domestic product and the influence
of interest rate on the rate of government loans on economic growth proposed in
[34], after appropriate transformation, is given by
dM
dt¼ FI
pb� mM; (1.68)
dQ
dt¼ Mf � F
p; (1.69)
dLG
dt¼ rGL
G þ FG � npF� nLsRL � nOðdP þ dBÞ; (1.70)
dp
dt¼ �a
Q
Mp; (1.71)
ds
dt¼ s
Dmax 0;
Rd � RS
RS
�;RL ¼ minfRd;RSg; (1.72)
Lp ¼ 1� xx
LG; (1.73)
dp ¼ 1� xx
br2LG; (1.74)
dB ¼ br2LG; (1.75)
x ¼ n1� d
1� snp
� �1�dd
!; (1.76)
50 1 Elements of Parametric Control Theory of Market Economic Development
Rd ¼ Mx; (1.77)
f ¼ 1� 1� 1� dn
x
� � 11�d
; (1.78)
F0 ¼ �0pMf ; (1.79)
FG ¼ ppMf ; (1.80)
FL ¼ ð1� nLÞsRd; (1.81)
FI ¼ 1� xxþ ð1� xÞnp
�ð1� npÞFG � �n0ðdB þ dPÞ þ npF0
� �nL � ð1� nLÞnp sRL�þ ðm þ rGÞLp;
(1.82)
F ¼ F0 þ FG þ FL þ FI; (1.83)
RS ¼ PA0 expðlptÞ
1
1þ no; o ¼ FL
pP0 expðlptÞ : (1.84)
Here М is the total productive capacity;
Q is the total stock-in-trade in the market with respect to some equilibrium state;
LG is the total public debt;
p is the level of prices;
s is the rate of wages;Lp is the indebtedness of production;dp and dB are the business and bank dividends, respectively;
RS and Rd are the supply and demand of the labor force;
D and v are the parameters of the function f(x),x is the solution to the equation f 0ðxÞ ¼ s
p ;
ФL and ФО are consumer expenditures of workers and owners, respectively;
ФI is the flow of investment;
ФG is the expenditure on consumers by the state;
x is the norm of reservation;
b is the ratio of the arithmetic mean return from business activity and the rate of
return of rentiers;
r2 is the deposit interest rate;rG is the interest rate of public bonds;
�0 is the coefficient of the propensity of owners to consume;
p is the share of consumer expenditure by the state in the gross domestic product;
1.5 Examples for Application of Parametric Control Theory 51
np, nО, and nL are payment flow, dividends, and income taxes of workers,
respectively;
b is the norm of fund capacity of the unit of power;
m is the coefficient of the power unit retirement as a result of degradation;
m* is the depreciation rate;
a is the time constant;
D is the time constant defining the typical time scale of the wage relaxation process;
P0 and P0A are the initial number of workers and total number capable, respectively;
lp > 0 is the set rate of population growth;
o is per capita consumption in the group of workers.
The equation and relations from mathematical model (1.68), (1.69), (1.70),
(1.71), (1.72), (1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79), (1.80), (1.81),
(1.82), (1.83), and (1.84) correspond to the respective expressions from [33]
possibly after some simple transformations. Thus, the differential equation (1.68)
results from Ref. [33]’s (3.2.18), (3.2.6); (1.69) results from (3.2.19) and (3.2.8);
(1.70) is derived from (3.2.26) by substituting the expression for (FGК – НG) from
(3.2.25); (1.71) represents (3.2.9); (1.72) represents (3.2.30); expression (1.73)
represents the expression from page 150 [33]; expressions (1.74) and (1.75) repre-
sent expressions from (3.2.39); expression (1.76) represents the solution of (3.2.10)
f 0ðxÞ ¼ sp , where function (1.78) is defined on page 157 of [33]; expression (1.77)
represents one of expressions (3.2.10); relation (1.79) is derived from (3.2.15) and
(3.2.8); relation (1.80) is derived from (3.2.16) and (3.12); relation (1.81) is derived
from (3.2.22); expression (1.82) represents (3.2.36); expression (1.83) is (3.2.11);
expressions (1.84) are derived from (3.2.12), (3.2.13), and (3.2.14).
The model parameters and the initial conditions for differential equations (1.68),
(1.69), (1.70), (1.71), (1.72), (1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79),
(1.80), (1.81), (1.82), (1.83), and (1.84) are obtained on the basis of the economic
data of the Republic of Kazakhstan for the years 1996–2000 [37] (r2 ¼ 0.12; rG ¼0.12; b ¼ 2; np ¼ 0.08; nL ¼ 0.12; s ¼ 0.1; nО ¼ 0.5; m ¼ m* ¼ 0.012; D ¼ 1) or
estimated by solving the parametric control problem [x ¼ 0.1136; p ¼ 0.1348; d ¼0.3; n ¼ 34; �О ¼ 0.05; b ¼ 3.08; a ¼ 0.008; Q(0) ¼ � 125,000].
As illustrated in Table 1.2, presenting the results of parametric identification, the
relative value of the mean square deviation of the calculated values of variables
from the respective observed values is less than 5%.
In Table 1.2: М*, М**, p*, p** are the respective values of the total productive
capacity and the product price, both measured and model (calculated) ones.
Table 1.2 Parametric
identification resultsYear M* M** p* p**
1998 144,438 158,576 1.071 1.09
1999 168,037 183,162 1.160 1.20
2000 216,658 212,190 1.310 1.29
52 1 Elements of Parametric Control Theory of Market Economic Development
1.5.4.2 Analysis of the Structural Stability of the Mathematical Modelof the National Economic System Subject to the Influence of the Shareof Public Expenses and Interest Rate of Government Loans WithoutParametric Control
Let’s analyze the robustness (structural stability) of model (1.26), (1.27), (1.28),
(1.29), (1.30), (1.31), (1.32), (1.33), (1.34), (1.35), (1.36), (1.37), (1.38), (1.39),
(1.40), (1.41), and (1.42) on the basis of the theorem establishing the sufficient
conditions of structural stability [67] within a compact region of the phase space.
Assertion 1.1 Let N be a compact set lying in the region ðM>0; Q<0; p>0Þ orðM>0; Q>0; p>0Þ of the phase space of the system of differential equationsderived from (1.26), (1.27), (1.28), (1.29), (1.30), (1.31), (1.32), (1.33), (1.34),(1.35), (1.36), (1.37), (1.38), (1.39), (1.40), (1.41), and (1.42), that is, the four-dimensional space of variables ðM; Q; p; LGÞ . Let the closure of the interior ofN coincide with N. Then the flux f defined by (1.26), (1.27), (1.28), (1.29), (1.30),(1.31), (1.32), (1.33), (1.34), (1.35), (1.36), (1.37), (1.38), (1.39), (1.40), (1.41), and(1.42) is weakly structurally stable on N.
One can choose N as, for instance, the parallelepiped with boundariesM ¼ Mmin;M ¼ Mmax; Q ¼ Qmin; Q ¼ Qmax; p ¼ pmin; p¼pmax; LG ¼ LGmin; LG ¼ LGmax.Here 0<Mmin<Mmax , Qmin<Qmax<0 or 0<Qmin<Qmax , 0<pmin<pmax , LGmin<
LGmax.
Proof We’ll first prove that semitrajectory of the flux f starting from any point of
the set N with some value t (t > 0) leaves N.Consider any semitrajectory starting in N. With t>0, the following two cases are
possible; namely, all the points of the semitrajectory remain in N, or for some t the
point of the semitrajectory does not belong to N. In the first case, from (1.29) dpdt
¼ �a QM p of the system, it follows that for all t>0; the variable p(t) has a derivative
greater than some positive constant with Q<0 or less than some negative constant
withQ>0; that is, p(t) increases infinitely or converges to zero for increasing valuesof t. Therefore, the first case is impossible, and the orbit of any point in N leaves N.
Since any chain-recurrent set Rðf ;NÞ lying within N is an invariant set of this
flux, it follows that when it is nonempty, it consists of only whole orbits. Hence, in
the considered case Rðf ;NÞ is empty. The assertion follows from Theorem A [67].
1.5.4.3 Choosing the Optimal Laws of Parametric Control of MarketEconomic Development on the Basis of the Mathematical Modelof a Country Subject to the Influence of the Share of PublicExpenses and the Interest Rate of Government Loans
Let’s now consider the ability of the realization of efficient public policy by
choosing the optimal control laws using the following parameters: the share of
consumers’ public expenses in the gross domestic product p, the interest rate of thegovernment loans rG, and the norm of reservation x.
1.5 Examples for Application of Parametric Control Theory 53
Evaluate the ability of choosing the optimal laws of parametric control in the
following order:
– Choosing the optimal control law at the level of one of the economic parameters
(x, p, rG)– Choosing the optimal pair of parametric control laws from the set of
combinations of two economic parameters out of three
– Choosing the optimal set of three parametric control laws for three economic
parameters
Choosing the optimal parametric control laws is carried out in the environment
of the following relations:
1:U1jðtÞ ¼ þk1jDMðtÞMðt0Þ þ constj;
2:U2jðtÞ ¼ �k2jDMðtÞMðt0Þ þ constj;
3:U3jðtÞ ¼ þk3jDpðtÞpðt0Þ þ constj;
4:U4jðtÞ ¼ �k4jDpðtÞpðt0Þ þ constj;
5:U5jðtÞ ¼ þk5jDMðtÞMðt0Þ þ
DpðtÞpðt0Þ
� �þ const
j;
6:U6jðtÞ ¼ �k6jDMðtÞMðt0Þ þ
DpðtÞpðt0Þ
� �þ const
j j:
(1.85)
Here Uij is the ith control law of the jth parameter; the case i ¼ 1; :::; 6; j ¼ 1; 2; 3,j¼ 1 corresponds to the parameter x; the case j¼ 2 corresponds to the parameter p;the case j ¼ 3 corresponds to the parameter rG; kij is the nonnegative adjusted
coefficient of the ith control law of the jth parameter; constj is a constant equal to
the estimation of the values of the jth parameter as a result of parametric
identification.
The problem of choosing the optimal parametric control law at the level of one
of the economic parameters (x, p, rG) can be formulated as follows:
On the basis of the mathematical model (1.68), (1.69), (1.70), (1.71), (1.72),
(1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and
(1.84), find the optimal parametric control law Uij in the environment of the set of
algorithms (1.75) minimizing the criterion
K ¼ 1
T
Zt0þT
t0
pðtÞdt ! minfUij;kijg
(1.86)
under the constraints
54 1 Elements of Parametric Control Theory of Market Economic Development
MðtÞ �MðtÞj j � 0; 09MðtÞ; ðMðtÞ; QðtÞ; LGðtÞ; pðtÞ; sðtÞÞ 2 X ;
0 � Uij � aj; i ¼ 1; :::; 4; j ¼ 1; 2; t 2 ½t0; t0 þ T�; (1.87)
where М**(t) is the value of the total production capacity without parametric
control, aj is the maximum possible value of the jth parameter, and X is the compact
set of possible values of the system variables.
The stated problem is solved in two stages:
– First, the optimal values of the coefficients kij are determined for each law Uij by
enumerating their values on the intervals ½0; kmij Þ quantized with a step equal to
0.01, minimizing K under constraints (1.87). Here kmij is the first value of the
coefficient violating (1.87).
– Second, the law of the optimal control of the specific parameter (out of three) is
chosen on the basis of the results of the first stage by the minimum value of the
criterion K (1.86).
The results of the numerical solution of the first stage of the stated problem for
{Uij} are presented in Table 1.3.
The analysis of Table 1.3 in accordance with the requirements of the second
stage of the stated problem solution makes it possible to propose at the level of one-
parameter control of the market economy the following law for the parameter p:
p ¼ �0:84Dp1
þ 0:1348;
Table 1.3 First stage of
the numerical solution of
the stated problem of
choosing the optimal law
of parametric control
Notation for
parametric
control laws
Optimal value
of coefficient
of law
Value of
criterion K
U11 0.220 1.0980
U21 0.000 1.1734
U31 0.156 1.0370
U41 0.000 1.1734
U51 0.160 1.0900
U61 0.000 1.1734
U12 0.000 1.1734
U22 0.110 1.0900
U32 0.000 1.1734
U42 0.840 1.0230
U52 0.000 1.1734
U62 0.080 1.0840
U13 0.000 1.1734
U23 0.290 1.1700
U33 0.000 1.1734
U43 0.390 1.1701
U53 0.000 1.1734
U63 0.230 1.1702
1.5 Examples for Application of Parametric Control Theory 55
which provides the minimum value of K ¼ 1.023 among all the laws Uij.
The problem of choosing the pair of optimal parametric control laws for simul-
taneous control of three parameters can be formulated as follows: Find the optimal
pair of parametric control laws (Uij, Uum) in the set of combinations of two
economic parameters out of three on the basis of the set of algorithms (1.85)
minimizing the criterion
K ¼ 1
T
Zt0þT
t0
pðtÞdt ! minðUij;kijÞ;ðUum;kumÞf g
;
i; v ¼ 1; :::; 6; j; m ¼ 1; 2; 3; j<m
(1.88)
under constraints (1.87).
The problem of choosing the optimal pair of laws is solved in two stages:
– In the first stage, the optimal values of the coefficients (kij, kum) are determined
for the chosen pair of control laws (Uij, Uum) by enumeration of their values from
the respective intervals quantized with the step equal to 0.01 minimizing K under
constraints (1.87).
– In the second stage, the optimal pair of parametric control laws is chosen
on the basis of the results of the first stage by the minimum value of the
criterion K.
The results of the numerical solution of the first stage of the stated problem of
choosing the optimal pair of parametric control laws are summarized in 18 tables
similar to Table 1.4, differing in the control law expression by at least one
parameter.
Choosing the optimal pair of parametric control laws according to the
requirements of the second stage, based on analysis of the content of 18 tables,
makes it possible to recommend the implementation of the control laws for the
Table 1.4 First-stage results of the numerical solution of the stated problem of choosing the
optimal pair of laws
Pairs of parametric control laws
Criterion
value
First law Second law
Law
denotation
Optimal
coefficient value
Law
denotation
Optimal
coefficient value
U21 0.000 U12 0.000 1.1734
U21 0.185 U22 0.123 0.9810
U21 0.000 U32 0.000 1.1734
U21 0.000 U42 0.840 1.0230
U21 0.000 U52 0.000 1.1734
U21 0.167 U62 0.167 0.9820
56 1 Elements of Parametric Control Theory of Market Economic Development
parameters (p, x) for the case of two-parameter control of the market economic
mechanism as follows:
x ¼ �0:185DMðtÞ139345
þ 0:1136; p ¼ �0:123DMðtÞ139345
þ 0:1348;
which provides the minimal value K ¼ 0.981 among all the pairs (Uij, Uum).
The problem of choosing the optimal set of three laws for simultaneous control
of the three parameters can be formulated as follows: Find the optimal set of
three-parameter control laws at the level of three parameters on the basis of the
set of algorithms (1.85) minimizing the criterion
K ¼ 1
T
Zt0þT
t0
pðtÞdt ! minðUi1;ki1Þ;ðUn2;kn2Þ;ðUg3;kg3Þf g;
i; n; g ¼ 1; :::; 6
(1.89)
under constraints (1.87).
This problem is solved in two stages:
– First, the optimal values of the coefficients are determined for the chosen set of
three control laws (Ui1;Un2;Ug3) by enumeration of their values from respective
intervals (quantized with the step equal to 0.01 for each coefficient) minimizing
K under constraints (1.87).
– Second, the optimal set of three parametric control laws is chosen on the basis of
the results of the first stage by the minimum value of the criterion K.
The results of the numerical solution of the first stage of the problem are featured
in 36 tables similar to Table 1.5 differing in the control law expression by at least
one parameter.
Table 1.5 First-stage results of the numerical solution of the stated problem of choosing optimal
set of three laws
Set of three parametric control laws
Criterion
value
First law of the set Second law of the set Third law of the set
Law
denotation
Optimal
coefficient
value
Law
denotation
Optimal
coefficient
value
Law
denotation
Optimal
coefficient
value
U21 0.185 U22 0.123 U13 0.00 0.981
U21 0.185 U22 0.123 U23 0.03 0.980
U21 0.185 U22 0.123 U33 0.00 0.981
U21 0.185 U22 0.123 U43 0.00 0.981
U21 0.185 U22 0.123 U53 0.00 0.981
U21 0.185 U22 0.123 U63 0.00 0.981
1.5 Examples for Application of Parametric Control Theory 57
The choice of the optimal set of three laws according to the requirements of the
second stage makes possible a recommendation for implementing the control laws
for the parameters x, p, rG:
xðtÞ ¼ �0:185DMðtÞ139345
þ 0:1136; pðtÞ ¼ �0:123DMðtÞ139345
þ 0:1348;
rGðtÞ ¼ �0:03DMðtÞ139345
þ 0:01;
providing the minimum value K ¼ 0.980 among all combinations (Ui1;Un2;Ug3).
Thus, this work shows one possible way of choosing efficient laws of parametric
control of market economy.
In addition, alternative formulations and solutions of the problem of choosing
the optimal set of laws have been considered.
Choosing optimal parametric control laws on the basis of model (1.68), (1.69),
(1.70), (1.71), (1.72), (1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79), (1.80),
(1.81), (1.82), (1.83), and (1.84) at the level of one of two parameters x (j¼ 1) and p(j ¼ 2) was carried out under the following set of assumptions:
1:U1jðtÞ ¼ k1jM �M0
M0
þ constj; 2:U2jðtÞ ¼ �k2jM �M0
M0
þ constj;
3:U3jðtÞ ¼ k3jp� p0p0
þ constj; 4:U4jðtÞ ¼ �k4jp� p0p0
þ constj:(1.90)
Here Uij is the ith control law of the jth parameter (i ¼ 1; :::; 4; j ¼ 1; 2); the case
j¼ 1 corresponds to the parameter x; the case j¼ 2 corresponds to the parameter p;kij is the adjusted coefficient of the ith control law of the jth parameter, kij 0 ;
constj is a constant equal to the estimate of the value of the jth parameter as a result
of parametric identification;M0, p0 are the initial values of the respective variables;the mention of (1.68), (1.69), (1.70), (1.71), (1.72), (1.73), (1.74), (1.75), (1.76),
(1.77), (1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and (1.84) means that the
functions Uij from (1.90) should be substituted into (1.68), (1.69), (1.70), (1.71),
(1.72), (1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79), (1.80), (1.81), (1.82),
(1.83), and (1.84) instead of parameter x or p.The problem of choosing the optimal parametric control law at the level of one
out of two economic parameters (x, p) can be formulated as follows: On the basis of
the mathematical model (1.68), (1.69), (1.70), (1.71), (1.72), (1.73), (1.74), (1.75),
(1.76), (1.77), (1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and (1.84), find the
optimal parametric control law at the level of one of two economic parameters
(x, p) in the environment of the set of algorithms (1.90); that is, find the optimal
control law from the set {Uij} and its adjusted coefficient maximizing the criterion
K ¼ 1
T
Zt0þT
t0
YðtÞdt; (1.91)
58 1 Elements of Parametric Control Theory of Market Economic Development
where Y ¼ Mf is the gross domestic product, under constraints
pijðtÞ � pðtÞ�� �� � 0:09pðtÞ; ðMðtÞ; QðtÞ; LGðtÞ; pðtÞ; sðtÞÞ 2 X ;
0 � uj � aj; i ¼ 1; :::; 4; j ¼ 1; 2; t 2 ½t0; t0 þ T�: (1.92)
Here aj is the maximum value of the jth parameter; pðtÞ is the model (calculated)
value of the price level without parametric control; pijðtÞ is the value of the price
level with the Uij th control law; X is the compact set of admissible values of the
given variables.
The problem formulated above is solved in two stages:
– First, the optimal values of the coefficients kij are determined for each law Uij by
enumerating their values on the intervals ½0; kmij Þ quantized with step size 0.01
minimizing K under constraints (1.92). Here kmij is the first value of the coefficient
violating (1.92).
– Second, the law of optimal control of the specific parameter (one of three) is
chosen on the basis of the results of the first stage by the maximum value of
criterion K (1.86).
Numerical solution of the problem of choosing the optimal law of parametric
control of a national economic system at the level of one economic parameter
shows that the best result K ¼ 177662 can be obtained by using the following
control law:
x ¼ �0:095M �M0
M0
þ 0:1136: (1.93)
Note that the criterion value without the use of parametric control isK ¼ 170784.
1.5.4.4 Parametric Control of Market Economic Development withVarying Objectives on the Basis of a Mathematical Model Subjectto the Influence of the Share of Public Expenses and the InterestRate of Governmental Loans
Let’s consider the parametric control of inflation processes in market economies on
the basis of themathematicalmodel (1.68), (1.69), (1.70), (1.71), (1.72), (1.73), (1.74),
(1.75), (1.76), (1.77), (1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and (1.84). One can
accept the level of prices as a feasible characteristic of the development of economic
processes, taking into account that for the period 1996–2000, the years included in the
research, the economy of Kazakhstan was on the rise. The level of prices can be used
as some measure of the efficiency of the production of goods and services and can be
considered as characterizing the presence of inflationary or deflationary processes.
Within the context of price-level variation, one can conditionally distinguish two
regions, namely, admissible and inadmissible regions of the price-level variation.
1.5 Examples for Application of Parametric Control Theory 59
The inadmissible region (В) of the price-level variation can be defined by the
inequalities pðtÞ � plðtÞ or pðtÞ puðtÞ , where plðtÞ is the admissible lower
bound of the price-level variation and puðtÞ is the admissible upper bound (pl(t)< pu(t), 0< t< T). Satisfying the inequalities pðtÞ � plðtÞ shows that there exists
some deflation process, whereas satisfying pðtÞ � puðtÞ indicates excessive infla-
tion. The admissible region (A) of the price-level variation can be defined by the
inequality plðtÞ<pðtÞ<puðtÞ; 0<t<T.Depending on the region, A or B, to which the price-level values belong, the
problem of choosing the optimal parametric control laws can be formulated as
the following problems:
– In region A, parametric control is not applied.
– In region B, we are interested in finding and realizing such parametric control
laws in the environment of some given set of algorithms that minimize the
criterion characterizing the transient performance under applied constraints on
the possible values of the respective indices of the economic state and control
parameters (block B).
The proposed approach is implemented as follows: First, the process of
simulating the economic system is begun based on the result of the parametric
identification problem. Regions A and B are determined as a preliminary to the
price-level values. The algorithm for computer simulation has a logical condition
determining the presence of the level of prices in one or another admissible region.
During this process, if it turns out that the value of p(t) is in region B, than block B is
switched on, solving the problem of taking the object out of inadmissible region Bto admissible region A. If the value of p(t) turns out to be in region A, the parametric
control is switched off.
Now consider the ability of implementing efficient public policy in the context
of block B by choosing the optimal control laws by the example of the following
economic parameters: the share of the state customers’ expenditure in the gross
domestic product ( p ); the interest rate of public bonds ( rG ); the norm of
reservation ( x ). These parameters are accepted for the research, taking into
consideration [40] and the analysis of the sensitivity matrix of the indices,
namely, the total production capacity (М), the volume of the public debt (LG),and the level of prices (p).
The algorithm for multiobjective control was tested for the model of an economy
of the Republic of Kazakhstan for the following bound of the price-level variation:
pl(t) ¼ 0.9 and pu(t) ¼ 1.1.
Let
dpðtÞ ¼0, if plðtÞ<pðtÞ<puðtÞ;pðtÞ � plðtÞ, if pðtÞ � plðtÞ;pðtÞ � puðtÞ; if pðtÞ puðtÞ:
8>><>>:
60 1 Elements of Parametric Control Theory of Market Economic Development
When the level of prices is in inadmissible region B, choosing the optimal
parametric control laws is carried out in the environment of the following relations
(control laws):
1: V1j ¼ �k1jdpðtÞpðt0Þ þ constj;
2: V2j ¼ � k2jt
Zt0þt
t0
dpðtÞpðt0Þ dtþ constj;
3: V3j ¼ �k3jdpðtÞpðt0Þ þ
1
t
Zt0þt
t0
dpðtÞpðt0Þdt
24
35þ constj:
(1.94)
Here the case j ¼ 1 corresponds to the parameter x; j ¼ 2 corresponds to the
parameter p; j ¼ 3 corresponds to the parameter rG; kij is the adjusted coefficient
of the ith control law of the j-parameter, kij 0; constj is a constant equal to the
estimate of the value of the jth parameter by the results of parametric identification.
Choosing the optimal laws of parametric control is carried out at the level of two
economic parameters from the set of three ðx; p; rGÞ.The problem of choosing the optimal pair of parametric control laws at the level
of two economic parameters from the triplet x; p; rGð Þ can be stated as follows:
Find the optimal pair of parametric control laws Vij;Vum� �
on the set of
combinations of two economic parameters out of three on the basis of the set of
algorithms (1.94) minimizing the performance criterion
K1 ¼Zt0þT
t0
dpðtÞ2dt ! minfVij;kijg
(1.95)
under the constraints
MðtÞ �MðtÞj j � 0:09MðtÞ; t 2 ½t0; t0 þ T�;0 � VijðtÞ � aj; 0 � VumðtÞ � am; i ¼ 1; 2; 3; m ¼ 1; 2; 3:
(1.96)
HereMðtÞ; pðtÞ are the values of the production capacity and level of prices with theuse of the parametric control, respectively; MðtÞ; pðtÞ are the values of the
production capacity and the level of prices without the parametric control, respec-
tively; aj, am are the maximum possible values of the respective control parameters.
Again, this problem is solved in two stages:
– First, the optimal values of the coefficients kij are determined for each pair of
laws Vij;Vum� �
by the enumeration of their values on the intervals from the
respective regions quantized with a sufficiently small step for each coefficient,
minimizing the value of the criterion K1 under constraints (1.96).
1.5 Examples for Application of Parametric Control Theory 61
– Second, the optimal pair of parametric control laws is chosen based on an
analysis of the results of the first stage by the minimum value of criterion K1.
The results of the numerical solution of the first and second problems allow for a
recommendation to implement the following laws of parametric control of the
parameters (p, x) for the case of the two-parametric control of the market economy
mechanism:
x ¼ �k1jdpðtÞpðt0Þ þ 0:1136; p ¼ �k1j
dpðtÞpðt0Þ þ 0:1348:
The optimal value of criterion K1 is equal to 0.0086.
Analysis of the results of computational experiments shows that the chosen and
implemented laws of parametric control of the reservation norm x and the share of
public consumers’ expenditures in the gross domestic product p ensure that the
value of the price level is taken out of the inadmissible region and into the
admissible one.
The results of computer simulation on the parametric control of the market
economy mechanisms by means of one control law and a pair of laws of parametric
control are presented in Table 1.6 and Fig. 1.9.
1.5.4.5 Analysis of the Structural Stability of the Mathematical Model of theCountry Subject to the Influence of the Share of Public Expenses andthe Interest Rate of Governmental Loans with Parametric Control
Let’s analyze the robustness of system (1.68), (1.69), (1.70), (1.71), (1.72), (1.73),
(1.74), (1.75), (1.76), (1.77), (1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and (1.84),
where the parameters x, p, and rG are determined in accordance with the solution of
the parametric control problems as expressions
1:U1j ¼ þk1jMðtÞ �Mð0Þ
Mð0Þ þ constj;
2:U2j ¼ �k2jMðtÞ �Mð0Þ
Mð0Þ þ constj;
3:U3j ¼ þk3jpðtÞ � pð0Þ
pð0Þ þ constj;
4:U4j ¼ �k4jpðtÞ � pð0Þ
pð0Þ þ constj;
5:U5j ¼ þk5jMðtÞ �Mð0Þ
Mð0Þ þ pðtÞ � pð0Þpð0Þ
� �þ const
j;
6:U6j ¼ �k6jMðtÞ �Mð0Þ
Mð0Þ þ pðtÞ � pð0Þpð0Þ
� �þ const
j;
(1.97)
62 1 Elements of Parametric Control Theory of Market Economic Development
with any values of the adjusted coefficients kij 0. Here constj is a constant equal to
the estimate of the jth parameter based on the results of parametric identification.
The application of the parametric control laws Uij i ¼ 1; :::; 6; i ¼ 1; 2; 3means substituting the respective functions for the parameters x (j ¼ 1), p (j ¼ 2),
Table 1.6 Values of price level p(t) with applied control of economic parameters
Months
Value of price
level p(t)without control
Value of price
level p(t) withcontrol of
parameter p
Value of price
level p(t) withcontrol of
parameter x
Value of price
level p(t) withcontrol of pair of
parameters (p,x)
1 1.10 1.10 1.10 1.10
2 1.11 1.11 1.11 1.11
3 1.12 1.12 1.12 1.12
4 1.12 1.12 1.12 1.12
5 1.13 1.13 1.13 1.13
6 1.14 1.14 1.14 1.14
7 1.15 1.15 1.15 1.15
8 1.16 1.15 1.16 1.15
9 1.16 1.16 1.16 1.16
10 1.17 1.17 1.17 1.17
11 1.18 1.17 1.18 1.17
12 1.19 1.18 1.18 1.18
13 1.19 1.19 1.19 1.18
14 1.20 1.19 1.20 1.18
15 1.21 1.19 1.20 1.18
16 1.22 1.20 1.21 1.19
17 1.22 1.20 1.21 1.18
18 1.23 1.20 1.21 1.18
19 1.24 1.20 1.22 1.18
20 1.24 1.20 1.22 1.18
21 1.25 1.19 1.22 1.17
22 1.26 1.19 1.22 1.16
23 1.26 1.19 1.22 1.15
24 1.27 1.18 1.22 1.14
25 1.27 1.17 1.22 1.13
26 1.28 1.16 1.22 1.12
27 1.28 1.15 1.21 1.10
28 1.29 1.14 1.21 1.09
29 1.29 1.13 1.20 1.07
30 1.30 1.12 1.20 1.05
31 1.30 1.10 1.19 1.03
32 1.31 1.08 1.18 1.01
33 1.31 1.07 1.18 0.99
34 1.31 1.05 1.17 0.97
35 1.31 1.03 1.16 0.94
36 1.32 1.01 1.15 0.92
1.5 Examples for Application of Parametric Control Theory 63
and rG (j ¼ 3) into the model equations (1.68), (1.69), (1.70), (1.71), (1.72), (1.73),
(1.74), (1.75), (1.76), (1.77), (1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and (1.84).
As a result of application of these laws, the following system is derived:
dM
dt¼ FI
pb� mM; (1.98)
dQ
dt¼ Mf � F
p; (1.99)
dLG
dt¼ Ui3L
G þ FG � npF� nLsRL � nOðdP þ dBÞ; (1.100)
dp
dt¼ �a
Q
Mp; (1.101)
ds
dt¼ s
Dmax 0;
Rd � RS
RS
�;RL ¼ minfRd;RSg; (1.102)
Lp ¼ 1� Ui1
Ui1LG; (1.103)
Months
Lev
el o
f pr
ices
Fig. 1.9 Values of price level p(t) with control of economic parameters. Notation: – values of
price level p(t) without control; – values of price level p(t) with control of parameter x; –
values of price level p(t) with control of pair of parameters ðp; xÞ
64 1 Elements of Parametric Control Theory of Market Economic Development
dp ¼ 1� Ui1
Ui1br2LG; (1.104)
dB ¼ br2LG; (1.105)
x ¼ n1� d
1� snp
� �1�dd
!; (1.106)
Rd ¼ Mx; (1.107)
f ¼ 1� 1� 1� dn
x
� � 11�d
; (1.108)
F0 ¼ �0pMf ; (1.109)
FG ¼ Ui2pMf ; (1.110)
FL ¼ ð1� nLÞsRd; (1.111)
FI ¼ 1� Ui1
Ui1 þ ð1� Ui1ðtÞÞnp�ð1� npÞFG � �n0ðdB þ dPÞ þ npF0
� �nL � ð1� nLÞnp
sRL�þ ðm þ Ui2ÞLp;
(1.112)
F ¼ F0 þ FG þ FL þ FI; (1.113)
RS ¼ PA0 expðlptÞ
1
1þ no;o ¼ FL
pP0 expðlptÞ : (1.114)
The proof of the weak structural stability of the mathematical model (1.68),
(1.69), (1.70), (1.71), (1.72), (1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79),
(1.80), (1.81), (1.82), (1.83), and (1.84), presented above and relying on (1.71),
indicates that the weak structural stability of the considered model will be preserved
with the use of each of the parametric control laws UijðtÞ in the form of the
following assertion.
Assertion 1.2 Let N be a compact set belonging to the region ðM>0; Q<0; p>0Þor ðM>0; Q>0; p>0Þ of the phase space of the system of differential equationsderived from (1.26), (1.27), (1.28), (1.29), (1.30), (1.31), (1.32), (1.33), (1.34),(1.35), (1.36), (1.37), (1.38), (1.39), (1.40), (1.41), and (1.42), that is, the four-dimensional space of variables ðM; Q; p; LGÞ. Let N coincide with the closure ofits interior. Then the flux f defined by (1.98), (1.99), (1.100), (1.101), (1.102),
(1.103), (1.104), (1.105), (1.106), (1.107), (1.108), (1.109), (1.110), (1.111),
(1.112), (1.113), and (1.114) is weakly structurally stable on N.
1.5 Examples for Application of Parametric Control Theory 65
1.5.4.6 Finding the Bifurcation Points of the Extremals of the VariationalCalculus Problem on the Basis of the Mathematical Modelof the Country Subject to the Influence of the Share of PublicExpenses and the Interest Rate of Governmental Loans
Let’s consider the ability to find the bifurcation point for the extremals of
the variational calculus problem of choosing the law of parametric control of the
market economic mechanism at the level of one economic parameter in the environ-
ment of a fixed finite set of algorithms on the basis of mathematical model (1.68),
(1.69), (1.70), (1.71), (1.72), (1.73), (1.74), (1.75), (1.76), (1.77), (1.78), (1.79),
(1.80), (1.81), (1.82), (1.83), and (1.84) of the national economic system.
The ability to choose the optimal law of parametric control at the level of one of
two parametersx (j¼ 1) and p (j¼ 2) on the time interval ½t0; t0 þ T� is considered inthe environment of the following algorithms (1.90):
1:U1jðtÞ ¼ k1jM �M0
M0
þ constj;
2:U2jðtÞ ¼ �k2jM �M0
M0
þ constj
3:U3jðtÞ ¼ k3jp� p0p0
þ constj;
4:U4jðtÞ ¼ �k4jp� p0p0
þ constj:
In the considered problem, criterion (1.91) is used (mean value of the gross
domestic products for the period of 1997–1999):
K ¼ 1
T
Zt0þT
t0
YðtÞdt;
where Y ¼ Mf .The closed set in the space of continuous vector functions of the output variables
of system (1.68), (1.69), (1.70), (1.71), (1.72), (1.73), (1.74), (1.75), (1.76), (1.77),
(1.78), (1.79), (1.80), (1.81), (1.82), (1.83), and (1.84) and regulating parametric
actions are determined by relations (1.92):
pijðtÞ � pðtÞ�� �� � 0:09pðtÞ; ðMðtÞ; QðtÞ; LGðtÞ; pðtÞ; sðtÞÞ 2 X ;
0 � uj � aj; i ¼ 1; :::; 4; j ¼ 1; 2; t 2 ½t0; t0 þ T�:
The following problems for finding the bifurcation points of the extremals of the
considered variational calculus problem were studied.
Problem 1 In this variational calculus problem, we consider its dependence on the
coefficient l ¼ r2 of the mathematical model with possible values on some interval
[a, b].
66 1 Elements of Parametric Control Theory of Market Economic Development
As a result of computer simulations, plots of the dependence of the optimal
values of criterion K on the deposit interest rate (in percentages) for the given set of
algorithms (Fig. 1.10) were obtained. As can be seen from Fig. 1.10, the conditions
of Theorem 1.8 are satisfied; for instance, for the interval [15.6; 21.6] since with
r2 ¼ 15:6 the optimal value of the criterion equal to 175,467 is attained with use of
the law U12. With r2 ¼ 21:6; the optimal value of the criterion equal to 171,309 is
attained with the use of another law U21. Using the proposed numerical algorithm
allows one to determine the bifurcation point of the extremal of the considered
problem r2 ¼ 18:0with an accuracy of up to 0.001. For this parameter, the lawsU21
and U12 are optimal, and the corresponding value of criterion K is 173,381
(monetary units per month).
Problem 2 Find the bifurcation point for the extremals of the variational calculus
problem of choosing the set of laws of parametric control of the market economic
mechanism subject to the influence of the public expenses at the level of two
economic parameters with one-parameter disturbance.
In this variational calculus problem, we consider its dependence on the coeffi-
cient l ¼ r2 of the mathematical model with possible values in some interval [a, b].As a result of computer simulations, plots a computer simulations, plots of
dependence of the optimal values of criterion K on the deposit interest rate (in
percentages) for all sets of algorithms (Fig. 1.11). As can be seen in Fig. 1.10, the
conditions of Theorem 1.8 are satisfied; for instance, for the interval [6; 9.6] since
with r2 ¼ 6 the optimal value of the criterion equal to 188803 is attained using laws
fU21;U32g. With r2 ¼ 9:6; the optimal value of the criterion equal to 190,831 is
attained with the use of other laws fU21;U12g . Using the proposed numerical
algorithm allows us to determine the bifurcation point of the extremal of the
considered problem r2 ¼ 0:075with an accuracy of up to 0.001. For this parameter,
two pairs of laws fU21;U32g and fU21;U12g are optimal, and the respective value of
criterion K is equal to 187,487 (monetary units per month).
Opt
imal
val
ues
of c
rite
rion
Fig. 1.10 Plots of dependencies of optimal criterion values on parameter of deposit interest rate
r2 . Notation: – U12 , – U32 , – U21 , – U41 , – without control
1.5 Examples for Application of Parametric Control Theory 67
Problem 3 Find the bifurcation point for the extremals of the variational calculus
problem of choosing the set of laws of parametric control of the market economic
mechanism subject to the influence of public expenses at the level of one economic
parameter with two-parameter disturbance.
In this variational calculus problem, we consider its dependence on the two-
dimensional coefficient l ¼ ðr2; nOÞ of the mathematical model with possible
values in some region (rectangle) L of the plane.
As a result of a computer simulations, plots of the dependence of the optimal
values of criterion K on the values of the parameters ðr2; nOÞ for each of 12 possiblelawsUij; i ¼ 1; :::; 6; j ¼ 1; 2, were obtained. Figure 1.12 presents the plots for thetwo lawsU21 andU41, maximizing the criterion in region L, the intersection curve ofthe respective regions, and the projection of this intersection curve to the plane of
the values l consisting of the bifurcation points of these two-dimensional
parameters. This projection divides the rectangle L into two parts. The control
law U21 is optimal in one of these parts, whereas U41 is optimal in the other part.
Both laws are optimal on the curve projection.
Problem 4 As a result of a computer simulation experiment, the plots of the
dependence of the optimal values of criterion (1.95) K1 on the values of uncon-
trolled parameters ðr2; nOÞ for each of nine possible laws (1.94) Vij; i ¼ 1; 2; 3;j ¼ 1; 2; 3 were obtained. Figure 1.13 presents these plots for the four laws (V11;V12; V21; V22), minimizing criterion K1 in region L, the intersection curves of the
respective surfaces, and the projections of these intersection curves to the plane of
the values l . This projection consists of the bifurcation points of the two-
dimensional parameter l dividing the rectangle L into two parts, inside which
only one control law is optimal. Two or three different laws are optimal on the
projection curves.
187487
168000
173000
178000
183000
188000
193000
2,4 6 9,6 13,2 16,8 20,4r2
Opt
imal
val
ues
of c
rite
rion
Laws U41 and U32
Laws U21 and U12
Laws U41 and U12
Laws U21 and U32
Fig. 1.11 Plots of the dependencies of optimal criterion values on the parameter of deposit
interest rate r2
68 1 Elements of Parametric Control Theory of Market Economic Development
Opt
imal
val
ues
of c
rite
rion
Fig. 1.12 Plots of the dependencies of optimal criterion values on the parameters of deposit
interest rate r2 and dividend tax rate nO
Opt
imal
val
ues
of c
rite
rion
Fig. 1.13 Plot of optimal values of criterion K1
1.5 Examples for Application of Parametric Control Theory 69
1.5.5 Mathematical Model of the National EconomicSystem Subject to the Influence of International Tradeand Currency Exchange on Economic Growth
1.5.5.1 Model Description
The mathematical model proposed in [33] for researching the influence of
the international trade and currency exchange on economic growth after the respec-
tive transformations can be expressed as the following system of differential
and algebraic equations (where i ¼ 1 or 2 is the number of states, and t is the
time variable):
dMi
dt¼ FI
i
pibi� miMi; (1.115)
dQi
dt¼ Mifi � Fi
pi; (1.116)
dLGidt
¼ rG iLGi þ FG
i � np iFi � nL isiRLi � nO iðdPi þ dBi Þ; (1.117)
dpidt
¼ �aiQi
Mipi; (1.118)
dsidt
¼ siDi
max 0;Rdi � RS
i
RSi
�;RL
i ¼ minfRdi ;R
Si g; (1.119)
LPi ¼ 1� xixi
LGi ; (1.120)
dPi ¼ 1� xixi
bir2 iLGi ; (1.121)
dBi ¼ bir2 iLGi ; (1.122)
xi ¼ ni1� di
1� sinipi
� �1�didi
!; (1.123)
Rdi ¼ Mixi; (1.124)
fi ¼ 1� 1� 1� dini
xi
� � 11�di
; (1.125)
FOi ¼ �0 ipiMifi; (1.126)
70 1 Elements of Parametric Control Theory of Market Economic Development
FGi ¼ pipiMifi; (1.127)
FLi ¼ ð1� nL iÞsiRd
i ; (1.128)
FIi ¼
1
1þ np i
kqiMifixi
� ð1� npiÞFGi þ n0 iðdBi þ dPi Þ þ np iFO
i þ
þ nL i þ ð1� nL iÞnp i�
siRLi þ npiðFji � yiFijÞ þ miLpi � rG iL
Pi
�; (1.129)
RSi ¼ PA
0 i expðlp itÞ1
1þ nioi;oi ¼ FL
i
ð1þ CLi yi
pjpiÞP0 iðlpitÞ
; j ¼ 3� i; (1.130)
F12 ¼CL1p2p1
1þ CL1y
p2p1
FL1 þ
CO1
p2p1
1þ CO1 y
p2p1
FO1 ; (1.131)
F21 ¼CL2p1p2
1þ CL21yp1p2
FL2 þ
CO2
p1p2
1þ CO2
1yp1p2
FO2 ; (1.132)
F1 ¼ FI1 þ FL
1 þ FO1 þ FG
1 þ F21 � yF12; (1.133)
F2 ¼ FI2 þ FL
2 þ FO2 þ FG
2 þ F12 � 1
yF21: (1.134)
Here
Мi is the total productive capacity;
Qi is the total stock-in-trade in the market with respect to some equilibrium state;
LGi is the total public debt;pi is the level of prices;si is the rate of wages;LPi is the indebtedness of production;
dPi and dBi are the business and bank dividends, respectively;
Rdi and RS
i are the demand and supply of the labor force;
di, ni are the parameters of the function fi;xi is the solution to the equation f 0i ðxiÞ ¼ si
pi;
FLi and F
Oi are the consumer expenditures of the workers and owners, respectively;
FIi is the flow of investment;
FGi is the consumer expenditure of the state;
Fij are the consumer expenses of the ith country of the product imported from the
jth country;
y is the exchange rate of the currency of the first country with respect to the
currency of the second country, y1 ¼ y; y2 ¼ 1=y;
1.5 Examples for Application of Parametric Control Theory 71
CLi ðCO
i Þ is the quantity of imported product items consumed by workers (owners) of
the ith country per domestic product item;
xi is the norm of reservation;
bi is the ratio of the arithmetic mean return from the business activity and the rate of
return of rentiers;
r2i is the deposit interest rate;rGi is the interest rate of public bonds;�Oi is the coefficient of the propensity of owners to consume;
pi is the share of consumer expenditure of the state in the gross domestic product;
nPi, nОi, nLi are the payments flow, dividends, andworkers’ income taxes, respectively;
bi is the norm of the fund capacity of the unit of power;
mi is the coefficient of the power unit retirement as a result of degradation;
m*i is the depreciation rate;
ai is the time constant;
Di is the time constant defining the typical time scale of the wage relaxation
process;
P0i;PA0iare, respectively, the initial number of workers and total number of those
capable;
oi is the per capita consumption in the group of workers;
lPi > 0 is rate of population growth;
kqi is the share of the gross domestic product of the country reserved in gold.
Among relations (1.115), (1.116), (1.117), (1.118), (1.119), (1.120), (1.121),
(1.122), (1.123), (1.124), (1.125), (1.126), (1.127), (1.128), (1.129), (1.130),
(1.131), (1.132), (1.133), and (1.134), Eqs. (1.131), (1.132), (1.133), and (1.134)
define the connection of the economic systems of two countries. Note that in the
case CL1 ¼ CL
2 ¼ CO1 ¼ CO
2 ¼ 0; there is no trade between these two countries, and
their economic systems are independent of one another.
For the purpose of analysis, the values of such parameters as bi, r2i, rGi, npi, nLi,bi, si, �0i, mi, mi
*, Di were taken from [37], [36]. Here we consider the case of
identical countries (i ¼ 1 and 2 correspond to statistical data of the Republic of
Kazakhstan) and the case of nonidentical countries (i ¼ 1 corresponds to the
Republic of Kazakhstan, i ¼ 2 corresponds to the Russian Federation).
For estimation of the remaining parameters of the model, xi, pi, di, ni, �Оi, bi, ai,Qi(0), the parametric identification problems were solved by the searching method
in the sense of the minimum of the squared discrepancies:
X2i¼1
XNj¼1
Mij �M
ij
Mij
!2
þ pij � pijpij
!224
35; (1.135)
where Mij*, Mij
**, pij*, pij
** are the respective values of the total product capacity
and product price of the ith country presented in [37], [36] and, calculated, N is the
number of observations, i ¼ 1; 2.
72 1 Elements of Parametric Control Theory of Market Economic Development
1.5.5.2 Analysis of the Structural Stability of the Mathematical Modelof the National Economic System Subject to the Influenceof International Trade and Currency Exchange WithoutParametric Control
Analysis of the robustness (structural stability) of model (1.115), (1.116), (1.117),
(1.118), (1.119), (1.120), (1.121), (1.122), (1.123), (1.124), (1.125), (1.126),
(1.127), (1.128), (1.129), (1.130), (1.131), (1.132), (1.133), and (1.134) is based
on the theorem on sufficient conditions of weak structural stability in the compact
set of the phase space.
Assertion 1.3 Let N be a compact set residing within the region ðM1>0; Q1<0;p1>0Þ or ðM1>0; Q1>0; p1>0Þ of the phase space of the system of differentialequations of mathematical model (1.115), (1.116), (1.117), (1.118), (1.119),(1.120), (1.121), (1.122), (1.123), (1.124), (1.125), (1.126), (1.127), (1.128),(1.129), (1.130), (1.131), (1.132), (1.133), and (1.134), that is, the eight-dimensional space of the variables ðMi; Qi; pi; LGiÞ, i ¼ 1; 2. Let the closure ofthe interior of N coincide with N. Then the flux f defined by the system of modeldifferential equations is weakly structurally stable on N.
One can choose N such as, for instance, the parallelepiped with boundary Mi
¼ Mimin; Mi ¼ Mimax; Qi ¼ Qimin; Qi ¼ Qimax; pi ¼ pimin; pi ¼ pimax; LGi ¼ LGimin; LGi ¼ LGimax: Here 0<Mimin<Mimax, Qimin<Qimax<0 or 0<Qimin<Qimax, 0<pimin<pimax, LGimin<LGimax.
Proof First, let’s prove the semitrajectory of flux f starting from any point of the set
N for some value of t (t > 0) leaving N.Consider any semitrajectory starting in N. With t>0, the following two cases are
possible; namely, all the points of the semitrajectory remain in N, or for some t apoint of the semitrajectory does not belong to N. In the first case, from (1.118), dp1dt¼ �a1
Q1
M1p1 , of the system it follows that for all t>0 , the variable p1(t) has a
derivative greater than some positive constant withQ1<0or less than some negative
constant with Q1>0 ; that is, p1(t) increases infinitely or tends to zero with
unbounded increase of t. Therefore, the first case is impossible, and the orbit of
any point in N leaves N.Since any chain-recurrent set Rðf ;NÞ lying within N is the invariant set of this
flux, then if it is nonempty, it consists of only whole orbits. Hence, in the considered
case, Rðf ;NÞ is empty. The assertion follows from Theorem A [67].
1.5.5.3 Choosing Optimal Laws of Parametric Control of MarketEconomic Development on the Basis of the Mathematical Modelof the Country Subject to the Influence of International Tradeand Currency Exchange
Choosing the optimal laws of parametric control of the parametersxi; pi; y is carriedout in the environment of the following relations:
1.5 Examples for Application of Parametric Control Theory 73
1: Ui1;b ¼ ki1;b
DMiðtÞMiðt0Þ þ constib;
2: Ui2;b ¼ �ki2;b
DMiðtÞMiðt0Þ þ constib;
3: Ui3;b ¼ ki3;b
DpiðtÞpiðt0Þ þ constib;
4: Ui4;b ¼ �ki4;b
DpiðtÞpiðt0Þ þ constib:
(1.136)
HereUia;b is the ath control law of the bth parameter of the ith country, a ¼ 1; :::; 4;
b ¼ 1; :::; 3. The case b ¼ 1 corresponds to the parameterxi; b ¼ 2 corresponds
to the parameter pi; b ¼ 3, corresponds to the parameter yDMiðtÞ ¼ Ma;b;iðtÞ �Mi
ðt0Þ;DpiðtÞ ¼ pa;b;iðtÞ � piðt0Þ; t0 is the control starting time, t 2 t0; t0 þ T½ �. HereMa;b;iðtÞ, pa;b;iðtÞ are the values of the product capacity and the price level of the ithcountry, respectively, with theUi
a;bth control law; kia;b is the adjusted coefficient of
the respective law (kia;b 0i); constib is a constant number equal to the estimate of
the values of the b-parameter by the results of parametric identification.
The problem of choosing the optimal parametric control law for the economic
system of the ith country at the level of one of the economic parameters xi; pi; yð Þcan be formulated as follows: On the basis of mathematical model (1.115), (1.116),
(1.117), (1.118), (1.119), (1.120), (1.121), (1.122), (1.123), (1.124), (1.125),
(1.126), (1.127), (1.128), (1.129), (1.130), (1.131), (1.132), (1.133), and (1.134),
find the optimal parametric control law in the environment of the set of algorithms
(1.136); that is, find the optimal law (and its coefficients kia;b) from the set {Uia;b}
minimizing the criterion
Ki ¼ 1
T
Zt0þT
t0
piðtÞdt ! minfkia;b;Ui
a;bg(1.137)
under the constraints
MiðtÞ�MiðtÞj j � 0:09Mi
ðtÞ;0�Ui
a;bðtÞ � aib; a¼ 1; :::;4;b¼ 1;2;3; piðtÞ 0; siðtÞ 0; where t 2 t0; t0 þ T½ � :
(1.138)
Here MiðtÞ are the values of the total production capacity of the ith country
without parametric control; aib is the bth parameter of the ith country.
The problem is solved in two stages:
1. In the first stage, the optimal values of the coefficients kia;b are determined for
each law Uia;b by enumerating their values on the respective intervals quantized
with step 0.01 minimizing K under constraints (1.138).
2. In the second stage, the law of optimal control of the specific parameter is chosen
on the basis of the results of the first stage by the minimum value of criterion Ki.
74 1 Elements of Parametric Control Theory of Market Economic Development
The problem of choosing the pair of optimal parametric control laws for the
simultaneous control of two parameters can be formulated as follows: Find the
optimal pair of parametric control laws (Uia;b,U
in;m) on the set of combinations of two
economic parameters from three parameters xi; pi; yð Þ on the basis of the set of
algorithms (1.136) minimizing the criterion
Ki ¼ 1
T
Zt0þT
t0
piðtÞdt ! minðUi
a;b; kia;bÞ;ðUi
n;m; kin;mÞ
� � ;a; n ¼ 1; :::; 4 ; b; m ¼ 1; 2; 3; b<m;
(1.139)
under constraints (1.138).
The problem of choosing the optimal pair of laws is solved in two stages:
1. In the first stage, the optimal values of the coefficients kia;b,kin;m are determined for
the chosen pair of the control laws (Uia;b ,U
in;m ) by enumeration of their values
from the respective intervals quantized with step equal to 0.01 minimizing Kt
under constraints (1.138);
2. In the second stage, the optimal pair of parametric control laws is chosen on the
basis of the results of the first stage by the minimum value of criterion Ki.
Here we present the results of numerical experiments on choosing efficient laws
of parametric control of the public consumers’ expenditures, the norm of reserva-
tion, and the currency exchange rate within the framework of the following part of
the research program:
– The estimation of the values of criterionKi on the basis of themathematical model
of the interaction between identical economic systems of two countries by foreign
trade (themodel coefficients are estimated by choosing and solving the parametric
identification problem with the data of one country, the Republic of Kazakhstan).
– On the basis of the mathematical model of the interaction between the identical
economic systems of two countries via foreign trade, choosing the optimal
parametric control law at the level of two of the economic parameters x1; p1; yð Þfor the economic system of the first country, and estimation of the values of
criterion K2 for the economic system of the second country.
– On the basis of the mathematical model of the interaction between identical
economic systems of two countries via foreign trade, choosing the optimal pair
of parametric control laws on the set of combinations of two economic parameters
from three parameters for the economic system of the first country and estimation
of the values of the criterion K2 for the economic system of the second country.
– The estimation of the values of criteria Ki i ¼ 1; 2ð Þ on the basis of the mathe-
matical model of the interaction between the nonidentical economic systems of
two countries (the Republic of Kazakhstan and the Russian Federation) via
foreign trade (the model coefficients are estimated by choosing and solving the
parametric identification problem for the data of two different countries).
– On the basis of the mathematical model of the interaction between the noniden-
tical economic systems of two countries via foreign trade, choosing the optimal
1.5 Examples for Application of Parametric Control Theory 75
law of parametric control of the currency exchange rate y for the economic
system of the first country and estimating the values of criterion K2 for the
economic system of the second country.
– On the basis of the mathematical model of the interaction between nonidentical
economic system of two countries via foreign trade, choosing the optimal pair of
parametric control laws on the set x1; yð Þ, p1; yð Þ for the economic system of the
first country and estimate the values of criterion K2 for the economic system of
the second country.
– On the basis of the mathematical model of the interaction between nonidentical
economic systems of two countries via foreign trade, choosing the optimal law
of parametric control of the currency exchange rate y2 for the economic system
of the second country and estimating the values of criterion K1 for the economic
system of the first country.
– On the basis of the mathematical model of the interaction between nonidentical
economic systems of two countries via foreign trade, choosing the optimal pair
of the parametric control laws on the set (x2; y2), (p2; y2) for the economic system
of the second country and estimating of values of criterion K1 for the economic
system of the second country.
– On the basis of the mathematical model of the interaction between nonidentical
economic systems of two countries, the estimation of the influence of the control of
the economic system of one country on the economic indices of another country
with simultaneous application of the optimal control laws at the level of one
economic parameter of three (z1; p1; y) and (z2; p2; y) in two countries. Simulta-
neous control of the currency exchange rate y by two countries is not considered.
Within the framework of the first intended stage of research, we estimate the
coefficients of the mathematical model of the interaction between the two identical
economic systems of two countries via foreign trade on the basis of the data of one
country [40]. The results of parametric identification show that the value of the
standard deviation from the measured values of the respective variables is 5%. The
values of criteria Ki are equal and are given by K1 ¼ K2 ¼ 1:145 with CL1 ¼ CL
2
¼ CO1 ¼ CO
2 ¼ 0:1 and y ¼ 1.
The results of the numerical solution of the first stage of the stated problem of
choosing the optimal law of parametric control at the level of one of the economic
parameters (x1, p1, y) for the economic system of the first country are presented in
Table 1.7. Analysis of Table 1.7 shows that the best resultK1 ¼ 0:99 is attained withuse of the control law
p1 ¼ �0:8Dp1ðtÞ
1þ 0:1348:
With such a control law, the criterion of optimality of the economic system of
the second country is K2 ¼ 1.144, differing slightly from the case without control.
The results of numerical solution of the first stage of the stated problem of choosing
the optimal pair of parametric control laws are presented in the eight tables in the form
of Table 1.8, differing in the control law expression by at least one parameter.
76 1 Elements of Parametric Control Theory of Market Economic Development
The choice of the optimal pair of the parametric control laws according to the
requirements of the second stage on the basis of analysis of the data from these
tables allows the recommendation to implement the control laws for the parameters
p1 and y given as follows:
p1 ¼ �0:8DP1ðtÞ
1þ 0:1348; y ¼ �1:6
DM1ðtÞ139; 435
þ 0:2:
The value of the criterion of the economic system of the first country is equal to
K1 ¼ 0:97, and the value of the criterion for the economic system of the second
country differs slightly from the case without control and K1 ¼ 1.144.
Further, we estimate the coefficients of the mathematical model of interaction
between the nonidentical economic systems of two countries via foreign trade on
the basis of the data of two different countries [37], [36]. The parametric identifica-
tion results show the admissible precision of the description. The values of the
criterion Ki i ¼ 1; 2ð Þ are, respectively, K1 ¼ 1:137; K2 ¼ 1:775; with C1 ¼ 0:15;C2 ¼ 0:015; y ¼ 0:2:
Table 1.7 Results of the numerical solution of the first stage
of the problem of choosing the optimal parametric control law
at the level of one parameter
Notation of laws Law coefficient Value of criterion K1
U111 0.2 1.072
U121 0.0 1.145
U131 2.1 1.009
U141 0.0 1.145
U112 0.0 1.145
U122 0.1 1.068
U132 0.0 1.145
U142 0.8 0.990
U113 0.0 1.145
U123 1.8 1.070
U133 0.0 1.145
U143 1.9 1.100
Table 1.8 First-stage results of the numerical solution of the problem of choosing the optimal pair
of laws
Pairs of parametric control laws
Value of
criterion K1
First law Second law
Law denotation
Optimal
coefficient value
Law
denotation
Optimal
coefficient value
U142 0.8 U1
13 0.0 0.99
U142 0.8 U1
23 1.6 0.97
U142 0.8 U1
33 0.0 0.99
U142 0.8 U1
43 0.0 0.99
1.5 Examples for Application of Parametric Control Theory 77
The solution of the problem of choosing the optimal law of parametric control of
the currency exchange y for the economic system of the first country on the basis of
the mathematical model of the interaction of the two nonidentical economic
systems of the two countries via foreign trade allows us to propose the law given by
y ¼ �1:2DM1ðtÞ139; 435
þ 0:2:
The application of this law to the control of the currency exchange rate of the
first country results in improving the criterion from 1.137 to 1.123. The criterion of
the second country goes down from 1.734 to 1.828.
The solution of the problem of choosing the optimal pair of parametric control
laws on the basis of the mathematical model of the interaction of the two noniden-
tical economic systems of the two countries via foreign trade allows us to propose
the following laws:
p1 ¼ 0:2DM1ðtÞ139; 435
þ 0:1136; y ¼ 1:5DM1ðtÞ139; 435
þ 0:2:
Criterion K2 is 1.83 for the economic system of the second country with
K1 ¼ 1.05.
In solving the problem of choosing the optimal parametric control law of the
second country from the given pair of countries, the following results are obtained:
The optimal control of the parameter y is realized by means of the law
y2 ¼ 1= �0:12Dp2ðtÞ þ 0:2ð Þ:
The value of criterion K2 improves from 1.775 to 1.73.
In solving the problem of choosing the optimal pair of parametric control laws
for the second country, the following pair of the laws is obtained:
y2 ¼ 1= �0:11Dp2ðtÞ þ 0:2ð Þ; p2 ¼ �0:01Dp2ðtÞ þ 0:1388:
With application of these control laws, the value of criterion K2 is equal to 1.66.
In both cases, the criterion of the first country K1 varies insignificantly (the increase
not exceeding 1%).
By carrying out the simultaneous control of the parameters of two countries, the
values of the criteria improve within the limits of 3% for each country in compari-
son with the control of each country separately. The optimal control of the first
country at the level of one parameter is implemented by means of lawU14;2; criterion
K1 is 0.99. The optimal control of the second country at the level of one parameter is
implemented by means of law U24;3 ; criterion K2 is 1.72. With the simultaneous
application of two control laws U14;2 and U2
4;3, for both countries the values of the
criteria turn out to be K1 ¼ 0.98 and K2 ¼ 1.66.
78 1 Elements of Parametric Control Theory of Market Economic Development
1.5.5.4 Analysis of the Structural Stability of the Mathematical Modelof the Country Subject to the Influence of International Tradeand Currency Exchange with Parametric Control
Let’s analyze the robustness of system (1.115), (1.116), (1.117), (1.118), (1.119),
(1.120), (1.121), (1.122), (1.123), (1.124), (1.125), (1.126), (1.127), (1.128),
(1.129), (1.130), (1.131), (1.132), (1.133), and (1.134), where the parameters xi;pi; y are defined in accordance with the solution of the parametric control
problems as the expressions
1: Ui1;b ¼ ki1;b
DMiðtÞMiðt0Þ þ constib;
2: Ui2;b ¼ �ki2;b
DMiðtÞMiðt0Þ þ constib;
3: Ui3;b ¼ ki3;b
DpiðtÞpiðt0Þ þ constib;
4: Ui4;b ¼ �ki4;b
DpiðtÞpiðt0Þ þ constib
(1.140)
for any values of the adjusted coefficients kiab 0. Here constib is a constant number
equal to the estimate of the values of the bth parameter of the ith country by the
results of parametric identification, i ¼ 1; 2; a ¼ 1; :::; 4; b ¼ 1; 2.The application of parametric control law Ui
a;b means the substitution of the
respective functions into model equations (1.115), (1.116), (1.117), (1.118),
(1.119), (1.120), (1.121), (1.122), (1.123), (1.124), (1.125), (1.126), (1.127),
(1.128), (1.129), (1.130), (1.131), (1.132), (1.133), and (1.134) for the parameters
xi (j ¼ 1), pi (j ¼ 2), and y (j ¼ 3).
As a result of the application of these laws to system (1.115), (1.116), (1.117),
(1.118), (1.119), (1.120), (1.121), (1.122), (1.123), (1.124), (1.125), (1.126),
(1.127), (1.128), (1.129), (1.130), (1.131), (1.132), (1.133), and (1.134), the follow-
ing system is derived:
dMi
dt¼ FI
i
pibi� miMi; (1.141)
dQi
dt¼ Mifi � Fi
pi; (1.142)
dLGidt
¼ rG iLGi þ FG
i � np iFi � nL isiRLi � nO iðdPi þ dBi Þ; (1.143)
dpidt
¼ �aiQi
Mipi; (1.144)
1.5 Examples for Application of Parametric Control Theory 79
dsidt
¼ siDi
max 0;Rdi � RS
i
RSi
�;RL
i ¼ minfRdi ;R
Si g; (1.145)
LPi ¼ 1� Uia;1
Uia;1
LGi ; (1.146)
dPi ¼ 1� Uia;1
Uia;1
bir2 iLGi ; (1.147)
dBi ¼ bir2 iLGi ; (1.148)
xi ¼ ni1� di
1� sinipi
� �1�didi
!; (1.149)
Rdi ¼ Mixi; (1.150)
fi ¼ 1� 1� 1� dini
xi
� � 11�di
; (1.151)
FOi ¼ �0 ipiMifi; (1.152)
FGi ¼ Ui
a;2piMifi; (1.153)
FLi ¼ ð1� nL iÞsiRd
i ; (1.154)
FIi ¼
1
1þ np i
kqiMifixi
� ð1� npiÞFGi þ n0 iðdBi þ dPi Þ þ np iFO
i þ
þ nL i þ ð1� nL iÞnp i�
siRLi þ npiðFji � Ui
a;3FijÞ þ miLpi � rG iLPi
�;
(1.155)
RSi ¼ PA
0 i expðlp itÞ1
1þ nioi; oi ¼ FL
i
ð1þ CLi U
ia;3
pjpiÞP0 iðlpitÞ
; j ¼ 3� i; (1.156)
F12 ¼CL1p2p1
1þ CL1U
ia;3
p2p1
FL1 þ
CO1
p2p1
1þ CO1U
ia;3
p2p1
FO1 ; (1.157)
F21 ¼CL2p1p2
1þ CL2
1Ui
a;3
p1p2
FL2 þ
CO2
p1p2
1þ CO2
1Ui
a;3
p1p2
FO2 ; (1.158)
80 1 Elements of Parametric Control Theory of Market Economic Development
F1 ¼ FI1 þ FL
1 þ FO1 þ FG
1 þ F21 � Uia;3F12; (1.159)
F2 ¼ FI2 þ FL
2 þ FO2 þ FG
2 þ F12 � 1
Uia;3
F21: (1.160)
The proof of the weak structural stability of the mathematical model indicates
that the weak structural stability of the considered model is maintained with the use
of each of the parametric control laws Uia;b in the form of the following assertion:
Assertion 1.4 Let N be a compact set belonging to region ðM1>0; Q1<0; p1>0Þor ðM1>0; Q1>0; p1>0Þ of the phase space of the model system of differentialequations (1.115), (1.116), (1.117), (1.118), (1.119), (1.120), (1.121), (1.122),(1.123), (1.124), (1.125), (1.126), (1.127), (1.128), (1.129), (1.130), (1.131),(1.132), (1.133), and (1.134), that is, the eight-dimensional space of variables ðMi
; Qi; pi; LGiÞ, i ¼ 1; 2. Let the closure of of the interior of N coincide with N. Thenthe flux f defined by system (1.141), (1.142), (1.143), (1.144), (1.145), (1.146),(1.147), (1.148), (1.149), (1.150), (1.151), (1.152), (1.153), (1.154),(1.155),(1.156), (1.157), (1.158), (1.159),and (1.160) is weakly structurally stable on N.
1.5.5.5 Finding the Bifurcation Points of the Extremals of the VariationalCalculus Problem on the Basis of the Mathematical Model of theCountry Subject to the Influence of International Trade andCurrency Exchanges
Besides the case considered above, the problem of choosing the optimal set of laws
was also solved in some other definition.
Choosing the optimal parametric control laws on the basis of model (1.115),
(1.116), (1.117), (1.118), (1.119), (1.120), (1.121), (1.122), (1.123), (1.124),
(1.125), (1.126), (1.127), (1.128), (1.129), (1.130), (1.131), (1.132), (1.133), and
(1.134) at the level of one of the two parameters xi; pi is carried out in the
environment of the following relations:
1: Ui1;b ¼ ki1;b
DMiðtÞMiðt0Þ þ constib;
2: Ui2;b ¼ �ki2;b
DMiðtÞMiðt0Þ þ constib;
3: Ui3;b ¼ ki3;b
DpiðtÞpiðt0Þ þ constib;
4: Ui4;b ¼ �ki4;b
DpiðtÞpiðt0Þ þ constib:
(1.161)
1.5 Examples for Application of Parametric Control Theory 81
Here Uia;b is the ath control law of the bth parameter of the ith country, a ¼ 1; :::; 4;
b ¼ 1; 2. The case b ¼ 1 corresponds to parameter xi; b ¼ 2� pi; DMiðtÞ ¼ Ma;b;i
ðtÞ �Miðt0Þ;DpiðtÞ ¼ pa;b;iðtÞ � piðt0Þ; t0 is the control starting time, t 2 t0; t0 þ T½ �.HereMa;b;iðtÞ,pa;b;iðtÞ are the values of the product capacity and the level of prices ofthe ith country, respectively, with the Ui
a;b th control law; kia;b is the adjusted
coefficient of the respective law (kia;b 0i); constib is a constant equal to the estimate
of the values of the bth parameter by the results of parametric identification.
The problem of choosing the optimal parametric control law for the economic
system of the ith country at the level of one of the economic parameters (xi, pi, y)can be formulated as follows: On the basis of the mathematical model (1.115),
(1.116), (1.117), (1.118), (1.119), (1.120), (1.121), (1.122), (1.123), (1.124),
(1.125), (1.126), (1.127), (1.128), (1.129), (1.130), (1.131), (1.132), (1.133), and
(1.134), find the optimal parametric control law in the environment of the set of
algorithms (1.136); that is, find the optimal law (and its coefficients kia;b) from the
set {Uia;b} maximizing the criterion
Ki ¼ 1
T
Zt0þT
t0
YiðtÞdt; (1.162)
where Yi ¼ Mifi . In computational experiments, we research the influence of the
parametric control of the first country (i ¼ 1).
A closed set in the space of the continuous vector functions of the output
variables of system (1.115), (1.116), (1.117), (1.118), (1.119), (1.120), (1.121),
(1.122), (1.123), (1.124), (1.125), (1.126), (1.127), (1.128), (1.129), (1.130),
(1.131), (1.132), (1.133), and (1.134) and regulating parametric actions is defined
by the following relations:
p1ðtÞ � p1 ðtÞ�� �� � 0:09p1 ðtÞ;ðMiðtÞ; QiðtÞ; LGiðtÞ; piðtÞ; siðtÞÞ 2 X ;
0 � Uiab � aib; a ¼ 1; :::; 4; b ¼ 1; 2; i ¼ 1; 2 t 2 ½t0; t0 þ T�:
(1.163)
Hereaib is the maximum possible value of the ath parameter of the ith country;pi ðtÞare the model (calculated) values of the price level of the ith country without
parametric control; X is the compact set of the admissible values of given variables.
In this variational calculus problem, we consider its dependence on the two-
dimensional coefficient l ¼ ðr2;1; yÞ of the mathematical model with possible
values in some region (rectangle) L in the plane.
As a result of a computer simulation experiment, the plots of the dependence of
the optimal value of criterion K on the values of the parameters ðr2;1; yÞ for each ofeight possible laws U1
a;b; a ¼ 1; :::; 4; b ¼ 1; 2 are established. Figure 1.14
presents the plots for the two laws U12;2 and U
14;2 maximizing the criterion in region
82 1 Elements of Parametric Control Theory of Market Economic Development
L , the intersection curve of the respective regions, and the projection of this
intersection curve to the plane of values l consisting of the bifurcation points of
this two-dimensional parameter. This projection divides rectangleL into two parts.
The control law U12;2 ¼ �k12;2
DM1ðtÞM1ðt0Þ þ const12 is optimal in one of these parts,
whereas U14;2 ¼ �k14;2
Dp1ðtÞp1ðt0Þ þ const12 is optimal in the other part. Both of the laws
are optimal on the curve projection.
1.5.6 Forrester’s Mathematical Model of Global Economy
1.5.6.1 Model Description
Forrester’s mathematical model of the “world dynamics” [29] is given by the
following system of ordinary differential and algebraic equations (here t is time):
P0ðtÞ ¼ PðtÞðBnðtÞ � DðtÞÞ; (1.164)
V0ðtÞ ¼ CVGPðtÞVMðMÞ � CVDVðtÞ; (1.165)
Z0ðtÞ ¼ CZPðtÞZVðVRÞ � ZðtÞ=TZðZRÞ; (1.166)
Opt
imal
val
ues
of c
rite
rion
Fig. 1.14 Plots of the dependencies of optimal criterion values on the parameters of deposit
interest rate r2;1 and currency exchange rate y
1.5 Examples for Application of Parametric Control Theory 83
R0ðtÞ ¼ �CRPðtÞRMðMÞ; (1.167)
S0ðtÞ ¼ ðCSSQQMðMÞSFðFÞ=QFðFÞ � SðtÞÞ=TS; (1.168)
MðtÞ ¼ VRðtÞð1� SðtÞÞERðRRÞ=½ð1� SNÞEN�; (1.169)
FðtÞ ¼ FSðSRÞFZðZRÞFPðPRÞFC=FN; (1.170)
BnðtÞ ¼ PðtÞCBBMðMÞBPðPRÞBFðFÞBZðZRÞ; (1.171)
DðtÞ ¼ PðtÞCDDMðMÞDPðPRÞDFðFÞDZðZRÞ; (1.172)
QðtÞ ¼ CQQMðMÞQPðPRÞQFðFÞQZðZRÞ; (1.173)
PRðtÞ ¼ PðtÞ=PN; (1.174)
VRðtÞ ¼ VðtÞ=PðtÞ; (1.175)
SRðtÞ ¼ VRðtÞSðtÞ=SN; (1.176)
RRðtÞ ¼ RðtÞ=R0; (1.177)
ZRðtÞ ¼ ZðtÞ=ZN: (1.178)
The model includes the following exogenous constants:
CQ is the standard quality of life;
CB is the normal rate of fertility;
CD is the normal rate of mortality;
FC is the nourishment coefficient;
CZ is normal pollution;
CR is the normal consumption of natural resources;
FN is the normal level of nourishment;
EN is the normal efficiency of the relative volume of funds;
CVD is the normal depreciation of funds;
CVG is the normal fund formation;
TS is the coefficient of pollution influence.
The exogenous functions of the model are as follows:
BM is the multiplier of the fertility dependence on the material standard of living;
BP is the coefficient of fertility dependence on the population density;
BF is the coefficient of fertility dependence on nourishment;
BZ is the coefficient of fertility dependence on the pollution;
DM is the coefficient of mortality dependence on the material standard of living;
DP is the coefficient of mortality dependence on the population density;
84 1 Elements of Parametric Control Theory of Market Economic Development
DF is the coefficient of mortality dependence on nourishment;
DZ is the coefficient of mortality dependence on pollution;
QM is the coefficient of life quality dependence on the material standard of living;
QP is the coefficient of life quality dependence on the population density;
QF is the coefficient of life quality dependence on nourishment;
QZ is the coefficient of life quality dependence on pollution;
FS is the food potential of the funds;
FZ is the coefficient of food production dependence on pollution;
FP is the coefficient of food production dependence on population density;
ER is the coefficient of dependence of the natural resources production cost;
ZV is the coefficient of pollution dependence on the specific volume of funds;
TZ is the time of the pollution decay (reflecting the difficulty of natural decay with
the growth of pollution);
RM is the coefficient of the natural resources’ production rate dependence on the
material standard of living;
SQ is the coefficient of the dependence of the fund share in agriculture on the relativequality of life;
SF is the coefficient of the dependence of the fund share in agriculture on the level ofnourishment;
RR is the share of the remaining resources;
PR is the relative population density;
VR is the specific capital;
ZR is the relative pollution;
SR is the relative volume of agriculture funds.
The endogenous variables of the model are as follows:
P is the world population;
V is the basic asset;
Z is the pollution level;
R is the remaining part of the natural resources;
S is the share of funds in agriculture (i.e., in the food supply industry);
M is the material standard of living;
F is the relative level of nourishment (quantity of food per head);
Q is the quality of life;
Bn is the rate of fertility;
D is the rate of mortality.
In [26], the following values of the coefficients and constants are used:
CB ¼ 0:04; CD ¼ 0:028; CZ ¼ 1; CR ¼ 1; CQ ¼ 1; FC ¼ 1; FN
¼ 1; EN ¼ 1; (1.179)
PN ¼ 3:6�109; ZN ¼ 3:6�109; SN ¼ 0:3; TS ¼ 15; TVD ¼ 40; CVG ¼ 0:05;
1.5 Examples for Application of Parametric Control Theory 85
as well as the following initial conditions for the differential equations:
P0 ¼ 1:65�109; V0 ¼ 0:4�109; S0 ¼ 0:2; Z0 ¼ 0:2�109; R0 ¼ 9�1011;
corresponding to the time starting point t0 ¼ 1900. These data were obtained on the
basis of observations for the years 1900–1970.
Here we accepted the values of the parametersCD,CZ,CR,CQ,TS, andTVD equal tothe above data from (1.179). The values of the parameters CB , CVG , and FC are
estimated again on the basis of information about the global population for the years
1901–2009 [61] and the data calculated by the state functionsVðtÞ, SðtÞ,RðtÞ, andZðtÞ (accepted as the measured functions in solving the parametric identification
problem) based on the model (1.164), (1.165), (1.166), (1.167), (1.168), (1.169),
(1.170), (1.171), (1.172), (1.173), (1.174), (1.175), (1.176), (1.177), and (1.178).
These values are determined by solving the parametric identification problem by
the searching method in the sense of the minimum of the criterion
K ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
545
X2009t¼1901
PðtÞPðtÞ � 1
� �2þ SðtÞ
SðtÞ � 1
� �2þ RðtÞ
RðtÞ � 1
� �2þ ZðtÞ
ZðtÞ � 1
� �2þ VðtÞ
VðtÞ � 1
� �2 !vuut :
Here PðtÞ and PðtÞ are the measured and modeled (calculated) values of the
population, respectively; VðtÞ, SðtÞ, RðtÞ, and ZðtÞ are the calculated data of system
(1.164), (1.165), (1.166), (1.167), (1.168), (1.169), (1.170), (1.171), (1.172),
(1.173), (1.174), (1.175), (1.176), (1.177), and (1.178). As a result of the solution
of the stated problem of parametric identification, the following are estimates of
the values of the estimated parameters: CB ¼ 0.042095, CVG ¼ 0.049644, and FC
¼ 1.078077. The relative value of the mean square deviation of the calculated
values of the variables from the respective measured values is approximately
100K ¼ 4.27%.
1.5.6.2 Analysis of the Structural Stability of Forrester’s MathematicalModel Without Parametric Control
Assertion 1.5 Let N be a compact set residing in the region fP>0; V>0; S>0; Z>0; R>0 g of the phase space of the system of the differential derived from (1.164),(1.165), (1.166), (1.167), (1.168), (1.169), (1.170), (1.171), (1.172), (1.173),(1.174), (1.175), (1.176), (1.177), and (1.178), that is, the five-dimensional spaceof the variables fP; V; S; Z; Rg. Let the closure of the interior of N coincide withN. Then the flux f defined by system (1.164), (1.165), (1.166), (1.167), (1.168),(1.169), (1.170), (1.171), (1.172), (1.173), (1.174), (1.175), (1.176), (1.177), and(1.178) is weakly structurally stable on N.One can choose N such as, for instance, the parallelepiped with the boundary P¼ Pmin; P ¼ Pmax; V ¼ Vmin; V ¼ Vmax; S ¼ Smin; S ¼ Vmax; Z ¼ Zmin; Z ¼ Zmax
86 1 Elements of Parametric Control Theory of Market Economic Development
R ¼ Rmin;R ¼ Rmax . Here 0<Pmin<Pmax, 0<Vmin<Vmax, 0<Smin<Smax, 0<Zmin<Zmax, 0<Rmin<Rmax.
Proof First, let’s prove that the semitrajectory of the flux f starting from any point
of the set N with some value of t (t > 0) leaves N.Consider any semitrajectory starting in N. With t>0, the following two cases are
possible; namely, all the points of the semitrajectory remain in N, or for some
t the point of the semitrajectory does not belong to N. In the first case, from (1.167)
R0ðtÞ ¼ �CRPðtÞRMðMÞ of the system, it follows that for all t>0; the variable R(t)has a derivative less than some negative constant number. That is, R(t) tends to
converge to zero with increasing t. Therefore, the first case is impossible, and the
orbit of any point in N leaves N.Since any chain-recurrent set Rðf ;NÞ lying within N is the invariant set of this
flux, in the case when it is nonempty, it consists of only whole orbits. Hence, in the
considered case, Rðf ;NÞ is empty. The assertion follows from Theorem A [67].
1.5.6.3 Choosing Optimal Laws of Parametric Control on Basisof the Forrester Model
Let’s consider working out the recommendations on choosing a rational scenario of
world policy development (in terms of the objective to maximize the mean value of
quality of life for the years 1971–2100) by choosing the optimal control laws for the
example of economic parameters FC (coefficient of nourishment, j ¼ 1) and CB
(normal fertility rate, j ¼ 2).
The problem of choosing the optimal parametric control law at the level of the
parameter is solved in the environment of the following relations:
1: U1j ¼ constj þ k1jðPðtÞ=Pðt0Þ � 1Þ;2: U2j ¼ constj � k2jðPðtÞ=Pðt0Þ � 1Þ;3: U3j ¼ constj þ k3jðRðtÞ=Rðt0Þ � 1Þ;4: U4j ¼ constj � k4jðPðtÞ=Pðt0Þ � 1Þ;5: U5j ¼ constj þ k5jðZðtÞ=Zðt0Þ � 1Þ;6: U6j ¼ constj � k6jðZðtÞ=Zðt0Þ � 1Þ;7: U71j ¼ constj þ k7jðVðtÞ=Vðt0Þ � 1Þ;8: U8j ¼ constj � k8jðVðtÞ=Vðt0Þ � 1Þ;9: U9j ¼ constj þ k9jðSðtÞ=Sðt0Þ � 1Þ;
10: U10j ¼ constj � k10jðSðtÞ=Sðt0Þ � 1Þ;11: U11j ¼ constj þ k11jðQðtÞ=Qðt0Þ � 1Þ;12: U12j ¼ constj � k12jðQðtÞ=Qðt0Þ � 1Þ:
(1.180)
1.5 Examples for Application of Parametric Control Theory 87
Here kij 0 is the adjusted coefficient of the respective law Uij i ¼ 1; :::; 12;ðj ¼ 1; 2Þ; constj is the base value (without parametric control) of the nourishment
coefficient FC(with j ¼ 1) or normal fertility rate C
B(with j ¼ 2), respectively.
The control starting time t0 corresponds to year 1971. Application of one of the
laws (1.180) means the substitution of the respective function for the right-hand
side of the corresponding relation (1.180) into (1.170) or (1.171) of system (1.164),
(1.165), (1.166), (1.167), (1.168), (1.169), (1.170), (1.171), (1.172), (1.173),
(1.174), (1.175), (1.176), (1.177), and (1.178) for the parameter FC or CB.
The problem of choosing the optimal parametric control law at the level of the
parameter FC in the environment of the algorithms (1.180) is stated as follows: On
the basis of mathematical model (1.164), (1.165), (1.166), (1.167), (1.168), (1.169),
(1.170), (1.171), (1.172), (1.173), (1.174), (1.175), (1.176), (1.177), and (1.178),
find the optimal parametric control law in the environment of algorithms (1.180);
that is, find the optimal law from this set of algorithms and its adjusted coefficient
that maximizes the criterion
K1 ¼ 1
130
X2100t¼1971
QðtÞ (1.181)
characterizing the mean the quality of life level on the interval of time 1971–2100
under the constraints
X2100t¼1971
ZðtÞ � �Z; FCðtÞ 2 ½0:9; 1:1�: (1.182)
Here �Z is the total value of the pollution levels for the years 1971–2100 without
parametric control.
The given problem is solved in two stages:
– In the first stage, the optimal values of the coefficients kij are determined for each
law (1.180) by the enumeration of their values on the intervals ½0; kmij Þ quantizedwith a sufficiently small step maximizing criterion K1 under constraints (1.182).
Here kmij is the first value of the coefficient violating (1.182).
– In the second stage, the law of optimal control of the specific parameter (of 12) is
chosen on the basis of the results of the first stage by the minimum value of
criterion K1.
The numerical solution of the problem of choosing the optimal parametric
control law of the economic system at the level of the stated economic parameter
shows that the best result K1 ¼ 0:70827 can be achieved with the application of thefollowing control law of type (8) from (1.180):
FC ¼ FC � 0:158ðVðtÞ=Vðt0Þ � 1Þ: (1.183)
88 1 Elements of Parametric Control Theory of Market Economic Development
Note that the value of criterion (1.180) without parametric control is equal to
K1 ¼ 0:6515: The increase of the criterion value with the given parametric control
in comparison with the base variant is equal to 5.025% (see Fig. 1.15).
The problem of choosing the optimal pair of parametric control laws at the level
of the parameters FC and CB in the environment of the set of algorithms (1.180) is
stated as follows: On the basis of mathematical model (1.164), (1.165), (1.166),
(1.167), (1.168), (1.169), (1.170), (1.171), (1.172), (1.173), (1.174), (1.175),
(1.176), (1.177), and (1.178), find the optimal pair of parametric control laws in
the environment of the set of algorithms (1.180); that is, find the optimal pair of
laws from this set of algorithms and its adjusted coefficients that maximize criterion
(1.181) under constraints (1.182).
The numerical solution of the problem of choosing the optimal pair of the
parametric control laws of the economic system at the level of two economic
parameters FC and CB shows that the best result K1 ¼ 0:703135 can be achieved
with the application of the following pair of control laws:
FC ¼ FC � 0:15ðVðtÞ=Vðt0Þ � 1Þ; CB ¼ C
B� 0:01ðPðtÞ=Pðt0Þ � 1Þ: (1.184)
In this case, the increase in the value of criterion K1 in comparison with the base
variant is equal to 7.93%.
Let’s compare the obtained results of the parametric control of the evolution of
dynamical system (1.164), (1.165), (1.166), (1.167), (1.168), (1.169), (1.170),
(1.171), (1.172), (1.173), (1.174), (1.175), (1.176), (1.177), and (1.178) with the
optimal laws found at the level of one (1.183) and two (1.184) parameters and the
results of the scenario consisting in the increase of the parameter FC by 25% in
Qua
lity
of life
Years
Tendency estimated by Forrester
Tendency obtained as a result of control of parameter of influence on agriculture production
Fig. 1.15 Trajectories characterizing the change in the quality of life Q
1.5 Examples for Application of Parametric Control Theory 89
comparison with the base solution (obtained for the following values of constants
CB ¼ 0:042095; CD ¼ 0:028; CZ ¼ 1; CR ¼ 1; CQ ¼ 1; FC ¼ 1:078077; FN ¼ 1;
EN ¼ 1; PN ¼ 3:6�109; ZN ¼ 3:6�109; SN ¼ 0:3; TS ¼ 15; TVD ¼ 40; CVG ¼ 0:049644and initial conditions for the differential equationsP0 ¼ 1:65�109;V0 ¼ 0:4�109;S0 ¼ 0:2; Z0 ¼ 0:2�109; R0 ¼ 9�1011).
A comparison shows that for the stated scenario (the increase of parameterFC by
25%), the mean value of the quality of life (criterion K1) on the time interval
1971–2100 decreases by 9.77% in comparison with the base variant, and the mean
value of the pollution 1130
P2100t¼1971
ZðtÞ increases by 4.97% in comparison with the base
variant. With use of optimal law (1.183) alongFC , the index of the quality of life
improves by 5.025% and the mean value of the pollution decreases by 3.5% in
comparison with the base value. Furthermore, the value of the nourishment coeffi-
cient FC by optimal law (1.183) changes by no more than 10% in comparison with
the base value of this coefficientFC ¼ 1.078077. With the use of the optimal pair of
laws (1.183), the quality of life index improves by 7.93%, and the mean pollution
decreases by 1% in comparison with the base variants.
1.5.6.4 Analysis of the Structural Stability of Forrester’s MathematicalModel Subject to Parametric Control
Application of the optimal laws of parametric control (1.180) means substitution of
the respective functions into (1.170), (1.171) for the parametersFC andCB, while the
other model equations remain unchanged.
The proof of the weak structural stability of the mathematical model presented
above and relying on (1.167) allows us to derive the following assertion:
Assertion 1.6 Let N be a compact set belonging to the region fP>0; V>0; S>0;Z>0; R>0 g of the phase space of the system of differential equations derived from(1.164), (1.165), (1.166), (1.167), (1.168), (1.169), (1.170), (1.171), (1.172),
(1.173), (1.174), (1.175), (1.176), (1.177), and (1.178), that is, the five-dimensionalspace of variables fP; V; S; Z; Rg:
Let the closure of the interior of N coincide with N. Then the flux f defined by
(1.164), (1.165), (1.166), (1.167), (1.168), (1.169), (1.170), (1.171), (1.172),
(1.173), (1.174), (1.175), (1.176), (1.177), and (1.178) and (1.183) or (1.184) is
weakly structurally stable on N.
1.5.6.5 Finding Bifurcation Points of Extremals of the Variational CalculusProblem on the Basis of Forrester’s Mathematical Model
Let’s analyze the dependence of the solution of the problem considered above of
choosing the optimal parametric control law on the values of the two-dimensional
parameter (CVG,CVD)with possible values belonging to region (rectangle)Lin theplane.
90 1 Elements of Parametric Control Theory of Market Economic Development
As a result of computer simulations, we generated plots of the dependence of the
optimal values of criterionK on the values of the parameter (CVG,CVD) for each of the
24 possible lawsUij; i ¼ 1; :::; 12; j ¼ 1; 2. Figure 1.16 demonstrates such plots for
the four laws U2;1,U6;1,U11;1,U8;1maximizing the values of criterionK in regionL, aswell as the intersection curves of the corresponding surfaces. The projection of these
curves to the plane (CVG , CVD ) consists of the bifurcation points of this two-
dimensional parameter. This projection divides rectangle L into two parts; inside
each of them only one parametric control law out of (U2;1,U6;1,U11;1,U8;1) is optimal.
Two or three different laws are optimal on the projection curves.
1.5.7 Turnovsky’s Monetary Model
1.5.7.1 Model Description
Turnovsky’s monetary model [72], after the corresponding transformations (for the
economic development scenario considered in [72], when the national deficit is
fully financed subject to invariable reserve of bonds per capita), is presented by the
system of the following differential and algebraic equations:
_p ¼ r½r� p� (1.185)
Fig. 1.16 Plot of dependencies of optimal values of criterion K on parameters (CVG , CVD ).
Here colorings correspond to parametric control laws as follows: –U2;1, –U6;1, –U11;1,
– U8;1
1.5 Examples for Application of Parametric Control Theory 91
_m ¼ g� uyþ bðreð1� uÞ þ pÞ � ðmþ bÞðnþpÞ; (1.186)
_z ¼ gð1� uÞg� 1
½rez� Rk� þ gg� 1
z
k� 1
� �þ 1
� �gðk � kÞ þ n
z� k
g� 1; (1.187)
_k ¼ gðk � kÞ; (1.188)
y ¼ Aka; (1.189)
re ¼ Aaka�1; (1.190)
R ¼ Aaka�1; (1.191)
k� ¼y� c ðy� RkÞð1� uÞ þ bþz
l2½ð1� l4Þm� l4ðbþ zÞ � l1yþ l3p� þ zn� mp
h i� nk þ g
l czk � cþ 1� ;
(1.192)
i ¼ lðk � kÞ þ nk: (1.193)
Here the derivatives with respect to time t in years are denoted by an overdot.
The output (endogenous) variables of the model are as follows:
p is the instant expected rate of inflation (1/year);
m is the nominal reserve of the foreign money per capita (tenge per capita; tenge is
the monetary unit of Kazakhstan);
z is the real share of stock per capita (tenge per capita) (the real indicators are
defined by the prices of year 2000);
k is the real fixed capital per worker (tenge per capita);
y is the real product output per capita [(tenge per capita)/year];re is the real rate of stock income before tax (1/year);
R is the limiting real physical product of the capital [(tenge per capita)/year2];
k is the desired real fixed capital per capita (tenge per capita);
i is the real investment per capita [(tenge per capita)/year].
The input (exogenous) time-dependent parameters of the model are as follows:
p is the consumer price index (1/year);
g is the real public expense per capita [(tenge per capita)/year] (g>0);
n is the rate of population growth (1/year);
g is the coefficient of the equation for fixed capital per worker (1/year) ð0<g<1Þ;A and a are the coefficients of the production function ðA>0; 0<a<1Þ;c is the consumption share of the available income ð0<a<1Þ (dimensionless);
l1; l2; l3; and l4 are the coefficients of the equation for real demand for money per
capita (l1>0; l2<0; l3>0; 0<l4<1) (dimensionless);
92 1 Elements of Parametric Control Theory of Market Economic Development
l is the coefficient of the equation of investment per capita (1/year);
u is the income taxation rate, 0<u<1 (dimensionless).
The input model parameters are as follows:
b is the nominal reserve of public bonds per capita; b > 0 (tenge per capita).
The input model parameters also incorporate the initial values (t ¼ 0) of the
output variables of dynamic equations (1.1), (1.2), (1.3), and (1.4) of model p0, m0,
z0, and k0. The values of the input model functions for the integer moments of time
t are also considered to be the input model parameters. All input model functions
are considered to be piecewise linearly continuous functions defined by their values
for integer moments of t.
1.5.7.2 Estimation of Turnovsky’s Model Parametersand Retrospective Forecast
Solving the problem to estimate input parameters (parametric identification) of a
model, we obtain the values of input functions and parameters p(t), g(t), n(t), g(t),A(t), a(t), l(t), c(t), l1(t), l2(t), l3(t), and l4(t), where t ¼ 0; :::; 9; as well as b, u, p0,m0, z0, and k0 by the search method for minimizing the sum of squares of
discrepancies of the output variables based on statistical information on evolution
of the economy of the Republic of Kazakhstan from 2000 to 2009. The values of the
input functions and parameters were searched within small intervals from the
middle in measured values (if available) of the respective functions and parameters.
(Please check this)
The criterion of parametric identification is given by
KI ¼ 1Pvj¼1
Pnt¼0 Mjt
Xv
j¼1
Xn
t¼0Mjt
xjðtÞ � xj ðtÞxj ðtÞ
!2
! min : (1.194)
Here v is the number of the output variables used in the estimation of the para-
meters equal to 5, j is the variable number; n + 1 is the number of measurements,
t ¼ 0 corresponds to the beginning of year 2000; and xjðtÞ are the computed values
of the output variables ðyðtÞ; kðtÞ; zðtÞ;pðtÞ; iðtÞÞ and for the corresponding time
moments. The symbol (*) corresponds to the measured values’ respective variables.
MJt are the positive weights’ fitted reasoning from the significance of their respec-
tive values of the output variables while solving the problem of model parametric
identification. In Table 1.9, we present the weights of criterion KI. The weightMjt is
situated at the intersection of the row j with column t.For solving the problem of parametric estimation, we shall apply the the
Runge–Kutta and Nelder–Mead algorithms [66]. The stated parametric identifica-
tion problem is solved with the use of statistical information from 2000 to 2007
(n ¼ 7). As a result of solving the parametric identification problem, the values of
1.5 Examples for Application of Parametric Control Theory 93
the mean square deviations of computed values of the model output variables from
their respective measured values ð100 ffiffiffiffiffiffiK1
p Þ do not exceed 1.2%.
For model verification, the problem of the retrospective forecast is solved. Using
values of input functions, parameters and initial values of output variables from
2000 to 2007 are determined by estimation (as well as the extrapolation of the
values of input functions for 2008–2009), between 2008 and 2009 and estimate
relative errors of the computed values of the model output variables relative to the
corresponding measured values. The results of the problem are presented in
Table 1.10. The symbol* corresponds to the measured values and “D” correspondsto the errors (percentage) of computed values of the corresponding measured
values.
The mean error of the variables indicated in Table 1.10 for the retro-forecasting
period is 3.7825%, hence showing the admissible accuracy of the description of
Kazakhstan’s economic evolution by the analyzed model.
Table 1.9 Weights Mjt of criterion KI
Year
Variable
2000,
t ¼ 0
2001,
t ¼ 1
2002,
t ¼ 2
2003,
t ¼ 3
2004,
t ¼ 4
2005,
t ¼ 5
2006,
t ¼ 6
2007,
t ¼ 7
yðtÞ; j ¼ 1 0.001 0.001 0.001 0.010 0.010 1.000 1.000 1.000
kðtÞ; j ¼ 2 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001
zðtÞ; j ¼ 3 0.001 0.001 0.001 0.010 0.010 1.000 1.000 1.000
pðtÞ; j ¼ 4 0.001 0.001 0.001 0.010 0.010 0.100 0.100 0.100
iðtÞ; j ¼ 5 0.001 0.001 0.001 0.010 0.010 0.100 0.100 0.100
Table 1.10 Measured and computed values of the model output variables and respective errors
(in percentage)
Year 2008 2009
y 336,140 334,680
y 333,843 333,015
Dy 0.68322 0.49744
z 1,117,488 1,305,937
z 1,228,469 1,483,599
Dz 9.93126 13.60410
k 747,806 771,832
k 675,539 762,518
Dk 9.66383 1.20669
p 0.11828 0.07525
p 0.11881 0.07639
Dp 0.45321 1.51652
i 136,501 151,534
i 136,424 151,855
Di 0.05666 0.21187
94 1 Elements of Parametric Control Theory of Market Economic Development
1.5.7.3 Analysis of the Structural Stability of Turnovsky’s Monetary Model
This analysis is carried out on the basis of Robinson’s Theorem A [67] on sufficient
conditions of weak structural stability.
In this algorithm, the mapping f is defined as the shift along the trajectories
of dynamical system (1.185), (1.186), (1.187), (1.188), (1.189), (1.190), (1.191),
(1.192), and (1.193) corresponding to the variation of time t by a year. The input
functions of the model are considered to be the constants (same as values of
year 2007).
While realizing the previously mentioned algorithm, we consider a parallelepi-
ped, f0 � p � 0:2; 0 � m � 8; 000, 0 � z � 50; 000, 0 � k � 830; 000g in the statespace of models (1)–(9) as the initial compact N.
The partitioning of the initial parallelepiped (and other cells obtained as a result
of the algorithm) into 16 parts is realized by dividing each of its edges into two equal
parts. As a result of these computations according to the mentioned algorithm, after
four iterations we generated graphGwith an empty set of nodes. This means that the
Turnovsky’s monetary model with the considered values of input parameters is
considered to be weakly structurally stable in the indicated compact N.
1.5.7.4 Estimation of the Parametric Sensitivity of Turnovsky’s Model
Evaluating the influences of input parameters and model functions on the values of
its output variables, we form a matrix whose rows are enumerated by all input
parameters and functions, and whose columns are enumerated by the values of the
output variables for t¼ 9, which corresponds to year 2009. This matrix incorporates
the coefficients of sensitivity of the indicated output values of the model by its input
values computed by the formula
FpjðtÞ ¼ 100xnj ðtÞ � xjðtÞ
xjðtÞ : (1.195)
Here, p is the variable input parameter or value of the input function; xjðtÞ is the
value of the jth output variable at time t obtained via running the model with values
of the input parameters and functions derived from the estimate of the parameters or
taken from the statistical sources (the base computation); xnj ðtÞ is the value of the
respective output variable obtained by increasing variable parameter p by 1%; other
values of the parameters and functions remain invariable in comparison to the base
computation.
The results from solving the problem of constructing the parametric sensitivity
matrix are presented in Table 1.11. For instance, increasing the value of the
parameter c(9) by 1% while keeping the value of c(8) yields the respective increaseof the function c(t), which is linear in the interval [8, 9]. In turn, this yields a
variation of the values of the output variables (and coefficients FpjðtÞ for t ¼ 9.
1.5 Examples for Application of Parametric Control Theory 95
The analysis of Table 1.11 shows that within the limits of the input parameters in
2009 presented in this table, the most influence on the values of the output variables
y(9), m(9), z(9), k(9), i(9) is exerted by the variation of coefficient a(9) of the
production function. The value of the output variable p(9) can be impacted only by
changing the price index p(9).
1.5.7.5 Finding Optimal Values of the Adjusted Parameters Basedon Turnovsky’s Model
Let’s now consider implementation of the effective public policy based on model
(1.185)–(1.193) via the synthesis of optimal values of economic parameters, public
expenses per capita g(t), and income taxation rate u(t) for 2010–2015.The problem of the synthesis of the optimal parametric control law at the level of
the previously mentioned parameters can be formulated as follows. Based on the
mathematical model (1.185)–(1.193), find such values g(t), u tð Þ; t ¼ 2010; . . . ;2015 that maximize criterion
K ¼ 1
6
X2015
t¼2010yðtÞ; (1.196)
which is the mean value of the real product output per capita in the years 2010–2015
under the following constraints on the model output variables and adjusted
parameters (here t 2 ½2010; 2015�):
mðtÞ>0; zðtÞ>0; kðtÞ>0; yðtÞ>0; reðtÞ>0;RðtÞ>0; k>0; iðtÞ>0; (1.197)
Table 1.11 Some elements of the model parametric sensitivity matrix for t ¼ 9
The output variableThe input
parameter y p M Z K I
C(9) �0.016370 0.0000 �0.01230 0.00309 �0.01650 1.693064
g(9) 0.280699 0.0000 0.26993 0.33196 0.28325 0.881013
n(9) �0.017260 0.0000 �0.02060 �0.02930 �0.01740 0.048407
p(9) �0.009930 0.5161 �0.05200 �0.01360 �0.01000 �0.397900
l(9) �0.038020 0.0000 0.02940 �0.08880 �0.03840 �2.592670
g(9) 0.136572 0.0000 �0.00200 �0.27080 0.13781 �0.455090
A(9) 0.790219 0.0000 �0.03340 �0.29850 �0.20960 �3.418850
a(9) 11.388490 0.0000 �0.31440 �2.22150 �2.60970 �53.240600
l1(9) �0.163710 0.0000 0.05780 �0.26970 �0.16520 �4.831100
l2(9) 0.009145 0.0000 0.01151 �0.00830 0.00923 �1.518530
l3(9) �0.043310 0.0000 �0.00520 �0.03870 �0.04370 0.849691
l4(9) �0.140340 0.0000 0.04360 �0.22460 �0.14160 �3.589840
96 1 Elements of Parametric Control Theory of Market Economic Development
gðtÞ>0; 0<uðtÞ<1: (1.198)
Note that for the base computation of the model until 2015 obtained from the
found values of the input parameters of the model and by the extrapolation of the
model input functions by the linear trend, the criterion value is K ¼ 437368 tenge
(in the prices of year 2000).
As a result of the numerical solution of the stated problem of finding optimal values
of parameters g(t) and u(t) of the economic system by the Nelder–Mead method [66],
we obtain the optimal valueK ¼ 511552. The increase in criterionKwith application
of the aforesaid parametric control in comparison with the base variant is 16.96%.
The diagrams of the computed values of the output variable of model y(t) (outputof real products per capita) without parametric control as well as with application of
the derived optimal parametric control law are presented in Fig. 1.17.
1.5.7.6 Analysis of the Dependence of Optimal Values of the ParametricControl Criterion on the Values of the Uncontrolled ParametersBased on Turnovsky’s Model
The optimization problem considered above was solved with fixed values of the
input parameters that do not participate in the control. Moreover, carrying out
the analysis, we realized the dependence of the optimal values of criterion K on
Fig. 1.17 Real product per capita
1.5 Examples for Application of Parametric Control Theory 97
the values of the model’s uncontrolled parameters by the example of the two-
dimensional parameter a ¼ c 9ð Þ; l 9ð Þð Þ incorporating the share of real consump-
tion in the real income available and the coefficient of the investment equation for
2009. The range of variation of these parameters was determined based on the
estimated values of c(9) and l(9) in the form of a rectangle A ¼ ½0:0820; 0:1090��½0:708; 0:719�.
In Fig. 1.18, we present some results of the analysis, namely, the diagrams of the
dependence of criterion K on parameter a ( a 2 A ) for the parametric control
problem considered above. The diagrams in Fig. 1.18 describe the base and optimal
(for the problem of finding public expense per capita and the income tax rate) value
of criterion K.
1.5.8 Endogenous Jones’s Model
1.5.8.1 Model Description
The mathematical model of technical progress and population [61], after some
transformations, is given by following difference and algebraic equations:
Nt ¼ Nt�1ð1þ ntÞ; (1.199)
LAt ¼ ltNtpt; (1.200)
LYt ¼ ltNtð1� ptÞ; (1.201)
Fig. 1.18 – the base variant, – adjustment of the public expense per capita and the income
taxation rate
98 1 Elements of Parametric Control Theory of Market Economic Development
At ¼ At�1 þ dLlAðt�1ÞA’t�1; (1.202)
Yt ¼ Ast L
bYtet (1.203)
wt ¼ YtptrLAt
(1.204)
Here time t is discrete, and the year number.
The endogenous variables of the model are as follows:
Nt is the population of the country (people);
LAt is the number of people involved in the production of ideas (people);
LYt is the number of people involved in the production of consumer goods (people);
At is the reserve of ideas in the economy (dimensionless);
Yt is the production of consumer goods expressed by the production function (tenge
per year);
wt is the wages per one worker in a unit of time (tenge/year).
The exogenous variables of the model are as follows:
nt is the rate of population growth (1/year);
lt is the share of people employed in the economy in the entire population
(dimensionless);
pt is the share of people involved in the production of ideas in the entire number of
people employed in the economy (dimensionless);
et is the exogenous production shock.
The exogenous parameters of the model are as follows:
d, l (dimensionless), and ’ (dimensionless) are parameters of the production
function of the production of ideas;
s and b are parameters of the production function of consumer goods production
(dimensionless);
r is the share of the wage fund in the entire production of consumer goods
(dimensionless).
It should be noted that this parameter is not present in [61]. It is assumed that the
entire output of consumer goods is intended for remuneration of workers’ labor.
The introduction of parameter r; 0 � r � 1; is made to obtain the model, which is
more adequate for real economic conditions.
The exogenous parameters of the model include initial values (with t ¼ 0) of the
endogenous variables of dynamical equations (1.199), (1.202) of models N0 and A0.
The values of the exogenous functions of the model with fixed time t are also
considered to be the exogenous parameters of this model.
1.5 Examples for Application of Parametric Control Theory 99
1.5.8.2 Evaluation of the Exogenous Parameters of the Jones’s Modeland Retrospective Forecast
Solving the problem of estimation of input parameters (parametric identification) of
the model, we obtain values of the input functions and parameters nt, lt, pt, et, d, l,j, s, b, r, N0, and A0 by the search method for minimization of the sum of squares
of discrepancies of output variables based on statistical data from economic evolu-
tion of the Republic of Kazakhstan from 2000–2009. The values of the input
functions and parameters were searched within small intervals with centers in the
measured values (if available) of the corresponding functions and parameters.
The criterion of parametric identification is given by
KI ¼ 1
vðnþ 1ÞXv
j¼1
Xn
t¼0
xtj � xtjxtj
!2
! min (1.205)
Here v ¼ 5 is the number of the output variables used in the estimation of
parameters; n + 1 is the number of measurements; t ¼ 0 corresponds to year
2000; xtj are the computed values of the output variables (Nt, LAt, LYt, Yt, and wt).
The symbol * corresponds to the measured values of the respective variables.
To solve the problem of parametric estimation, we apply the Nelder–Mead
algorithm [66]. This parametric identification problem is solved with the use of
statistical information from 2000–2007 (n ¼ 7). As a result of solving the problem,
the relative values of the mean square deviations of computed values of the model
output variables from the respective measured values (100ffiffiffiffiffiKI
p) do not exceed 1%.
For model verification, the following problem of the retrospective forecast is
solved. Using the values of the input functions, parameters, and initial values of the
output variables from 2000–2007 determined by the estimation (as well as the
extrapolation of the values of the input functions for 2008–2009), within the period
of 2008–2009, estimate the relative errors of the computed values of the model
output variables relative to the corresponding measured values. The results of the
problem solution are presented in Table 1.12. Here the symbol * corresponds to the
measured values, and D corresponds to the errors (percentage) of the computed
values of the respective measured values.
The mean error of the variables indicated in Table 1.12 for the retro-forecasting
period is equal to 3.49%, thereby indicating admissible accuracy of the description
of Kazakhstan’s economic evolution by the analyzed model.
1.5.8.3 Analysis of the Structural Stability of the MathematicalJones’s Model
Let’s analyze the robustness (structural stability) of model (1.199), (1.200), (1.201),
(1.202), (1.203), and (1.204) on the basis of the mentioned theorem on sufficient
conditions of weak structural stability under the additional assumption on the
100 1 Elements of Parametric Control Theory of Market Economic Development
constant character of all the exogenous functions of the model. We’ll consider the
values of these functions as well as values of all exogenous parameters of the
analyzed model to be some constant positive numbers. In this case, (1.199), (1.200),
(1.201), (1.202), (1.203), and (1.204) define the homeomorphism f (and the respec-tive cascade) determined in the plane (N, A).
Assertion 1.7 Let M be a compact set in the region ðN>0;A>0Þ of state space ofthe cascade f defined by (1.199), (1.200), (1.201), (1.202), (1.203), and (1.204). Theclosure of the interior of M coincides with M. Then, cascade f is weakly structurallystable in M.
Choose M to be a rectangle with boundaries N ¼ Nmin ,N ¼ Nmax , A ¼ Amin;A ¼ Amax. Here 0<Nmin<Nmax, 0<Amin<Amax.
Proof Let’s first determine that the semitrajectory of cascade f,which begins at anypoint of setM with some value of t (t> 0), leavesM. It follows from (1.199) with nt> 0 since in that case the sequence of Nt is the increasing geometric progression.
As any chain-recurrent set R(f, M) lying within M is the invariant set of the
cascade, if it is nonempty, it consists only of the whole orbits. Hence, in our case,
set R(f, M) is empty. This statement follows from Robinson’s Theorem A [67].
1.5.8.4 Estimation of the Parametric Sensitivity of Jones’s Model
Solving the problem of evaluating the influences of input parameters and model
functions on the values of its output variables, we form a matrix whose rows are
enumerated by all the input parameters and functions and columns are enumerated
by the values of the output variables for t¼ 9, which corresponds to year 2009. This
matrix incorporates the coefficients of sensitivity of the indicated output values of
the model by its input values computed by the formula
Table 1.12 Measured
and computed values of
the model output variables
and the respective errors
(in percentage terms)
Year 2008 2009
Y:t 2.895 � 1013 3.450 � 1013
Yt 2.988 � 1013 3.660 � 1013
DYt 3.210 6.092
N:e 16.068 � 106 16.470 � 106
Nt 15.748 � 106 15.927 � 106
DNt 1.992 3.298
LAt 18,219 18,674
LAt 17,975 18,179
DLAt 1.335 2.650
Lt 7.844 � 106 8.041 � 106
Lyt 7.699 � 106 7.786 � 106
DLYt 1.847 3.155
wt 741.751 � 103 863.110 � 103
Wt 767.743 � 103 930.701 � 103
Dwt 3.504 7.831
1.5 Examples for Application of Parametric Control Theory 101
Fpj ¼ 100xntj � xtj
xtj: (1.206)
Here p is the variable input parameter or value of the input function; xtj is the valueof the ith output variable at time t obtained via running the model with values of
input parameters and functions derived from the estimation of parameters or taken
from statistical sources (the base computation); xntj is the value of the respective
output variable obtained by increasing the variable parameter p by 1%, while
the other values of the parameters and functions remain invariable in comparison
to the base computation.
The results of the construction of the parametric sensitivity matrix are partially
presented in Table 1.13.
The analysis of Table 1.13 shows that the most influence on the values of
endogenous variables Y9 and W9 is exerted by the variation of the coefficient
of the production function s, and the value of the endogenous variable A9 is
exerted by the variation of the coefficient of the equation of the reserve of ideas
in economy j.
1.5.8.5 Finding the Optimal Values of the Adjusted ParametersBased on Jones’s Model
Let’s now consider implementation of the effective public policy based on model
(1.199), (1.200), (1.201), (1.202), (1.203), and (1.204) by via the synthesis of
optimal values of economic parameters, i.e., the shares of people involved in the
production of ideas in the number of people employed from 2010 to 2014. This
ability is in particular grounded by the analysis of the sensitivity matrix obtained in
the above subsection.
The problem of the synthesis of the optimal parametric control law at the level of
parameters pt can be formulated as follows. Based on mathematical models (1.199),
(1.200), (1.201), (1.202), (1.203), and (1.204) , find the optimal law of parametric
control of the parameters; i.e., find such values pt, t ¼ 10, . . ., 14, that maximize
criterion
K ¼ 1
5
X14t¼10
Yt; (1.207)
which is the mean value of the consumer goods output in the aforesaid period under
the following constraints:
Nt 0; LAt 0; LYt 0; Yt 0;wt 0; t ¼ 10; :::; 14; (1.208)
0:0012<pt � 0:0480; t ¼ 10; :::; 14: (1.209)
102 1 Elements of Parametric Control Theory of Market Economic Development
Note that for the base computation of the model until 2014 obtained for the
values found from the input parameters of the model and by extrapolating the model
input functions by a linear trend, the criterion value is K ¼ 8:128 � 1013 tenge.As the result of the numerical solution of the stated problem of finding the
optimal values of parameters pt of the economic system by the Nelder–Mead
method [66], we obtain the optimal value K ¼ 8:462 � 1013; which exceeds the
base value of the criterion by 4.1%.
Table 1.13 Some elements
of the model parametric
sensitivity matrix
Variable
Parameter Y9 W9 A9
A0 0.158903 0.158903 0.072459
N0 0.108772 0.007974 0.003775
s 4.075839 4.075839 0.000000
b 3.891727 3.891727 0.000000
d 0.198968 0.198968 0.089890
l 0.095282 0.095282 0.044117
j 2.071020 2.071020 0.702719
l0 0.000877 0.000877 0.000416
l1 0.000879 0.000879 0.000417
l2 0.000880 0.000880 0.000417
l3 0.000881 0.000881 0.000418
l4 0.000882 0.000882 0.000418
l5 0.000888 0.000888 0.000421
l6 0.000891 0.000891 0.000422
l7 0.000887 0.000887 0.000421
l8 0.000885 0.000885 0.000420
l9 0.100000 �9.53 � 10–10 0.000000
p0 0.000877 0.000877 0.000416
p1 0.000879 0.000879 4.17 � 10–4
p2 0.000880 0.000880 4.17 � 10–4
p3 0.000881 0.000881 4.18 � 10–4
p4 0.000882 0.000882 0.000418
p5 0.000888 0.000888 4.21 � 10–4
p6 0.000891 0.000891 4.22 � 10–4
p7 0.000887 0.000887 4.21 � 10–4
p8 0.000885 0.000885 0.000420
p9 –0.000210 –0.000210 0.000000
n0 –0.000260 �2.01 � 10–5 �9.56 � 10–6
n1 –0.00010 �7.23 � 10–6 �3.43 � 10–6
n2 0.000112 6.84 � 10–6 3.25 � 10–6
n3 0.000596 3.14 � 10–5 1.49 � 10–5
n4 0.000858 3.80 � 10–5 1.80 � 10–5
n5 0.000985 3.53 � 10–5 1.67 � 10–5
n6 0.001186 3.21 � 10–5 1.52 � 10–5
n7 0.001142 2.08 � 10–5 9.86 � 10–6
n8 0.001311 1.20 � 10–5 5.70 � 10–6
n9 0.001619 �1.62 � 10–11 0.000000
1.5 Examples for Application of Parametric Control Theory 103
The obtained optimal values of the adjusted parameters pt are presented in
Table 1.14.
The diagrams of the computed values of the endogenous variable of the model,
namely, output of the consumer products Yt without parametric control as well as
with application of the determined parametric control law, are presented in
Fig. 1.19.
1.5.8.6 Analysis of the Dependence of the Optimal Values of theParametric Control Criterion on the Values of UncontrolledParameters Based on Jones’s Model
The optimization problem considered above was solved with fixed values of the
input parameters that do not participate in control. Moreover, in carrying out
the analysis, we discovered the dependence of the optimal values of criterion K(9) on the values of the model’s uncontrolled parameters via the example of the
two-dimensional parameter a ¼ d; sð Þ and by incorporating the parameters of
Table 1.14 Optimal values
of the adjusted parameters ptAdjusted parameter Optimal value
p10 0.0480
p11 0.0480
p12 0.0291
p13 0.0123
p14 0.0012
Fig. 1.19 Consumer products’ output (Y)
104 1 Elements of Parametric Control Theory of Market Economic Development
the equation from ideas of production and the production function of the consumer
goods output.
The range of variation for these parameters was determined based on the estimated
values of d and s in the form of rectangle A ¼ ½0:615; 0:769� � ½2:11; 3:16�.In Fig. 1.20, we present some results of the analysis, namely, the diagrams of the
dependence of criterion K on parameter a (where a 2 A) for the parametric control
problem considered before.
The diagrams in Fig. 1.20 describe the base and optimal (for the solved problem
of finding the shares of the whole wage fund, which is used for rewarding the
attempts of innovations) values of criterion K.
1.5.8.7 Appendix. Proof of Statements Given in Chapter 1
Proof of Theorem 1.1 From constraints (1.9), it follows that for any admissible
control signal u, the respective optimality criterion K(u) given by (1.7) makes sense
and there exists an upper bound of such values of function K(u). Therefore, thereexists a maximizing sequence such that sequence {uk} of the elements of set Wad
converge as
KðukÞ ! supu2Wad
KðuÞ: (1.210)
Obviously, set W is bounded in the space H1ðt0; t0þTÞ�q½ . Then, applying
Banach–Alaoglu’s theorem, one can choose a subsequence in {uk} such that
Fig. 1.20 – the base variant, – the adjustment of the shares of people involved in the
production of ideas in the number of people employed in the economy
1.5 Examples for Application of Parametric Control Theory 105
(keeping the initial notation for brevity) the convergence uk ! u takes place in the
weak topology of H1ðt0; t0þTÞ�q½ ; Taking into account the closeness of setW, we’ll
determine the inclusion u 2 W . According to Rellikh–Kondrashov’s theorem, the
embedding H1ðt0; t0þTÞ�q � ðC½t0; t0þT�Þ½ :q is compact. Then, after a possible
choice of a subsequence, we determine the convergence
uk ! uin C t0; t0þT�½ Þq:ð (1.211)
Let’s denote x as the solution of problems (1.1) and (1.2) corresponding to the
limiting value of u. The following relations hold true:
_xkðtÞ�x: ðtÞ¼ f ðxkðtÞ; ukðtÞ; lÞ � f ðxðtÞ; uðtÞ; lÞ; xkðt0Þ � xðt0Þ¼ 0;
where xk denotes the solution of system (1.1), (1.2) corresponding to the control law
uk. As a result, we obtain the following inequality for t 2 ½0; T�:
jxkðtÞ � xðtÞj �Z t
t0
jf ðxkðtÞ; ukðtÞ; lÞ � f ðxðtÞ; uðtÞ; lÞjdt
� Lf
Z t
t0
jxkðtÞ � xðtÞjdtþ Lf
Z t
t0
jukðtÞ � uðtÞjdt
� Lf
Z t
t0
jxkðtÞ � xðtÞjdtþ Lf Tjjuk � ujj;
where Lf is the Lipschitz constant of function f. Hereinafter, the norm is understood
as the respective power (in general difference for the control and state function) of
space C½t0; t0 þ T� . Applying Gronwall’s lemma, we’ll show that there exists a
constant c > 0 independent of k such that the following estimate holds:
jjxk � xjj � cjjuk � ujj: (1.212)
Let’s proceed to the limit, taking into account closeness of set X(t). We deter-
mine that inclusion (1.8) holds true for the limiting function x; only then is control
law u admissible.
The following inequality holds true:
Fðxk�� ðtÞÞ � FðxðtÞÞj � LFjxkðtÞ � xðtÞj;
where LF is the Lipschitz constant of function F.Taking into account condition (1.212), we obtain the estimate
Z t0þT
t0
jFðxkðtÞÞ � FðxðtÞÞjdt � cTLf jjuk � ujj:
106 1 Elements of Parametric Control Theory of Market Economic Development
Hence, passing to the limit, we obtain
Z t0þT
t0
FðxkðtÞÞdt !Z t0þT
t0
FðxkðtÞÞdt:
Taking into account that the left-hand side of this expression is KðukÞ, and by
virtue of (1.210), we conclude that
Z t0þT
t0
FðxðtÞÞdt ¼ supu2Wad
KðuÞ:
Hence, control law u is optimal. And Theorem 1.1 is proved.
Proof of Theorem 1.2 Since the set of the control laws is finite, we suffice to prove
solvability of the problem of maximizing function Kj on set Vjad for the selected
number j ðj ¼ 1; :::; rÞ. By virtue of the constraint on the growth rate of function Fand condition (1.19) with the bounded set
SXðtÞ
t2ðt0;t0þT�, there exists an upper bound of
function Kj; hence, the corresponding maximizing subsequence {vk} of the
elements of set Vjad exists such that it converges as
KjðvkÞ ! supv2Vj
ad
KjðvÞ: (1.213)
Since set V is close and bounded, after choosing the subsequence, we obtain
convergence
vk ! v in Rl; (1.214)
and v 2 V.From (1.19) and (1.20), we obtain equalities
_xvkj ðtÞ � _xvj ðtÞ ¼ f ðxvkj ðtÞ;Gjðvk; xvkj ðtÞÞ; lÞ � f ðxvj ðtÞ;Gjðv; xvj ðtÞÞ; lÞ;
xvkj ðt0Þ � xvj ðt0Þ ¼ 0:
We also obtain the following inequalities for t 2 ðt0; t0 þ T� :
jxvkj ðtÞ � xvj ðtÞj �Z t
t0
jf ðxvkj ðtÞ;Gjðvk; xvkj ðtÞÞ; lÞ � f ðxvj ðtÞ;Gjðv; xvj ðtÞÞ; lÞjdt
� Lf
Z t
t0
jxvkj ðtÞ � xvj ðtÞjdtþ Lf
Z t
t0
jGjðvk; xvkj ðtÞÞ � Gjðv; xvj ðtÞÞjdt
� Lf ð1þ LjÞZ t
t0
jxvkj ðtÞ � xvj ðtÞjdtþ Lf LjTjvk � vj:
1.5 Examples for Application of Parametric Control Theory 107
Here Lf and Lj are the Lipschitz constants for functions f and Gj, respectively. The
norm is understood as the space Rl. Applying Gronwall’s lemma and the last
inequality, we’ll show that there exists a constant c > 0 independent of k, j, l,and t; we thus have the estimate
jxvkj ðtÞ � xvj ðtÞj � cjvk � vj; (1.215)
from which it follows that
xvkj ðtÞ ! xvj ðtÞ: (1.216)
The condition vk 2 Vjad yields the inclusion
xvkj ðtÞ 2 XðtÞ; t 2 ðt0; t0 þ T�:
Hence, by virtue of (1.216) and the closeness of the set X(t), the following
condition holds true:
xvj ðtÞ 2 XðtÞ; t 2 ðt0; t0 þ T�: (1.217)
The Lipschitz condition for the function Gj yields the inequality
jGjðvk; xvkj ðtÞÞ ! Gjðvk; xvj ðtÞÞj � Ljðjvk � vj þ jxvkj ðtÞ � xvj ðtÞjÞ:
From the above inequality together with (1.215), we conclude that there exists a
positive constant c1 independent of any parameter such that the estimate
jGjðvk; xvkj ðtÞÞ ! Gjðv; xvj ðtÞÞj � c1jvk � vj
takes place. From the above estimate, the convergence
Gjðvk; xvkj ðtÞÞ ! Gjðv; xvj ðtÞÞ
follows. Taking into account the inclusionGjðvk; xvkj ðtÞÞ 2 UðtÞ and the closeness ofset U(t), we obtain Gjðv; xvj ðtÞÞ 2 UðtÞ; from which it follows that v 2 Vj
ad.
The proof of the convergence
ðt0þT
t0
F xvkj ðtÞ� �
dt !ðt0þT
t0
F xvkj ðtÞ� �
dt
based on the convergence in (1.216) can be made in the same way as the similar
property in Theorem 1.1. Then, from condition (1.213), it follows that the adjusted
coefficient v is optimal. Theorem 1.2 is proved.
108 1 Elements of Parametric Control Theory of Market Economic Development
Proof of Theorem 1.3 Obviously, the set of the admissible values of function K is
bounded from above (the boundaries of a continuous function in a compact set).
There exists an upper bound sup K of this set. Therefore, the sequence of the vector
functions ðxk; ukÞ exists such that the relations
xkðtþ 1Þ ¼ f ðxkðtÞ; ukðtÞ; lÞ; t ¼ 0; :::; n� 1; (1.218)
xkð0Þ ¼ x0; (1.219)
xkðtÞ 2 XðtÞ; t ¼ 1; :::; n; (1.220)
ukðtÞ 2 UðtÞ; t ¼ 0; :::; n� 1; (1.221)
hold and the convergence
KðxkÞ ! supK (1.222)
takes place. Since sets X(t) and U(t) are uniformly bounded, sequences {xk} and
{uk} are bounded as well. Then, by virtue of Bolzano–Weierstrass’s theorem, after
selecting a convergent subsequence, we obtain convergences uk ! u and xk ! x.Taking into account that sets X(t) and U(t) are closed, we conclude that the limiting
pair (u, x) satisfies inclusions (1.23) and (1.24). Proceeding to the limit in relations
(1.218) and (1.219), we determine that relations (1.20) and (1.21) hold and
hence the pair (u, x) is admissible. Since function K is continuous, the convergence
KðxkÞ ! supKðxÞ takes place. From this convergence and condition (1.222), it
follows that value K(x) coincides with the upper bound sup K in the set of admissi-
ble pairs of the system. Theorem 1.3 is proved.
Proof of Theorem 1.4 The boundaries of setsVjad follow from the boundaries of V.
The closeness of these sets follows from the closeness and boundaries of sets X(t)and U(t) and the continuity of functions f and Gj and the theorem about the
closeness of the complete preimage of a compact set with continuous mapping V! ðXðtÞ;Uðtþ 1ÞÞ given by the definition of set Vj
ad. The statement of the theorem
follows fromWeierstrass’s theorem about attaining the maximal value of a continu-
ous function in a compact set.
Proof of Theorem 1.5 By virtue of Weierstrass’s theorem, a continuous function
has a maximum on a nonempty closed bounded set. Thus, we suffice to show that
the multivariable function K ¼ K(u) defined by (1.34) is continuous and set Uad is
closed and bounded. Recall that set Uad is nonempty by the conditions of the
theorem.
Let’s show that there exist some expectations of the values entering state-space
constraint (1.36). Indeed, from (1.33), we have
E[xðtþ 1Þ] = E½f ðxðtÞ; uðtÞ; lÞ� þ E½xðtÞ�:
1.5 Examples for Application of Parametric Control Theory 109
The second term on the right-hand side of this equation makes sense by virtue of
the theorem conditions; the first one is calculated by the formula
E[f ðxðtÞ; uðtÞ; lÞ� ¼ZRn
f ðo; uðtÞ; lÞpxðtÞðoÞdo:
The last integral converges absolutely [here px(t) denotes the probability density
function of a random variable x(t)]. This is really so because of the constraints on
the growth of function f and the existence of the expectation of x(t) for any t ¼ 1,
. . ., n (this fact is determined by induction).
The constraints on the growth of the function Ft and the existence of the
expectation of x(t) yield the existence of the expectation on the right-hand side of
(1.34). Let the convergence of the vectors uk ! u; uk 2 Uadtake place. From (1.33),
it follows that
jxkðtþ 1Þ � xðtþ 1Þj ¼ jf ðxkðtÞ; ukðtÞ; lÞ � f ðxðtÞ; uðtÞ; lÞj;
where xk and x are the solutions to problems (1.33) and (1.33) with control laws ukand u, respectively. Then the following relation holds:
jxkðtþ 1Þ � xðtþ 1Þj � Lf ½jxkðtÞ � xðtÞj þ jukðtÞ � uðtÞj�;
where Lf is the Lipschitz constant of function f. By the same reasoning and taking
into account that xkð0Þ ¼ xð0Þ and by (1.33), we have
jxkðtþ 1Þ � xðtþ 1Þj � ðLf Þ2jxkðt� 1Þ � xðt� 1Þj þ
þðLf Þ2jukðt� 1Þ � uðt� 1Þj þ Lf jukðtÞ � uðtÞ �
�Xts¼0
ðLf Þsþ1jukðt� sÞ � uðt� sÞj � ek;
where ek ! 0 as k ! 1.
Here LF is the maximum of the Lipschitz constants of functions Ft, for t¼ 1, . . .,n, and we obtain the estimate.
jFt½xkðtÞ� � Ft½xðtÞ�j � LFek
Taking the expectations of both parts of this inequality, we obtain
EfjFt½xkðtÞ� � Ft½xðtÞ�jg � LFek;
from which it follows that EfjFt½xkðtÞ� � Ft½xðtÞ�jg ! 0 and the convergence of the
considered sequence EfFt½xkðtÞg ! EfFt½xðtÞ�g takes place. By virtue of (1.34),
this yields the continuity of function K ¼ K(u).
110 1 Elements of Parametric Control Theory of Market Economic Development
The boundaries of the set Uad follow from the boundaries of U(t). The closenessof set Uad follows from the continuity of the mapping Uad ! X given by definition
of set Uad and the compactness of set X (the theorem about the closeness of a
complete preimage of a compact set under a continuous mapping). Then the
existence of the solution of the considered problem follows from Weierstrass’s
theorem. Theorem 1.5 is proved.
Proof of Theorem 1.6 We suffice to determine that all functions Kj are continu-
ous, and all sets Vjad are closed and bounded, where j ¼ 1; :::; r. The existence of all
expectations used below is proved similarly to the proof of Theorem 1.5.
Taking into account the additivity of the expectation, let’s find the values
Kj ¼ KjðvÞ ¼Xnt¼1
E Ft xvj ðtÞh in o
from which we obtain the inequalities for v;w 2 Vjad :
KjðvÞ � KjðwÞ�� �� ¼ Xn
t¼1
E Ft xvj ðtÞh in o
�Xnt¼1
E Ft xwj ðtÞh in o�����
������Xnt¼1
jEf Ft½xvj ðtÞ�g � EfFt[xwj ðtÞ�gj:
(1.223)
From relations (1.42) and (1.43), it follows that
xvj ðtþ 1Þ � xwj ðtþ 1Þ��� ��� ¼ f xvj ðtÞ;Gj v; xvj ðtÞ
� �; l
� �� f xwj ðtÞ;Gj w; xwj ðtÞ
� �; l
� ���� ���� Lf xvj ðtÞ � xwj ðtÞ
��� ���þ Gj v; xvj ðtÞ� �
� Gj w; xwj ðtÞ� ���� ���h i
; t ¼ 0; . . . ; n� 1
(1.224)
where Lf is the Lipschitz constant of function f. Denoting by LG the maximum
Lipschitz constant of function Gj, we obtain the inequality
xvj ðtþ 1Þ � xwj ðtþ 1Þ��� ��� � Lf ð1þ LGÞ xvj ðtÞ � xwj ðtÞ
��� ���LfLGjv� wj; t ¼ 0; . . . ; ; n� 1
Taking into account the equality xvj ð0Þ � xwj ð0Þ��� ��� ¼ 0; we obtain the estimate
xvj ðtþ 1Þ � xwj ðtþ 1Þ��� ��� � Lf LG
Xtl¼0
Lf ð1þ LG Þ� ljv� wj
� bjv� wj8v;w 2 Vjad; (1.225)
1.5 Examples for Application of Parametric Control Theory 111
where
b= LfLG
Xtl¼0
Lf ð1þ LG Þ� l:
Denoting by LF, the maximum Lipschitz constant of function Ft, we have
Ft xvj ðtÞh i
� Ft xwj ðtÞh i��� ��� � LF xvj ðtÞ
��� � xwj ðtÞ��� � LF bjv� wj8v;w 2 Vj
ad: (1.226)
So, in the case of a sufficiently small difference in the adjusted coefficients v
and w, the values xvj ðtÞ and xwj ðtÞ (as well as Ft½xvj ðtÞ� and Ft xwj ðtÞh i
Þ are arbitrarilyclose to one another. Let’s define the convergent sequence w ¼ vk ! v. Then, aftertaking the expectations of the left- and right-hand sides of inequality (1.226), we
obtain the inequality
E Ft xvj ðtÞh i
� Ft xvkj ðtÞh i��� ���h i
� LFbjv� vkj;
which yields the convergence
E Ft xvkj ðtÞh in o
! E Ft xvj ðtÞh in o
from which it follows that function Kj(v) is continuous.Since Vj
ad 2 V , sets Vjad are bounded. The closeness of the sets Vj
ad follows
from the closeness of the sets U(t) and X(t+1), the continuity of the mappings
v ! E xvj ðtþ 1Þh i
; v ! E Gj v; xvj ðtÞ� �ih
proved above, and the definition of setVjad
as the complete preimage of the sets under continuous mappings. The theorem
statement follows from Weierstrass’s theorem about the attainment of the upper
bound by a continuous function in a compact set.
Proof of Lemma 1.1 Let the convergence of some sequence be ak ! a0, whereak 2 A: Let’s denote by uk the maximum point of function Kak in set Uak, and let u0denote the maximum point of function Ka0 in set Ua0 . Taking into account the
K-continuity of the family of sets {Ua} and the continuity of the function fa(x) on A� U, we conclude that for any e > 0, there exists a number k0 such that for k > k0there exists such a point u0k 2 Ua0 that satisfies неравенства Kakðu0kÞ � KakðukÞj j� e and, moreover, maxy2U KakðyÞ�j Ka0ðyÞj � e.
As a result, with k>k0; we obtain the following inequalities:
Ka0ðu0Þ Ka0ðu0kÞ Kakðu0kÞ � e KakðukÞ � 2e: (1.227)
112 1 Elements of Parametric Control Theory of Market Economic Development
By the same reasoning, one can verify the relation
KakðukÞ Ka0ðu0Þ � 2e: (1.228)
From (1.227) and (1.228), it follows that for sufficiently large k, the inequality
KakðuÞ � Ka0ðu0Þj j � 2e holds true. This inequality ensures the convergence of
sequence KakðuÞ ! Ka0ðu0Þ. The lemma is proved.
Proof of Theorem 1.7 Without loss of generality, by locality of the problem, set Lcan be considered to be compact. As noted above, the considered problems of the
synthesis of parametric control laws reduce to finding the maximal values of the
continuous functions (the optimization criterion K(u, l) in the compact sets of
admissible values of the adjusted parameters: maxu2Uad;l
Kðu; lÞ. HereUad; l is the
set of admissible values of the adjusted parameters of the corresponding problem for
somefixed values of the uncontrolled parameter l 2 L. The compactness of these sets
is already proved in Theorems 1.1, 1.3, and 1.5. From the continuity of mappings ðu;lÞ ! xu;lððu; lÞ ! E½xu;l�Þ for the stochastic cases), ðu; lÞ ! Kðu; lÞ, taking intoaccount the compactness of setsU andL, it follows that the set Uad;l : l 2 L
� �isK-
continuous. Then the required statement follows from Lemma 1.1.
Proof of Theorem 1.8 Just as in the previous theorem, the setL is considered to be
compact. For the given value of j, the problems of choice of the optimal control
laws reduce to finding the maximum values of continuous functions [the optimiza-
tion criterion Kjðu; lÞÞ in the compact sets of the admissible values of the adjusted
coefficients: maxv2Vj
ad; l
Kjðv; lÞ . Here Vjad; l is the set of the admissible values of the
adjusted coefficients of the respective problem for the chosen number j of the law and
value of the uncontrolled parameter l 2 L. The compactness of these sets is already
proved in Theorems 1.2, 1.4, and 1.6. From the continuity of mappings ðv; lÞ ! xv;lj
�ððu; lÞ ! E[xv;lj � for the stochastic cases), ðv; lÞ ! Gj v; xv;lj
� ��
ððu; lÞ ! E Gj v; xv;lj
� �h ifor the stochastic cases), and ðu; lÞ ! Kjðu; lÞ , taking
into account the compactness of setsV andA, it follows that the set Vjad;l : l 2 L
n oKj
is continuous. Then the required statement follows from Lemma 1.1.
Proof of Corollary 1.1 Since the optimal value of the criterion of the indicated
problems is defined as the maximal among the optimal values of the criteria Kj over
all possible parametric control laws from the given set, it can be written as
KðlÞ ¼ maxj¼1;...;r maxv2Vjad;l
Kjðv; lÞ:
1.5 Examples for Application of Parametric Control Theory 113
As Kð lÞ is the maximum value from the finite set of the continuous function,
this function is also continuous in L.
Proof of Theorem 1.9 Let’s connect points l0 and l1 by a smooth curve S lying in
the region L : S ¼ lðsÞ; s 2 ½0; 1�f g; l(0) = l0; l(1) = l1 . We’ll denote by
Kj(s) the optimal value of criteria Kj of Problem 1.2, 1.4, 1.6 for the selected
parametric control law Gj and value l ¼ lðsÞ of the uncontrolled parameter.
From Theorem 1.9, it follows that the function y ¼ KjðsÞ is continuous in the
interval [0,1]. Hence, for the function y ¼ maxj¼1;:::;r
KjðsÞ ¼ KðsÞ, solving considered
Problem 1.2, 1.5, or 1.6 is also continuous in interval [0,1]. Denote by DðjÞ � ½0; 1�the set of all values of parameter s that satisfy KjðsÞ ¼ KðsÞ. This set is closed as a
preimage of a closed set for a continuous function y ¼ KjðsÞ � KðsÞ. The set DðjÞcan be empty. As a result, the interval [0, 1] is represented as the following finite
sum of sets consisting at a minimum of two nonempty closed sets (see the theorem
conditions):
½0; 1� ¼ [j¼1;:::;rDðjÞ:
Hence, by the theorem conditions 0 2 Dðj0Þ and 1=2Dðj0Þ; it follows that the point s0exists at the boundary of setDðj0Þ; which lies within the interval (0, 1) [let’s assume
that s0 is the lower bound of such boundary points of the set Dðj0Þ]. The point s0 isalso the boundary point of some other set Dðj2Þ and belongs to it. For the found
value of s0, pointl s0ð Þ is the desired bifurcation point, since forl ¼ l s0ð Þ there existat least two optimal laws (Gj0 andGj2), and for l ¼ lðsÞ, 0 � s � s0 there exists oneoptimal law Gj0 , Theorem 1.9 is proved.
Proof of Lemma 1.2 Setting the right-hand sides of the equations of system (1.49)
equal to zero, we obtain relations (1.50). Obviously, k>0; c>0. Let’s construct
the determinant of the Jacobian matrix for the right-hand sides of (1.50) at the point
ðkþ; cþÞ:
D ¼ a� 1
ð1� bÞa ð dþ pÞ ðnþ dÞð1� aÞ þ p� nð Þ:
Since D < 0 for all indicated values of parameters A, a, b, d, n, and r of the
mathematical model, the obtained singularity point ðk; cÞ is a saddle point of
system (1.49).
Proof of Theorem 1.10 Let’s show that system (1.51) does not have cyclic
trajectories in region O1. Assume the contrary: There exists a cyclic trajectory in
region O1. Then it must include at least one singularity point, and the sum of the
Poincare’s indices of the singularity points belonging to the interior of this cycle is
equal to zero [12, p. 117]. But the region O1 includes only one saddle point with
index �1. Thus, we have a contradiction.
114 1 Elements of Parametric Control Theory of Market Economic Development
Let’s show that the stable and unstable separatrices of the saddle point x; yð Þdonot form one trajectory in region O1. Assume the contrary: The stable and unstable
separatrices of the saddle point x; yð Þ form one singular trajectory g lying in O1.
Then this trajectory (or, if it exists, the second trajectory consists of some other
stable and unstable separatrices) together with the singularity point are the bound-
ary of the bounded cellO2, which lies in regionO1. Let’s consider the semitrajectory
L+ originating from some point ðx1; y1Þ, where ðx1; y1Þ is the interior point of O2.
Then, since there are no cyclic trajectories and the equilibrium state is unique, only
the boundary of cellO2 can be the limiting points of L+ [the point ðx1; y1Þ cannot bethe single limiting point of L+ since this is the saddle point] [12, p. 49]. Then, let’sconsider the semitrajectory L� originating from the point ðx1; y1Þ in the opposite
direction of L+. Obviously, the boundary of O2 cannot be the limiting points of L�.Since there are no other singularity points and singular trajectories in the regionO2,
we have a contradiction.
By virtue of [12, p. 146], this statement is proved.
Proof of Assertion 1.1 Let’s first show that the semitrajectory of flux f originatingfrom any point of the set N for some t (t > 0) leaves N.
Let’s consider any semitrajectory originating in N. With t>0; the following twocases are possible: All points of the semitrajectory remain in N or for some value of
t, a point of the semitrajectory does not belong to N. In the first case, from (1.71)dpdt ¼ �a Q
M p of the system, it follows that for all t>0; the variable p(t) has a
derivative that is greater than some positive constant with Q<0 or less than some
negative constant with Q>0; i.e., p(t) grows unboundedly or tends to converge to
zero with unbounded growth of t. Therefore, the first case is impossible and the
orbit of any point in N leaves N.Since any chain-recurrent set Rðf ;NÞ lying inside N is the invariant set of this
flux, then, if it is nonempty, it only consists of the whole orbits. Hence, in our case
Rðf ;NÞ is empty. The statement follows from Robinson’s Theorem A [67].
1.5 Examples for Application of Parametric Control Theory 115
Chapter 2
Methods of Macroeconomic Analysisand Parametric Control of EquilibriumStates in a National Economy
Conducting a stabilization policy on the basis of the results of macroeconomic
analysis of a functioning market economy is an important economic function of the
state.
The AD–AS, IS, LM, IS-LM, IS–LM–BPmodels, as well as the Keynesian model
of common economic equilibrium for a closed economy and the model of a small
country for an open economy [39], include one of the efficient instruments of
macroeconomic analysis of the functioning of a national economy.
In previously published literature, we can see how these models are used for
carrying out a macroeconomic analysis of the conditions of equilibrium in national
economic markets. But there are no published results in the context of the estima-
tion of optimal values of the economic instruments on the basis of the Keynesian
model of common economic equilibrium and the model of an open economy of a
small country in the sense of some criteria, as well as an analysis of the dependence
of the optimal criterion value on exogenous parameters.
Based on the dependence of a solution to a system of algebraic equations on its
coefficients, we propose an approach to parametric control of the static equilibrium
of a national economy that reduces to the estimation of optimal values of economic
instruments as a solution of some respective mathematical program.
In this chapter, we construct IS, LM, and IS-LM models of the Keynesian all-
economic equilibrium and a small open national economy. We also present results
of macroeconomic analysis and parametric control of the static equilibrium of a
national economy.
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2_2,# Springer Science+Business Media New York 2013
117
2.1 Macroeconomic Analysis of a National Economic StateBased on IS, LM, and IS-LM Models, KeynesianAll-Economy Equilibrium. Analysis of the Influenceof Instruments on Equilibrium Solution
This section elaborates the construction of the IS, LM, and IS-LMmodels as well as
the Keynesian model of common economic equilibrium using an example of the
economy of the Republic of Kazakhstan, analysis of the influence of economic
instruments to the equilibrium conditions in the respective markets, as well as the
estimation of optimal values of economic instruments on the basis of the Keynesian
mathematical model of common economic equilibrium [39].
2.1.1 Construction of the IS Model and Analysisof the Influence of Economic Instruments
Let’s introduce the notation for economic indices used for model construction: Т is
the tax proceeds (to the state budget, in billions of tenge); S is the net savings,
billions of tenge; I is the investment in the capital asset, billions of tenge; G is the
public expenses, billions of tenge; Y is the gross national income, billions of tenge;
and C is the household consumption, billions of tenge.
Macroestimation of the equilibrium conditions in the wealth market can be done
on the basis of the IS model [39, p. 76] represented as
T þ S ¼ I þ G: (2.1)
The tax proceeds T to the state budget represented by the expressionT ¼ TyY hasthe following econometric estimation based on statistical information for the years
2000–2008:
T ¼ 0:2207 Y:
ð0:000Þ (2.2)
The statistical characteristics of model (2.2) are as follows: the determination
coefficient R2 ¼ 0.986; the standard error Se ¼ 209.5; the approximation coeffi-
cient A ¼ 10.47%; the Fisher statistics F ¼ 581.66. The statistical significance of
the coefficient of regression (2.2), as well as the regressions estimated below, is
given within parentheses under the respective coefficients of the regressions in the
form of p-values.
118 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
The net savings S represented by the expression S ¼ aþ SyY has the following
econometric estimation:
S ¼ �366:055þ 0:222 Y:
ð0:000Þ 0:000ð Þ (2.3)
The statistical characteristics of model (2.3) are as follows: The determination
coefficient R2 ¼ 0.994; the standard error Se ¼ 69.2; the approximation coefficient
A ¼ 11.47%; the Fisher statistics F ¼ 1287.2; the Durbin–Watson statistics DW ¼1.96.
The investment in capital assets represented by the expression I ¼ aþ Iiiafter estimating the parameters of this model using statistical information becomes
the following:
I ¼ 1367:9� 81:3 iþ 0:2751 Ymean:
ð0:02Þ ð0:03Þ 0:00ð Þ (2.4)
The statistical characteristics of model (2.4) are as follows: The determination
coefficient R2 ¼ 0.99; the standard error Se ¼ 126.8; the approximation coefficient
A ¼ 4.2%; the Fisher statistics F ¼ 326.48; the Durbin–Watson statistics DW ¼1.72. Substituting into (2.4) the value of the mean nominal gross national income
for the years 2000–2008 in billions of tenge, Ymean ¼ 6662:7 , finally yields the
following model for the investment:
I ¼ 3202� 81:3 i: (2.5)
Substituting expressions (2.2), (2.3), and (2.5) into (2.1), we obtain the IS model
representation in the following form:
�366:055þ 0:222Y þ 0:2207Y ¼ 3202� 81:3 iþ G200X; (2.6)
which allows us to determine the equilibrium value of i for the given values of Y and
G200Х. From macroeconomic theory, a method [39, p. 77] of plotting the IS curve
exists, which is the set of combinations of the equilibrium values of Y and i(Fig. 2.1).
From the model IS2007 (Fig. 2.1), it follows that the equilibrium GNI2007 with
interest rate 13.6% equals 11,602.75 billion tenge, and the real GNI2007 with interest
rate 13.6% equals 11,371 billion tenge, which shows a lack of wealth in the
considered market. From the model IS2008 (Fig. 2.1), it follows that the equilibrium
GNI2008 with interest rate 15.3% equals 13,957.91 billion tenge, while the real
GNI2008 with the interest rate 15.3% equals 13,734 billion tenge, which also
shows a lack of wealth within that market.
2.1 Macroeconomic Analysis of a National Economic State Based on IS. . . 119
To estimate the multiplicative effects [41, p. 78] of the economic instruments Tyand G, we’ll construct an econometric model of the consumption of households C,which, on the basis of statistical information for the years 2000–2008, is given by
C ¼ 428:68þ 0:552Yv;
0:000ð Þ 0:000ð Þ
where Yv ¼ Y � TyY;CYv ¼ 0:552: The statistical characteristics of this model are
as follows: The determination coefficient R2 ¼ 0.999; the standard error
SE ¼ 68.92; the approximation coefficient A ¼ 1.78%; the Fisher statistics
F ¼ 5394; the Durbin–Watson statistics DW ¼ 1.53.
Table 2.1 presents the expressions and values of the multipliers [39, p. 83] of the
instruments Ty and G derived on the basis of the IS model (2.6).
Let’s estimate the multiplicative effects of the instruments Ty and G based on the
data for the year 2008. According to that data, we have G ¼ 3859.98,
Y ¼ 13,734.3, and Ty ¼ 0.2207. Now, let’s change G to DG ¼ 579. This change,
in accordance with the multiplier of DG, results in an increment of GNI by the
value DY ¼ 1308.54.
0
5
10
15
20
25
30
0,00 3000,00 6000,00 9000,00 12000,00 15000,00 18000,00
i (interest rate)
Y (Gross National Income)
actual point 2007: GNI = 11371.07; i=13.6 actual point 2007: GNI = 11374.29; i=15.3
IS 2007 IS 2008
Fig. 2.1 Plots of IS2007 and IS2008 models
Table 2.1 Consequences of changing public expenses and taxation
Action
Consequence Public expensesincrease by DG Taxes decrease by DТ
National income increases by1
Tyþ SyDG ¼ 2.26 DG
Cyv
Tyþ SyDТ ¼ 1.3 DТ
Budgeted deficit increases by 1� Ty
Tyþ Sy
� �DG ¼ 0.5 DG 1� TyCyv
Tyþ Sy
� �DТ ¼ 0.7 DТ
120 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
Also, from the data of year 2008, we have G ¼ 3859.98, Y ¼ 13,734.3, and
Ty ¼ 0.2207. Let’s replace Ty by DTy ¼ �0.01. This change, in accordance with
the multiplier ofDTy, results in an increment of GNI by the value DY ¼ 328.37. The
derived results agree with the macroeconomic theory that considers the influence of
the economic instruments on the changes in the domestic national income, which is
represented by Table 2.1, “Consequences of changing public expenses and taxa-
tion” [39, p. 83].
2.1.2 Macroeconomics of Equilibrium Conditionsin the Money Market
The macroestimation of equilibrium conditions in the money market can be realized
on the basis of the LM model represented as follows [41, p. 111]:
M ¼ lpr þ ltr; (2.7)
where М is the money supply, in billions of tenge; lpr is the volume of property
(deposits in deposit organizations by sectors and currencies), in billions of tenge; lpris the volume of transaction [the volume of credits given by second-level banks
(SLB) taking into account the money velocity], in billions of tenge.
To estimate the money velocity, let’s use the Fisher equation [39, p. 112]:
MV ¼ Y;
where V is the money velocity, Y is the nominal GNI, and the money aggregateM3is accepted in the Fisher equation as the active money volume M.
Estimation of the money velocity by the expression V ¼ Y/M on the basis of the
statistical information for the years 2007–2008 is presented in Table 2.2.
The value of the money supply represented in the Fisher equation by the
aggregate M3 can be checked again through its estimation determined by yearly
values of the money base and the money multiplier m.The money multiplier m is defined by the following relation [39, p. 99]:
m ¼ 1þ gð1� a� bÞaþ bþ gð1� a� bÞ ;
Table 2.2 Value of the money aggregate M3 and the velocity of money
Year GNI Value of money aggregate М3 V, velocity of money
2007 11,371 4,629.8 2.5
2008 13,734 6,266.4 2.2
2.1 Macroeconomic Analysis of a National Economic State Based on IS. . . 121
where a ¼ RR/D is the normative of minimal reserve;
b ¼ ER/D is the coefficient of cash remainders of the commercial banks;
g ¼ CM/K is the share of money in cash in the total sum of credits of the
commercial banks;
RR the minimal reserves;
D is the check (current) deposits (we used the information about deposits in
the deposit organizations by sectors and currencies);
ER is the excess reserves;
K is the credits of the commercial banks accepted in accordance with the
expression K1/V;К1 is the statistical information about the given credits;
СM is the active money in cash.
Estimates of the money supply M by the money bases for the years 2007–2008
and values of m for the same period are respectively equal to the following: For the
year 2007,М ¼ mН ¼ 4519.9 billion tenge; for the year 2008,М ¼ mН ¼ 5343.6
billion tenge.
Table 2.4 presents the calculated values of the money supply and the values of
the money aggregate М3 by years. Table 2.4 shows that the calculated values ofMand values of the money aggregateM3 are of the same order and close to each other.
Taking into consideration this fact, together with the result on the money velocity
derived above, in this specific analysis we accept the calculated values as the money
supply, and the actual values of credits of the second-level banks are corrected
subject to the money velocity.
The property demand represented by the expression lpr ¼ eaþlii has the following
econometric estimate:
lpr ¼ 438883:3� 0:66i:
0:000ð Þ 0:01ð Þ (2.8)
The regression coefficients are statistically significant, although we have the
coefficient of determination R2 ¼ 0.33, the standard error Se ¼ 0.6, and the Fisher
Table 2.3 Values
of multipliersValues of multipliers
Year a b g Deposit Credit Money
2007 0.143 0.043 0.250 2.565 2.087 3.087
2008 0.045 0.069 0.252 2.969 2.632 3.632
Table 2.4 Calculated
values of money supply
and values of money
aggregate
Years
Calculated values
of money supply
Values of money
aggregate М3
2007 4,519.9 4,629.8
2008 5,343.6 6,266.4
122 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
statistics F ¼ 67. The demand for money for transactions represented by the
expression ltr ¼ a + bY describes the following econometric estimation:
ltr ¼ �1062:85þ 0:326 Y:
0:0005ð Þ 0:0000ð Þ (2.9)
The statistical characteristics of model (2.9) are as follows: The determination
coefficient R2 ¼ 0.965, the standard error SE ¼ 267, and the Fisher statistics
F ¼ 193.7.
Substituting expressions (2.8) and (2.9) into (2.7), we obtain the representation
of the LM model in the following form:
M200X ¼ 438883:3� 0:66i � 1062:85þ 0:326 Y; (2.10)
which allows us to determine the equilibrium value of i for the given values of Y and
M200X . In macroeconomic theory, a method exists [39, p. 113] to plot the LM curve,
which is the set of combinations of the equilibrium values of Y and i. Figure 2.2
presents the plots of the LM models for the years 2007 and 2008.
In accordance with the obtained results and plotted LM2007, LM2008, one can
conclude that the actual values of Y and i for the years 2007–2008 are situated abovethe respective curves LM2007, LM2008, which shows the relatively low demand for
the monetary assets.
The alarming aspect is that the actual state in which the money market found
itself in the year 2008 corresponds to a higher mean market interest rate than in the
year 2007, whereas the whole line LM for 2008 is situated below and to the right of
0
5
10
15
20
25
30
0,00 3000,00 6000,00 9000,00 12000,00 15000,00 18000,00 21000,00
i (interest rate)
Y (Gross National Income)
actual point 2007: GNI = 11371.07; i=13.6 actual point 2007: GNI = 11374.29; i=15.3
LM 2007 LM 2008
Fig. 2.2 Plots of LM2007 and LM2008 models
2.1 Macroeconomic Analysis of a National Economic State Based on IS. . . 123
the respective line for 2007; that is, the same volume of GNI corresponds to a lower
equilibrium interest rate than that of a year before. This is an indirect indicator that
the government has regulated the money market based on the necessity of making
money cheaper, but the second-level banks reacted to those signals in the opposite
way, raising the commercial rate. Exactly the same situation occurred in 2008 in
most developed countries on the threshold of the economic crisis.
2.1.3 Macroestimation of the Mutual Equilibrium Statein Wealth and Money Markets. Analysis of the Influenceof Economic Instruments
On the basis of the derived IS and LMmodels, the model for macroestimation of the
joint equilibrium state in the wealth and money markets can be represented by the
following system:
�366:055þ 0:222Y þ 0:2207Y ¼ 3202� 81:3iþ G200x;M200x ¼ 438833:3� 0:66i � 1062:85þ 0:326Y:
�(2.11)
The results of solving system (2.11) to estimate the joint equilibrium state in the
wealth and money market for the years 2007 and 2008 are presented in Table 2.5.
The plots of the IS and LM models in the same period are shown in Fig. 2.3.
From Fig. 2.3, it follows that the coordinates of the effective demand point for
years 2007 and 2008 are respectively represented by Y*2007 ¼ 11,670.89;
i*2007 ¼ 13.23, and Y*2008 ¼ 14,327.31; i*2008 ¼ 13.29. The points of the actual
state of the economy of the Republic of Kazakhstan in 2007 and 2008 are respec-
tively situated to the left of the corresponding IS2007 and IS2008 plots and above the
respective LM2007 and LM2008 plots. Such location of the points of the actual
economic state means a respective lack of wealth market and an excess of money
in the money market in 2007 and 2008.
Let’s estimate the influence of the instruments G andM on the joint equilibrium
conditions using the data for the year 2008.
By the results of the solution of system (2.11), on the basis of the data from 2008,
we have that G ¼ 3,859.98 andM ¼ 5,343.6. Let’s now increase G by DG ¼ 579.
With unchangedM, this fluctuation results in an increase of the Keynesian effective
demand – GNI up to 15,522 billion tenge and an increase of the interest rate up to
Table 2.5 Joint equilibrium and actual values of Y and i
Actual values Joint equilibrium conditions
i, interest rateof SLB, %
Y, gross domestic
income, billion tenge i*Y*, Keynesian effective
demand
2007 13.6 11,371.1 13.23 11,670.89
2008 15.3 13,734.3 13.29 14,327.31
124 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
13.9% due to the shift of IS to the right as a result of the multiplicative effect from
increasing the public expenses.
Let’s now increaseМ2008 by DМ ¼ 534. With unchanged G2008, this fluctuation
results in an increase of GNI up to 14,438.6 billion tenge and a decrease of the
interest rate to 12.7% due to the shift of IS to the right as a result of the multiplica-
tive effect from increasing the money supply.
The obtained results also agree with the macroeconomic theory on the influence
of the economic instruments in the wealth and money markets [39, pp. 78, 114].
2.1.4 Macroestimation of the Equilibrium State on the Basis ofthe Keynesian Model of Common Economic Equilibrium.Analysis of the Influence of Economic Instruments
The Keynesian mathematical model of common economic equilibrium on the basis
of the IS and LM models, as well as the econometric function of the labor supply
price and the econometric expression of the production function, is given by the
following [39, p. 223]:
TðYÞ þ SðYÞ ¼ IðiÞ þ G; ð2:12ÞM ¼ lðY; iÞ; ð2:13ÞWSðN;PÞ ¼ PYN; ð2:14ÞY ¼ YðNÞ; ð2:15Þ
ð2:12Þ
8>>>>><>>>>>:
0
5
10
15
20
25
30
0,00 5000,00 10000,00 15000,00 20000,00 25000,00
i (interest rate)
Y (Gross National Income)
IS 2007LM 2008
actual point 2007: GNI = 11371.07; i=13.6 actual point 2007: GNI = 11374.29; i=15.3IS 2008 LM 2007
Fig. 2.3 Plots of IS2007, LM2008, LM2007, and LM2008 models
2.1 Macroeconomic Analysis of a National Economic State Based on IS. . . 125
where Ws (N, P) is the function representing the labor supply price, YN is the
derivative of the production function, and Y(N) is the production function.
Equations (2.12) and (2.13) of the common economic equilibrium model are
given by the respective IS and LM equations (2.11).
The econometric representation of the labor supply price using the statistical
data for the years 2000–2008 is given by
Ws N;Pð Þ ¼ 60:12P� 0:007N;
ð0:000Þ ð0:000Þ (2.16)
where P is the level of prices for year 2000, and N is the busy population in
thousands per capita. The respective p-values (of t-statistics) in the equation in
Ws are presented in parentheses below the regression coefficients. The results of the
analysis of the statistical significance of the model for Ws are as follows: The
determination coefficient R2 ¼ 0.99, the standard error Se ¼ 3.37, the Fisher
statistics F ¼ 522.6, and the approximation coefficient A ¼ 7.4%.
The econometric representation of the production function Y(N) using the
statistical data for the years 2000–2008 is given by
Y ¼ �5:654 N þ 0:0009 N2:
0:000ð Þ 0:000ð Þ (2.17)
The results of analysis of statistical significance of the model for Ws are as
follows: The determination coefficient R2 ¼ 0.98, the standard error Se ¼ 1227,
and the Fisher statistics F ¼ 172.
The Keynesian model of common economic equilibrium on the basis of relations
(2.11), (2.16), and (2.17) is given by
�366:055þSyYþTyY¼ 3202� 81:30iþG200X;
M200X¼ 438883:3� 0:66i � 1062:85þ 0:326Y;
60:12P� 0:00698N¼�5:65Pþ0:0018NP;
Y¼�5:65Nþ0:0009N2:
8>>>><>>>>:
(2.18)
In this system describing the behavior of the macroeconomic subjects, the
exogenously given parameters include the value of public expenses G and the
nominal values of the money in cashM. The values of five endogenous parameters,
Y*, i*, P*, N*, and W*, that result in attaining equilibrium simultaneously in all
three mentioned markets are determined from the solution of this system of
equations.
Substituting the actual values of G200X and М200X of the respective year and
solving system (2.18), we obtain the values of variables that are in equilibrium
simultaneously in all three markets.
126 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
Table 2.6 presents the equilibrium values of the endogenous parameters by using
the solution of system (2.18) on the basis of the data for the years 2007 and 2008.
Let’s estimate the influence of instruments G and M on the Keynesian common
economic equilibrium from the data from 2008.
Increasing G by DG ¼ 579 while keeping the values of M results in an increase
of the GNI to 15,522.6 billion tenge and a decrease in the interest rate to 13.9%,
while at the same time unemployment drops by 1.6% and the level of prices
increases to 1.12.
Increasing М2008 by DМ ¼ 534.4 while keeping the values of G results in an
increase of the GNI to 14,438.56 billion tenge and a decrease of the interest rate to
12.68%, while unemployment is reduced by 0.15%, and the level of prices increases
insignificantly to 1.105.
Increasing G by DG ¼ 579 and increasing М2008 by DМ ¼ 534.4 result in an
increase of the GNI to 15,658.85 billion tenge and a decrease of the interest rate to
13.15%, while unemployment is reduced by 1.77% and the level of prices increases
to 1.13.
2.1.5 Parametric Control of an Open Economic State Basedon the Keynesian Model
Estimation of the optimal values of the instruments M and G for the given external
exogenous parameters Sy, Тy on the basis of model (2.18) for the year 2008 in the
sense of the GNI criterion gives
Y ! max : (2.19)
This estimate can be obtained by solving the following mathematical program-
ming problem.
Problem 1On the basis of mathematical model (2.18), find the values of (M, G) maximizing
criterion (2.19) under the constraints
Table 2.6 Comparative
analysis of actual and
equilibrium values of GNI,
interest rate, level of prices,
busy population
Y i P N
2007 Actual 11,371.10 13.60 1.789 7,631.10
Equilibrium 11,670.89 13.23 1.050 7,751.60
Deviation 2.64% �0.37 �0.740 1.58%
2008 Actual 13,734.30 15.30 1.959 7,857.20
Equilibrium 14,327.30 13.30 1.103 8,048.80
Deviation 4.32% �2.00 �0.900 2.44%
2.1 Macroeconomic Analysis of a National Economic State Based on IS. . . 127
M �M�j j � 0:1M�;
G� G�j j � 0:1G�;
N � N�j j � 0:1N�;
P� P�j j � 0:1P�;
i� i�j j � 0:1i�;
Y � Y�j j � 0:1Y�:
8>>>>>>>>>>><>>>>>>>>>>>:
(2.20)
Here M* and G* are the respective actual values of the money and public expenses
supplies in 2008. The symbol (*) for the unknown variables of system (2.20)
corresponds to the equilibrium values of these variables with fixed values of
M* and G*.For Problem 1, the optimal values of the parameters are M ¼ 5877.96,
G ¼ 4245.98, which ensure attaining the maximum value of the criterion
Y ¼ 15,255.9. The value of this criterion without control is equal to 14,327.3. For
the optimal values of the instruments M and G that were obtained, the equilibrium
values of the other endogenous variables turn out to be N ¼ 8148.539, P ¼ 1.1210,
and i ¼ 12.986. Here we should also note that solving this optimization problem
results in an increase of the working segment of the population by approximately
100,000 people.
On the basis of Problem 1, we carry out the analysis of the dependence of the
optimal values of criterion Y on the pair of the exogenous parameters Ty; Sy� �
given
in their respective regions. The obtained plot of the optimal values of criterion
(2.19) is presented in Fig. 2.4.
Fig. 2.4 Plot of dependence of optimal values of criterion Y on parameters Ty, Sy
128 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
2.2 Macroeconomic Analysis and Parametric Controlof the National Economic State Based on the Modelof a Small Open Country
Ensuring a double equilibrium, that is, a common economic equilibrium in
conditions of full employment with a planned (assumed zero) balance of payments,
is an urgent problem in the conditions of an open economy, when the country is
engaged in the exchange of goods and capital with the outside world.
All of the remaining states in the national economy differing from the double
equilibrium represent various kinds of nonequilibrium states. Hence, unemployment
remains the same in spite of an excess in the balance of payments. Unemployment
can be accompanied by an excess in the balance of payments. The excess of
employment can be accompanied by both the excess and deficiency of the balance
of payments. Therefore, public economic policy aims at attainment of a double
equilibrium. The estimation of the equilibrium conditions for an open economy can
be partially considered on the basis of the model of a small country [39, p. 433].
This section is devoted to the construction of a mathematical model of an open
economy of a small country using the example of the Republic of Kazakhstan, to
the analysis of the influence of economic instruments on the conditions of common
economic equilibrium and state of the balance of payments, and to the estimation of
the optimal values of the economic instruments on the basis of the model of an open
economy of a small country, as well as an analysis of the dependencies of the
optimal values of the criteria on the values of one, two, and three parameters from
the set of the external economic parameters given in the respective regions.
2.2.1 Construction of the Model of an Open Economyof a Small Country and Estimationof Its Equilibrium Conditions
Let’s introduce the following notation for the economic indices used for the model
construction: Y is the gross national income (GNI); C is the household consump-
tion; I is the investment in capital assets; G is public expenses; NE is the net export
of wealth; P is the level of prices of RK; Pz is the level of prices abroad; l is the realcash remainder; I is the interest rate of second-level banks; N is the number of
employed; dY/dN is the derivative of the gross national income as a function of the
number of employed; WS is the level of wages; NKE is the net capital export; e isthe rate of exchange of the national currency; ее is the expected rate of exchange ofthe national currency; e
_eis the expected rate of increase of the exchange rate of the
national currency [39, p. 121];M is the money supply determined from [39, p. 412]
by the formula М ¼ mН, where Н is the money base of each year; m is the money
multiplier calculated from the balance equations of the banking system and defined
by the formula
2.2 Macroeconomic Analysis and Parametric Control of the National Economic. . . 129
m ¼ 1þ g 1� a� bð Þð Þ= aþ bþ g 1� a� bð Þð Þ; (2.21)
where a ¼ RR/D is the norm of the minimal reserve;
b ¼ ER/D is the coefficient of the cash remainder of the second-level banks;
g ¼ СМ/К is the share of cash in the whole sum of the credits of second-level
banks;
RR is the minimal reserve;
ER is the excessive reserve;
D is the check deposits;
CM is the active money in cash;
K is the credits of second-level banks corrected subject to the velocity of
money.
Let’s begin to construct a mathematical model of an open economy of a small
country by estimating the money multiplier, real cash remainders, and economic
functions characterizing the national economic state.
The estimations of values of the money multiplier calculated by formula (2.21)
using the statistical data for the period of years 2006–2008 are presented below:
Year 2006 2007 2008
m 2.372 3.087 3.632
The real cash remainder l is determined by the formula
l ¼ lpr þ ltr; (2.22)
where lpr is the property volume [deposits in the deposit organizations (by sectors
and kinds of currency)], billions of tenge, and ltr is the volume of the transaction (the
volume of the credits given by second-level banks subject to the money velocity),
billions of tenge.
The estimation of the money velocity is calculated by the Fisher equation [42]:
MV ¼ Y;
where V is the money velocity, and M is the quantity of the active money usually
represented by the money aggregate M3 in the Fisher equation.
From the latter formula, the estimates of the money velocity calculated by the
formula V ¼ Y=M on the basis of the statistical information for 2006–2008 [37] are
presented in Table 2.7.
In the macroeconomic theory, the behavior of the national economy is
characterized by the following functions constructed by econometric methods [1]
on the basis of official statistical information.
The consumption C represented by the expression C ¼ aþ CYY has the follow-
ing econometric estimation derived on the basis of the statistical information of the
Republic of Kazakhstan for the period 2000–2008:
C ¼ 555:8þ 0:4101Y:
0:00ð Þ 0:00ð Þ (2.23)
130 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
The statistical characteristics of the constructed model of the consumption C are
as follows: The determination coefficient R2 ¼ 0.994, and the approximation
coefficient A ¼ 1.8%. The statistical significance of the coefficients of regression
(2.23), as well as of the regressions estimated below, is presented in parentheses
under the respective regression coefficients as the p-values.The consumption of the imported wealth Qim is represented by the regression
equation
Qim ¼ a1Y þ b1ePZ=P
or, in estimated form,
Qim ¼ 0:4076Y � 2:6059ePZ=P
0;00ð Þ 0;17ð Þ (2.24)
with the determination coefficient R2 ¼ 0.975 and the approximation coefficient
A ¼ 10%.
The model of the demand of the real cash remainder is given by l ¼ a2 þ b2Yþ b3iþ b4e or, after estimating the parameters of this model using the statistical
information,
l ¼1:3320Y � 138:1i � 22:6e:
0;01ð Þ 0;08ð Þ 0;03ð Þ (2.25)
In constructing model (2.25), the values of l calculated in accordance with
formula (2.22) are accepted as the data for the left-hand side. The determination
coefficient is given by R2 ¼ 0.999, and the approximation coefficient is A ¼ 0.2%.
The statistical insignificance of the latter model concerns the fact that in the model
there are correlated factors.
The model of the labor supply price is given by WS ¼ b5N þ b6Pmean; where
Pmean ¼ 1� að ÞPþ aePZ=e0 has the following econometric estimation derived on
the basis of the statistical information:
WS ¼ �0:0219N þ 156:9 Pcp;
0;00ð Þ 0;00ð Þ (2.26)
Table 2.7 Values of GNI (billions of tenge), money aggregateM3 (billions of tenge), and money
velocity V
Year GNI М3 V
2007 11,371 4,629.8 2.5
2008 13,734 6,266.4 2.2
2.2 Macroeconomic Analysis and Parametric Control of the National Economic. . . 131
where Pmean ¼ 0:6Pþ 0:4ePz=e0; e0 is the currency exchange rate within the base
period (year 2000), and a is the share of the imported goods in their entire volume
accepted at the level of 0.4. We also have the determination coefficient R2 ¼ 0.98
and the approximation coefficient for 2007–2008 at the 5% level.
The model of the net capital export is given by NKE ¼ b7eðiZ þ e_e � iÞ or, after
estimating the parameters of this model by using the statistical information,
NKE ¼ � 0:3349eðiZ þ e_e � iÞ;
0;00ð Þ (2.27)
with the determination coefficient R2 ¼ 0.51.
The production function is represented in the regression pair Y ¼ a3 + b8 N or,
in the estimated form,
Y ¼ � 17;409:0þ 3:0866N;
0;00ð Þ 0;00ð Þ (2.28)
with the determination coefficient R2 ¼ 0.93 and the coefficient approximated
based on 2007–2008 data not exceeding A = 3.5%.
The model of investment in capital assets is given by
It ¼ a4 þ b9Yt�1 þ b10it;
where It and it are the values of the investments in the current period, and Yt-1 is thevalue of the gross national income in the preceding period.
After estimating the latter model parameters by the statistical data, the following
expression is derived:
It ¼ 862:8þ 0:3122 Yt�1 � 48:4it:
0:34ð Þ 0:00ð Þ 0:41ð Þ (2.29)
At that, the determination coefficient R2 ¼ 0.93 and the approximation coeffi-
cient A ¼ 5%.
Substituting the value Yt�1 ¼ Y2007 to (2.29), finally we obtain the following
model of investment in the year 2008:
I2008 ¼ 2;846:7� 48:4i: (2.30)
Similarly, substituting the value Yt�1 ¼ Y2006 into (2.29) for investment in 2007,
we obtain the following model:
I2007 ¼ 2;737:3� 48:4i: (2.31)
132 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
The wealth export model is a regression of the form Qex ¼ b11ePZ P= . After
estimating the parameters, this model becomes
Qex ¼ 17:87ePZ P= :
0; 01ð Þ (2.32)
The determination coefficient is R2 ¼ 0.65.
On the basis of derived econometric estimate (2.23), (2.24), (2.25), (2.26),
(2.27), (2.28), (2.29), (2.30), (2.31) and (2.32) characterizing the state of the
national economy, let’s proceed to the construction of a model of an open economy
of a small country for year 2008.
Within the framework of the IS curve, we constructed the function Y ¼ Cþ IþGþ Qex � Qim; which, subject to (2.23), (2.24), (2.29), (2.30), (2.31) and (2.32),becomes
Y ¼ 555:8þ 0:4101 Y þ 2;846:7� 48:4iþ Gþ 20:47e PZ=P� 0:4076 Y or
Y ¼ 3;410:9� 48:4iþ 20:52ePZ P= þ 1:0024G:
(2.33)
The equation of the LM line M/P ¼ l subject to the econometric model (2.25)
becomes
M=P ¼ 1:3320Y � 138:1iþ 22:6e;
from which one can derive the following relation:
i ¼ �0:1640eþ 0:0096Y � 0:0072M=P (2.34)
Substituting (2.34) into (2.33), we obtain the value of the aggregate demand YD:
YD ¼ 2;324:0þ 5:42eþ 13:98 ePZ P= þ 0:6830Gþ 0:2392M P= : (2.35)
Let’s substitute (2.33) into (2.34) and determine the function of the domestic
commercial interest rate:
i ¼ 22:42� 0:1117e� 0:0049M=Pþ 0:1349ePZ=Pþ 0:0065G (2.36)
The condition of equilibrium in the labor market is given by Р dY/dN ¼ WS [41,
p. 435], which, subject to the econometric functions (2.26) and (2.28), can be
represented by the expression
7; 5P ¼ �0:0219N þ 156:87ð0:6Pþ 0:4ePZ=e0Þ: (2.37)
2.2 Macroeconomic Analysis and Parametric Control of the National Economic. . . 133
From (2.37) we obtain the following relation for N:
N ¼ 4;165:8Pþ 20:20ePZ (2.38)
Substituting expression (2.38) into the production function (2.28), we obtain the
function of the aggregate supply:
YS ¼ �17; 408:6þ 12; 858:2Pþ 62:36ePZ (2.39)
The balance of payments has a zero balance if the net wealth export equals the
net capital export, i.e., NE ¼ NKE/P, is valid. The econometric representation of
the latter equality on the basis of (2.24), (2.27), and (2.32) is given by
17:87ePZ P= � 0:4076Y � 2:6059ePZ P=� � ¼ �0:3349eðiZ þ e
_e � iÞ=P:
Substituting the value of the domestic interest rate (2.36) into the latter equality,
after some transformation we obtain the following equation of the curve of the zero
balance of payments:
Y ZBO ¼ 50:23 ePZ P= � 0:8215 eiZ P= � 0:8215 ee P= þ 19:24 e P=
� 0:0918 e2 P= � 0:0041 eM P= 2 þ 0:111 e2PZ P=2 þ 0; 005 eG P= :
(2.40)
Thus, the model of an open economy of a small country in the year 2008 is given
by the following system of equations:
YD ¼ 2324:0þ 5:42eþ 13:98ePZ
Pþ 0:6830Gþ 0:2392
M
P;
YS ¼ �17408:6þ 12858:2Pþ 62:36ePZ;
YZBO ¼ 50:23ePZ
P� 0:8215
eiZ
P� 0:8215
ee
Pþ 19:24
e
P� 0:0918
e2
P�
�0:0041eM
P2þ 0:111
e2PZ
P2þ 0:005
eG
P;
YD ¼ YS ¼ YZBO:
8>>>>>>>>>>>>><>>>>>>>>>>>>>:
(2.41)
Similarly to (2.41), a model of an open economy of a small country in 2007 can
be constructed.
134 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
Solving system (2.41) with prescribed values of the external economic indexes
PZ, iZ, ее and the economic instruments M and G, we’ll determine the equilibrium
conditions of the gross national income Y� ¼ YD ¼ YS ¼ YZBO, level of prices P*,
and exchange rate of the national currency е*. The equilibrium values of the credit
interest rate of the second-level banks i* and the number of employed are calculated
by formulas (2.36) and (2.38), respectively.
The following equilibrium values of the endogenous variables are obtained by
solving system (2.41) for the given external uncontrolled economic indices PZ, iZ, ее
and the controlled economic instruments M and G:
– by year 2007: Y� ¼ 6;383:1; P� ¼ 1:2054; e� ¼ 110:2; i� ¼ 16:4; N� ¼ 7;708:0– by year 2008: Y� ¼ 6;785:4; P� ¼ 1:2099; e� ¼ 114:7; i� ¼ 14:7; N� ¼ 7;838:4:
Figure 2.5 presents the double equilibrium state, where the point of intersection
of the IS–LM–ZBO curves corresponds to a simultaneous equilibrium in the wealth,
money, and labor markets with full employment and zero balance of payments in
2007. All combinations of the values of the national income and interest percent,
except i ¼ 14.7%, Y ¼ 6785.4, offer different types of non-equilibrium states.
In 2008 Kazakhstan, as per the presented figure, also has unemployment and deficit
of balance payments. In the graph of Fig. 2.2.1 this situation is represented by point
A (Y2008 ¼ 70098.0; i2008 ¼ 15.3%). However one can note that according to
official statistics in 2008 Kazakhstan had a surplus of balance payments.
Taking into account the obtained equilibrium values, the equilibrium values of
the economic indices C, I, and others calculated by econometric models are
constructed above. We present the results of comparison of the equilibrium
indices with the actual values of these indices in 2007. Table 2.8 shows similar
results for 2008.
−150
−100
−50
0
50
100
150
200
IS LM ZBO A
0 2000 4000 6000
A
8000 10000 12000 14000
Fig. 2.5 Double equilibrium by 2008
2.2 Macroeconomic Analysis and Parametric Control of the National Economic. . . 135
2.2.2 Influence of Economic Instruments on EquilibriumSolutions and Payment Balance States
Below, we’ll estimate the influence of economic instruments, namely, the money
supply and public expenses, on the conditions of common economic equilibrium
and the state of the balance of payments using the following algorithm:
1. Changing the value М2008 by DM ¼ 0.01 M2008 while keeping the values G2008
and iZ2008, PZ2008, ее2008 unchanged, define the values ðMDY�Þ ðY�DMÞ= ;
ðMDP�Þ ðP�DMÞ= ; ðMDe�Þ ðe�DMÞ= ; and ðMDi�Þ ði�DMÞ= that show the per-
centage by which the equilibrium values of the indicesY�; P�; e� i� change withvariation of М2008 by 1%.
2. Changing the value G2008 by DG ¼ 0:01G2008 while keeping the values M2008
and iZ2008, PZ2008, ее2008 unchanged, define the values ðGDY�Þ ðY�DGÞ= ;
ðGDP�Þ ðP�DGÞ= ; ðGDe�Þ ðe�DGÞ= ; and ðGDi�Þ ði�DGÞ= that show the percent-
age by which the equilibrium values of the indices Y�; P�; e�; i� change with
variation of G2008 by 1%.
3. Changing the value М2008 by DM ¼ 0.01M2008 and the value G2008 by DG ¼0:01G2008 while keeping the values iZ2008; PZ
2008; ee2008 unchanged, define thevalues 100DY� Y�= ; 100DP� P�= ; 100De� e�= ; and 100Di� i�= that show the
percentage by which the equilibrium values of the indices Y�; P�; e�; i� changewith simultaneous variation of М2008 and G2008 by 1%.
The results of computations carried out by the above algorithm are given in
Tables 2.9, 2.10, and 2.11.
According to the proposed algorithm, first we estimate the influence of the
economic instruments, namely, the money supply and public expenses, on the
conditions of the common economic equilibrium and the state of the balance of
payment individually. From Tables 2.9 and 2.10, it follows that increasing G2008 by
DG while keeping the value М2008 results in growth of the national income and an
Table 2.8 Equilibrium and actual values of indices in 2008
2008
Indices
Equilibrium
value of Y*Equilibrium
value of Yactual
Deviation Yactual � Y∗
Absolute %
Level of prices P 1.2099 1.96 0.7501 38.8
Currency exchange rate e 114.6 120.3 5.6687 4.7
Interest rate of SLB i 14.7 15.3 0.6246 4.1
National income Y 6785.4 7009.8 224.4 3.2
Consumption C 3338.2 3395.1 56.9 1.7
Import Qim 2467.5 2326.4 �141.2 �6.1
Investment I 2137.043 2149.2 12.2 0.6
Export Qex 2045.384 4370.6 2325.2 53.2
136 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
increase in the interest rate, whereas increasingМ2008 byDМwhile keeping the value
G2007 also results in growth of the common economic equilibrium of the GNI, but it
also results in a decrease in the interest rate. Also, from the tables it follows that the
growth in public expenses has a stronger influence on the national income growth,
whereas the money supply growth affects the currency exchange rate more strongly.
Here Y�; P�; e�; i� are the equilibrium solutions for the year 2008, DY� ¼ YM�
� Y�; DP� ¼ PM� � P�; De� ¼ eM
� � e�; Di� ¼ iM� � i�; where YM
*, PM*, eM
*,
iM* are the equilibrium solutions corresponding to M ¼ M2008 þ DM:According to the macroeconomic theory, the money supply growth shows the
following influence on the equilibrium solutions of system (2.41): The national
income, level of prices, and national currency exchange must increase, whereas the
interest rate must decrease. The results of the influence of the money supply
instrument on the equilibrium state of the national economy in 2008 presented in
Table 2.9 coincide with the theoretical assumptions except the price-level index,
which in this case decreases.
Here DY� ¼ YG� � Y�; DP� ¼ PG
� � P�; De� ¼ eG� � e�; Di� ¼ iG
� � i�;where YG
�; PG�; eG
�; iG� are the equilibrium solutions corresponding to
G ¼ G2008 þ DG:According to macroeconomic theory, the public expenses growth exerts the
following influence on the equilibrium solutions of system (2.41): The national
income, level of prices, national currency exchange rate, and interest rate must
grow. The results of the money supply instrument influence on the equilibrium state
of the national economy in 2008 presented in Table 2.9 completely coincide with
these theoretical assumptions.
Table 2.9 Influence of the money supply instrument on the equilibrium state of national economy
in 2007 for DM ¼ 0:01M2008 %ð ÞðMDY�Þ ðY�DMÞ= ðMDP�Þ ðP�DMÞ= ðMDe�Þ ðe�DMÞ= ðMDi�Þ ði�DMÞ=
0.359 �0.1128 0.4827 �1.4692
Table 2.10 Influence of the public expenses instrument on the equilibrium state of national
economy in 2007 for DG ¼ 0:01G2008 %ð ÞðGDY�Þ ðG�DMÞ= ðGDP�Þ ðP�DGÞ= ðGDe�Þ ðe�DGÞ= ðGDi�Þ ði�DGÞ=
0.1892 0.0345 0.0865 0.8017
Table 2.11 Influence of money supply and public expenses instruments on the equilibrium state
of the national economy in 2007 for DM ¼ 0:01M2008 and DG ¼ 0:01G2008 %ð Þ100DY� Y�= 100DP� P�= 100De� e�= 100Di� i�= :
0.5477 �0.0780 0.5675 �0.6434
2.2 Macroeconomic Analysis and Parametric Control of the National Economic. . . 137
Here DY� ¼ YMG� � Y�; DP� ¼ PMG
� � P�; De� ¼ eMG� � e�; Di� ¼ iMG
� � i�;where YMG
�; PMG�; eMG
�; iMG� are the equilibrium solutions corresponding to
M ¼ M2008 þ DM and G ¼ G2008 þ DG:Figures 2.7 and 2.8 present the plots of the IS, LM, and ZBO curves from the
derived econometric models for the actual statistical information for 2008.
As stated above (Fig. 2.5), the country has cyclical unemployment and a deficit
in balance of payments from the constructed models. In Fig. 2.6, such a situation is
represented by point Е0. According to the macroeconomic theory, the balance of
payments deficit can be eliminated by applying a restrictive monetary policy by
means of shifting the LM curve to the left up to its intersection with the IS curve at
point C, or the counteractive fiscal policy by means of the IS curve to the left up to
its intersection with the LM curve at point D.
EoC
D
−150
−100
−50
0
50
100
150
200
0 2000 4000 6000 8000 10000 12000 14000
IS LM ZBO Eo C D
Fig. 2.6 Plots IS–LM–ZBO by actual values of P, e for 2008
Fig. 2.7 Plot of the dependence of optimal values of criterion Qimp on pair PZ ; ee
138 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
2.2.3 Parametric Control of an Open Economy State Basedon a Small Country Model
Estimate the optimal values of instruments M and G given the external exogenous
parameters ee, iZ, PZ on the basis of model (2.41) for the year 2008 in the sense of
the criteria
Qex ¼ aePZ=P ! max (2.42)
and
Qimp ¼ bYS þ cePZ=P ! min: (2.43)
Such an estimate can be obtained by solving the following problems of mathe-
matical programming:
Problem 1On the basis of mathematical model (2.41), find the values (M, G) maximizing
criterion (2.42) under the constraints
Fig. 2.8 Plot of the dependence of optimal values of criterion Qex on pair PZ ; ee
2.2 Macroeconomic Analysis and Parametric Control of the National Economic. . . 139
M �M�j j � 0:1M�;
G� G�j j � 0:1G�;
P� P�j j � 0:1P�;
e� e�j j � 0:1e�;
i� i�j j � 0:1i�;
Y � Y�j j � 0:1Y�:
8>>>>>>>>>>><>>>>>>>>>>>:
(2.44)
Here M� and G� are the actual values of the money supply and public expenses in
the year 2008.
Problem 2On the basis of mathematical model (2.41), find the values (M, G) minimizing
criterion (2.42) under constraints (2.44).
Solving Problems 1 and 2 by the iterative technique [66] given the values
ee ¼ 120:3; iZ ¼ 1:32; PZ ¼ 1:2002; the following results are obtained:
For Problem 1, the optimal values of the parameters are M ¼ 5877.96,
G ¼ 4246, providing the attainment of the maximum value Qex ¼ 3122:74: Thevalue of this criterion without control is 3023.01.
For Problem 2, the optimal values of the parameters are M ¼ 4809.234,
G ¼ 3474, providing the attainment of the minimum value Qimp ¼ 4010:64:The value of this criterion without control is 4183.73.
On the basis of Problems 1 and 2, we carried out the analysis of the dependencies
of the optimal values of the criteriaQex andQimp on the one pair and one set of three
of the parameters from the set of the external parameters ee; iZ;PZf g given within
the respective regions. The plots of the dependencies of the optimal values of
criteria (2.42) and (2.43) for the single cases including that on the pair of the
parameters ðPZ; eeÞ and iZ; eeð Þ are shown in Figs. 2.7 and 2.8.
140 2 Methods of Macroeconomic Analysis and Parametric Control of Equilibrium. . .
Chapter 3
Parametric Control of Cyclic Dynamicsof Economic Systems
The theory of market cycles is an important part of modern macroeconomic
dynamics. This theory is based on mathematical models [39] proposed for describ-
ing the evolution of business activity as an oscillatory processes. Readers can find a
number of mathematical models of market cycles in [20]. In this context, the main
factors causing oscillations in market tendencies are considered. Nevertheless,
issues of the structural stability of such models of parametric control of develop-
ment of the economic systems on the basis of mathematical models of business
cycles are not under consideration.
Developing a theory of business cycles is of great interest, including estimation
of the structural stability of mathematical models of business cycles and parametric
control of the evolution of economic systems based on the proposed mathematical
models.
This chapter is devoted to results in the theory of business cycles based on
mathematical models, namely, the Kondratiev cycle model [17] and the Goodwin
model [5, 39].
3.1 Mathematical Model of the Kondratiev Cycle
3.1.1 Model Description
This model [17] combines descriptions of nonequilibrium economic growth and
nonuniform scientific and technological advancement. The model is described by
the following system of equations, including two differential equations and one
algebraic equation:
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2_3,# Springer Science+Business Media New York 2013
141
nðtÞ ¼ AyðtÞa;dx=dt ¼ xðtÞðxðtÞ � 1Þðy0n0 � yðtÞnðtÞÞ;
dy=dt ¼ nðtÞð1� nðtÞÞyðtÞ2 xðtÞ � 2þ mþ l0n0y0
� �;
n0 ¼ Ay0a:
8>>>>>><>>>>>>:
(3.1)
Here t is the time (in months), x is the efficiency of innovations, y is the capital
productivity ratio, y0 is the capital productivity ratio corresponding to the equilib-
rium trajectory, n is the rate of savings, n0 is the rate of saving corresponding to theequilibrium trajectory, m is the coefficient of withdrawal of funds, l0 is the job
growth rate corresponding to the equilibrium trajectory, and A and a are some
model constants.
Preliminary estimation of the model parameters is carried out based on statistical
information from the Republic of Kazakhstan for the years 2001–2005 [24].
The deviations in the observed statistical data and the calculated data do not exceed
1.9% within the considered period.
As a result of solving the problem of the preliminary estimation of parametric
identification, the following values of the exogenous parameters are obtained:
a ¼ �0:0046235 , y0 ¼ 0:081173 , n0 ¼ 0:29317 , m ¼ 0:00070886 , l0 ¼ 0:00032161; andxð0Þ ¼ 1:91114.
A preliminary prediction for 2006 and 2007 is characterized by errors equal to
6.1% and 12.1%, respectively, for the capital productivity ratio, and 2.3% and 11%,
respectively, for the rate of savings.
The respective cyclic phase trajectory of the Kondratiev cycle model is
presented in Fig. 3.1.
The period of cyclic trajectory corresponding to the statistical information of the
Republic of Kazakhstan for the given years is estimated to be 232 months.
Phase trajectory
Fig. 3.1 Cyclic phase trajectory of the Kondratiev cycle model
142 3 Parametric Control of Cyclic Dynamics of Economic Systems
3.1.2 Estimating the Robustness of the Kondratiev CycleModel Without Parametric Control
The estimation of structural stability (robustness) of the mathematical model is
carried out according to Sect. 1.4 on parametric control theory (in Ch. 1) in the
chosen compact set of the model state space.
Figure 3.2 presents an estimate of the chain-recurrent set R f ;Nð Þ obtained by
the application of the chain-recurrent set estimation algorithm for the region N ¼½1:7; 2:3� � ½ 0:066; 0:098� of the phase plane Oxy of system (3.1). Since the set Rf ;Nð Þ is not empty, we can draw no conclusions about the weak structural stability
of the Kondratiev cycle model in N on the basis of Robinson’s theorem. However,
since there is a nonhyperbolic singular point in N, namely, the center
x0 ¼ 2� mþ l0n0y0
; y0
� �[16], then system (3.1) is not weakly structurally
stable in N.
Chain-recurrent set
Fig. 3.2 Chain-recurrent set for the Kondratiev cycle model
3.1 Mathematical Model of the Kondratiev Cycle 143
3.1.3 Parametric Control of the Evolution of EconomicSystems Based on the Kondratiev Cycle Model
Choosing the optimal laws of parametric control is carried out in the environment of
the following four relations:
1. n0ðtÞ ¼ n0� þ k1
yðtÞ � yð0Þyð0Þ ;
2. n0ðtÞ ¼ n0� � k2
yðtÞ � yð0Þyð0Þ ; (3.2)
3. n0ðtÞ ¼ n0� þ k3
xðtÞ � xð0Þxð0Þ ;
4. n0ðtÞ ¼ n0� � k4
xðtÞ � xð0Þxð0Þ :
Here ki is the scenario coefficient, and n0� is the value of the exogenous parameter
n0 obtained as a result of the preliminary estimation of parameters.
The problem of choosing the optimal law of parametric control at the level of the
econometric parameter n0 can be formulated as follows.
On the basis of mathematical model (3.1), find the optimal parametric control
law in the environment of the set of algorithms (3.2) ensuring the attainment of
optimal values of the following criteria:
1. K1 ¼ 1
36
X36t¼1
yðtÞ ! max;
2. K2 ¼ 1
36
X36t¼1
xðtÞ ! max; (3.3)
3. K3 ¼ 1
36
P36t¼1
xðtÞxð0Þ þ
P36t¼1
yðtÞyð0Þ
0BBB@
1CCCA! max;
4. K4 ¼ 1
T
XTt¼1
xðtÞ � x0x0
� �2þ yðtÞ � y0
y0
� �2 !! min
(here T ¼ 232 is the period of one cycle) under the constraints
0 � yðtÞ � 1; 0 � nðtÞ � 1; 0 � xðtÞ: (3.4)
The base values of the criteria (without parametric control) are as follows:
K1 ¼ 0:06848;K2 ¼ 2:05489;K3 ¼ 2:08782;K4 ¼ 0:0307:
144 3 Parametric Control of Cyclic Dynamics of Economic Systems
The values of all criteria for the control law, that is, optimal in the sense of the
criterion, from (3.2) represented before are obtained by solving the problems
formulated above through application of the parametric control approach to the
evolution of the economic system. The results are presented in Table 3.1.
The values of the model’s endogenous variables without applying parametric
control and with use of the optimal parametric control laws for each criterion are
presented in graphic form in Figs. 3.3, 3.4, 3.5, 3.6, and 3.7.
Table 3.1 Values of coefficients and criteria for optimal laws
Criterion Optimal law Coefficient value Criterion value
1 3 0.2404966 0.068890
2 3 0.4766800 2.230337
3 4 0.0718620 2.196740
4 4 0.3005190 0.007273
Months
no scenario scenario 3
Fig. 3.3 Capital productivity ratio without parametric control and with use of law 3, optimal in the
sense of criterion 1
Months
no scenario scenario 3
Fig. 3.4 Capital productivity ratio without parametric control and with use of law 3, optimal in the
sense of criterion 2
3.1 Mathematical Model of the Kondratiev Cycle 145
Months
no scenario scenario 3
Fig. 3.5 Capital productivity ratio without parametric control and with use of law 4, optimal in the
sense of criterion 3
Years
no control with optimal control law
Fig. 3.6 Capital productivity ratio without parametric control and with use of law 4, optimal in the
sense of criterion 4
Years
no control with optimal control law
Fig. 3.7 Efficiency of innovations without parametric control and with use of law 4, optimal in the
sense of criterion 4
146 3 Parametric Control of Cyclic Dynamics of Economic Systems
3.1.4 Estimating the Structural Stability of the KondratievCycle Mathematical Model with Parametric Control
To carry out this analysis, the expressions for optimal parametric control laws (3.2)
with the obtained values of the adjusted coefficients are substituted into the right-
hand side of the second and third equations of system (3.1) for the parametern0 .Then, by using a numerical algorithm to estimate the weak structural stability of the
discrete-time dynamical system for the chosen compact set N determined by the
inequalities 1:7 � x � 2:3, 0:066 � y � 0:098 in the state space of the variables
x; yð Þ , we obtain the estimation of the chain-recurrent set Rðf ;NÞ as the empty
(or one-point) set. This means that the Kondratiev cycle mathematical model with
optimal parametric control law is estimated as weakly structurally stable in the
compact set N.
3.1.5 Analysis of the Dependence of the Optimal Valueof Criterion K on the Parameter for the VariationalCalculus Problem Based on the Kondratiev CycleMathematical Model
Now we’ll analyze the dependence of the optimal value of criterion K on the
exogenous parameters m (share of withdrawal of capital production assets per
month) and a for parametric control laws (3.2) with the obtained optimal values
of the adjusted coefficients ki, where the values of the parameters m; að Þ belong to
the rectangle L ¼ ½0:00063; 0:00147� � ½�0:01; 0:71� in the plane.
Plots of dependencies of the optimal values of criterion K (for parametric control
laws 0 and 2, yielding the maximum criterion values) on the uncontrolled parameters
(see Fig. 3.8) were obtained by computational experimentation. The projection of the
intersection line of the two surfaces in the plane m; að Þ consists of the bifurcation
points of the extremals of the given variational calculus problem.
3.2 Goodwin Mathematical Model of MarketFluctuations of a Growing Economy
3.2.1 Model Description
The Goodwin model describing market fluctuations in a growing economy is
presented in [19, 41].
3.2 Goodwin Mathematical Model of Market Fluctuations. . . 147
The model is described by the following system of two differential equations:
d0ðtÞ ¼ ðalðtÞ � a0ÞdðtÞ;l0ðtÞ ¼ ð�bdðtÞ þ b0ÞlðtÞ:
((3.5)
Here d is the percentage of employed in the entire population; l is the percentage ofsupply for consumption in the GDP; a, a0, b, and b0 are constants in the model.
The estimation of the model parameters a, a0, b, b0 is carried out using the
statistical information of the Republic of Kazakhstan for the years 2001–2005 [40],
for which the deviations of the observed statistical data from the calculated results
do not exceed 4.93% during the period under consideration. In solving the
parametric identification preliminary estimation problem, we obtained the follow-
ing exogenous parameters:
a ¼ 0:1710; a0 ¼ 0:08; b ¼ 0:00211; b0 ¼ 0:001:
The calculated period was one cycle in this case, T ¼ 706:27 months.
The model relies on an assumption of invariability of the following economic
parameters:
Opt
imal
crite
rion
val
ues
Fig. 3.8 Plots of the dependencies of the optimal value of criterionK on exogenous parameters m, a
148 3 Parametric Control of Cyclic Dynamics of Economic Systems
k is the capital output ratio, 0 < k < 1;
n is the population growth rate, n >�1;
g is the labor productivity growth rate, g >�1.
It is also assumed that the percentage of employed s depends linearly on the
wage growth rate o:
s ¼ s0 þ bo; 0 < s0 < 1; b > 0:
The constant parameters of model (3.5) are derived by the following relations:
a ¼ 1
bð1þ gÞ > 0; a0 ¼ s0bð1þ gÞ > 0; b ¼ 1
kð1þ gÞð1þ nÞ > 0;
b0 ¼ 1� kðgþ nþ ngÞkð1þ gÞð1þ nÞ :
Let’s also assume that gþ nþ ng < 1, in which case b0 > 0.
Let’s consider the solutions of system (3.5) in some closed, simply connected
region Owith boundary defined by a simple closed curve lying in the first quadrant
of the phase plane R2þ ¼ fd > 0; l > 0g: dð0Þ ¼ d0; lð0Þ ¼ l0; ðd0; l0Þ 2 O.
It is a well-known fact that in region R2þ , system (3.5) has only the following
state-space trajectories:
– The stationary singular point
l� ¼ a0=a; d� ¼ b0=b; 0 < l� < 1; 0 < d� < 1; (3.6)
– The nonstationary cyclic trajectories lying in R2þ and caused by the initial
conditions ðd0; l0Þ 6¼ ðd�; l�Þ . The singular point ðd�; l�Þ lies inside these
cycles.
3.2.2 Analysis of the Structural Stability of the GoodwinMathematical Model Without Parametric Control
Let’s estimate the structural stability of the Goodwin model without parametric
control in closed regionsO, being guided by the theorem on necessary and sufficient
conditions for structural stability [12]. First, let’s prove the following assertion.
Lemma 3.1The singular point ðd�; l�Þ of system (3.5) lies in the center.
ProofWith (3.6) in mind, let’s write down the Jacobian for the right-hand sides of system
(3.5) at the point ðd�; l�Þ:
3.2 Goodwin Mathematical Model of Market Fluctuations. . . 149
A ¼ al� � a0 ad�
�bl� b0 � bd�
� �¼ 0 ab0=b
�ba0=a 0
� �:
It is obvious that this matrix has imaginary eigenvalues � iffiffiffiffiffiffiffiffiffia0b0
p; . Therefore,
the point ðd�; l�Þ is the structurally unstable center point (nonhyperbolic point).
Assertion 3.1System (3.5) is structurally unstable in the closed regionO ðO � R2
þÞwith boundarya simple closed curve containing the point ðd�; l�Þ of the form (3.6) for any fixedvalues of the parameters k; n; g; l0; b; each taken from their respective regionalvalues.
System (3.5) is structurally stable in the closed regionOðO � R2þÞwith boundary
a simple closed curve not containing the point ðd�; l�Þof the form (3.6) for any fixedvalues of parameters k; n; g; l0; b, each taken from their respective regionalvalues.
ProofLet the closed regionO � R2
þ contain the singular point ðd�; l�Þ. A neighborhood of
this system of points (3.5) is locally structurally unstable. Therefore, it is structur-
ally unstable in region O.Let the closed region O � R2
þ not contain the singular point ðd�; l�Þ. In this case,the region O does not contain any cycle, since at least one singular point must be
inside any cycle. Therefore, considering this case, system (3.5) is structurally stable
in region O.
3.2.3 Problem of Choosing Optimal Parametric Control Lawson the Basis of the Goodwin Mathematical Model
It should be noted that the estimates for parameters k; n; g; s0; b derived using
statistical information from the Republic of Kazakhstan for the period 2000–2008
do not describe the economy of the Republic of Kazakhstan with acceptable
accuracy. Therefore, choosing optimal parametric control laws is presented below
for conventional values of the given parameters.
Now, let’s consider implementing an efficient public policy by choosing optimal
control laws with the example of economic parameter k (capital output ratio). Thegoal of the economic policy is to reduce the magnitudes of fluctuations of the
indices ðd; lÞ of the evolution of the national economic system.
Choosing the optimal laws of parametric control is carried out using the follow-
ing sets of relations:
1:U1ðtÞ ¼ c1dðtÞ � d0
d0þ k0; 2:U2ðtÞ ¼ �c2
dðtÞ � d0d0
þ k0;
3:U3ðtÞ ¼ c3lðtÞ � l0
l0þ k0; 4:U4ðtÞ ¼ �c4
lðtÞ � l0l0
þ k0:
(3.7)
150 3 Parametric Control of Cyclic Dynamics of Economic Systems
Here Ui is the ith control law of parameter k ( i ¼ 1; :::; 4 ); ci is the adjusted
coefficient of the ith control law, ci > 0; k0 is a constant equal to the base value
of parameter k. Application of control lawUi means the substitution of the function
from the right-hand side of (3.7) to system (3.5) for parameter k; t ¼ 0 is the control
system’s starting time, t 2 ½0; T�.The problem of choosing the optimal parametric control law at the level of the
economic parameter k can be stated as follows: On the basis of mathematical model
(3.5), find the optimal law of parametric control of the economic parameter k fromthe set of algorithms (3.7); that is, find the optimal law from the set Ui minimizing
the criterion characterizing the mean distance from the trajectory points to the
singular point ðd�; l�Þ of the system:
K ¼ 1
T
ðT0
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiðdðtÞ � d�Þ2 þ ðlðtÞ � l�Þ2
qdt ! min
fUi; cig(3.8)
under the constraints
0 � k � 1; 0 � l � 1; 0 � d � 1; t 2 ½0; T�: (3.9)
Here T is the period of the controlled cyclic trajectory of system (3.5), and criterion
K characterizes the mean distance from the points of this trajectory to the stationary
point (3.6).
This problem is solved in two stages:
– In the first stage, the optimal values of the coefficients ci for each law Ui are
determined by enumerating their values in the respective intervals (quantized
with a small step size), minimizing K under constraints (3.9).
– In the second stage, the optimal law regulating parameter k is chosen based on
the outcome of the first stage using the minimum value of criterion K.
The problem is solved:
• with given values of parameters b ¼ 10=13, g ¼ 0:5, d0 ¼ 0:4, and l0 ¼ 0:5,• with a fixed value of the uncontrolled parameter n ¼ 0:3,• and with the base value of the controlled parameter k0 ¼ 10=19.
These values of the parameters yield the system stationary point with the
coordinates l� ¼ 0:5; d� ¼ 0:5.A numerical solution to the problem of choosing the optimal parametric control
law shows that the best result for K ¼ 0.03215307 can be achieved using the
following law:
k ¼ 4:28lðtÞ � 0:5
0:5þ 10=19: (3.10)
Note that the criterion value K without parametric control is 0.0918682.
3.2 Goodwin Mathematical Model of Market Fluctuations. . . 151
Results of computational experiments allow us to determine the following facts:
– A decrease in the value of criterion K in comparison to the case without control
is obtained only with use of laws U1ðtÞ and U3ðtÞ from (3.7).
– Using laws of type U1ðtÞ , we can observe that cyclic character of the phase
trajectory of system (3.5) as seen in Fig. 3.9 is preserved.
– Using laws of type U3ðtÞ instead of a cyclic trajectory, we can observe
trajectories approaching the stable singular point of system (3.5) with parametric
control as t ! 1 approaches infinity (Fig. 3.10).
Fig. 3.9 Curve 1 corresponds
to the market cycle without
control; curve 2 corresponds
to the market cycle applying
control law U1ðtÞ
Fig. 3.10 Curve 1
corresponds to the market
cycle without control; curve
2 corresponds to the market
cycle applying control law
U3ðtÞ
152 3 Parametric Control of Cyclic Dynamics of Economic Systems
3.2.4 Analysis of the Structural Stability of the GoodwinMathematical Model with Parametric Control
Let’s analyze the structural stability of system (3.5) using a parametric control law
of kindU3ðtÞ orU4ðtÞ from the set of algorithms (3.7) for any admissible fixed value
of the adjusted coefficient c 6¼ 0.
These laws are given by
k ¼ clðtÞ � l0
l0þ k0: (3.11)
Here k0 is a constant equal to the base value of parameter k. First, let’s find the
singular points of system (3.5) using parametric control. Substituting the respective
expression for k into the right-hand sides of the equations of system (3.5) and setting
them equal to zero, we obtain the following unknown variables ðd; lÞ (with the
remaining fixed admissible values of variables and constants):
ðal� a0Þd ¼ 0;
� 1
kðlÞð1þ gÞð1þ nÞ dþ1� kðlÞðgþ nþ ngÞkðlÞð1þ gÞð1þ nÞ
� �l ¼ 0:
8><>: (3.12)
Here kðlÞ ¼ c l�l0l0
þ k0: We only use values of c such that 0 < kðlÞ < 1 .
System (3.8) has a unique solution in R2þ:
l� ¼ a0=a;
d� ¼ 1� kðl�Þðgþ nþ ngÞ;
((3.13)
where 0 < l� < 1; 0 < d� < 1:Now, let’s write the Jacobian for the left-hand side of system (3.12) at point
(3.13):
A¼
al��a0 ad�
�bl�cd�
ðc1l��l0l0
þk0Þ2
l0ð1þgÞð1þnÞ� c
ðc1l��l0l0
þk0Þ2
l0ð1þgÞð1þnÞ
0BBB@
1CCCAl�þð�bd�þb0Þ
0BBBBB@
1CCCCCA
¼
0 ad�
�bl�cðd��1Þl�
ðcl��l0l0
þk0Þ2
l0ð1þgÞð1þnÞ
0BBB@
1CCCA:
3.2 Goodwin Mathematical Model of Market Fluctuations. . . 153
The eigenvalues of matrix A are the roots of the equation
m2 þ cð1� d�Þl�
ðc l� � l0l0
þ k0Þ2
l0ð1þ gÞð1þ nÞmþ 1� kðl�Þðgþ nþ ngÞ
kðl�Þð1þ gÞð1þ nÞ a0 ¼ 0:
Denoting the coefficients of this equation by p and q, we obtain the quadratic
equation
m2 þ pmþ q ¼ 0; (3.14)
where q > 0, and the sign of p coincides with the sign of the coefficient c.The following cases are possible:
1. If the discriminant of Eq. (3.14) is
D ¼ cð1� d�Þl�
ðc l� � l0l0
þ k0Þ2
l0ð1þ gÞð1þ nÞ
0BBB@
1CCCA
2
� 41� kðl�Þðgþ nþ ngÞkðl�Þð1þ gÞð1þ nÞ a0 < 0;
then the singularity point ðd�; l�Þ of system (3.5) with parametric control (3.11)
is the focus, which is stable with c > 0 and unstable with c < 0.
2. If D 0, then singular point (3.13) of system (3.5) with parametric control
(3.11) is the node, and this node is stable with c>0 and unstable with c < 0.
Assertion 3.2System (3.5) with parametric control (3.11) is locally structurally stable in anysufficiently small closed region O ðO � R2
þÞ, small with the boundary as a simpleclosed curve containing the point ðd�; l�Þ of form (3.13) for any fixed values ofparameters c; n; g; l0; b from their respective regional values.
System (3.5) is structurally stable in any closed region OðO � R2þÞ with
boundary a simple closed curve not containing the point ðd�; l�Þ of form (3.13) forany fixed values of parameters c; n; g; l0; b from their respective regional values.
ProofLet the singular point ðd�; l�Þ not belong to the closed regionO � R2
þ. In this case,by the same reasoning as in the proof of Assertion 3.1, we obtain that system (3.5),
(3.11) is structurally stable in the region O.Now, let the singular point ðd�; l�Þ belong to the closed region O � R2
þ .
Since this point is hyperbolic (node or focus), then system (3.5), (3.11) is locally
structurally stable in its neighborhood.
Let’s analyze the structural stability of system (3.5) using a parametric control
law of type U1ðtÞ or U2ðtÞ from the set of algorithms (3.7) for any fixed admissible
value of the adjusted coefficient c 6¼ 0.
154 3 Parametric Control of Cyclic Dynamics of Economic Systems
These laws are given by
k ¼ cdðtÞ � d0
d0þ k0: (3.15)
First, let’s find the singular points of system (3.5) with parametric control.
Substituting this expression for k into the right-hand sides of the equations of
system (3.5) and equating them to zero, we obtain the following system in the
unknown variables ðd; lÞ (with the remaining fixed admissible values of variables
and constants):
ðal� a0Þd ¼ 0;
� 1
kðdÞð1þ gÞð1þ nÞ dþ1� kðdÞðgþ nþ ngÞkðdÞð1þ gÞð1þ nÞ
� �l ¼ 0:
8><>: (3.16)
Here kðdÞ ¼ cd� d0d0
þ k0. System (3.16) has a unique solution
l� ¼ a0=a;
d� ¼ 1þ ðc� k0Þðgþ nþ ngÞ1þ cðgþ nþ ngÞ=d0 :
8><>: (3.17)
We use only the values of c such that 0< kðdÞ< 1; 0< l� < 1; 0< d� < 1:Now, the Jacobian for the left-hand side of the system (3.12) at point (3.13) is
A¼
0 ad�
cðd� �1Þ
ðcd� � d0d0
þ k0Þ2
d0ð1þ gÞð1þnÞ� 1
ðcd� � d0d0
þ k0Þd0ð1þ gÞð1þnÞ
0BBB@
1CCCAl� 0
0BBBB@
1CCCCA:
It is obvious that this matrix has imaginary eigenvalues. Therefore, singularity
point (3.17) is at the center. Applying methods from [11], it can be proved that
all phase trajectories of system (3.5) with parametric control (3.15) are the cycles in
R2þ except at point (3.13). The following assertion can be proved similarly to
Assertion 3.1.
Assertion 3.3System (3.5) with parametric control (3.15) is structurally unstable in the closedregion O ðO � R2
þÞ with boundary a simple closed curve containing the point ðd�;l�Þof the form (3.17) for any fixed values of parameters c; k; n; g; l0; b from theirrespective regional values.
System (3.5) with parametric control (3.15) is structurally stable in the closedregion OðO � R2
þÞ with the boundary as a simple closed curve not containing thepoint ðd�; l�Þof form (3.17) for any fixed values of the parameters c; k; n; g; l0; bfrom their respective regional values.
3.2 Goodwin Mathematical Model of Market Fluctuations. . . 155
3.2.5 Analysis of the Dependence of the Optimal ParametricControl Law on Values of the Uncontrolled Parameterof the Goodwin Mathematical Model
Let’s consider the dependence of the results of choosing the optimal parametric
control law at the level of parameter k on the uncontrolled parameter n (population
growth rate) with values in the interval ½0; 0:4�.The results of computational experiments are presented in Table 3.2 and
Fig. 3.11. These results reflect the dependence of the optimal value of criterion
K on the values of parameter n for each of four possible laws (3.7).
An analysis of Table 3.2 shows that for all considered values of parameter n, thecontrol lawU3ðtÞ is optimal; that is, for the given interval of values of parameter n, abifurcation point of the extremals of the given variational calculus problem does not
exist.
Table 3.2
Parameter n 0 0.1 0.2 0.3 0.4
Control law Optimal value of the criterion for this law
U3ðtÞ 0.130000 0.093165 0.060932 0.032153 0.006379
U1ðtÞ 0.210856 0.167352 0.121062 0.069768 0.014642
U2ðtÞ, U4ðtÞ 0.336324 0.251121 0.151368 0.091868 0.018441
Opt
imal
crite
rion
val
ues
Fig. 3.11 Plots of the dependencies of the optimal values of criterionK on uncontrolled parameter n.Notation: – U3, – U1, – U2, U4
156 3 Parametric Control of Cyclic Dynamics of Economic Systems
Chapter 4
Macroeconomic Analysis and ParametricControl of Economic Growth of a NationalEconomy Based on Computable Modelsof General Equilibrium
As is well known [39], in the context of implementing economic policy, one must
estimate values of economic instruments that will ensure uniform growth (dynamic
equilibrium) in order to provide economic development such that supply and
demand in macroeconomic markets increasing from one period to another are
always equal when labor and capital are fully employed. To a certain extent, this
is a requirement of the mathematical models used for estimating rational values of
economic instruments of public policy in the field of economic growth.
The problem of economic growth is covered at present by a large number of
phenomenological and econometric models [45].
Using the basic regression equation for estimating the determinants of economic
growth,
g ¼ a0 þXl
alxl þXp
bpzp þXr
crSLVr þ e
[where g is the rate of the economic growth of the main indexes of the gross national
product (GDP, GNP) in the country, a0 is a constant, al are the coefficients of the
economic variables, хl are the economic variables, bp are the coefficients of
additional variables, zp are additional variables (political, social, geographical,
etc.), cr are the coefficients of the slack variables, SLVr are the slack variables
reflecting the group effect, e is the random component], we are able to derive
various econometric models of dependencies of economic growth on various
types of determinants intended to estimate a wide spectrum of hypotheses and
assumptions about their influence on economic growth, econometric dynamic
interbranch models, as well as econometric macroeconomic model [57, 59, 68].
These models are mainly intended to give an estimate and do not meet the
aforementioned requirements. A wide range of phenomenological models [45],
starting with the mathematical model of neoclassical theory of Solow [69] and
Swan [71] [complemented by dynamical optimization models on the basis of the
Ramsey problem to such mathematical models of endogenous economic growth
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2_4,# Springer Science+Business Media New York 2013
157
that represent, for example, production of innovations as a product of a particular
economic sector (e.g., the Grossman and Helpman model [60]); activity aimed at
the development of people themselves (e.g., Robert Lucas model [65]); interna-
tional trade and dissemination of technologies (e.g., Lucas model [64]); and others],
answer the questions about economic growth sources, but do not meet the afore-
mentioned requirements of a mathematical model for estimating rational values of
the economic instruments for public policy in the field of economic growth.
In the context of the balance model [63], where the interbranch connections are
represented via a system of material balances for some set of products constituting
in aggregate the entire national economy, one can note that the system of material
balances expressing the interbranch connections is formed without market relations
between the agents. They also do not include descriptions of such prime agents as
the state, banking sector, and aggregate consumer. Therefore, the balance models
meet the aforementioned requirement to a lesser degree.
In [27], a number of computable models of general equilibrium are proposed.
These models to a greater degree meet the aforementioned requirement for mathe-
matical models applied for estimating rational values of the economic instruments
of the public policy in the field of economic growth.
In this chapter we present results of national economic growth control based on
computable models of general equilibrium subject to constraints on the level of
prices. To a certain extent, this allows us to take the requirements of an antiinflation
policy into consideration.
4.1 National Economic Evolution Control Basedon a Computable Model of General Equilibriumof Economic Branches
4.1.1 Model Description, Parametric Identification,and Retrospective Prediction
The considered CGE model [27] can be written using a system of relations
decomposed into the following subsystems:
1. A subsystem of difference equations connecting the values of the endogenous
variables for two consecutive years,
xtþ1 ¼ Fðxt; yt; zt; u; lÞ: (4.1)
Here t is a year number, in discrete time; t ¼ 0; :::; ext ¼ xt; yt; ztð Þ 2 Rm is
a vector of endogenous variables of the system; xt ¼ x1t ; x2t ; . . . ; x
m1t
� � 2 X1;yt ¼ y1t ; y
2t ; . . . ; y
m2t
� � 2 X2 , zt ¼ z1t ; z2t ; . . . ; z
m3t
� � 2 X3 m1 þ m2 þ m3 ¼ m:Here, variables xt include the values of the key assets, account balances of
158 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
the agents, etc.; yt includes the values of demand and supply of the agents in
various markets; zt includes various types of market prices and shares
of budgets in markets with governmental prices for various economic agents;
u 2 U � Rq and l 2 ^ � Rs are the vectors of the exogenous parameters
(controlled and uncontrolled, respectively); X1;X2;X3; and U are compact
sets with nonempty interiors IntðXiÞ; i ¼ 1; 2; 3; and Int(WÞ, respectively; andF : X1 � X2 � X3 �W � ^ ! Rm1 is a continuous function.
2. A subsystem of algebraic equations describing the behavior and interaction of
the agents in various markets during a considered year. These equations allow us
to express the variables yt via the exogenous parameters and other endogenous
variables
yt ¼ Gðxt; zt; u; lÞ: (4.2)
Here G : X1 � X2 � X3 � U � ^ ! Rm2 is a continuous function.
3. A subsystem of recurrent relations for iterative computations of the equilibrium
values of market prices in various markets and shares of budget in the markets
with governmental prices for various economic agents
zt½Qþ 1� ¼ Zðyt½Q�; zt½Q�; L; u; lÞ: (4.3)
Here Q ¼ 0; 1; 2; . . . is the number of iterations. L is a set of positive numbers
(adjustable iteration constants). As L is decreased, the economic system goes into
the equilibrium state faster, although it increases the danger of the prices going to
the negative region. Here Z:X2 � X3 � ð0;þ1Þm3 � U � ^ ! Rm3 is a continuous
mapping (being contractive with fixed xt 2 X1; u 2 U , l 2 ^, and some fixed L.In this case, mapping Z has a unique stationary point, in which iterative process
(4.2), (4.3) converges.
For some fixed values of the exogenous parameters, for each moment of time
t, the CGE model of general equilibrium (4.1), (4.2) and (4.3) defines values of the
endogenous variables ext corresponding to the equilibrium of demand and supply
prices in markets with no governmental prices and shares of budget in markets
with governmental prices of the agents within the limits of the following
algorithm.
1. In the first step, we assume t ¼ 0 and give some initial values x0 to the variables.2. In the second step, for the current value of t, we give initial values zt½0� for the
variables in various markets and for different agents; then values yt½0� ¼ Gðxt;zt½0�; u; lÞ are computed (initial values of demand and supply of the agents in the
markets of goods and services) by (4.2).
3. In the third step, for the current value of time t, we start the iterative process
(2.12). For any Q, current values of supply and demand are found using (2.11):
yt½Q� ¼ Gðxt; zt½Q�; u; lÞ via adjustment of market prices and shares of the
budgets in the markets with the governmental prices of economic agents.
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 159
Now, the iterative process is terminated if the values of supply and demand in
various markets are equal. As a result, we determine the equilibrium values of
market prices in every market and the shares of budget in the markets with
governmental prices for various economic agents. The index Q for such equilib-
rium values of the endogenous variables is omitted.
4. For the last step, using the derived equilibrium solution for moment t, wecompute the values of variables xtþ1 for the next moment of time via difference
equations (4.1). Then the value of t is incremented by 1. Then we proceed to
step 2.
A number of iterations for steps 2–4 are defined in accordance with the problems
of calibration, prediction, and control in the same time intervals as those chosen
before.
The CGE model can be represented in the form of a continuous mapping f : X�U � ^ ! Rm defining a transformation of values of the system’s endogenous
variables for year 0 to the respective values of the next year according to the
algorithm described above. Here the compact set X in the state space of the
endogenous variables is defined by the set of possible values of variables x(the compact set X1 with nonempty interior) and respective equilibrium values of
the variables y and z computed by relations (4.2) and (4.3).
Let’s suppose that for a chosen point x0 2 Int(X1Þ, the inclusion ext ¼ f tðex0Þ X1j
2 Int(X1Þ: is correct with fixed u 2 Int ðUÞ and l 2 ^ for t ¼ 0; . . . ; T (T is any fixed
natural number). This mapping f defines a discrete-time dynamic system (semi-
cascade) in set X with the following initial condition to its trajectories:
f t; t ¼ 0; 1; . . .f g; x t¼0j ¼ x0: (4.4)
4.1.1.1 Model Agents
Model [27] describes the behavior and interaction in 46 product markets and 16
labor markets of the following 20 economic agents (sectors).
Economic agent no. 1 represents agriculture, hunting, and forestry.
Economic agent no. 2 represents fishery and fish breeding.
Economic agent no. 3 is the mineral resource industry.
Economic agent no. 4 is the manufacturing and process industry.
Economic agent no. 5 is the production and distribution of electric power, gas, and
water.
Economic agent no. 6 is construction.
Economic agent no. 7 represents trade, repair services for cars, and goods for
household use.
Economic agent no. 8 represents hotels and restaurants.
Economic agent no. 9 represents transport and communications.
Economic agent no. 10 corresponds to financial activity.
160 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Economic agent no. 11 represents real estate activities, rent, and services for
various businesses.
Economic agent no. 12 is public administration.
Economic agent no. 13 represents education.
Economic agent no. 14 is public health and social services.
Economic agent no. 15 represents other public utilities, social and personal
services.
Economic agent no. 16 represents housekeeping services.
A part of the products from economic agents producing goods and services
(economic agent nos. 1–16) are used in production, while the other part is spent to
invest and the remainder is sold to households. The producing agents deal in the
intermediate and investment products with each other.
Economic agent no. 17 is the aggregate consumer joining households.
The aggregate consumer purchases consumer goods produced by the producing
agents. Moreover, it purchases imported goods offered by the outer world.
Economic agent no. 18 is the government represented by the aggregate of the
central, regional, and local governments, as well as the off-budget funds. The
government establishes the taxation rates and defines the sum of the subsidies to
the producing agents, as well as the volumes of social transfers to the
households. Additionally, this sector includes the nonprofit organizations servic-
ing the households (the political parties, trade unions, public associations, etc.).
Economic agent no. 19 is the banking sector, including the central bank and
commercial banks.
Economic agent no. 20 is the outer world
The considered model includes 698 endogenous variables and 2045 exogenous
parameters to be estimated.
The following system of notations is used here for the constants and variables of
the CGE models:
<Type > <Parameter > _ < Price and its code > _ < Number of economic
agent and market code > [<instant time (t) or number of iteration (Q)>].
Here < Type > can take on two values; namely, С is the exogenous parameter,
and V is the endogenous variable.
<Parameter > corresponds to the action realized by the agent. The examples of
such actions can be given by S (product supply), D (product demand), O (determin-
ing the share of budget by the agent), and others.
For example, the notation CO p3 1z 0½ � corresponds to the exogenous parame-
ter, which is the share of budget of the first sector (agriculture, hunting, and
forestry) for purchasing the intermediate product produced by the third branch
(mineral resource industry) at the price of P 3z for the intermediate products of
the third branch in the year (2000).
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 161
4.1.1.2 Exogenous Variables of the Model
The exogenous parameters include the following:
– The coefficients of the production functions of the sectors;
– The various shares of the budgets of the sectors;
– The shares of the products for selling in the various markets;
– The depreciation rates of the capital assets and shares of the retired capital
assets;
– The deposit interest rates;
– The various taxation rates;
– The coefficients reflecting the level of nonpayments to the producing agents;
– The depreciation rates of the capital assets;
– The shares of the retired capital assets;
– The coefficient reflecting the level of wage liabilities for the employees in all
branches;
– The export prices and governmental prices of the goods, services, and labor
force, etc.
The list of the exogenous model parameters is given in Table 4.1.
Table 4.1 Exogenous variables of the computable model of economic branches
Producing agents nos. 1–16
CO pi il The share of the budget of the ith branch spent for paying the labor force at the
price of P il
CO pj iz The share of the budget of the ith branch spent for purchasing the intermediate
products produced by the branches j ¼ 1; :::; 16 at the price of P jz
CO p in The share of the budget of the ith branch spent for purchasing investment
products at the price of P n
CE pi iz The share of product produced by the ith branch for selling in the markets of the
intermediate products at the price of P iz
CE p ic The share of product produced by the ith branch for selling in the markets of final
products at the price of P ic
CE p in The share of product produced by the ith branch for selling in the markets of
investment products at the price of P in
CE pexi ic The share of product produced by the ith branch for selling in the markets of the
exported products at the price of P exi
CA r i The empirically determined coefficient of dimension of the production function
CA z ji The coefficients of the production function with the intermediate products
j ¼ 1; :::; 16 consumed by the ith branch
CA k i The coefficient of the capital in the production function
CA l i The coefficient of the labor in the production function
CO y i The coefficient reflecting the level of nonpayments to the producing agents
CA n The depreciation rate of the capital assets
CO w i The coefficient reflecting the level of arrears of wages to the employees in all
branches
CR i The share of the retired capital assets
(continued)
162 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.1.1.3 Endogenous Variables of the Model
The endogenous variables include the following:
– The budgets of the sectors and their various shares;
– The remainders of the agents’ budgets;
– The produced values-added of the producing sectors;
– Demand and supply of various products and services;
– The gains of the sectors;
– The capital assets of the producing sectors;
– The number of employees employed in Sectors 1–16;
– The wages of the employees;
– The various types of expenditures of the consolidated budget;
– The various types of prices of the products, services, and the labor force;
– The subsidies to the producing sectors;
– The social transfers to the citizens;
– The gross production of goods and services;
– The volume of production of the intermediate products;
– The volume of production of final products;
– The GDP of a country.
The list of the endogenous variables of the model is given in Table 4.2.
Table 4.1 (continued)
17 Households
CO p 17c The share of the budget of the aggregate consumer spent for purchasing final
products at the price of P c
CO b 17 The share of the budget deposited at the banks
CS pi 17l The number of employees employed in Sectors 1–16
18 Government
CT vad The VAT rate
CT pr The organization profit tax rate
CT pod The rate of physical body income tax
CT esn The rate of single social tax
CO s i 18 The shares of the consolidated budget for backing the producing agents
CO tr 18 The share of the consolidated budget for payment of social transfers to the
inhabitants
CO f 18 The share of off-budget funds for payment of pensions, welfare payments, etc.
CB other 18 The sum of the tax proceeds (not included into the considered ones), nontax
income, and other incomes of the consolidated budget
19 Banking sector
CP bpercent The deposit interest rate for enterprises
CP h bpercent The deposit interest rate for physical bodies
General part of the model
CP exi The price of the exported product produced by the ith branch
Technical parameters
Ceta 1 The iteration constant applied in the case of the exogenous price
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 163
Table 4.2 Endogenous variables of the computable model of economic branches
Producing agents nos. 1–16
VO tc i The share of the budget of the producing agent spent for discharging the taxes to the
consolidated budget
VO tf i The share of the budget of the producing agent spent for discharging the taxes to the
off-budget funds
VO s i The remainder of the agent’s budget
VD pi il The demand of labor power in the ith branch at the price of P il
VD pj iz The demand of the intermediate products produced by the branches j ¼ 1; :::; 16 in theith branch at the price of P jz
VD p in The demand of investment products in the ith branch at the price of P in
VY i Production of products and services in the prices of the base period
VY g i The value-added produced by the ith branch
VK i The capital assets of the producing agent
VS pi iz The supply of intermediate products
VS p ic The supply of final products
VS p in The supply of investment products
VS pex ic The supply of exported products
VY p i The gain of the producing agent
VY r i The profit of the producing agent
VB i The budget of the producing agent
VB b i The balance of banking accounts of the producing agent
17 Households
VO tc 17 The share of the budget of the aggregate consumer for discharging the taxes to the
consolidated budget
VO s 17 The remainder of the budget of Sector 17
VD p 17c The household demand for final products
VW i The wages of employees in Sectors 1–16
VB 17 The budget of households
VB b 17 The money of the households in banking accounts
18 Government
VO s 18 The share of the retained consolidated budget
VO s 18f The share of the retained off-budget funds
VG s i 18 Subsidies to the producing sectors
VG tr 18 Social transfers to the inhabitants
VG f 18 The off-budget funds allocated for the inhabitants
VB 18 The consolidated budget
VB b 18 The surplus (deficit) of the consolidated budget
VF 18 The monetary assets of the off-budget funds
VF b 18 The remainder of monetary assets of the off-budget funds
Integral indices of the model
VY The gross production of goods and services (in prices of the base period)
VS z The volume of production of the intermediate products (in prices of the base period)
VS c The volume of production of final products (in prices of the base period)
VY g The GDP
VP The consumer price index
(continued)
164 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.1.1.4 The Model Markets
The equilibrium prices are formed in 50 markets of the model as a result of leveling
the supply and demand of various types of products, services, and labor force.
The described model has
– 16 markets of the intermediate products and services produced and rendered by
the producing agents;
– A market of investment products; and
– A market of final products.
In addition, the model also includes the following:
– 16 foreign markets of exported products produced by the producing agents; and
– 16 markets of the labor force.
The total number of markets in this model is 46. The governmental and market
mechanisms of pricing are used in domestic markets. The prices of foreign markets
enter the model exogenously. We’ll now consider the formulas reflecting the
process of changes in prices in domestic markets (below, i is the agent number).
The price of labor force in the ith branch is given by
VP il Qþ 1½ � ¼ VP il Q½ � � Abs VD ps il t½ �=VS ps il t½ �ð Þ: (4.5)
The price of the intermediate product produced by the ith branch is as follows:
VP iz Qþ 1½ � ¼ VP iz Q½ � � Abs VD ps iz t½ �=VS ps iz t½ �ð Þ: (4.6)
The price of investment products is determined by
VP n Qþ 1½ � ¼ VP n Q½ � � Abs VD ps n t½ �=VS ps n t½ �ð Þ: (4.7)
Table 4.2 (continued)
General part of model
VP il The price of the labor force in the ith branch
VP iz The price of the intermediate product produced by the ith branch
VP n The price of investment products
VP c The price of consumer products
VP The consumer price index
VD ps il The total demand for the labor force at the price of P il
VD ps iz The total demand for intermediate products at the price of P iz
VD ps n The total demand for investment products at the price of P n
VD ps c The total demand for consumer products at the price of P c
VS ps il The total supply of the labor force at the price of P il
VS ps iz The total supply of intermediate products at the price of P iz
VS ps n The total supply of investment products at the price of P n
VS ps c The total demand for consumer products at the price of P c
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 165
The price of consumer products is as follows:
VP c Qþ 1½ � ¼ VP c Q½ � � Abs VD ps c t½ �=VS ps c t½ �ð Þ: (4.8)
We now have 16þ 16þ 1 ¼ 33 prices of products sold in domestic markets for
the given model.
The notations of prices in foreign markets are given below.
The price of the exported product produced by the ith branch is
P exi: (4.9)
Thus, the total number of prices in this model is 33þ 16 ¼ 49.
Let’s now proceed to the formulas describing the mechanism of the formation of
the demand and supply of products produced by agent nos. 1–16 at governmental
and market prices.
The final formulas determining the demand and supply of each economic agent
in the product markets included in the model are presented below.
The total demand for the labor force at the price of VP il t½ � is given by
VD ps il t½ � ¼ VD pi il t½ �: (4.10)
For simplicity, we do not consider the demand for the labor force in the ithbranch from the other branches. In this connection, the total demand for the labor
force at the price of VP il t½ � is defined by the demand in the single ith branch.
The total supply of the labor force at the price of VP il t½ � is as follows:
VS ps il t½ � ¼ CS pi 17l: (4.11)
The total demand for the intermediate product at the price of VP jz t½ � producedby the jth branch is determined as
VD ps jz t½ � ¼ SUM VD pj iz t½ �ð ÞÞ: (4.12)
Hereafter, SUM X ið Þ corresponding toP16i¼1
X i, i ¼ 1; :::; 16; is the economic
agent number.
As can be seen, the total demand for the intermediate product at the price
of VP jz t½ � includes the demand for the intermediate products in the jth branch
j ¼ 1; :::; 16 from the direction of all 16 branches.
The total supply of the intermediate product at the price of VP iz t½ � is given by
VS ps iz t½ � ¼ VS pi iz t½ �: (4.13)
The total demand for investment products at the price of VP__n[t] is
VD ps n t½ � ¼ SUMðVD p in t½ �Þ: (4.14)
166 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The total supply of investment products at the price of VP n t½ � is
VS ps n t½ � ¼ SUM VS p in t½ �ð Þ: (4.15)
The total demand for consumer products at the price of VP c t½ � is
VD ps c t½ � ¼ VD p 17c t½ �: (4.16)
The total supply of consumer products at the price of VP c t½ � is
VS ps c t½ � ¼ SUM VS p ic t½ �ð Þ: (4.17)
Thus, we have 32þ 32þ 2þ 2 ¼ 68 formulas for determining the total supply
of and demand for products in domestic markets.
Now we’ll present the notations defining the total supply of and demand for
exported products:
The total demand for exported products at the price of CP pex ic t½ � (given) is
VD pex ic t½ �: (4.18)
The total supply of exported products at the price of CP pex ic t½ � is
VS pex ic t½ �: (4.19)
Finally, we derive 68þ 32 ¼ 100 formulas for determining the total supply of
and demand for all products used in the model.
Next, we’ll describe the activities of the economic agents participating in this
model.
4.1.1.5 Economic Agent Nos. 1–16 Producing Products and Services
The equation of the production function of an economic producing agent is given by
VY i tþ 1½ � ¼ CA r i t½ � � Exp VD p1 iz t½ � � CA z 1i t½ �ð Þ � Exp VD p2 iz t½ � � CA z 2i t½ �ð Þ� Exp VD p3 iz t½ � � CA z 3i t½ �ð Þ � Exp VD p4 iz t½ � � CA z 4i t½ �ð Þ� Exp VD p5 iz t½ � � CA z 5i t½ �ð Þ � Exp VD p6 iz t½ � � CA z 6i t½ �ð Þ� Exp VD p7 iz t½ � � CA z 7i t½ �ð Þ � Exp VD p8 iz t½ � � CA z 8i t½ �ð Þ� Exp VD p9 iz t½ � � CA z 9i t½ �ð Þ � Exp VD p10 iz t½ � � CA z 10i t½ �ð Þ� Exp VD p11 iz t½ � � CA z 11i t½ �ð Þ � Exp VD p12 iz t½ � � CA z 12i t½ �ð Þ� Exp VD p13 iz t½ � � CA z 13i t½ �ð Þ � Exp VD p14 iz t½ � � CA z 14i t½ �ð Þ� Exp VD p15 iz t½ � � CA z 15i t½ �ð Þ � Exp VD p16 iz t½ � � CA z 16i t½ �ð Þ� Power VK i t½ �þVK i tþ1½ �ð Þ=2ð Þ;CA k i t½ �ð Þ�Power VD pi il t½ �;CA l i t½ �ð Þ:
(4.20)
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 167
Here CA r i;CA z ji, CA k i;CA l i are the parameters of the production function,
Power(X, Y) corresponds to XY, and Exp(X) corresponds to eX.The following formulas determine the demand for the production factors by the
ith agent.
The demand for the labor force is
VD pi il t½ � ¼ CO pi il� VB i t½ �ð Þ=VP il t½ �: (4.21)
The demand for intermediate products produced by all the producing agents is
VD p1 iz t½ � ¼ VO p1 iz t½ � � VB i t½ �ð Þ=CP 1z t½ �; (4.22)
VD p2 iz t½ � ¼ CO p2 iz� VB i t½ �ð Þ=VP 2z t½ �; (4.23)
VD p3 iz t½ � ¼ CO p3 iz� VB i t½ �ð Þ=VP 3z t½ �; (4.24)
VD p4 iz t½ � ¼ CO p4 iz� VB i t½ �ð Þ=VP 4z t½ �; (4.25)
VD p5 iz t½ � ¼ CO p5 iz� VB i t½ �ð Þ=VP 5z t½ �; (4.26)
VD p6 iz t½ � ¼ CO p6 iz� VB i t½ �ð Þ=VP 6z t½ �; (4.27)
VD p7 iz t½ � ¼ CO p7 iz� VB i t½ �ð Þ=VP 7z t½ �; (4.28)
VD p8 iz t½ � ¼ CO p8 iz� VB i t½ �ð Þ=VP 8z t½ �; (4.29)
VD p9 iz t½ � ¼ CO p9 iz� VB i t½ �ð Þ=VP 9z t½ �; (4.30)
VD p10 iz t½ � ¼ CO p10 iz� VB i t½ �ð Þ=VP 10z t½ �; (4.31)
VD p11 iz t½ � ¼ CO p11 iz� VB i t½ �ð Þ=VP 11z t½ �; (4.32)
VD p12 iz t½ � ¼ CO p12 iz� VB i t½ �ð Þ=VP 12z t½ �; (4.33)
VD p13 iz t½ � ¼ CO p13 iz� VB i t½ �ð Þ=VP 13z t½ �; (4.34)
VD p14 iz t½ � ¼ CO p14 iz� VB i t½ �ð Þ=VP 14z t½ �; (4.35)
VD p15 iz t½ � ¼ CO p15 iz� VB i t½ �ð Þ=VP 15z t½ �; (4.36)
VD p16 iz t½ � ¼ ðCO p16 iz� VB i t½ �Þ=VP 16z t½ �: (4.37)
The demand for investment products is
VD p in t½ � ¼ CO p in� VB i t½ �ð Þ=VP n t½ �: (4.38)
168 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The following formulas determine the supply of products and services produced
by the producing agent.
The supply of the intermediate products is
VS pi iz t½ � ¼ CE pi iz� VY i t½ �: (4.39)
The supply of final products is
VS p ic t½ � ¼ CE p ic� VY i t½ �: (4.40)
The supply of investment products is
VS p in t½ � ¼ CE p in� VY i t½ �: (4.41)
The supply of the exported products is
VS pex ic t½ � ¼ CE pexi ic� VY i t½ �: (4.42)
The following formula calculates the gain of the producing agent:
VY p i t½ � ¼ VS pi iz t½ � � VP iz t½ � þ VS p ic t½ � � VP c t½ � þ VS p in t½ � � VP n t½ �þ VS pex ic t½ � � CP exi t½ �:
(4.43)
The profit of the producing agent is
VY r i t½ � ¼ CO y i� VY p i t½ � � ðVD p1 iz t½ � þ VD p2 iz t½ � þ VD p3 iz t½ �þ VD p4 iz t½ � þ VD p5 iz t½ � þ VD p6 iz t½ � þ VD p7 iz t½ � þ VD p8 iz t½ �þ VD p9 iz t½ � þ VD p10 iz t½ �þVD p11 iz t½ �þVD p12 iz t½ �þVD p13 iz t½ �þ VD p14 iz t½ � þ VD p15 iz t½ � þ VD p16 iz t½ �þ VW i t½ � � CO w ið Þ þ CA n t½ � � ðVK i t½ � � VP n t½ �ÞÞ:
(4.44)
Here CO y i is the coefficient reflecting the level of nonpayments; CA n is the
depreciation rate of capital assets. Here we calculate the profit of the sector
consisting of the gain corrected by the level of nonpayments. The assets spent for
the intermediate product, wages (without taking into account the debt, the coeffi-
cient CO w i), and amortization of capital assets are subtracted.
The value-added produced by the ith sector is given by
VY g i t½ � ¼ VY r i t½ � þ VW i t½ �: (4.45)
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 169
The value-added consists of the profit received in the current period and wages
paid to the employee sector.
The budget of the producing agent is as follows:
VB i t½ � ¼ VB b i t� 1½ � � 1þ CP bpercent t� 1½ �ð Þ þ CO y i� VY p i t½ � þ VG s i18 t� 1½ �:(4.46)
The agent budget consists of the following:
1. The funds in banking accounts (taking into consideration interests on deposits);
2. The gain received in the current period;
3. The subsidies received from the consolidated budget VG s i18.
The dynamics of the banking account balance of the producing agent is as
follows:
VB b i t½ � ¼ VO s i t½ � � VB i t½ �: (4.47)
The capital assets are determined by
VK i tþ 1½ � ¼ 1� CR i t½ �ð Þ � VK i t½ � þ VD p in t½ �: (4.48)
This formula calculates the volume of the capital assets, taking their retirement
into retirement. The asset put into operation enters the formula with the plus sign.
The share of the budget of the producing agent for discharging the taxes to the
consolidated budget is given by
VO tc i t½ � ¼ VY g i t½ � � CT vad t½ �ð Þ=VB i t½ �þ VY r i t½ � � CT pr t½ �ð Þ=VB i t½ �: (4.49)
This formula takes into consideration the value-added tax (VAT) and profit tax.
The share of the budget for discharging the single social tax to the off-budget
funds is described as
VO tf i t½ � ¼ VW i t½ � � CT esn t½ �ð Þ=VB i t½ �: (4.50)
The remainder of the budget of the producing agent is given by
VO s i t½ � ¼ 1� ðCO pi ilþ CO p inþ VO tc i t½ � þ VO tf i t½ � þ VO p1 iz t½ �þ CO p2 izþ CO p3 izþ CO p4 izþ CO p5 izþ CO p6 iz
þ CO p7 izþ CO p8 izþ CO p9 izþ CO p10 izþ CO p11 iz
þ CO p12 izþ CO p13 izþ CO p14 izþ CO p15 izþ CO p16 izÞ:(4.51)
170 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.1.1.6 Economic Agent No. 17: Aggregate Consumer
Let’s proceed to the formulas that determine the behavior of the aggregate
consumer.
The household demand for final products is given by
VD p 17c t½ � ¼ CO p 17c� VB 17 t½ �ð Þ=VP c t½ �: (4.52)
The wages of the employees of Sectors 1–16 are
VW i t½ � ¼ VD pi il t½ � � VP il t½ �: (4.53)
The budget of households is determined as follows:
VB 17 t½ � ¼ VB b 17 t� 1½ � � 1þ CP h bpercent t� 1½ �ð Þ þ VB 17 t� 1½ �� VO s 17 t� 1½ � þ VG tr 18 t� 1½ � þ VG f 18 t� 1½ � þ SUM VW i t½ �ð Þ:
(4.54)
The agent’s budget is formed using the following:
1. Funds in the banking accounts (subject to interests on deposits);
2. Retained money (in cash) kept from the preceding period;
3. Pensions, welfare payments, and subsidies received from the off-budget funds;
4. Wages received from the producing agent nos. 1–16.
The dynamics of the banking account balance of households is as follows:
VB b 17 t½ � ¼ CO b 17� VB 17 t½ �: (4.55)
The share of the budget for discharging income tax is given by
VO tc 17 t½ � ¼ tS� CT pod t½ �ð Þ=VB 17 t½ �: (4.56)
The remainder of the money in cash is calculated as follows:
VO s 17 t½ � ¼ 1� CO p 17c� VO tc 17 t½ � � CO b 17: (4.57)
4.1.1.7 Economic Agent No. 18 Government
As shown above, this economic agent is represented by the aggregate of the federal,
regional, and local governments, and as well as the off-budget funds. Additionally,
it includes the nonprofit organizations servicing the households.
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 171
Now we’ll move on to the formulas determining the behavior of economic agent
no. 18.
The consolidated budget is given by
VB 18 t½ � ¼ SUM VO tc i t½ � � VB i t½ �ð Þ þ VO tc 18 t½ � � VB 18 t½ �þ CB other 18 t½ � þ VB b 18 t½ � � 1þ CP bpercent t� 1½ �ð Þ:
(4.58)
This formula sums up money collected as taxes from the producing agents, as
well as from the inhabitants. The value CB_other_18 entering the model exoge-
nously is the sum of other taxes (not included in the list of taxes considered in
the model), nontaxable income, and other income of the consolidated budget. The
obtained sum is incremented by the funds in banking accounts (subject to the
deposit interests).
The dynamics of the banking account balance of the consolidated budget is
determined by
VB b 18 tþ 1½ � ¼ VO s 18 t½ � � VB 18 t½ �: (4.59)
The cash assets of off-budget funds are as follows:
VF 18 t½ � ¼ SUM VO tf i t½ � � VB i t½ �ð Þ þ VF b 18 t½ �� 1þ CP bpercent t� 1½ �ð Þ: (4.60)
This formula calculates the sum collected from the producing agents in the form
of the single social tax entering the accounts of the following off-budget funds: the
pension fund, the social insurance fund, and the federal and territorial funds of
obligatory medical insurance. The derived sum is added by the funds on the banking
accounts (subject to the deposit interests).
The dynamics of the banking account balance of the off-budget funds is deter-
mined by
VF b 18 tþ 1½ � ¼ VO s 18f t½ � � VF 18 t½ �: (4.61)
The subsidies to the producing sectors are as follows:
VG s i18 t½ � ¼ CO s i18� VB 18 t½ �: (4.62)
The social transfers to the inhabitants are
VG tr 18 t½ � ¼ CO tr 18� VB 18 t½ �: (4.63)
172 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The assets of the off-budget funds made available for the inhabitants are
VG f 18 t½ � ¼ CO f 18� VF 18 t½ �: (4.64)
This includes the assets of the pension fund and social insurance fund for paying
pensions and welfare payments.
4.1.1.8 Integral Indexes of the Model
Let’s present the formulas for calculating some integral indexes of the economy of
Kazakhstan.
The gross production of goods and services (in prices of the base period) is
VY t½ � ¼ SUM VY i t½ �ð Þ: (4.65)
The total supply of the intermediate products (in prices of the base period) is
VS z t½ � ¼ SUM VS pi iz t½ �ð Þ: (4.66)
The total supply of final products (in prices of the base period) is
VS c t½ � ¼ SUM VS p ic t½ �ð Þ: (4.67)
The GDP of Kazakhstan is
VY g t½ � ¼ SUM VY g i t½ �ð Þ=VP c 0½ �: (4.68)
The consumer price index is
VP t½ � ¼ 100� VP c t½ �=VP c t� 1½ �ð Þ: (4.69)
The considered model is presented in the context of the common relations:
Relations (4.1) are presented by m1 ¼ 67 expressions.
Relations (4.2) are presented by m2 ¼ 597 expressions.
Relations (4.3) are presented by m2 ¼ 34 expressions.
4.1.1.9 Parametric Identification of the Model and Retrospective Prediction
The problem of parametric identification of the considered macroeconomic mathe-
matical model requires estimation of unknown parameters that minimize the value
of the objective function that characterizes the deviations of values of the output
model’s variables from the corresponding measured values (known statistical data).
This problem reduces to finding the minimal value of the function in several
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 173
variables (parameters) in some close region Ω of the Euclidean space under
constraints ex 2 eX imposed onto values of the model’s endogenous variables. In
the case of large dimensions of a region of possible parameter values, the standard
method for finding the extremum of a function is often inefficient due to the
objective function’s several local minimum points. Proposed here is an algorithm
taking into account peculiarities of the problem of parametric identification of
macroeconomic models that allows us to avoid “local extrema”.
The regionO ¼ Qqþsþm1
i¼1
½ai; bi�where ½ai; bi� is the interval of possible values of theparameteroi; i ¼ 1; . . . ; ðqþ sþ m1Þ, is considered the regionO � U � ^ � X1 for
estimating possible values of the exogenous parameters. The estimate of parameters
with their respective measured values available was searched within the intervals
½ai; bi�with centers in the respective measured values (in case of one such value) or
within some intervals covering the measured values (in case of several values).
Other intervals ½ai; bi� for searching the parameters were chosen by indirect
estimation of their possible values. For searching the minimal values of the contin-
uous function in severable variables F : O ! R with the additional constraints on
the endogenous variables, we applied the Nelder–Mead directed search algorithm.
Using this algorithm for an initial point, o1 can be interpreted as a sequence
o1;o2;o3; . . .f g [converging to the local minimum o0 ¼ arg minO;ex2ex F oð Þ offunction F], sequence {o1, o2, o3,…} where F ojþ1
� � � F oj
� �;oj 2 O; j ¼ 1; 2;
. . . . While describing the next algorithm, we consider that the point o0 can be
found with sufficient precision.
For parametric identification of the considered computable model based on the
obvious assumption about disagreement (in general case) of the minimum points of
two different functions, the following two criteria are proposed:
KAðoÞ ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
naðt2 � t1 þ 1ÞXt2t¼t1
XnAi¼1
aiyit � yi�tyi�t
� �2
vuut ;
KBðoÞ ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1
nbðt2 � t1 þ 1ÞXt2t¼t1
XnBi¼1
biyit � yi�tyi�t
� �2
vuut :
(4.70)
Here ft1; . . . ; t2g is the identification time interval; yit; yi�t are computed and
measured values of the model output variables; KAðoÞ is the auxiliary criterion;
KBðoÞ is the main criterion; nB > nA ; ai > 0 and bi > 0 are some weighting
coefficients whose values are calculated during parametric identification of the
dynamic system; andPnA
i¼1 ai ¼ na;PnB
i¼1 bi ¼ nb:The algorithm for the parametric identification problem is as follows:
1. For some vector of initial values of parameter o1 2 O, solve problems A and Bsimultaneously. Then find the minimum pointsoA0 andoB0 of criteriaKA andKB,
respectively.
2. If KB oB0ð Þ<e for some sufficiently small number e, then the model parametric
identification problem is solved.
174 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
3. Otherwise, choose the point oB0 as the initial point o1, solve problem A, and,choosing the point oA0 as the initial point o1, solve problem B. Go to step 2.
After a sufficiently large number of iterations of stages 1, 2, and 3, searched
values of the parameters might leave neighborhoods of the nonglobal minima in one
criterion with the help of the other and thereby solve the parametric identification
problem.
As a result of a joint solution of problems A and B using the Nelder–Mead
algorithm, we obtain values KA ¼ 0:015 and KB ¼ 0:0063. The relative values ofdeviations in the computed variables used in the main criterion turn out to be less
than 0.63%.
The results of computation and retrospective prediction of the model for 2008
presented in Table 4.3 show computed VY;VY�g;VP� �
; measured values, and
deviations of the computed values of the main output variables from their respec-
tive measured values. Year 2000–2007 corresponds to the period of model
parametric identification; year 2008 is the retrospective prediction period; VY�gis the gross output ( �1012 tenge in prices of year 2000); VY�g is GDP ( �1012 in
prices of year 2000); VP is the consumer price index in percentage with respect to
the preceding year; the symbol «*» H= «D» corresponds to the measured values; the
symbol «*» corresponds to the deviations (in percentage) of the computed values
from their respective measured values.
The mean square error of the retrospective prediction of all 491 endogenous
variables of the model for the years 2008–2009 is 5.86%.
The results of verification show acceptable adequacy of the CGE model of
economic branches.
4.1.2 Macroeconomic Analysis on the Basis of the ComputableModel of General Equilibrium of Economic Branches
Rational implementation of the discretionary public policy determines carrying out
macroeconomic analysis on the basis of available instruments.
Table 4.3 Measured and computed values of the model output variables and respective deviations
Year 2000 2001 2002 2003 2004 2005 2006 2007 2008
VY� 5.44 6.32 6.47 6.86 7.72 8.52 9.25 9.69 9.84
VY 5.38 6.32 6.47 6.86 7.72 8.52 9.27 9.64 9.82
DVY �1.22 �0.02 0.00 0.00 0.05 0.08 0.21 �0.51 �0.26
VY g� 2.45 2.78 3.05 3.36 3.72 4.09 4.55 5.01 5.18
VY g 2.47 2.78 3.05 3.35 3.72 4.09 4.55 5.01 5.20
DVY g 0.88 0.07 �0.04 �0.02 �0.02 �0.02 �0.04 �0.15 0.38
VP� 106.40 106.60 106.80 106.70 107.50 108.40 118.80 109.50
VP 107.60 106.80 106.90 106.70 107.30 108.20 118.60 109.40
DVP 1.13 0.18 0.08 �0.05 �0.23 �0.22 �0.24 �0.05
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 175
In this section we present some results of macroeconomic analysis on the basis
of the computable model of general equilibrium of economic branches after esti-
mation of its parameters using statistical data from the Republic of Kazakhstan for
the years 2000–2009.
We also analyzed the dynamics (and trends) of the exogenous and endogenous
indicators, the rates (and their trends) of the endogenous indicators of economic
agents, and the entire national economy from 2000 to 2009 and middle-term
prediction (perspective) from 2010 to 2015.
The list of the exogenous and endogenous indicators of the economic agents and
national economy forming the basis for retrospective and perspective macroeco-
nomic analysis is given in Tables 4.1 and 4.2.
4.1.2.1 Retrospective and Perspective Analysis of the Economic Agents’Indicators and National Economy on the Basis of the ComputableModel of General Equilibrium of Economic Branches
Below we present several results of the retrospective analysis at the level of the
branches of the national economy.
Results of the Retrospective Analysis of the Exogenous Variables
at the Level of Economic Branches
The retrospective analysis of the exogenous indicators by economic branches
for 2000–2009 shows that within this period
– The number employed in all branches except branch 16 (housekeeping services)
increased on average. The maximum average growth (computed by linear trend)
is observed for branch 6 (construction) and is equal to 4:2� 104 people per year.
The negative growth (decrease by 2:3� 103 people per year on average) was
observed in branch 16 (housekeeping services).
– The trend of the indicator “Share of intermediate products in branch’s output”
had a positive slope for branches 8, 11, 10, 2, 4, 1, and a negative slope for
branches 16, 9, 13, 14, 5, 12, 6, 7, 15, 3. The maximum growth of a trend that is
4.4% per year is observed in branch 8 (hotels and restaurants). The maximum
decrease by 3.4% per year is observed in branch 3 (mineral resource industry).
– The trend of the indicator “Share of final products in branch’s output” had a
positive slope for branches 5, 12, 16, 1, 4, 14, 11, 13 and a negative slope for
branches 3, 6, 15, 10, 9, 7, 2, 8. The maximum growth of a trend by 4.9% per
year is observed in branch 5 (production and distribution of electric power, gas,
and water). The maximum decrease by 3.8% per year is observed in branch
8 (hotels and restaurants).
– The trend of the indicator “Share of export in branch’s output” had a positive
slope for branches 7, 3, 4, 9, 15, 10, 2, 5 and a negative slope for branches 6, 13,
176 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
8, 11, 12, 1. In branches 14 and 16, the export component is approximately zero.
The maximum growth of trend by 3.6% per year is observed in branch 7 (trade,
repair services for cars, and goods for household use). The maximum decrease
by 1.8% per year is observed in branch 1.
– The trend of the indicator “Price of export supplies of branch’s products” had a
positive slope for all in branch 16. The maximum growth of the trend by 62% per
year is observed in branch 5 (production and distribution of electric power, gas,
and water). The minimum increase by 5.9% per year is observed in branch 13
(education).
– The trend of the indicator “Share of final products in branch’s output” had a
positive slope for branches 15, 13, 7, 4, 6, 5, 9, 18, 2 and a negative slope for
branches 14, 10, 11, 3, 16. The maximum growth of the trend by 1.9% per year is
observed in branch 15 (other public, social, and personal services). The maxi-
mum decrease by 1.4% per year is observed in branch 16 (housekeeping
services).
– The trend of the indicator “Share of budget of branch spent for purchasing
investment products” had a positive slope for branches 12, 6, 11, 5, 4, 1, 9, 14,
2, 7, 10 and a negative slope for branches 13, 15, 8, 3. This share was zero for
branch 16. The maximum growth of the trend by 10% per year is observed in
branch 12 (public administration). The maximum decrease by 2% per year is
observed in branch 3 (mineral resource industry).
– The indicator “Share of investment product in branch’s output” was nonzero in
8 branches out of 16. In branches 6, 7, 2, 1, 9 (hereinafter the branches are
ordered by descending trend angular coefficient), this indicator had a positive
slope of linear trend (the maximum value 2.9% per year was observed in branch
6, construction). This share was equal to 78% for branch 6 in 2008. Branches 15,
4, and 11 demonstrated negative slopes of linear trends. The maximum decrease
of the trend is observed in branch 11 (real estate activities, rent, and services for
business) and constitutes close to 0.94% per year.
In Fig. 4.1 we give a graphic illustration of the dynamics (and trend) of one of the
economic agents’ indicators for retrospective macroanalysis by the example of the
variable “Share of investment products in the output of economic agent 1 (Agricul-
ture, hunting, and forestry).”
The analysis of Fig. 4.1 shows a positive slope of linear trend of the considered
indicator, thereby demonstrating growth (on average) of the share of investment
products in the output of economic agent 1 (Agriculture, hunting, and forestry)
within the considered period.
In the context of the retrospective analysis of the indicators of economic
branches on the basis of the considered model, we also obtained the measured
and computed values (and linear trends) of such exogenous indicators for each of
the16 economic branches as CO pj iz (share of budget of the ith branch spent for
purchasing the intermediate products produced by the branch j; i; j ¼ 1; . . . ; 16).Particularly, retrospective analysis of the mentioned indicators of the branches for
2000–2009 shows
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 177
– The trend of the indicator CO pj 1z (share of budget of branch 1 that is
agriculture, hunting, and forestry) spent for purchasing intermediate products
produced by the jth branch had a positive slope for j ¼ 1, 8, 2, 13 and a negative
slope for j ¼ 11 , 3, 15, 14, 10, 6, 12, 5, 7, and 9. Branch 1 did not buy
intermediate goods produced by economic branch 16. The maximum growth
trend of 1.7% per year is observed for j ¼ 1. The maximum decrease of 0.63%
per year is observed for j ¼ 3 (mineral resource industry). The maximum value
of the indicator CO pj 1z of 36% was noted for j ¼ 1.
– The trend of the indicator CO pj 2z (share of budget of branch 2 that is fishery
and fish breeding) spent for purchasing intermediate products produced by the
jth branch had a positive slope for j ¼ 9, 11, 5, 10, 3, 6, 8 and a negative slope for
j ¼ 14, 15, 1, 12, 4, 7, and 2. Branch 2 did not buy intermediate goods produced
by economic branches 13 and 16. The maximum growth of the trend by 1.6% per
year is observed for j ¼ 9. The maximum decrease by 3.2% per year is observed
for j ¼ 2 (mineral resource industry). The maximum value of the indicator COpj 2z equal to 18% was noted for j ¼ 4 (manufacturing and process industry) in
2008.
– The trend of the indicator CO pj 3z (share of budget of branch 3 that is mineral
resource industry) spent for purchasing the intermediate products produced by
the jth branch had a positive slope for j ¼ 11, 10, 7, 12, 1, 2, 16 and a negative
slope for j ¼ 8, 15, 13, 4, 14, 6, 5, 3, and 9. Branch 3 did not buy intermediate
goods produced by economic branch 16. The maximum growth of the trend by
1.3% per year is observed for j ¼ 11 (real estate activities, rent, and services for
business). The maximum decrease by 1.1% per year is observed for j ¼ 9
(transport and communications). The maximum value of the indicator CO pj 3z equal to 19% was observed for j ¼ 11 (real estate activities, rent, and services
for business) in 2008.
Linear trend of parameter
Measured parameter values Estimated parameter values
Fig. 4.1 Measured and estimated values (and linear trend) of exogenous variable “Share of
investment products in the output of economic agent 1”
178 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
– The trend of the indicator CO pj 4z (share of budget of branch 4 that is
manufacturing and process industry) spent for purchasing intermediate products
produced by the jth branch had a positive slope for j ¼ 4, 3, 11, 1, 10, 6, 8, 2 and a
negative slope for j ¼ 14 , 12, 13, 15, 5, 9, and 7. Branch 4 did not buy
intermediate goods produced by economic branch 16. The maximum growth
of the trend by 0.46% per year is observed for j¼ 4. The maximum decrease by
0.46% per year is observed for j ¼ 7 (trade, repair services for cars, and goods for
household use). The maximum value of the indicatorCO pj 4z equal to 25% was
noted for j ¼ 4 (manufacturing and process industry) in 2008.
– The trend of the indicatorCO pj 5z (share of budget of branch 5 that is productionand distribution of electric power, gas, and water) spent for purchasing the
intermediate products produced by the jth branch had a positive slope for j ¼ 4,
5, 11, 3, 10, 8, 14, 1, 12, 2, 16 and a negative slope for j ¼ 15, 13, 6, 9, and 7.
The maximum growth of the trend by 2.0% per year is observed for j ¼ 4
(manufacturing and process industry). The maximum decrease by 1.0% per
year is observed for j ¼ 7 (trade, repair services for cars, and goods for household
use). The maximum value of the indicator CO pj 5z equal to 19% was noted for
j ¼ 4 (manufacturing and process industry) in 2008.
– The trend of the indicator CO pj 6z (share of budget of branch 6 that is
construction) spent for purchasing the intermediate products produced by the
jth branch had a positive slope for j ¼ 4, 6, 10, 8, 2, 14 and a negative slope for
j ¼ 13, 12, 1, 15, 11, 5, 7, 9, and 3. Branch 6 did not buy intermediate goods
produced by economic branch 16. The maximum growth of the trend by 2.3%
per year is observed for j ¼ 4 (manufacturing and process industry). The
maximum decrease by 0.79% per year is observed for j ¼ 3 (mineral resource
industry). The maximum value of the indicatorCO pj 6z equal to 37% was noted
for j ¼ 4 (manufacturing and process industry) in 2008.
– The trend of the indicator CO pj 7z (share of budget of branch 7 that is trade,
repair services for cars, and goods for household use) spent for purchasing
the intermediate products produced by the jth branch had a positive slope for
j ¼ 11, 3, 10, 8, 13 and a negative slope for j ¼ 2, 14, 12, 15, 1, 5, 6, 9, 4, and 7.
Branch 7 did not buy the intermediate goods produced by economic branch 16.
The maximum growth of the trend by 2.3% per year is observed for j ¼ 11
(manufacturing and process industry). The maximum decrease by 0.79% per
year is observed for j ¼ 7 (real estate activities, rent, and services for business).
Themaximumvalue of the indicatorCO pj 7zequal to 17%was noted for j ¼ 11
(real estate activities, rent, and services for business) in 2008.
– The trend of the indicator CO_pj_8z (share of budget of branch 8 that is hotels
and restaurants) spent for purchasing the intermediate products produced by the
jth branch had a positive slope for j ¼ 1, 9, 10, 6 and a negative slope for j ¼ 3,
14, 13, 12, 4, 2, 15, 7, 5, 8, and 11. Branch 8 did not buy intermediate goods
produced by economic branch 16. The maximum growth of the trend by 0.92%
per year is observed for j ¼ 1 (agriculture, hunting, and forestry). The maximum
decrease by 0.62% per year is observed for j ¼ 11. The maximum value of the
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 179
indicator CO_pj_8z equal to 20% was noted for j ¼ 4 (manufacturing and
process industry) in 2008.
– The trend of the indicatorCO pj 9z (share of budget of branch 9 that is transportand communications) spent for purchasing the intermediate products produced
by the jth branch had a positive slope for j ¼ 9, 11, 10, 8, 2 and a negative slope
for j ¼ 14, 13, 15, 12, 1, 3, 6, 5, 4, and 7. Branch 8 did not buy intermediate
goods produced by economic branch 16. The maximum growth of the trend by
0.69% per year is observed for j ¼ 9. The maximum decrease by 1.0% per year
is observed for j ¼ 7 (trade, repair services for cars, and goods for household
use). The maximum value of the indicator CO pj 9z equal to 15% was noted for
j ¼ 4 (manufacturing and process industry) in 2008.
– The trend of the indicator CO_pj_10z (share of budget of branch 10 that is
financial activity) spent for purchasing the intermediate products produced by
the jth branch had a positive slope for j ¼ 4, 9, 11, 8, 1, 6, 5, 15, 13, 2 and a
negative slope for branches j ¼ 16, 14, 12, 3, 7, and 10. The maximum growth of
the trend by 0.15% per year is observed for j ¼ 4 (manufacturing and process
industry). The maximum decrease by 0.98% per year is observed for j ¼ 10. The
maximum value of the indicator CO pj 10z equal to 16% was noted for j ¼ 4
(manufacturing and process industry) in 2008.
– The trend of the indicator CO pj 11z (share of budget of branch 11 that is real
estate activities, rent, and services for business) spent for purchasing the inter-
mediate products produced by the jth branch had a positive slope for j ¼ 4, 7, 9,5, 13, 14, 8, 2 and a negative slope for j ¼ 16, 12, 1, 15, 3, 6, and 11. The
maximum growth of the trend by 1.6% per year is observed for j ¼ 4
(manufacturing and process industry). The maximum decrease by 0.70% per
year is observed for j ¼ 11. The maximum value of the indicator CO pj 11zequal to 25% was noted for j ¼ 11 in 2008.
– The trend of the indicator CO pj 12z (share of budget of branch 12 that is real
estate activities, rent, and services for businesses) spent for purchasing the
intermediate products produced by the jth branch had a positive slope for
j ¼ 8, 10, 6, 11, 15, 1, 14, 13 and a negative slope for j ¼ 2, 9, 3, 5, 4, 7, and
12. Branch 12 did not buy intermediate goods produced by economic branch 16.
The maximum growth of the trend by 0.92% per year is observed for j ¼ 8
(hotels and restaurants). The maximum decrease by 0.76% per year is observed
for j ¼ 11. The maximum value of the indicator CO pj 12z equal to 18% was
noted for j ¼ 4 (manufacturing and process industry) in 2008.
– The trend of the indicator CO pj 13z (share of budget of branch 13 that is
education) spent for purchasing the intermediate products produced by the jthbranch had a positive slope for j ¼ 4, 6, 9, 15, 1, 8 and a negative slope for j ¼ 2,
10, 3, 12, 14, 7, 13, 5, and 11. Branch 13 did not buy intermediate goods
produced by economic branch 16. The maximum growth of the trend by 2.9%
per year is observed for j ¼ 4 (manufacturing and process industry). The
maximum decrease by 0.66% per year is observed for j ¼ 11 (real estate
activities, rent, and services for business). The maximum value of the indicator
180 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
CO pj 13z equal to 24% was noted for j ¼ 4 (manufacturing and process
industry) in 2008.
– The trend of the indicator CO pj 14z (share of budget of branch 14 that is publichealth and social services) spent for purchasing the intermediate products pro-
duced by the jth branch had a positive slope for j ¼ 4, 6, 1, 7, 8 and a negative
slope for j ¼ 2, 3, 12, 11, 10, 15, 9, 13, 5, and 14. Branch 14 did not buy
intermediate goods produced by economic branch 16. The maximum growth of
the trend by 3.2% per year is observed for j ¼ 4 (manufacturing and process
industry). The maximum decrease by 1.1% per year is observed for j ¼ 14. The
maximum value of the indicator CO pj 14z equal to 37% was noted for j ¼ 4
(manufacturing and process industry) in 2008.
– The trend of the indicator CO pj 15z (share of budget of branch 15 that is other
public utilities, social and personal services) spent for purchasing the intermedi-
ate products produced by the jth branch had a positive slope for j ¼ 4, 8, 10, 5, 7,2, 16 and a negative slope for branches j ¼ 6, 1, 3, 13, 12, 9, 14, 15, and 11.
Branch 15 did not buy intermediate goods produced by economic branch 16. The
maximum growth of the trend by 0.22% per year is observed for j ¼ 4
(manufacturing and process industry). The maximum decrease by 0.83% per
year is observed for j ¼ 11 (real estate activities, rent, and services for business).
The maximum value of the indicator CO pj 15z equal to 21% was noted for
j ¼ 4 (manufacturing and process industry) in 2008.
– The trend of the indicator CO pj 16z (share of budget of branch 16 that is
housekeeping services) spent for purchasing the intermediate products produced
by the jth branch had a positive slope for branches j ¼ 11, 4 and a negative slope
for j ¼ 1, 14, 16, 8, 15, 6, 9, and 13. Branch 16 did not buy intermediate goods
produced by economic branches 2, 3, 7, 10, 12. The maximum growth of the
trend by 1.2% per year is observed for j ¼ 11 (real estate activities, rent, and
services for business). The maximum decrease by 0.41% per year is observed for
j ¼ 13 (education). Branch 16 did not buy the intermediate goods in 2008.
In Fig. 4.2 we give a graphic illustration of the initial data for retrospective
macroanalysis at the level of economic branches within the framework of the
exogenous variable “Share of budget of branch 1 (agriculture, hunting, and forestry)
spent for purchasing the intermediate products produced by branch 4 (manu-
facturing and process industry).” The negative slope of the linear trend in Fig. 4.2
is evidence of a decrease (on average) of the mentioned share of expenses within the
considered time interval.
Also, we carried out the retrospective analysis of the endogenous variables at the
level of economic branches. In Figs. 4.3 and 4.4, we give a graphic illustration of
the dynamics, its rates, and trends of one of the economic agents’ indicators for
retrospective macroanalysis presented by the example of the indicator “Gross
values added in prices of year 2000” of the branch manufacturing and process
industry.
Note that the computed and measured values of the variable at Fig. 4.2 actually
coincide with each other. It is explained by the quality of parametric identification
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 181
of the considered model. An analysis of data presented in Figs. 4.3 and 4.4
(a positive slope of linear trends of both variable “Gross value-added of branch
4” and its rate) shows accelerated growth of the mentioned indicator within the
interval 2000–2009.
Retrospective Comparative Analysis
Within the framework of retrospective comparative analysis of economic branch
indicators for each year from 2000 to 2008, we also carried out the ranking of
branches with respect to the following indicators:
Linear trend of parameter
Measured parameter values
Estimated parameter values
Fig. 4.2 Exogenous variable “Share of budget of branch 1 (agriculture, hunting, and forestry)
spent for purchasing the intermediate products produced by branch 4 (manufacturing and process
industry)”, its trend, and measured values
Linear trend of variable
Measured variable values Computed variable values
Fig. 4.3 Gross value added of branch 4 (in prices of year 2000), its trend, and measured values
182 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
– Gross value-added (GVA) of a branch (in prices of year 2000);
– Fixed assets of a branch (in prices of year 2000);
– Capital productivity of a branch;
– Demand of labor force in a branch;
– Labor productivity in a branch.
An analysis of the results of ranking with respect to the mentioned indicators
allows us to conclude the following:
– The maximum GVA over all economic branches is observed within the branch
in 2000–2002, in branch 11 (real estate activities, rent, and services for business)
in 2003–2004, and in branch 3 (mineral resource industry) in 2005–2008.
– The maximum fixed assets were discovered in branch 3 (mineral resource
industry) for the entire period of observations.
– The maximum capital productivity was observed in branch 7 (trade, repair
services for cars, and goods for household use) in 2000–2005, and in branch
10 (financial activity) in 2006–2008.
– The maximum number of employees (about 30% of all employees) was noted in
branch 1 (agriculture, hunting, and forestry) for all periods of observations.
– Branch 1 demonstrated the minimum labor productivity in 2000–2008. The
maximum labor productivity was demonstrated by branch 3 (mineral resource
industry) for practically the entire considered period, with a growing gap with
respect to other branches.
The results of the Kazakhstan Republic’s economic branches’ ranking within
the framework of the given indicators in year 2008 are presented in Tables 4.4, 4.5,
4.6, 4.7, and 4.8.
Linear trend of variable rateVariable rate values
Fig. 4.4 Rate of the variable “Gross value added of branch 4 (in constant prices)” and its trend
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 183
Results of the Retrospective Analysis of the Main Integral Indicators of a Country’s
Development, Indicators of Public Administration and Households
Also, we present the results of retrospective analysis of the mentioned indicators for
2000–2009.
Table 4.4 Gross value-added
of branch
No. No. of branch
Value of indicator GVA
in year 2008 (tenge,
in prices of 2000)
1 3 9.92923 � 1011
2 11 7.94163 � 1011
3 7 6.49315 � 1011
4 4 6.24578 � 1011
5 9 5.83432 � 1011
6 6 4.29188 � 1011
7 10 2.79863 � 1011
8 1 2.78125 � 1011
9 13 1.49005 � 1011
10 12 9.00073 � 1010
11 5 8.88712 � 1010
12 15 8.88331 � 1010
13 14 7.84887 � 1010
14 8 4.36384 � 1010
15 16 4.40276 � 109
16 2 4.14457 � 109
Table 4.5 Fixed assets
of branch No. No. of branch
Value of indicator “Fixed assets of a branch”
in year 2008 (tenge, in prices of 2000)
1 3 1.55078 � 1012
2 11 1.15043 � 1012
3 4 6.43257 � 1011
4 9 6.41518 � 1011
5 12 3.21504 � 1011
6 6 2.94698 � 1011
7 5 2.75815 � 1011
8 1 2.33045 � 1011
9 7 2.21753 � 1011
10 13 9.67223 � 1010
11 14 8.86825 � 1010
12 10 7.66333 � 1010
13 15 5.98872 � 1010
14 8 3.36888 � 1010
15 2 1.44711 � 109
16 16 0
184 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
In Figs. 4.5 and 4.6, we present an example of initial data for the retrospective
macroanalysis of the national economic indicators in the context of the endogenous
variable “Volume of production of investment products.”
Analysis of the trends presented in Figs. 4.6 and 4.7 allows us to conclude that
from 2000 to 2009, one can observe an average growth of the real output of
investment products in a country with some average deceleration of its growth.
Table 4.6 Capital
productivity of branch No. No. of branch
Value of indicator “Capital productivity
of a branch” in year 2008
1 10 3.651970
2 7 2.928106
3 2 2.864025
4 13 1.540538
5 15 1.483340
6 6 1.456368
7 8 1.295338
8 1 1.193439
9 4 0.970963
10 9 0.909455
11 14 0.885053
12 11 0.690321
13 3 0.640272
14 5 0.322213
15 12 0.279957
16 16 0.000000
Table 4.7 Branch demand
for labor force No. No. of branch
Value of indicator “Branch demand
for labor force” in year 2008
1 1 2,349,730
2 7 1,150,307
3 13 754,299
4 9 588,847
5 4 572,886
6 6 548,895
7 11 378,129
8 12 352,489
9 14 347,329
10 15 205,425
11 3 200,283
12 5 164,816
13 8 103,048
14 10 96,165
15 16 24,360
16 2 19,971
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 185
Results of Prospective Analysis of the Endogenous Variables
at the Level of Economic Branches
The perspective analysis of the endogenous variables of the Kazakhstan Republic
was carried out on the basis of a computation of economic branches model until
year 2015 with use of extrapolation of the measured values of the exogenous
variables of the model until 2015.
Table 4.8 Labor
productivity in branch No. No. of branch
Value of indicator “Labor productivity
in branch” in year 2008 (in prices of 2000)
1 3 4,957,609.0
2 10 2,910,234.0
3 11 2,100,238.0
4 4 1,090,231.0
5 9 990,804.0
6 6 781,913.2
7 7 564,471.0
8 5 539,213.9
9 15 432,435.0
10 8 423,477.3
11 12 255,348.0
12 14 225,978.1
13 2 207,533.5
14 13 197,540.5
15 16 180,739.0
16 1 118,364.8
Linear trend of variableVariable values
Fig. 4.5 Endogenous variable “Volume of production of investment products” and its trend
186 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The perspective analysis of the endogenous indicators for 2000–2015 by the
branches shows that
The trends of the indicator “Budget of a branch’s enterprises (in prices of year
2000)” increase for all branches. The maximum growth of 7.37 � 1010 tenge per
year is observed in branch 10, whereas the minimum growth of 3.3 � 108 tenge per
year is observed in branch 14. The trends of the rates of the considered indicator
have negative slopes for almost all branches except branches 12 and 4. This shows
that all branches except 12 and 4 are expected to demonstrate a drop in the growth
Linear trend of variable rate
Variable rate values
Fig. 4.6 Rate of endogenous variable “Volume of production of investment products” and its
trend
Linear trend of variable
Computed values of variableMeasured variable values
Fig. 4.7 Gross value added by branch 4 (in prices of year 2000), its trend, and measured values
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 187
rate of budgets (maximum drop by 3% per year by trend was observed in
branch 14).
The trends of the indicator “Gross value-added by branch (in prices of year
2000)” increased for almost all branches (except branches 12 and 14). The maxi-
mum growth of 7.4 � 1010 tenge per year is observed in branch 4, whereas the
maximum drop by 2.2 � 109 tenge per year is observed in branch 14. The trends of
the rates of the considered indicator have negative slopes for almost all branches
except 5, 15, 13, 2, 12, and 14. This shows that all branches except 5, 15, 13, 2, 12,
and 14 are expected to demonstrate a drop in the growth rate of the GVA (maximum
drop by 18% per year by trend was observed in branch 16; maximum growth of
1.6% was observed in branch 14).
The trends of the indicator “Output of branch (in prices of year 2000)” increased
for almost all branches (except branches 16, 2, 3, 14, 11, and 9). The maximum
growth of 1.6 � 1011 tenge per year is observed in branch 6, whereas the maximum
drop of 3.4 � 1010 tenge per year is observed in branch 9. The trends of the rates of
the considered indicator have negative slopes for almost all branches except 4, 5,
12, and 15. This shows that all branches except 4, 5, 12, and 15 are expected to
demonstrate a drop in the growth rate of the output of the branch (a maximum drop
by 19% per year by trend was observed in branch 16; a maximum growth of 1.1%
was observed in branch 4).
The trends of the indicator “Fixed assets of branch (in prices of year 2000)”
increased for almost all branches (except branches 2, 8, 14, 1, and 5). The maximum
growth of 5.5 � 1010 tenge per year is observed in branch 6, whereas the maximum
drop by 6.2 � 109 tenge per year is observed in branch 9. The trends of the rates of
the considered indicator have negative slopes for almost all branches except 10, 2,
8, 14, 1, and 5. This shows that all branches except 10, 2, 8, 14, 1, and 5 are expected
to demonstrate a drop in the growth rate of the fixed assets (a maximum drop by 1%
per year by trend was observed in branch 12; a maximum growth of 0.58% was
observed in branch 10).
The trends of the indicator “Supply of investment products of branch (in prices
of year 2000)” increased for branches 6, 4, 7, 1, 9, 15, and 2. The maximum growth
of 1.1 � 1011 tenge per year is observed in branch 6, whereas the maximum drop of
8.8 � 108 tenge per year is observed in branch 11. The trends of the rates of the
considered indicator have negative slopes for almost all branches or almost zero
slopes for branches 6, 4, 7, 1, 9, 15, and 2. This shows that all these branches except
branch 11 are expected to demonstrate a drop in the growth rate of investment
products. Investment products are not produced for all other branches that are not
mentioned in this paragraph.
The trends of the indicator “Supply of final products of branch (in prices of year
2000)” increased for almost all branches (except branches 11, 2, 14, and 9). The
maximum growth of 3.1 � 1010 tenge per year is observed in branch 4, whereas the
maximum drop of 6.9 � 109 tenge per year is observed in branch 9. The trends of
the rates of the considered indicator have negative slopes for almost all branches
except 15 and 7. This shows that all branches except 15 and 7 are expected to
demonstrate a drop in the growth rate of the supply of final products (the maximum
188 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
drop of 15% per year by trend was observed in branch 12; the maximum growth of
0.28% was observed in branch 15).
The trends of the indicator “Supply of intermediate products of branch (in prices
of year 2000)” increased for almost all branches (except branches 16, 2, 14, 3, and 9).
The maximum growth of 6.6 � 1010 tenge per year is observed in branch 4, whereas
the maximum drop of 1.0 � 1010 tenge per year is observed in branch 9. The trends
of the rates of the considered indicator have negative slopes for almost all branches
except 5 and 15. This shows that all branches except 5 and 15 are expected to
demonstrate a drop in the growth rate of the supply of intermediate products (the
maximum drop, with considerable jumps, was observed in branch 16; the maximum
growth of 2.9% was observed in branch 5).
The trends of the indicator “Supply of export products of branch (in prices of
year 2000)” increased for almost all branches (except branches 2, 13, 11, 1, and 9).
The maximum growth of 3.0 � 1010 tenge per year is observed in branch 4,
whereas the maximum drop of 1.6 � 109 tenge per year is observed in branch 9.
The trends of the rates of the considered indicator have negative slopes for almost
all branches except 12, 8, 13, and 1. This shows that all branches except 12, 8, 13,
and 1 are expected to demonstrate a drop in the growth rate of the supply of export
products (the maximum drop of 22% per year by trend was observed in branch 15;
the maximum growth of 3.3% was observed in branch 13). The export products are
not produced by branches 14 and 16.
The trends of the indicator “Profit of branch (in prices of year 2000)” increased
for almost all branches (except branches 16, 14, 12, and 13). The maximum growth
of 6.6 � 1010 tenge per year is observed in branch 4, whereas the maximum drop of
9.7 � 109 tenge per year is observed in branch 13. The trends of the rates of the
considered indicator have negative slopes for almost all branches except 4, 5, 1, 15,
2, and 14. This shows that all branches except 4, 5, 1, 15, 2, and 14 are expected to
demonstrate a drop in the growth rate of the profit of branch (the maximum drop of
15% per year by trend was observed in branch 12; the maximum growth of 4.0%
was observed in branch 15).
In Figs. 4.7 and 4.8, we present an example of the initial data for the retrospec-
tive macroanalysis at the level of economic branches on the basis of the endogenous
variables and in the context of the endogenous variable “Gross value-added (in
prices of year 2000) of economic agent 4 (manufacturing and process industry),” its
rate, and their trends.
Analysis of information presented in Figs. 4.7 and 4.8 and comparison of those
figures with Figs. 4.2 and 4.3 show that in perspective up until year 2015, one can
expect a growth (on average) in GVA of branch 4 with an increase (on average) in
the rate of that growth. That is, the tendencies similar to those of Figs. 4.2 and 4.3
noted before remain the same (according to the base prediction) for 2010–2015.
Nevertheless, the angular coefficient of the trend in Fig. 4.7 is somewhat higher
than the respective angular coefficient in Fig. 4.2. This is evidence to show an
increase (on average) in the rate of that growth in the perspective period. This is
also seen by a greater (on average) value of the trend in Fig. 4.8 in comparison with
its values for the retrospective period (Fig. 4.3).
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 189
Within the framework of perspective analysis of economic branch indicators on
the basis of the considered model, we also obtained the computed prediction values
(and their linear trend) of the indicators “Demand for intermediate products of jthbranch i (in constant prices of year 2000)” (i, j ¼ 1, . . ., 16) for each of the 16
economic branches.
The perspective analysis of the mentioned indicators of branches for 2000–2015
shows that
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 1 (agriculture, hunting, and forestry) (in constant prices of year 2000)”
had a positive slope for branches j ¼ 4, 5, 16, 10, 7, 12, 8, 16 and a negative
slope for branches j ¼ 13, 2, 15, 14, 3, 11, and 9. Branch 1 does not purchase
intermediate products of economic branch 16. The maximum increase of the
trend by 2.3 � 109 tenge per year is observed for j ¼ 4, whereas the maximum
decrease by 2.3 � 109 tenge per year is observed for j ¼ 9. The trend of the rate
of the considered indicator increases only for products of branch 5 and decreases
for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 2 (fishery and fish breeding) (in constant prices of year 2000)” had a
positive slope for branches j ¼ 4, 5, 9, 10, 11, 12, 6, 3, 7, and 8 and a negative
slope for branches j ¼ 15, 14, 1, and 2. Branch 2 did not purchase intermediate
products from economic branches 13 and 16. The maximum increase of the trend
by 5.1 � 107 tenge per year is observed for j ¼ 4, whereas the maximum
decrease by 7.7 � 107 tenge per year is observed for j ¼ 2. The trend of the
rate of the considered indicator increases only for products of branches 12, 6, and
8 and decreases for all other branches.
Linear trend of variable rate
Variable rate values
Fig. 4.8 Rate of variable “Gross value added by branch 4 (in fixed prices)” and its trend
190 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 3 (mineral resource industry) (in constant prices of year 2000)” had a
positive slope for branches j ¼ 5, 4, 10, 7, 6, 8, 12, 15, and 11 and a negative
slope for branches j ¼ 2, 1, 16, 13, 14, 3, and 9. The maximum increase of the
trend by 4.8 � 109 tenge per year is observed for j ¼ 5, whereas the maximum
decrease of 3.1 � 109 tenge per year is observed for j ¼ 9. The trend of the rate
of the considered indicator increases only for products of branches 3 and 16 and
decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of jth branch in
branch 4 (manufacturing and process industry) (in constant prices of year 2000)”
had a positive slope for branches j ¼ 4, 5, 10, 6, 11, 8, and 12 and a negative
slope for branches j ¼ 16, 15, 14, 13, 2, 1, 3, 7, and 9. The maximum increase of
the trend by 1.5 � 1010 tenge per year is observed for j ¼ 4, whereas the
maximum decrease by 2.3 � 109 tenge per year is observed for j ¼ 9. The
trend of the rate of the considered indicator increases only for products of
branches 4, 5, 6, 12, and 16 and decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 5 (manufacturing and process industry) (in constant prices of year 2000)”
had a positive slope for branches j ¼ 5, 4, 6, 10, 3, 7, 11, 12, 15, 8, 1, 2 and a
negative slope for branches j ¼ 16, 9, 13, and 14. The maximum increase of the
trend by 1.2 � 1010 tenge per year is observed for j ¼ 5, whereas the maximum
decrease by 1.3 � 107 tenge per year is observed for j ¼ 14. The trend of the
rate of the considered indicator increases only for products of branches 7, 2, and
16 and decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 6 (construction) (in constant prices of year 2000)” had a positive slope for
branches j ¼ 4, 6, 5, 7, 10, 11, 8, 12, 1, 15, 13 and a negative slope for branches
j ¼ 2, 14, 3, and 9. Branch 6 does not purchase intermediate products of
economic branch 6. The maximum increase of the trend by 1.0 � 1010 tenge
per year is observed for j ¼ 4, whereas the maximum decrease by 3.7 � 108
tenge per year is observed for j ¼ 9. The trend of the rate of the considered
indicator decreases for all 15 branches.
– The trend of the indicator “Demand for intermediate products of jth branch in
branch 7 (trade, repair services for cars, and goods for household use) (in
constant prices of year 2000)” had a positive slope for branches j ¼ 10, 4, 5,
11, 8, 7, 12 and a negative slope for branches j ¼ 13, 2, 14, 15, 1, 3, and 9.
Branch 7 does not purchase intermediate products of economic branch 16. The
maximum increase of the trend by 3.8 � 109 tenge per year is observed for
j ¼ 10, whereas the maximum decrease by 1.9 � 109 tenge per year is observed
for j ¼ 9. The trend of the rate of the considered indicator increases only for
products of branches 4, 5, 6, 15, and 1 and decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 8 (hotels and restaurants) (in constant prices of year 2000)” had a positive
slope for branches j ¼ 4, 5, 10, 1, 6, 7, 8, 9, 15, 11, 12, 3 and a negative slope for
branches j ¼ 14, 13, and 2. Branch 8 does not purchase intermediate products of
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 191
economic branch 16. The maximum increase of the trend by 1.0 � 109 tenge per
year is observed for j ¼ 4, whereas the maximum decrease by 1.2 � 107 tenge
per year is observed for j ¼ 2. The trend of the rate of the considered indicator
increases only for products of branches 4, 8, 15, and 12 and decreases for all
other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 9 (transportation and communications) (in constant prices of year 2000)”
had a positive slope for branches j ¼ 4, 5, 10, 6, 8, 12, 11 and a negative slope
for branches j ¼ 16, 2, 15, 14, 13, 1, 7, 3, and 9. The maximum increase of the
trend by 4.5 � 109 tenge per year is observed for j ¼ 4, whereas the maximum
decrease by 1.2 � 109 tenge per year is observed for j ¼ 9. The trend of the rate
of the considered indicator increases only for products of branches 5, 16, 15, and
7 and decreases for other branches.
– The trend of the indicator “Demand for intermediate products of branch j inbranch 10 (financial activity) (in constant prices of year 2000)” had a positive
slope for all branches except 2 and 4. The maximum increase of the trend of
1.1 � 1010 tenge per year is observed for j ¼ 10, whereas the maximum
decrease by 4.8 � 105 tenge per year is observed for j ¼ 14. The trend of the
rate of the considered indicator increases only for products of branches 16 and
2 and decreases for other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 11 (real estate activities, rent, and services for business) (in constant
prices of year 2000)” had a positive slope for branches j ¼ 4, 10, 5, 6, 7, 8, 12, 9,
15, 13 and a negative slope for branches j ¼ 18, 2, 11, 1, 14, and 3. The
maximum increase of the trend of 5.3 � 109 tenge per year is observed for
j ¼ 4, whereas the maximum decrease by 3.0 � 108 tenge per year is observed
for j ¼ 3. The trend of the rate of the considered indicator increases only for
products of branch 15 and decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 12 (public administration) (in constant prices of year 2000)” had a
positive slope for branches j ¼ 5, 4, 10, 6, 8, 12, 15, 1, 11, 13 and a negative
slope for branches j ¼ 2, 14, 3, 7, and 9. Branch 12 does not purchase interme-
diate products of economic branch 16. The maximum increase of the trend of
3.2 � 109 tenge per year is observed for j ¼ 5, whereas the maximum decrease
by 2.8 � 108 tenge per year is observed for j ¼ 9. The trend of the rate of the
considered indicator increases only for products of branches 6, 12, and 7 and
decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 13 (public administration) (in constant prices of year 2000)” had a
positive slope for branches j ¼ 5, 4, 6, 15, 10, 13, 12, 8, 1 and a negative
slope for branches j ¼ 2, 3, 9, 7, 14, and 11. Branch 13 does not purchase
intermediate products of economic branch 16. The maximum increase of the
trend of 4.2 � 108 tenge per year is observed for j ¼ 5, whereas the maximum
decrease by 2.8 � 108 tenge per year is observed for j ¼ 11. The trend of the
192 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
rate of the considered indicator increases only for products of branch 13 and
decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 14 (public health and social services) (in constant prices of year 2000)”
had a positive slope for branches j ¼ 4, 5, 6, 7, 10, 15, 8, 12, 13 and a negative
slope for branches j ¼ 2, 3, 11, 1, 14, and 9. Branch 14 does not purchase
intermediate products of economic branch 16. The maximum increase of the
trend of 5.1 � 109 tenge per year is observed for j ¼ 4, whereas the maximum
decrease by 7.0 � 108 tenge per year is observed for j ¼ 5. The trend of the rate
of the considered indicator increases only for products of branch 13 and
decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of branch j inbranch 15 (other public utilities, social and personal services) (in constant prices
of year 2000)” had a positive slope for branches j ¼ 4, 5, 10, 15, 7, 6, 8, 12, 11,
13, 1 and a negative slope for branches j ¼ 16, 2, 3, 9, and 14. The maximum
increase of the trend of 2.8 � 109 tenge per year is observed for j ¼ 4, whereas
the maximum decrease by 6.3 � 107 tenge per year is observed for j ¼ 14. The
trend of the rate of the considered indicator increases only for products of
branches 4 and 16 and decreases for all other branches.
– The trend of the indicator “Demand for intermediate products of the jth branch inbranch 16 (housekeeping services) (in constant prices of year 2000)” had a
positive slope for branches j ¼ 6, 4, 8 and a negative slope for branches j ¼ 1,
16, 14, 15, 9, 11, and 13. Branch 16 does not purchase intermediate products of
economic branches 10, 5, 12, 7, and 2. The maximum increase of the trend of
8.8 � 106 tenge per year is observed for j ¼ 6, whereas the maximum decrease
by 1.0 � 107 tenge per year is observed for j ¼ 13. The trend of the rate of the
considered indicator increases only for products of branches 4, 16, and 15 and
decreases for all other branches.
In Figs. 4.9 and 4.10, we present an example of initial data for the perspective
macroanalysis of the dynamics (and its trend) and rate (and its trend) of economic
branches in the context of the endogenous variable “Demand for intermediate
products of branch 4 (manufacturing and process industry) in branch 1 (agriculture,
hunting, and forestry) (in prices of year 2000)”.
A positive slope of the linear trend of the indicator (Fig. 4.9) argues for an
increase (on average) of the considered branch demand within the considered time
interval. At the same time, a negative slope of the trend of the rate of this indicator
argues for a decrease (on average within the period of 2000–2015) of the indicator
growth rate.
Perspective Comparative Analysis of the Indicators of Economic Branches
Within the framework of perspective comparative analysis of economic branches’
indicators for each year from 2009 to 2015, we also carried out a ranking of the
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 193
branches with respect to the same indicators just as in the case of retrospective
macroanalysis:
– Gross value-added (GVA) of a branch (in prices of year 2000);
– Fixed assets of a branch (in prices of year 2000);
– Capital productivity of a branch;
– Demand for labor force in a branch;
– Labor productivity in a branch.
Linear trend of variable rateComputed variable values
Fig. 4.9 Demand of intermediate products of branch 4 in branch 1 (in prices of year 2000) and
its trend
Linear trend of variable rateVariable rate values
Fig. 4.10 Rate of the variable “Demand of intermediate products of branch 4 in branch 1 (in
prices of year 2000)” and its trend
194 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Analysis of the results of ranking with respect to the mentioned indicators allows
us to conclude, e.g., the following:
– The maximum real GVA over economic branches is expected in branch 4
(manufacturing and process industry) in 2010–2015. The growth of the GVA
in this branch in 2015 will be equal to about 86% in comparison with 2009.
– The maximum fixed assets over economic branches is expected in branch 3
(mineral resource industry) for the entire period of retrospective and perspective
prediction. The growth of the fixed assets in this branch in 2015 will be equal to
about 9.2% in comparison with 2009.
– The maximum capital productivity over economic branches is expected in
branch 2 (fishery and fish-breeding) in 2010–2015. The growth of capital
productivity in this branch in 2015 will be more than 60 times the capital
productivity of branch 12 (public administration).
– The maximum number of employees is expected in branch 1 (agriculture,
hunting, and forestry) in 2000–2015. In 2015, the demand for the labor force
in this branch will constitute 2,581,607 people. The minimum number of people
starting from 2012 is expected in branch 16 (housekeeping services); in 2015, it
will constitute 24,361.
– The maximum labor productivity is expected in branch 10 (financial activity) in
2010–2015. In 2015, the branch labor of productivity will be 23 times greater
than branch 14 (public health and social services).
The results of the Kazakhstan Republic’s economic branches’ ranking within the
framework of the mentioned indicators in year 2015 are presented in Tables 4.9,
4.10, 4.11, 4.12, and 4.13.
Table 4.9 Gross value-added
of branch
No. No. of branch
Value of indicator GVA
in year 2015 (tenge,
in prices of 2000)
1 4 1.45122 � 1012
2 6 6.65949 � 1011
3 3 6.57340 � 1011
4 7 5.87018 � 1011
5 11 5.44444 � 1011
6 10 4.38485 � 1011
7 1 4.25092 � 1011
8 9 4.12240 � 1011
9 5 3.05231 � 1011
10 15 2.09079 � 1011
11 13 1.99452 � 1011
12 8 9.18216 � 1010
13 12 7.56724 � 1010
14 14 5.51708 � 1010
15 2 7.43652 � 109
16 16 5.15982 � 109
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 195
Perspective Analysis of the Main Integral Indicators of a Country’s Development,
Indicators of Public Administration and Households
The perspective analysis of the aforementioned indicators in 2000–2015 shows the
following:
– GVA of the country (in prices of year 2000) in the considered period grows with
an average rate of the trend equal to 2.5 � 1011 tenge per year. The rate of the
GVA indicator decreases (by its trend) by 0.84% per year.
Table 4.10 Fixed assets of branch
No. No. of branch
Value of indicator “Fixed assets of branch”
in year 2015 (tenge, in prices of 2000)
1 3 1.72577 � 1012
2 11 1.33253 � 1012
3 9 7.32356 � 1011
4 12 7.25682 � 1011
5 4 6.67990 � 1011
6 6 3.87501 � 1011
7 5 2.88265 � 1011
8 7 2.78329 � 1011
9 1 2.08751 � 1011
10 10 1.41096 � 1011
11 13 1.16788 � 1011
12 14 7.83038 � 1010
13 15 7.73435 � 1010
14 8 2.76785 � 1010
15 2 1.16724 � 1010
16 16 0
Table 4.11 Capital productivity of branch
No. No. of branch
Value of indicator “Capital productivity
of a branch” in year 2015
1 2 6.371005
2 8 3.317435
3 10 3.107695
4 15 2.703250
5 4 2.172512
6 7 2.109079
7 1 2.036358
8 6 1.718572
9 13 1.707822
10 5 1.058857
11 14 0.704573
12 9 0.562895
13 11 0.408581
14 3 0.380897
15 12 0.104278
16 16 0.000000
196 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
– Gross products’ output (in prices of year 2000) in the considered period grows
with average rate of the trend equal to 4.0 � 1011 tenge per year. The rate of this
indicator decreases (by its trend) by 0.45% per year.
– Investment products’ output (in prices of year 2000) in the considered period
grows with an average rate of the trend equal to 1.2 � 1011 tenge per year. The
rate of this indicator decreases (by its trend) by 2.0% per year.
– Final products’ output (in prices of year 2000) in the considered period grows
with an average rate of the trend equal to 9.9 � 1010 tenge per year. The rate of
this indicator decreases (by its trend) by 0.46% per year.
Table 4.12 Branch demand
for labor force No. No. of branch
Value of indicator “Branch demand
for labor force” in year 2015
1 1 2,581,608
2 7 1,235,941
3 13 963,120
4 6 841,841
5 9 604,098
6 4 595,420
7 11 516,757
8 12 406,500
9 14 401,955
10 3 247,131
11 15 225,833
12 5 184,281
13 8 150,238
14 10 139,326
15 2 27,907
16 16 24,361
Table 4.13 Labor
productivity in branch No.
No. of
branch
Value of indicator “Labor productivity
in branch” in year 2015 (in prices of 2000)
1 10 3,147,177.0
2 3 2,659,884.0
3 4 2,437,295.0
4 5 1,656,329.0
5 11 1,053,578.0
6 15 925,811.3
7 6 791,063.1
8 9 682,406.4
9 8 611,171.8
10 7 474,956.1
11 2 266,466.3
12 16 211,806.2
13 13 207,089.8
14 12 186,155.8
15 1 164,661.7
16 14 137,256.4
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 197
– Intermediate products’ output (in prices of year 2000) in the considered period
grows with an average rate of the trend equal to 1.7 � 1011 tenge per year. The
rate of this indicator decreases (by its trend) by 1.2% per year.
– Export products’ output (in prices of year 2000) in the considered period grows
with an average rate of the trend equal to 5.1 � 1010 tenge per year. The rate of
this indicator decreases (by its trend) by 1.4% per year.
– State budget (in prices of year 2000) in the considered period grows with an
average rate of the trend equal to 1.0 � 1011 tenge per year. The rate of this
indicator decreases (by its trend) by 1.5% per year.
– Social transfers to households (in prices of year 2000) in the considered period
grows with an average rate of the trend equal to 1.0 � 1011 tenge per year. The
rate of this indicator decreases (by its trend) by 1.5% per year.
– Fixed assets to households (in prices of year 2000) in the considered period
grows with an average rate of the trend equal to 2.2 � 1010 tenge per year. The
rate of this indicator decreases (by its trend) by 0.79% per year.
– Capital productivity in the considered period grows with an average rate of the
trend equal to 1.4 � 10�2 1/(year)2. The rate of this indicator decreases (by its
trend) by 0.69% per year.
– The total number employed within the mentioned period grows with the average
rate of trend equal to 1.9 � 105 people per year. The rate of this indicator
decreases (by its trend) by 0.14% per year.
– Labor productivity in the mentioned period grows with the average rate of trend
equal to 1.9 � 104 tenge/(people � year2). The rate of this indicator decreases
(by its trend) by 0.68% per year.
– The consumer price index in the mentioned period fluctuates within the limits of
6–8% per year along with an insignificant drop in the trend of this indicator with
the rate of 0.008% per year.
– Real prices of investment products in the considered period drop with the
average rate of trend equal to 1.2% per year. The rate of this indicator drops
(by its trend) by 0.16% per year.
– Money of households in banking accounts (in prices of year 2000) in the
mentioned period increases, with the average rate of the trend equal to
4.3 � 109 tenge per year. The rate of this indicator drops (by its trend) by
0.047% per year.
In Figs. 4.11 and 4.12, we carry out the analysis of initial data for the perspective
macroanalysis of the national economy indicators in the context of the endogenous
variable “Investment products output.”
Analysis of the trends presented in Figs. 4.11 and 4.12 allows us to conclude that
in 2000–2015, just as in the retro-analysis period (Figs. 4.6 and 4.7) , one can expect
average growth of the real output of investment products in the country with some
average deceleration of this growth. However, the decrease of a negative slope of
the rate trend (Fig. 4.12) for the period of 2000–2015 in comparison with the period
of retro-analysis (Fig. 4.6) argues for some increase of the rate of growth of
investment products by the end of the considered period.
198 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.1.2.2 Analysis of Elasticity Coefficients of Endogenous Variablesof Computable Model of General Equilibriumof Economic Branches
Within the limits of the retrospective analysis of economic branches, we obtained
the estimates of the elasticity coefficients of 18 endogenous variables for each of 16
economic branches, as well as 14 integral indicators of the country economy, public
Linear trend of variable rateComputed variable values
Fig. 4.11 Investment products output (in prices of year 2000) and its trend
Linear trend of variable rateVariable rate values
Fig. 4.12 Rate of variable “Investment products output (in prices of year 2000)” and its trend
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 199
administration, and households in 2009 over 341 exogenous indicators of the model
for 2008–2014. These elasticity coefficients form a matrix of size 302 � 341 with
the elements defined by
FpjðtÞ ¼ 100xnj ðtÞ � xjðtÞ
xjðtÞ : (4.71)
Here p is the varied exogenous parameter; xjðtÞ is the value of the jth endogenousvariable for t ¼ 2009 obtained from base computation of the model; xnj ðtÞ is the
value of the respective endogenous variable obtained from the increase in the
values of the varied exogenous parameter p from 2000 to 2008; other values of
the exogenous parameters remain unchanged in comparison with the base
computation.
For example, the computed elasticity coefficient of the households’ budget by
the share of budget of branch 1 spent for paying the labor force appears to be
0.0229. This means that if the share of budget of branch 1 spent for paying the labor
force in each of 2000–2008 is increased by 1% in comparison with the base variant,
then in 2009, the households’ budget will increase by 0.0229 in comparison with the
base variant.
Within the limits of the perspective analysis of economic branches, we also
obtain the similar elasticity coefficients of endogenous indicators in 2015 by
exogenous indicators in 2009.
Analysis of influence of the exogenous parameters and, in particular, the
parameters CO_pj_iz, “Share of budget of the ith branch spent for purchasing
intermediate products produced by the jth branch,” on the country’s GVA shows
that
– The mean value of absolute values of the GVA elasticity coefficients by the
shares CO_pj_iz appears to be rather high and equal to 4.7450; and
– The mean value of absolute values of the GVA elasticity coefficients by shares
of budgets of other branches appears to be 2.379.
This remark allows recommendation of parameters CO_pj_iz for use in
parametric regulation of the national economic growth. We shall consider
experiments with the parametric regulation in 4.1.3.
4.1.2.3 Analysis of Economic Growth Sources on the Basisof the Computable Model of General Equilibriumof Economic Branches
Let’s now proceed to the analysis of economic growth sources by economic
branches on the basis of the retrospective data for the years 2000–2015. For this
purpose, using expressions (4.16) for the production functions, we’ll estimate the
influence of variation of these functions’ arguments on the rate of GVA growth of
200 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
branches VY__i[t + 1] under assumption about the constancy of the coefficients of
the intermediate products consumed by the branch CA_z_ji[t], coefficients of thecapital CA_k_i[t], and coefficients of labor CA_l_i[t]. Such an assumption is
applied in extrapolation of the mentioned functions within the prediction period
2010–2015 of the model computation.
Taking the logarithm of both parts of (4.20), finding the total increment of the
function ln(VY__i), and discarding the terms of higher infinitesimal order, we
obtain the following estimate of the growth rate yi of the real GVA of the ith branchin dependence on the growth rates of the production function arguments: CA_r_i,VD_pj_iz, Kiср ¼ (VK__i[t] + VK__i[t + 1])/2, and VD_pi_il[t]:
yi ¼DVY i
VY i¼ DCA r i
CA r iþX16j¼1
CA z ji� VD pj izð ÞDVD pj iz
VD pj iz
þ CA k iDKimKim
þ CA l iDVD pi il
VD pi il: (4.72)
Let ai ¼ DCA r iCA r i denote the rate of the technological progress in the ith branch; let
zij ¼ DVD pj izVD pj iz denote the rate of the intermediate products of the jth branch consumed
in the ith branch; letki ¼ DKimKim
denote the rate of capital accumulation in the ith branch;
let li ¼ DVD pi ilVD pi iz denote the rate of labor inputs growth in the ith branch, where the
sign “D” means the variable increment. The time in (4.72) is omitted for brevity.
The coefficients on the right-hand side of formula (4.72) with the rates indicated
above characterize the degree of influence of the considered factors on the eco-
nomic growth and allow us to compare their influence with that of the technological
progress with the coefficient equal to 1. Denoting these coefficients by
aij ¼ CA z ji� VD pj iz , bi ¼ CA k i , gi ¼ CA l i , from (4.72) we obtain its
abbreviated notation:
yi ¼ ai þX16j¼1
aijzij þ biki þ gili: (4.73)
Now we’ll present the values of the coefficients defining the contributions of the
sources of economic growth of the branches on the basis of the considered model in
2008 (Table 4.14). The coefficients in the table show the percentage of increase of
the GVA growth rate as a result of increase of growth factors (fixed assets, labor or
demand for intermediate products of the economic agents) by 1%.
Analysis of Table 4.15 shows that, except for the rate of technological progress
with equal influence on the rate of growth of all branches, which is the same in the
considered model, the maximum influence on the growth of GVA of economic
branches 1, 2, 5, 7, 12, 13, and 16 is exerted by the growth rate of labor inputs, GVA
of branches 4, 6, 8, 9, 10, 11, and 15 are exerted by the rate of the capital
accumulation, GVA of the branches 3 and 14 are exerted by the rate of intermediate
products of other branches consumed by the branch.
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 201
Table
4.14
Coefficientscharacterizinginfluence
offactorsofeconomic
growth
Number
ofbranch
ib i
g ia i1
a i2
a i3
a i4
a i5
a i6
a i7
10.3089
0.9051
1.345�10
�12
9.480�10
�02
2.897�10
�14
2.171�10
�13
1.602�10
�14
2.028�10
�14
1.345�10
�12
20.2426
2.4964
1.590�10
�16
7.308�10
�01
7.087�10
�16
9.884�10
�15
1.390�10
�15
1.120�10
�15
1.590�10
�16
30.9650
0.6886
2.886�10
�15
0.000
2.970�10
�12
8.269�10
�13
9.894�10
�14
1.478
2.886�10
�15
41.2900
0.0805
2.227�10
�13
1.343�10
�18.634�10
�13
9.989�10
�13
9.029�10
�03
2.641�10
�16
2.227�10
�13
51.0
�10�1
02.4083
3.199�10
�17
5.537�10
�17
8.913�10
�14
6.720�10
�45.406�10
�35.300�10
�43.199�10
�17
60.9343
0.7721
2.078�10
�16
9.186�10
�18
3.313�10
�14
3.940�10
�13
2.467�10
�15
1.324�10
�14
2.078�10
�16
71.0
�10�1
01.8792
8.133�10
�16
1.061�10
�15
3.115�10
�14
2.660�10
�13
4.124�10
�14
2.508�10
�13
8.133�10
�16
81.0691
0.4706
2.003�10
�02
1.498�10
�15
4.343�10
�17
6.171�10
�14
3.375�10
�15
7.288�10
�15
2.003�10
�02
90.8660
0.2153
4.289�10
�16
2.265
1.666�10
�13
8.753�10
�15
3.057�10
�14
4.005�10
�14
4.289�10
�16
10
0.6702
0.5492
0.000
0.000
0.000
4.397�10
�15
5.602�10
�15
8.811�10
�15
0.000
11
1.2022
0.1006
2.929�10
�05
2.267
5.064�10
�14
1.030�10
�12
1.664�10
�04
5.308�10
�13
2.929�10
�05
12
1.0
�10�1
02.5822
5.415�10
�14
0.000
0.000
4.255�10
�13
2.573�10
�14
0.000
5.415�10
�14
13
0.2635
1.7177
2.370�10
�14
0.000
0.000
8.847�10
�13
1.142�10
�13
5.739�10
�15
2.370�10
�14
14
0.0227
1.7814
6.736�10
�14
2.018�10
�16.280�10
�15
1.288�10
�12
1.919�10
�13
4.809�10
�13
6.736�10
�14
15
0.9304
0.2173
1.041�10
�15
1.408�10
�37.681�10
�15
3.712�10
�13
3.926�10
�14
1.579�10
�13
1.041�10
�15
16
01.9372
00
00
00
0
202 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table
4.15
Number
ofbranch
ia i8
a i9
a i10
a i11
a i12
a i13
a i14
a i15
a i16
14.325�10
�15
4.595�10
�21.807�10
�14
1.177�10
�14
01.681�10
�16
1.178�10
�15
9.304�10
�18
0
28.054�10
�17
1.087�10
�14
2.616�10
�17
2.280
00
00
0
34.633�10
�14
2.221�10
�21.436�10
�13
4.626�10
�12
01.581�10
�15
4.991�10
�4
1.571�10
�14
0
48.038�10
�15
1.444�10
�13
4.353�10
�14
6.338�10
�16
01.247�10
�16
3.379�10
�3
5.445�10
�16
0
51.642�10
�15
2.760�10
�16.132�10
�3
2.512�10
�14
03.273�10
�52.026�10
�17
5.494�10
�40
67.977�10
�15
1.117�10
�26.525�10
�16
2.014�10
�14
08.436�10
�17
1.554�10
�17
1.089�10
�15
0
76.067�10
�14
4.548�10
�13.407�10
�13
1.505�10
�12
05.394�10
�16
8.576�10
�5
4.404�10
�16
0
81.528�10
�15
8.267�10
�15
3.053�10
�14
1.507�10
�14
02.339�10
�17
1.338�10
�17
2.244�10
�16
0
93.346�10
�14
1.002�10
�12
3.203�10
�13
4.339�10
�13
03.657�10
�15
4.803�10
�16
1.091�10
�15
0
10
1.991�10
�14
3.270�10
�01
2.862�10
�13
5.408�10
�15
01.347�10
�14
08.805�10
�18
0
11
6.664�10
�14
6.596�10
�13
2.965�10
�13
2.646�10
�12
09.260�10
�14
1.661�10
�03
9.599�10
�06
0
12
4.185�10
�1
1.548�10
�19.506�10
�13
7.964�10
�30
2.126�10
�14
1.408�10
�15
2.886�10
�15
0
13
4.820�10
�14
5.807�10
�11.794�10
�14
1.331�10
�13
03.939�10
�15
7.738�10
�17
1.835�10
�14
0
14
1.891
2.118�10
�13
4.749�10
�17
2.743�10
�13
02.219�10
�16
1.782�10
�14
6.646�10
�14
0
15
6.285�10
�14
2.139�10
�19.739�10
�14
8.436�10
�14
09.538�10
�17
1.416�10
�17
1.222�10
�13
0
16
00
00
00
00
0
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 203
4.1.2.4 Estimation of the Macroeconomic Theory Provisions on the Basisof the Computable Model of General Equilibriumof Economic Branches
In this section we describe the results of computational experiments focused on the
estimation of some provisions of the macroeconomic theory on cyclic oscillations
of macroeconomic indicators as a result of shock changes of demands of final and
investment products.
Within the limits of the macroeconomic theory, cyclic oscillations of economic
processes can appear in the presence of [40]
– Linear dependence between the volume of consumer expenses and current
income;
– Linear dependence between investments and increase of income.
To verify these provisions, we carried out a number of computational
experiments aimed at computation of the following scenarios of variation of the
mentioned demands of final and investment products:
(a) O1½t� ¼ O1½t� 1� þ aðYg½t� � Yg½t� 1�Þ; (4.74)
(b) O2i½t� ¼ O2i½t� 1� þ bðYg½t� 3� � 2Yg½t� 2� þ Yg½t� 1�Þði ¼ 1; . . . ; 16Þ; (4.75)
(c) Joint application of scenarios (a) and (b);
(d) Increase of the shares O1ðtÞ by k times in comparison with the base variant;
(e) Increase of the shares O2iðtÞ ði ¼ 1; . . . ; 16Þ by l times in comparison with the
base variant;
(f) Joint application of scenarios (d) and (e).
Here t ¼ 2010; . . . ; 2015 is the time in years;O1½t� ¼ CO p 17c½t� is the share ofbudget of the households spent for purchasing final products (exogenous function);
O2iðtÞ is the share of the budget of the ith branch spent for purchasing investment
products ði ¼ 1; . . . ; 15Þ (exogenous function); Yg½t� ¼ VY g½t� is the sum of
branches’ GVA in constant prices (endogenous variable); a, b, l, and k are some
positive constants (l>1; k>1).
The application of scenarios (a) and (b) means the model computation with the
values of the parameters O1 and O2i defined by formulas (4.74) and (4.75),
respectively, starting from t ¼ 2010 . The values of all other exogenous model
parameters for scenarios (a)–(f) correspond to its base variant.
As a result of computational experiments on the base of the model aimed at
realization of scenarios (a)–(f), one can observe cyclic oscillations of the variable P(t) (consumer price index) (see Figs. 4.13, 4.14, 4.15, 4.16, 4.17, and 4.18). We did
not reveal the oscillation phenomenon for the values of the real indicators (in
particular, Yg[t]).The experimental results presented in Figs. 4.13, 4.14, 4.15, 4.16, 4.17, and 4.18
prove the respective provisions of the macroeconomic theory [40].
204 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.1.3 Finding Optimal Parametric Control Laws on the Basisof the CGE Model of Economic Branches
4.1.3.1 Attenuation of Cyclic Oscillations of Macroeconomic Indicatorsby Parametric Control Methods
In computational experiments we consider the following problem of attenuation of
cyclic oscillations of the consumer price level arising in application of scenario (c)
of the economic system development taking into account the linear dependence
Base Scenario
Fig. 4.13 Values of the consumer price index with application of scenario (a) with a ¼ 3 � 10�12
and for the base variant
Base Scenario
Fig. 4.14 Values of the consumer price index with application of scenario (b) with b ¼ 10�13 and
for the base variant
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 205
between the volume of consumer expenses and current income, as well as the linear
dependence between investments and income growth.
On the basis of the computable model of economic branches with application of
scenario (c) (see Sect. 4.1.2.4, where a ¼ 3 � 10�13, b ¼ 10�13), find the values of
the shares Oji½t� ¼ CO pj iz½t� of budgets of the j-th producing agent spent for
purchasing goods and services produced and rendered by the i-th producing agent
in 2010–2015, i; j ¼ 1; . . . ; 16, which provide the lower bound of the following
Base Scenario
Fig. 4.15 Values of the consumer price index with application of scenario (c) with a ¼ 3 � 10�13,
b ¼ 10�13 and for the base variant
Base Scenario
Fig. 4.16 Values of the consumer price index with application of scenario (d) with k ¼ 1.2 and
for the base variant
206 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
criterion KP which characterizes the deviations of the computed values of the
consumer price index VP½t� from the respective desired values of P½t�:
KP ¼X2015t¼2010
VP½t� � P½t�P½t�
� �2:
Here we use the computed base values of the model consumer price index
without parametric control as the desired values of P[t].
Base Scenario
Fig. 4.17 Values of the consumer price index with application of scenario (e) with l ¼ 1.4 and for
the base variant
Base Scenario
Fig. 4.18 Values of the consumer price index with application of scenario (f) with k ¼ 1.2,
l ¼ 1.4 and for the base variant
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 207
The constraints on the adjusted parameters are as follows:
0:5 � Oji½t� O
j
i
.� 1:5;
P16i¼1 O
jiðtÞ � 1; where i; j ¼ 1; . . . ; 16; t ¼ 2010; :::; 2015.
Here Oj
i are the base values of the mentioned shares obtained as a result of
solving the parametric identification problem by data of 2000–2008.
The constraints on the growth of the macroeconomic indicator:Yr½t� 0; 95YC½t�.Here YC½t� are the computed scenario values of the sum of GVA of the branches
without parametric control; Yr½t� are the computed values of the sum of GVA of the
branches with parametric control.
The value of the criterion KP without parametric control is KP ¼ 0; 424. Theoptimal values of the criterion KP with application of the parametric control law
appears to be equal to KP ¼ 0; 000844.The base values of macroeconomic indicators are YðtÞ and PðtÞ; values obtained
with application of scenario (c) and values obtained with application of the optimal
law of parametric control are presented in Figs. 4.19 and 4.20.
Analysis of the results of the computational experiments presented in Figs. 4.19
and 4.20 shows that with application of the derived optimal parametric control law,
the consumer price index within the controlled period practically coincides with the
desired values, while the values of the sum of economic branches’ GVAs appear to
be less than the respective base values except for 2015.
4.1.3.2 Finding Optimal Parametric Control Laws on the Basisof the Stochastic CGE Model of Economic Branches
The stochastic computable model of economic branches was derived from the
respective deterministic model (with estimates of values of the exogenous
Base Scenario Attenuation
Fig. 4.19 Computed values of sums of GVA of economic branches
208 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
parameters found from solving the parametric identification problem) via adding
the discrete Gaussian noise with independent constituents to the right-hand sides of
all of model dynamic equations (4.1). These equations include the following
equations for computing the following endogenous variables:
– Gross output of products and services in prices of the base period (VY i, i ¼ 1;. . . ; 16) of 16 producing sectors by means of the respective production functions;
– Fixed assets (VK i,i ¼ 1; . . . ; 15) of 15 producing sectors;
– Annual budgets (VB i,i ¼ 1; . . . ; 18) of 16 economic branches, households, and
consolidated budget.
The additive noises inserted into the expressions with respect toVB i can initiatethe respective cyclic oscillations caused by quick shifts (shocks) in development of
the technological progress and random changes of the population growth rate. The
additive noises inserted or added to the expressions with respect toVK icharacterizethe random changes of the shares of budgets of the producing agents spent for
purchasing investment goods and random character of the coefficients of the retire-
ment funds. The additive noises added to the expressions with respect to VB idescribe the random character of the income obtained by the sector within the
current period.
In this work, the estimates of mean square deviations of the generated Gaussian
random values defining the given noise were obtained on the basis of analysis of the
respective statistical data for economic development of the Kazakhstan Republic
from 2000 to 2008 as follows. For each time series of the measured values of the
aforementioned variables, we computed some selected mean square deviations of
differences between the measured values and trends of those values. The values
obtained in such a way were accepted to be the estimated mean square deviations of
the components of the generated discrete Gaussian noise xðtÞ added to the right-
hand sides of 50 dynamic equations indicated above.
Base Scenario Attenuation
Fig. 4.20 Computed values of the consumer price index
4.1 National Economic Evolution Control Based on a Computable Model of General. . . 209
In computational experiments with the stochastic computable model of the
economic branch, we use the following optimization criterion:
Ks ¼ E1
6
X2015
t¼2010VY g½t�
� �! max : (4.76)
Here and below “E” denotes expectation. Here Ks is the expectation value of the
gross output of the country in the prices of year 2000 from years 2010–2015.
In computational experiments, we compute the criterion Ks as follows. We simulate
N realizations of random process xðtÞ by the Monte Carlo method and, after Ncomputations of the model for all these realizations used consecutively in equations
of type (1.33), we compute the criterion Ks as the arithmetic mean of the values of
the expressions 16
P2015t¼2010 YðtÞ over these N realizations. Similarly, we check the
condition of type (1.36) of the expectation values of the endogenous variables
belonging to given regions of the model state space.
The value of the criterion Ks for the base computation variant (with use of the
values of the exogenous parameters obtained as a result of the model parametric
identification) equals Ks ¼ 0:9891:1013.In experiments with the optimization criterion (4.76), we applied the constraints
on the growth of the consumer prices in the following form:
EðVPr½t�ÞÞ � 1:09E(VP[t]),.t = 2010, . . . ; 2015:
Here VP½t� is the computed level of the consumer prices in the model without
parametric control, and VPr½t� is the consumer price level with parametric control.
In the computational experiments, we realize the regulation of 1536 exogenous
parameters, shares of budgets of the jth producing agents spent for purchasing
goods and services produced by the ith producing agent from 2010 to 2015: OjiðtÞ;
t ¼ 2010; . . . ; 2015; i; j ¼ 1; . . . ; 16. HereP16
i¼1 OjiðtÞ � 1 for the mentioned values
of t. The base values of these shares obtained as a result of solving the model
parametric identification problem from data of 2000–2008 will be denoted by Oj
i;
i; j ¼ 1; . . . ; 16.We considered the problem of finding optimal values of the adjusted parameters’
vectors. On the basis of the stochastic computable model of economic branches,
find the mentioned values of the shares of the producing agents’ budgets OjiðtÞ ,
which provide the upper bound of the criterion Ks with additional constraints on
these shares of the following form:
0:5 � OjiðtÞ O
j
i
.� 2; i; j ¼ 1; . . . ; 16; t ¼ 2010; . . . ; 2015:
These optimization problemswere solved by applying theNelder–Mead algorithm.
After application of parametric control of the shares of the stochastic model budgets,
210 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
the value of the criterion appears to be equal to Ks ¼ 1:2453 � 1013 ; its value
increases by 25.89% in comparison with the base variant.
A similar parametric control problem with the respective constraints was also
solved on the basis of the deterministic CGE model of economic branches with the
use of the deterministic analog of criterion (4.76):
Kd ¼ 1
6
X2015t¼2010
Yt ! max:
After application of parametric control of the shares of budgets of producing
agents, the value of the criterion for the deterministic model appears to be equal to
Kd ¼ 1:6283 � 1013 . The criterion value increases by 33.14% in comparison with
the base variant.
The comparison of the results of solution of variational calculus problem on the
basis of stochastic and deterministic computable models of general equilibrium
shows that the computed value of the functional of the variational calculus problem
decreases while taking into account the disturbing violations in the deterministic
computable model of general equilibrium in the form of additive noise.
4.2 National Economic Evolution Control Basedon the Computable Model of General Equilibriumwith the Knowledge Sector
4.2.1 Model Description, Parametric Identification,and Retrospective Prediction
4.2.1.1 Model Agents
The considered model [27, 10] describes the behavior and interaction in nine
product markets and two labor markets for the following seven economic agents:
Economic agent no. 1 is the science and education (knowledge) sector rendering theservices on education of students and the production of knowledge. These include
educational institutions (public and private) rendering the services of higher educa-
tion, and as well as scientific (research) organizations.
This sector renders the services distributed among the following three areas:
1. The services for the innovation sector (mainly carrying out research and devel-
opment) and other sectors of the economy (mainly carrying out research and
development too), as well as services for economic agent no. 5, including, in
accordance with the methodology of National Economic Accounting (NEA),
4.2 National Economic Evolution Control Based on the Computable Model. . . 211
the services of nonmarket science. Additionally, a part of the services on
providing the knowledge is consumed by the sector itself.
2. The services for economic agent no. 5 (including, in accordance with methodol-
ogy of NEA, services of free education), services of paid education for the
innovation sector and other branches of the economy and households. Moreover,
a part of the educational services is consumed by the sector itself.
3. The services for the outer world, carrying out the works by scientific grants.
Economic agent no. 2 is the innovation sector, which is an aggregate of innovation-active enterprises and organizations. The sector produces the product distributed
between the following two areas:
1. The innovative products for the domestic market. The innovative products are
understood to be final products manufactured on the basis of various technologi-
cal and other innovations. This index corresponds to the volume of shipped
innovative products. The products manufactured by the sector are consumed by
all producing sectors (including this sector itself) as the costs of research and
development, as well as costs of the technological innovations, and by economic
agent no. 5 (this means government financing of the innovation activity).
2. The innovative products of the outer world.
Economic agent no. 3 is other branches of the economy.
The other branches of the economy produce the products distributed among the
following four areas:
1. Final products for households, including consumer goods of current consump-
tion (foodstuffs, etc.), durable products (home technical equipment, motor
vehicles, etc.), as well as services;
2. Final products for economic agent no. 5, including the following:
(a) Final products for public institutions (according to the NEA’s methodology,
expenditures of the public institutions on acquiring final products), including
free services for the inhabitants rendered by the enterprises and
organizations in the field of public health, culture (this does not include
the educational services, because they are rendered by economic agent no.
1); services satisfying the needs of the entire society, i.e., the general public
administration, protection of law and order, national defense, housing,
economy, etc.;
(b) Final products for nonprofit organizations servicing households, including
the free services of a social character;
3. Investment products, i.e., expenditures on improvement of produced and
nonproduced tangible assets (in other words, the expenditures on the creation
of the capital assets). In accordance with the NEA’s methodology, this type of
product is determined as the sum of gross saving in capital assets and change of
reserves of material circulating assets minus the cost of acquired new and
existing capital assets (with the deduction of withdrawal).
212 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4. Export products. Since imported products are one of the constituent parts of the
products considered above, then, to avoid double counting, the exported
products include only the net export (i.e., export minus import).
To produce products and services, producing agents nos. 1–3 purchase the
following production factors:
1. The labor force (by governmental and market prices);
2. Investment products;
3. Innovative products;
4. Services for providing knowledge (e.g., R&D sector);
5. Educational services (paid education).
Economic agent no. 4 is the aggregate consumer joining households. The agent
purchases final products produced by other branches of the economy. Furthermore,
the households use the paid educational services as well. Also, this sector forms the
labor force.
Economic agent no. 5 is the government, establishing taxation rates, determining
the shares of budget for financing the producers and social transfers, and spending
its budget for purchasing final products produced by other branches of the economy.
Economic agent no. 6 is the banking sector determining the interest rate for the debt
deposits.
Economic agent no. 7 is the outer world.
4.2.1.2 Exogenous Parameters of Model
This model includes 86 exogenous parameters and 110 endogenous variables. The
exogenous parameters include the following:
– The coefficients of the production functions of the sectors;
– The various shares of the budgets of the sectors;
– The shares of the products for selling in the various markets;
– The depreciation rates of capital assets and shares of the retired capital assets;
– The deposit interest rates;
– The various taxation rates;
– The export prices and governmental prices of goods, services, and labor force,
etc.
The list of the exogenous model parameters is given in Table 4.16.
4.2.1.3 Endogenous Variables of the Model
The endogenous variables include the following:
– The budgets of the sectors and their various shares;
– The produced values-added;
4.2 National Economic Evolution Control Based on the Computable Model. . . 213
Table 4.16 Exogenous variables of the computable model with sector of knowledge
Economic agent 1: Knowledge and education sector
CO_p1_1l The share of the budget for purchasing labor force at the price of P__1l
CO_p1_1z The share of the budget for purchasing knowledge provisional services at the
price of P__1z
CO_p1_1r The share of the budget for purchasing educational services at the price of P__1r
CO_p1_1n The share of the budget for purchasing innovative products at the price of P__1n
CO_p1_1i The share of the budget for purchasing investment products at the price of P__1i
CE_p1_1z The share of the produced product for selling in the markets of knowledge-
provisional services at the price of P__1z
CE_p2_1z The share of the produced product for selling in the markets of knowledge-
provisional services at the price of P__2z
CE_p1_1r The share of the produced product for selling in the markets of educational
services at the price of P__1r
CA_r_1 The dimension coefficient of the production function
CA_k_1 The coefficient of capital assets of the production function
CA_l_1 The coefficient of the labor of the production function
Calpha__1 The coefficient of the costs of the knowledge-provisional services of the
production function
Cbeta__1 The coefficient of the costs of the educational services of the production function
Cgamma__1 The coefficient of the costs of the innovative products of the production function
CA_0_1 The rate of depreciation for the capital assets
CR__1 The share of the retired capital assets
Economic agent 2: The innovation sector
CO_p1_2l The share of the budget for purchasing the labor force at the price of P__1l
CO_p1_2z The share of the budget for purchasing knowledge-provisional services at the
price of P__1z
CO_p1_2r The share of the budget for purchasing educational services at the price of P__1r
CO_p1_2n The share of the budget for purchasing innovative products at the price of P__1n
CO_p1_2i The share of the budget for purchasing investment products at the price of P__1i
CE_p1_2n The share of the produced product for selling in the market of innovative
products at the price of P__1n
CE_p2_2n The share of the produced product for selling in the market of innovative
products at the price of P__2n
CA_r_2 The dimension coefficient of the production function
CA_k_2 The coefficient of the capital assets of the production function
CA_l_2 The coefficient of the labor of the production function
Calpha__2 The coefficient of the costs of knowledge-provisional services of the production
function
Cbeta__2 The coefficient of the costs of educational services of the production function
Cgamma__2 The coefficient of the costs of innovative products of the production function
CA_0_2 The rate of depreciation for capital assets
CR__2 The share of retired capital assets
Economic agent 3: Other branches of the economy
CO_p1_3l The share of the budget for purchasing the labor force at the price of P__1l
CO_p1_3z The share of the budget for purchasing knowledge-provisional services at the
price of P__1z
CO_p1_3r The share of the budget for purchasing educational services at the price of P__1r
(continued)
214 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table 4.16 (continued)
CO_p1_3n The share of the budget for purchasing innovative products at the price of P__1n
CO_p1_3i The share of the budget for purchasing investment products at the price of P__1i.
CE_p1_3c The share of the produced product for selling in the markets of final products at
the price of P__1сCE_p1_3g The share of the produced product for selling in the markets of final products for
the economic agent no. 5 at the price of P__1g
CE_p1_3i The share of the produced product for selling in the markets of investment
products at the price of P__1i
CE_p2_3c The share of the produced product for selling in the markets of exported products
at the price of P__2сCA_r_3 The dimension coefficient of the production function
CA_k_3 The coefficient of capital assets of the production function
CA_l_3 The coefficient of labor of the production function
Calpha__3 The coefficient of the costs of knowledge-provisional services of the production
function
Cbeta__3 The coefficient of the costs of educational services of the production function
Cgamma__3 The coefficient of the costs of innovative products of the production function
CA_0_3 The rate of depreciation for capital assets
CR__3 The share of the retired capital assets
Economic agent 4: The aggregate consumer
CO_p1_4c The share of the budget for purchasing final products at the price of P__1c
CO_p1_4r The share of the budget for purchasing educational services at the price of P__1r
CO_b_4 The share of the budget for saving in bank deposits
CS_p3_4l The supply of the labor force at the price of P__3l
CS_p1_4l The supply of the labor force at the price of P__1l
Economic agent 5: Government
CT_vad The VAT rate
CT_pr The organization profit tax rate
CT_pod The rate of physical body income tax
CT_esn The rate of single social tax
CO_p1_5g The share of the consolidated budget for purchasing the final goods at the price of
P__1g
CO_p1_5z The share of the consolidated budget for purchasing knowledge-provisional
services at the price of P__1z
CO_p1_5r The share of the consolidated budget for purchasing educational services at the
price of P__1r
CO_p1_5n The share of the consolidated budget for purchasing innovative products at the
price of P__1n
CO_s1_5 The share of the consolidated budget for backing Sector 1
CO_s2_5 The share of the consolidated budget for backing Sector 2
CO_s3_5 The share of the consolidated budget for backing Sector 3
CO_tr_5 The share of the consolidated budget for payment of social transfers to the
inhabitants
CO_f4_5 The share of off-budget funds for payment of pensions, welfare payments, etc.
CO_s_5b The share of the retained consolidated budget
CO_s_5f The share of the retained off-budget funds
(continued)
4.2 National Economic Evolution Control Based on the Computable Model. . . 215
– The demand for and supply of various products and services;
– The gains of the sectors;
– The capital assets of the sectors;
– The wages of employees;
– The various types of expenditures of the consolidated budget;
– The various types of prices of the products, services, and the labor force.
The list of the endogenous model variables is given in Table 4.17.
4.2.1.4 Model Markets
As a result of leveling the demand for and supply of various types of products,
services, and labor force, the equilibrium prices are formed in the following
markets:
– The market for final products for households;
– The market for exported final products;
– The market for final products for economic agent no. 5;
– The market for investment products;
– The market for the labor force paid by privately owned enterprises;
Table 4.16 (continued)
CB_other_5 The sum of the tax proceeds (not included into the considered ones), nontax
income, and other incomes of the consolidated budget
Economic agent 6: The banking sector
CP__bpercent The deposit interest rate for enterprises
CP_h_bpercent The deposit interest rate for physical bodies
General parts of the model
CP__3l The governmental price of labor force
CP__2z The export price of knowledge-provisional services
CP__2n The export price of innovative products
CP__2c The export price of final products
CD_p2_sz The total demand for knowledge-provisional services at the export prices
CD_p2_sn The total demand for innovative products at the export prices
CD_p2_sc The total demand for final products at the export prices
Technical parameters
CC__1l The iteration constant applied in the case of the equilibrium price
CC__1c The iteration constant applied in the case of the equilibrium price
CC__1g The iteration constant applied in the case of the equilibrium price
CC__1n The iteration constant applied in the case of the equilibrium price
CC__1i The iteration constant applied in the case of the equilibrium price
CC__1r The iteration constant applied in the case of the equilibrium price
CC__1z The iteration constant applied in the case of the equilibrium price
Ceta__1 The iteration constant applied in the case of the exogenous price
Ceta__2 The iteration constant applied in the case of the exogenous price
Ceta__3 The iteration constant applied in the case of the exogenous price
216 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table 4.17 Endogenous variables of the computable model with sector of knowledge
Economic agent 1: Knowledge and education sector
VO_p3_1l The share of the budget for purchasing labor force at the price of P__3l
VO_t_1 The share of the budget for paying the taxes to the consolidated budget
VO_f_1 The share of the budget for paying the taxes to the off-budget funds
VO_s_1 The share of the retained budget
VY__1 The value-added produced by the sector
VS_p1_1z The supply of knowledge-provisional services at the price of P__1z
VS_p2_1z The supply of knowledge-provisional services at the price of P__2z
VS_p1_1r The supply of educational services at the price of P__1r
VD_p3_1l The demand for the labor force at the price of P__3l
VD_p1_1l The demand for the labor force at the price of P__1l
VD_p1_1z The demand for knowledge-provisional services at the price of P__1z
VD_p1_1r The demand for educational services at the price of P__1r
VD_p1_1n The demand for innovative products at the price of P__1n
VD_p1_1i The demand for investment products at the price of P__1i
VY_p_1 The gain in current prices
VB__1 The budget of the sector
VB_b_1 The balance of banking accounts
VK__1 The capital assets of the sector
Economic agent 2: Innovation sector
VO_p3_2l The share of the budget for purchasing the labor force at the price of P__3l
VO_t_2 The share of the budget for paying the taxes to the consolidated budget
VO_f_2 The share of the budget for paying the taxes to the off-budget funds
VO_s_2 The share of the retained budget
VY__2 The value-added produced by the sector
VS_p1_2n The supply of innovative products at the price of P__1n
VS_p2_2n The supply of innovative products at the price of P__2n
VD_p3_2l The demand for the labor force at the price of P__3l
VD_p1_2l The demand for the labor force at the price of P__1l
VD_p1_2z The demand for knowledge-provisional services at the price of P__1z
VD_p1_2r The demand for educational services at the price of P__1r
VD_p1_2n The demand for innovative products at the price of P__1n
VD_p1_2i The demand for investment products at the price of P__1i
VY_p_2 The gain in current prices
VB__2 The budget of the sector
VB_b_2 The balance of banking accounts
VK__2 The capital assets of the sector
Economic agent 3: Other branches of the economy
VO_p3_3l The share of the budget for purchasing the labor force at the price of P__3l
VO_t_3 The share of the budget for paying the taxes to the consolidated budget
VO_f_3 The share of the budget for paying the taxes to the off-budget funds
VO_s_3 The share of the retained budget
VY__3 The value-added produced by the sector
VS_p1_3c The supply of final products at the price of P__1c
VS_p1_3g The supply of final products at the price of P__1g
VS_p1_3i The supply of investment products at the price of P__1i
(continued)
4.2 National Economic Evolution Control Based on the Computable Model. . . 217
Table 4.17 (continued)
VS_p2_3c The supply of final products at the price of P__2c
VD_p3_3l The demand for the labor force at the price of P__3l
VD_p1_3l The demand for the labor force at the price of P__1l
VD_p1_3z The demand for knowledge-provisional services at the price of P__1z
VD_p1_3r The demand for educational services at the price of P__1r
VD_p1_3n The demand for innovative products at the price of P__1n
VD_p1_3i The demand for investment products at the price of P__1i
VY_p_3 The gain in current prices
VB__3 The budget of the sector
VB_b_3 The balance of banking accounts
VK__3 The capital assets of the sector
Economic agent 4: The aggregate consumer
VO_tax_4 The share of the budget for discharging the income tax
VO_s_4 The share of the retained budget
VD_p1_4c The household demand for final products at the price of P__1c
VD_p1_4r The household demand for educational services at the price of P__1r
VW_3_1 The wages of the employees of Sector 1 (state-owned enterprises)
VW_1_1 The wages of the employees of Sector 1 (privately owned enterprises)
VW_3_2 The wages of the employees of Sector 2 (state-owned enterprises)
VW_1_2 The wages of the employees of Sector 2 (privately owned enterprises)
VW_3_3 The wages of the employees of Sector 3 (state-owned enterprises)
VW_1_3 The wages of the employees of Sector 3 (privately owned enterprises)
VB__4 The budget of the households
VB_b_4 The balance of banking accounts
Economic agent 5: Government
VD_p1_5g The demand for final products at the price of P__1g
VD_p1_5z The demand for knowledge-provisional services at the price of P__1z
VD_p1_5r The demand for educational services at the price of P__1r
VD_p1_5n The demand for innovative products at the price of P__1n
VG_s_1 The expenditures of the consolidated budget aimed at backing Sector 1
VG_s_2 The expenditures of the consolidated budget aimed at backing Sector 2
VG_s_3 The expenditures of the consolidated budget aimed at backing Sector 3
VG_tr_4 The social transfers to the inhabitants from the consolidated budget
VG_f_4 The off-budget funds made available for the inhabitants
VB__5 The consolidated budget
VB_b_5 The remainder of the consolidated budget
VF__5 The monetary assets of the off-budget funds
VF_b_5 The remainder of the monetary assets of the off-budget funds
General parts of model
VP__1l The price of the labor force
VP__1c The price of final products for the households
VP__1g The price of final products for the economic agent no. 5
VP__1n The price of innovative products
VP__1i The price of investment products
VP__1r The price of educational services
VP__1z The price of knowledge-provisional services
(continued)
218 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
– The market for the labor force paid from the funds of the national state budget;
– The market for innovative products;
– The market for exported innovative products;
– The market for knowledge;
– The market for exported knowledge;
– The market for educational services.
The formula used in the model and determining the deficiency indicator for the
labor force market with governmental regulation of prices is given by
VI l t½ � ¼ VS p3 sl t½ �=VD p3 sl t½ �: (4.77)
The model formulas that describe the process of changing the prices for all these
markets are as follows:
The labor force price:
VP 1l Qþ 1½ � ¼ VP 1l Q½ � þ VD p1 sl t½ � � VS p1 sl t½ �ð Þ=CC 1l: (4.78)
Table 4.17 (continued)
VD_p3_sl The total demand for the labor force at the price of P__3l
VD_p1_sl The total demand for the labor force at the price of P__1l
VD_p1_sc The total demand for final products for the households at the price of P__1c
VD_p1_sg The total demand for final products for economic agent no. 5 at the price of P__1g
VD_p1_sn The total demand for innovative products at the price of P__1n
VD_p1_si The total demand for investment products at the price of P__1i
VD_p1_sr The total demand for educational services at the price of P__1r
VD_p1_sz The total demand for knowledge-provisional services at the price of P__1z
VS_p3_sl The total supply of the labor force at the price of P__3l
VS_p1_sl The total supply of the labor force at the price of P__1l
VS_p1_sc The total supply of final products the households at the price of P__1c
VS_p2_sc The total supply of final products for at the price of P__2c
VS_p1_sg The total supply of final products economic agent no. 5 at the price of P__1g
VS_p1_sn The total supply of innovative products for at the price of P__1n
VS_p2_sn The total supply of innovative products for at the price of P__2n
VS_p1_si The total supply of investment products for at the price of P__1i
VS_p1_sr The total supply of educational services for at the price of P__1r
VS_p1_sz The total supply of knowledge-provisional services for at the price of P__1z
VS_p2_sz The total supply of knowledge-provisional services for at the price of P__2z
Integral indices
VY GDP (in base period prices)
VY_p GDP (in current prices)
VP The consumer price index
VK Capital assets
Technical variable
VI__l The deficiency indicator for the labor force market
4.2 National Economic Evolution Control Based on the Computable Model. . . 219
The price of final products for households:
VP 1c Qþ 1½ � ¼ VP 1c Q½ � þ VD p1 sc t½ � � VS p1 sc t½ �ð Þ=CC 1c: (4.79)
The price of final products for economic agent no. 5:
VP 1g Qþ 1½ � ¼ VP 1g Q½ � þ VD p1 sg t½ � � VS p1 sg t½ �ð Þ=CC 1g: (4.80)
The price of innovative products:
VP 1n Qþ 1½ � ¼ VP 1n Q½ � þ VD p1 sn t½ � � VS p1 sn t½ �ð Þ=CC 1n: (4.81)
The price of investment products:
VP 1i Qþ 1½ � ¼ VP 1i Q½ � þ VD p1 si t½ � � VS p1 si t½ �ð Þ=CC 1i: (4.82)
The price of educational services:
VP 1r Qþ 1½ � ¼ VP 1r Q½ � þ VD p1 sr t½ � � VS p1 sr t½ �ð Þ=CC 1r: (4.83)
The price of knowledge-provisional services:
VP 1z Qþ 1½ � ¼ VP 1z Q½ � þ VD p1 sz t½ � � VS p1 sz t½ �ð Þ=CC 1z: (4.84)
We’ll now present the formulas determining the total demand for and supply of
the products for each of the prices used in the model. The final formulas determin-
ing the demand for and supply of a specific economic agent are given in the
respective items.
The total supply of and demand for the labor force at the governmental and market
prices:
VD p3 sl t½ � ¼ VD p3 1l t½ � þ VD p3 2l t½ � þ VD p3 3l t½ �; (4.85)
VD p1 sl t½ � ¼ VD p1 1l t½ � þ VD p1 2l t½ � þ VD p1 3l t½ �; (4.86)
VS p3 sl t½ � ¼ CS p3 4l t½ �; (4.87)
VS p1 sl t½ � ¼ CS p1 4l t½ �: (4.88)
The total supply of and demand for final products for households at market prices:
VD p1 sc t½ � ¼ VD p1 4c t½ �; (4.89)
VS p1 sc t½ � ¼ VS p1 3c t½ �: (4.90)
220 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The total supply of and demand for final products for economic agent no. 5 at
market prices:
VD p1 sg t½ � ¼ VD p1 5g t½ �; (4.91)
VS p1 sg t½ � ¼ VS p1 3g t½ �: (4.92)
The total supply of and demand for innovative products at market prices:
VD p1 sn t½ � ¼ VD p1 1n t½ � þ VD p1 2n t½ � þ VD p1 3n t½ �þ VD p1 5n t½ �; (4.93)
VS p1 sn t½ � ¼ VS p1 2n t½ �: (4.94)
The total supply of and demand for investment products at market prices:
VD p1 si t½ � ¼ VD p1 1i t½ � þ VD p1 2i t½ � þ VD p1 3i t½ �; (4.95)
VS p1 si t½ � ¼ VS p1 3i t½ �: (4.96)
The total supply of and demand for educational services at market prices:
VD p1 sr t½ � ¼ VD p1 1r t½ � þ VD p1 2r t½ � þ VD p1 3r t½ �þ VD p1 4r t½ � þ VD p1 5r t½ �; (4.97)
VS p1 sr t½ � ¼ VS p1 1r t½ �: (4.98)
The total supply of and demand for knowledge-provisional services at market
prices:
VD p1 sz t½ � ¼ VD p1 1z t½ � þ VD p1 2z t½ � þ VD p1 3z t½ �þ VD p1 5z t½ �; (4.99)
VS p1 sz t½ � ¼ VS p1 1z t½ �: (4.100)
Thus, in total we have 16 formulas determining the total supply of and demand
for the products considered in this model.
Notations determining the total supply of and demand for exported products and
services are as follows:
The total supply of and demand for knowledge-provisional services (scientific
grants) at the export prices:
CD p2 sz t½ � is given; (4.101)
VS p2 sz t½ � ¼ VS p2 1z t½ �: (4.102)
4.2 National Economic Evolution Control Based on the Computable Model. . . 221
The total supply of and demand for innovative products at the export prices:
CD p2 sn t½ � is given; (4.103)
VS p2 sn t½ � ¼ VS p2 2n t½ �: (4.104)
The total supply of and demand for final products at the export prices:
CD p2 sc t½ � is given; (4.105)
VS p2 sc t½ � ¼ VS p2 3c t½ �: (4.106)
Finally, we have 16 + 6 ¼ 22 formulas for determining the total supply of and
demand for all products used in this model.
Let’s now describe the activity of economic agents participating in this model.
4.2.1.5 Economic Agent No. 1: Science and Education Sector
As presented above, leveling of the total supply and demand in the markets with
governmental prices is realized by correcting the share of budget VO_p3_1. Thisprocess is described by the following formula:
VO p3 1l Qþ 1½ � ¼ VO p3 1l Q½ � � Ceta 1þ VO p3 1l Q½ �� VI l t½ � � 1� Ceta 1ð Þ: (4.107)
Here Q is the iteration step and 0<Ceta 1<1 is the model constant. With its
increase, the process of attaining equilibrium is slower. Nevertheless, the equation
system becomes more stable.
Let’s proceed to the formulas that determine the behavior of the science and
education sector.
The production function equation is given by
VY 1 tþ 1½ � ¼ CA r 1� Power VK 1 t½ � þ VK 1 tþ 1½ �ð Þ=2ð Þ;CA k 1ð Þ� PowerððVD p1 1l t½ � þ VD p3 1l t½ �Þ;CA l 1Þ� ExpðCalpha 1� VD p1 1z t½ � þ Cbeta 1� VD p1 1r t½ �þ Cgamma 1� VD p1 1n t½ �Þ:
(4.108)
Here Power(X, Y) corresponds to XY ; ExpðXÞ corresponds to eX.The following formulas determine the demand for production factors in the
science and education sector.
222 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The demand for the labor force at the governmental prices:
VD p3 1l t½ � ¼ VO p3 1l t½ � � VB 1 t½ �ð Þ=CP 3l t½ �: (4.109)
The demand for the labor force at market prices:
VD p1 1l t½ � ¼ ðCO p1 1l t½ � � VB 1 t½ �Þ=VP 1l t½ �: (4.110)
The demand for knowledge-provisional services:
VD p1 1z t½ � ¼ CO p1 1z t½ � � VB 1 t½ �ð Þ=VP 1z t½ �: (4.111)
The demand for educational services:
VD p1 1r t½ � ¼ CO p1 1r t½ � � VB 1 t½ �ð Þ=VP 1r t½ �: (4.112)
The demand for innovative products:
VD p1 1n t½ � ¼ CO p1 1n t½ � � VB 1 t½ �ð Þ=VP 1n t½ �: (4.113)
The demand for investment products:
VD p1 1i t½ � ¼ CO p1 1i t½ � � VB 1 t½ �ð Þ=VP 1i t½ �: (4.114)
The following formulas determine the supply of the services rendered by the
science and education sector.
The supply of knowledge-provisional services at market prices:
VS p1 1z t½ � ¼ CE p1 1z� VY 1 t½ �: (4.115)
The supply of knowledge-provisional services at export prices:
VS p2 1z t½ � ¼ CE p2 1z� VY 1 t½ �: (4.116)
The supply of educational services:
VS p1 1r t½ � ¼ CE p1 1r � VY 1 t½ �: (4.117)
The following formula calculates the gain of the science and education sector
from the supplied services:
VY p 1 t½ � ¼ VS p1 1z t½ � � VP 1z t½ � þ VS p2 1z t½ � � CP 2z t½ �þ VS p1 1r t½ � � VP 1r t½ �: ð4:118Þ
4.2 National Economic Evolution Control Based on the Computable Model. . . 223
The budget of the science and education sector is determined as follows:
VB 1 t½ � ¼ VB b 1 t½ � � 1þ CP bpercent t� 1½ �ð Þ þ VY p 1 t½ �þ VG s 1 t� 1½ �: (4.119)
The agent’s budget is formed from the following:
1. The funds in the banking accounts (subject to the interests on deposits);
2. The gain received in the current period;
3. Bounties received from the consolidated budget VG s 1 t� 1½ �:
The dynamics of the banking account balance in the science and education sector
is as follows:
VB b 1 tþ 1½ � ¼ VO s 1 t½ � � VB 1 t½ �: (4.120)
The capital assets are determined by
VK 1 tþ 1½ � ¼ 1� CR 1 t½ �ð Þ � VK 1 t½ � þ VD p1 1i t½ �: (4.121)
This formula calculates the volume of the capital assets, taking their retirement
into account. The asset put into operation enters the formula with the plus sign.
The share of the budget of the science and education sector for discharging the
taxes to the consolidated budget is given by
VO t 1 t½ � ¼ VY p 1 t½ � � CT vad t½ �ð Þ=VB 1 t½ � þ ððVY p 1 t½ � � VW 3 1 t½ �� VW 1 1 t½ � � VK 1 t½ � � CA 0 1 t½ �Þ � CT pr t½ �Þ=VB 1 t½ �: ð4:122Þ
This formula takes into consideration the value-added tax (VAT) and profit tax.
While calculating the share of the budget for discharging the profit tax, the gain is
subtracted by the costs of the labor force of the state-owned VW 3 1 t½ �ð Þand privately owned VW 1 1 t½ �ð Þ enterprises, as well as the depreciation charges
VK 1 t½ � � CA 0 1 t½ �. The share of the budget for discharging the single social tax
to the off-budget funds is described as
VO f 1 t½ � ¼ VW 3 1 t½ � þ VW 1 1 t½ �ð Þ � CT esn t½ �ð Þ=VB 1 t½ �: (4.123)
The remainder of the budget of the science and education sector is given by
VO s 1 t½ � ¼ 1� CO p1 1l t½ � � VO p3 1l t½ � � CO p1 1z t½ � � CO p1 1r t½ �� CO p1 1n t½ � � CO p1 1i t½ � � VO t 1 t½ � � VO f 1 t½ �:
(4.124)
224 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.2.1.6 Economic Agent No. 2: Innovation Sector
As presented above, the leveling of the total supply and demand in the markets with
governmental prices is realized by correcting the share of budget VO_p3_2l. Thisprocess is described by the following formula:
VO p3 2l Qþ 1½ � ¼ VO p3 2l Q½ � � Ceta 2þ VO p3 2l Q½ �� VI l t½ � � 1� Ceta 2ð Þ: (4.125)
Here Q is the iteration step and 0 < Ceta__2 < 1 is the model constant. With its
increase, the process of attaining equilibrium is slower. Nevertheless, the equation
system becomes more stable. We’ll now examine the formulas determining the
behavior of the innovations sector.
The production function equation is given by
VY 2 tþ 1½ � ¼ CA r 2� Power VK 2 t½ � þ VK 2 tþ 1½ �ð Þ=2ð Þ;CA k 2ð Þ� PowerððVD p1 2l t½ � þ VD p3 2l t½ �Þ;CA l 2Þ� ExpðCalpha 2� VD p1 2z t½ � þ Cbeta 2� VD p1 2r t½ �þ Cgamma 2� VD p1 2n t½ �Þ:
(4.126)
The following formulas determine the demand for the production factors in the
innovations sector:
The demand for the labor force at governmental prices:
VD p3 2l t½ � ¼ VO p3 2l t½ � � VB 2 t½ �ð Þ=CP 3l t½ �: (4.127)
The demand for the labor force at market prices:
VD p1 2l t½ � ¼ CO p1 2l t½ � � VB 2 t½ �ð Þ=VP 1l t½ �: (4.128)
The demand for knowledge-provisional services:
VD p1 2z t½ � ¼ CO p1 2z t½ � � VB 2 t½ �ð Þ=VP 1z t½ �: (4.129)
The demand for educational services:
VD p1 2r t½ � ¼ CO p1 2r t½ � � VB 2 t½ �ð Þ=VP 1r t½ �: (4.130)
The demand for innovative products:
VD p1 2n t½ � ¼ CO p1 2n t½ � � VB 2 t½ �ð Þ=VP 1n t½ �: (4.131)
4.2 National Economic Evolution Control Based on the Computable Model. . . 225
The demand for investment products:
VD p1 2i t½ � ¼ CO p1 2i t½ � � VB 2 t½ �ð Þ=VP 1i t½ �: (4.132)
The following formulas determine the supply of the products produced by the
innovations sector:
The supply of innovative products at market prices:
VS p1 2n t½ � ¼ CE p1 2n� VY 2 t½ �: (4.133)
The supply of innovative products at export prices:
VS p2 2n t½ � ¼ CE p2 2n� VY 2 t½ �: (4.134)
The following formula calculates the gain in the innovations sector:
VY p 2 t½ � ¼ VS p1 2n t½ � � VP 1n t½ � þ VS p2 2n t½ � � CP 2n t½ �: (4.135)
The budget of the innovations sector is determined as follows:
VB 2 t½ � ¼ VB b 2 t½ � � 1þ CP bpercent t� 1½ �ð Þ þ VY p 2 t½ �þ VG s 2 t� 1½ �: (4.136)
The agent’s budget is formed from the following:
1. The funds on the banking accounts (subject to the interests on deposits);
2. The gain received in the current period;
3. Bounties received from the consolidated budget VG s 2:
The dynamics of the banking account balance of the innovations sector is as
follows:
VB b 2 tþ 1½ � ¼ VO s 2 t½ � � VB 2 t½ �: (4.137)
The capital assets are determined by
VK 2 tþ 1½ � ¼ 1� CR 2 t½ �ð Þ � VK 2 t½ � þ VD p1 2i t½ �: (4.138)
This formula calculates the volume of the capital assets, taking their retirement
into account. The asset put into operation enters the formula with a plus sign.
The share of the budget of the innovations sector for discharging taxes to the
consolidated budget is given by
226 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
VO t 2 t½ � ¼ VY p 2 t½ � � CT vad t½ �ð Þ=VB 2 t½ � þ ððVY p 2 t½ � � VW 3 2 t½ �� VW 1 2 t½ � � VK 2 t½ � � CA 0 2 t½ �Þ � CT pr t½ �Þ=VB 2 t½ �: ð4:139Þ
This formula takes into consideration the VAT and profit tax. While calculating
the share of the budget for discharging the profit tax, the gain is subtracted by the
costs of the labor force of the state-owned VW 3 2ð Þ and privately owned VW 1 2ð Þenterprises, as well as the depreciation charges VK 2 t½ � � CA 0 2 t½ �:
The share of the budget for discharging the single social tax to the off-budget
funds is described as
VO f 2 t½ � ¼ VW 3 2 t½ � þ VW 1 2 t½ �ð Þ � CT esn t½ �ð Þ=VB 2 t½ �: (4.140)
The remainder of the budget of the innovations sector is given by
VO s 2 t½ � ¼ 1� CO p1 2l t½ � � VO p3 2l t½ � � CO p1 2z t½ � � CO p1 2r t½ �� CO p1 2n t½ � � CO p1 2i t½ � � VO t 2 t½ � � VO f 2 t½ �: ð4:141Þ
4.2.1.7 Economic Agent No. 3: Other Branches of the Economy
As presented above, the leveling of the total supply and demand in the markets with
governmental prices is realized by correcting the share of budget VO_p3_3l. Thisprocess is described by the following formula:
VO p3 3l Qþ 1½ � ¼ VO p3 3l Q½ � � Ceta 3þ VO p3 3l Q½ � � VI l t½ �� 1� Ceta 3ð Þ: (4.142)
Here Q is the iteration step and 0<Ceta 3<1 is the model constant.
We’ll now look at the formulas determining the behavior of the other branches of
economy.
The production function equation is given by
VY 3 tþ 1½ � ¼ CA r 3� Power VK 3 t½ � þ VK 3 tþ 1½ �ð Þ=2ð Þ;CA k 3ð Þ� Power VD p1 3l t½ � þ VD p3 3l t½ �ð Þ;CA l 3ð Þ� ExpðCalpha 3� VD p1 3z t½ � þ Cbeta 3
� VD p1 3r t½ � þ Cgamma 3� VD p1 3n t½ �Þ: ð4:143Þ
The following formulas determine the demand of the production factors in other
branches of the economy:
The demand for the labor force at governmental prices:
VD p3 3l t½ � ¼ VO p3 3l t½ � � VB 3 t½ �ð Þ=CP 3l t½ �: (4.144)
4.2 National Economic Evolution Control Based on the Computable Model. . . 227
The demand for the labor force at market prices:
VD p1 3l t½ � ¼ CO p1 3l t½ � � VB 3 t½ �ð Þ=VP 1l t½ �: (4.145)
The demand for knowledge-provisional services:
VD p1 3z t½ � ¼ CO p1 3z t½ � � VB 3 t½ �ð Þ=VP 1z t½ �: (4.146)
The demand for educational services:
VD p1 3r t½ � ¼ CO p1 3r t½ � � VB 3 t½ �ð Þ=VP 1r t½ �: (4.147)
The demand for innovative products:
VD p1 3n t½ � ¼ CO p1 3n t½ � � VB 3 t½ �ð Þ=VP 1n t½ �: (4.148)
The demand for investment products:
VD p1 3i t½ � ¼ CO p1 3i t½ � � VB 3 t½ �ð Þ=VP 1i t½ �: (4.149)
The following formulas determine the supply of products produced by the other
branches of the economy:
The supply of final products for households:
VS p1 3c t½ � ¼ CE p1 3c� VY 3 t½ �: (4.150)
The supply of final products for economic agent no. 5:
VS p1 3g t½ � ¼ CE p1 3g� VY 3 t½ �: (4.151)
The supply of investment products:
VS p1 3i t½ � ¼ CE p1 3i� VY 3 t½ �: (4.152)
The supply of exported products:
VS p2 3c t½ � ¼ CE p2 3c� VY 3 t½ �: (4.153)
The following formula calculates the gain of the other branches of the economy:
VY p 3 t½ � ¼ VS p1 3c t½ � � VP 1c t½ � þ VS p1 3g t½ � � VP 1g t½ �þ VS p1 3i t½ � � VP 1i t½ � þ VS p2 3c t½ � � CP 2c t½ �: (4.154)
228 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The budget of the other branches of the economy is determined as follows:
VB 3 t½ � ¼ VB b 3 t½ � � 1þ CP bpercent t� 1½ �ð Þ þ VY p 3 t½ �þ VG s 3 t� 1½ �: (4.155)
The agent’s budget is formed from the following:
1. The funds in banking accounts (subject to the interests on deposits);
2. The gain received in the current period;
3. Bounties received from the consolidated budget VG s 3.
The dynamics of the banking account balance of the other branches of the
economy is as follows:
VB b 3 tþ 1½ � ¼ VO s 3 t½ � � VB 3 t½ �: (4.156)
The capital assets are determined by
VK 3 tþ 1½ � ¼ 1� CR 3 t½ �ð Þ � VK 3 t½ � þ VD p1 3i t½ �: (4.157)
This formula calculates the volume of the capital assets, taking their retirement
into account. The asset put into operation enters the formula with the plus sign.
The share of the budget of the other branches of the economy for discharging the
taxes to the consolidated budget is given by
VO t 3 t½ � ¼ VY p 3 t½ � � CT vad t½ �ð Þ=VB 3 t½ � þ ððVY p 3 t½ � � VW 3 3 t½ �� VW 1 3 t½ � � VK 3 t½ � � CA 0 3 t½ �Þ � CT pr t½ �Þ=VB 3 t½ �:
(4.158)
This formula takes into consideration the VAT and profit tax. While calculating
the share of the budget for discharging the profit tax, the gain is subtracted by the
costs of the labor force of state-owned VW 3 3ð Þ and privately owned VW 1 3ð Þenterprises, as well as the depreciation charges.
The share of the budget for discharging the single social tax to the off-budget
funds is described as
VO f 3 t½ � ¼ VW 3 3 t½ � þ VW 1 3 t½ �ð Þ � CT esn t½ �ð Þ=VB 3 t½ �: (4.159)
The remainder of the budget of other branches of the economy is given by
VO s 3 t½ � ¼ 1� CO p1 3l t½ � � VO p3 3l t½ � � CO p1 3z t½ � � CO p1 3r t½ �� CO p1 3n t½ � � CO p1 3i t½ � � VO t 3 t½ � � VO f 3 t½ �: ð4:160Þ
4.2 National Economic Evolution Control Based on the Computable Model. . . 229
4.2.1.8 Economic Agent No. 4: The Aggregate Consumer (Households)
Let’s now proceed to the formulas determining the behavior of the aggregate
consumer.
The household demand for final products is given by
VD p1 4c t½ � ¼ CO p1 4c t½ � � VB 4 t½ �ð Þ=VP 1c t½ �: (4.161)
The household demand for educational services:
VD p1 4r t½ � ¼ CO p1 4r t½ � � VB 4 t½ �ð Þ=VP 1r t½ �: (4.162)
The wages of the employees of state-owned enterprises in the science and education
sector:
VW 3 1 t½ � ¼ VD p3 1l t½ � � CP 3l t½ �: (4.163)
The wages of the employees of privately owned enterprises in the science and
education sector:
VW 1 1 t½ � ¼ VD p1 1l t½ � � VP 1l t½ �: (4.164)
The wages of the employees of state-owned enterprises in the innovations sector:
VW 3 2 t½ � ¼ VD p3 2l t½ � � CP 3l t½ �; (4.165)
The wages of the employees of privately owned enterprises in the innovations
sector:
VW 1 2 t½ � ¼ VD p1 2l t½ � � VP 1l t½ �: (4.166)
The wages of the employees of state-owned enterprises in the other branches of the
economy:
VW 3 3 t½ � ¼ VD p3 3l t½ � � CP 3l t½ �: (4.167)
The wages of the employees of privately owned enterprises in the other branches of
the economy:
VW 1 3 t½ � ¼ VD p1 3l t½ � � VP 1l t½ �: (4.168)
230 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The budget of households is determined as follows:
VB 4 t½ � ¼ VB b 4 t� 1½ � � 1þ CP h bpercent t� 1½ �ð Þþ VB 4 t� 1½ � � VO s 4 t� 1½ � þ VW 3 1 t½ � þ VW 1 1 t½ � þ VW 3 2 t½ �þ VW 1 2 t½ � þ VW 3 3 t½ � þ VW 1 3 t½ � þ VG f 4 t� 1½ �þ VG tr 4 t� 1½ �:
(4.169)
The agent’s budget is formed from the following:
1. The funds in the banking accounts (subject to the interests on deposits);
2. The gain received in the current period;
3. Wages received from the three producing agents;
4. Pensions, welfare payments, and subsidies received from the off-budget funds.
The dynamics of the banking account balance of the households is as follows:
VB b 4 t½ � ¼ CO b 4 t½ � � VB 4 t½ �: (4.170)
The share of the budget for discharging income tax is given by
VO tax 4 t½ � ¼ ððVW 3 1 t½ � þ VW 1 1 t½ � þ VW 3 2 t½ � þ VW 1 2 t½ � þ VW 3 3 t½ �þ VW 1 3 t½ �Þ � CT pod t½ �Þ=VB 4 t½ �: ð4:171Þ
The remainder of the money in cash is as follows:
VO s 4 t½ � ¼ 1� CO p1 4c t½ � � CO p1 4r t½ � � VO tax 4 t½ � � CO b 4 t½ �: (4.172)
4.2.1.9 Economic Agent No. 5: Government
Let’s now review the formulas determining the behavior of economic agent no. 5.
The consolidated budget is given by
VB 5 t½ � ¼ VO t 1 t½ � � VB 1 t½ � þ VO t 2 t½ � � VB 2 t½ � þ VO t 3 t½ � � VB 3 t½ �þ VO tax 4 t½ � � VB 4 t½ � þ CB other 5þ VB b 5 t½ �� 1þ CP bpercent t� 1½ �ð Þ: ð4:173Þ
This formula sums up money collected as taxes from the producing agents, as
well as from inhabitants. The valueCB other 5 entered in the model exogenously is
the sum of other taxes (not included in the list of taxes considered in this model),
4.2 National Economic Evolution Control Based on the Computable Model. . . 231
nontaxable income, and other income included in the consolidated budget. The
obtained sum is to be supplemented by the funds in banking accounts (subject to the
deposit interests).
The dynamics of the banking account balance of the consolidated budget is
determined by
VB b 5 tþ 1½ � ¼ CO s 5b t½ � � VB 5 t½ �: (4.174)
The cash assets of off-budget funds are as follows:
VF 5 t½ � ¼VO f 1 t½ � � VB 1 t½ � þ VO f 2 t½ � � VB 2 t½ � þ VO f 3 t½ � � VB 3 t½ �þ VF b 5 t½ � � 1þ CP bpercent t� 1½ �ð Þ:
(4.175)
This formula calculates the sum collected from the producing agents in the form
of the single social tax entering the accounts of the following off-budget funds:
– The pension fund;
– The social insurance fund;
– The federal and territorial funds of obligatory medical insurance.
The derived sum is supplemented by the funds on the banking accounts (subject
to the deposit interests).
The dynamics of the banking account balance of the off-budget funds is deter-
mined by
VF b 5 tþ 1½ � ¼ CO s 5f t½ � � VF 5 t½ �: (4.176)
The demand for final products:
VD p1 5g t½ � ¼ CO p1 5g t½ � � VB 5 t½ �ð Þ=VP 1g t½ �: (4.177)
The knowledge-provisional service payment:
VD p1 5z t½ � ¼ CO p1 5z t½ � � VB 5 t½ �ð Þ=VP 1z t½ �: (4.178)
The educational service payment:
VD p1 5r t½ � ¼ CO p1 5r t½ � � VB 5 t½ �ð Þ=VP 1r t½ �: (4.179)
The demand for innovative products:
VD p1 5n t½ � ¼ CO p1 5n t½ � � VB 5 t½ �ð Þ=VP 1n t½ �: (4.180)
232 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The subsidies to the producing sectors are as follows.
The science and education sector:
VG s 1 t½ � ¼ CO s1 5 t½ � � VB 5 t½ �: (4.181)
The innovations sector:
VG s 2 t½ � ¼ CO s2 5 t½ � � VB 5 t½ �: (4.182)
Other branches of the economy:
VG s 3 t½ � ¼ CO s3 5 t½ � � VB 5 t½ �: (4.183)
Social transfers to inhabitants:
VG tr 4 t½ � ¼ CO tr 5 t½ � � VB 5 t½ �: (4.184)
The assets of the off-budget funds made available for the inhabitants:
VG f 4 t½ � ¼ CO f4 5 t½ � � VF 5 t½ �: (4.185)
This includes assets of the pension and social insurance fund for paying out
pensions and welfare payments.
4.2.1.10 Integral Indices of Model
In this subsection we present the formulas for calculating some integral indices of
the economy of the Russian Federation.
The GDP (in prices of the base period):
VY t½ � ¼ VY 1 t½ � þ VY 2 t½ � þ VY 3 t½ �: (4.186)
The GDP (in current prices):
VY p t½ � ¼ VY p 1 t½ � þ VY p 2 t½ � þ VY p 3 t½ �: (4.187)
The consumer price index:
VP t½ � ¼ 100� VP 1c t½ �=VP 1c t� 1½ �ð Þ: (4.188)
Capital assets:
VK t½ � ¼ VK 1 t½ � þ VK 2 t½ � þ VK 3 t½ �: (4.189)
4.2 National Economic Evolution Control Based on the Computable Model. . . 233
In this model:
– Relations (1.12) are represented by 12 expressions for finding the gross value-
added (GVA) of sectors by means of the production functions, capital assets of
the sectors, and balance of the banking accounts;
– Relations (1.13) are represented by 88 expressions for finding the supply and
demand for various products and services of the sectors, the budgets and shares
of budgets of the sectors, subsidies to the sectors from the consolidated budget,
etc.;
– Relations (1.14) are represented by 10 expressions for finding the equilibrium
market prices and shares of budgets of the sectors in the markets with the
exogenous prices.
The values of exogenous functions of the considered model were determined by
solving the problem of parametric identification of the model with the use of the
available statistical data at the Republic of Kazakhstan covering the period of 2000
to 2008. The validity of the model and identification process is ensured by the
following facts:
1. The identification criterion includes statistical information on basic macroeco-
nomic indices (the GDP and GVA of the sectors, the capital assets of the
sectors, etc.).
2. The estimations of the exogenous parameters and initial values of the difference
equations, which have measured values, are found in the intervals with their
centers in the respective measured values or those covering several measured
values.
3. The intervals for estimating other parameters are determined by indirect factors.
The values of parameters varying with years are found under assumptions on the
insignificance of their variations.
As a result of solving the parametric identification problem, the value of the
relative mean square deviation of the calculated values of the endogenous
variables from the respective measured values (statistical information) is less
than 1%.
4. The validity test of the model for the purpose of assessing its capability to
produce the precise prediction values was carried out via the retrospective
prediction. To do this, after solving the parametric identification problem
using the statistical information of the Republic of Kazakhstan [37] from
2000 to 2007, the values of all exogenous variables of the model were extended
to 2008, and model computation were performed for the test period without
performing additional parametric identification.
As a result of the joint solution of problems A and B (see Sect. 4.1.1), with
application of the Nelder–Mead algorithm [66], we obtain value KB ¼ 0.0073. The
relative value of deviations of the computed values of the variables used in the main
criterion from the respective measured values is less than 0.73%.
The results of computation and retrospective prediction of the model for
2008 partially presented in Table 4.18 demonstrate computed VY; VY 1; VY 2;ð
234 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table
4.18
Measuredandcomputedvalues
ofthemodel
outputvariablesandrespectivedeviations
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
VY*
2.60
2.95
3.24
3.54
3.88
4.26
4.72
5.14
5.30
VY
2.60
2.95
3.24
3.54
3.88
4.26
4.72
5.15
5.26
DVY
�0.04
0.00
0.04
0.03
0.04
�0.01
0.07
0.26
�0.73
VY
1*
0.107
0.102
0.116
0.125
0.134
0.157
0.164
0.165
0.200
VY
10.109
0.102
0.116
0.126
0.134
0.158
0.164
0.164
0.199
DVY
11.80
�0.76
�0.01
1.32
0.46
0.49
�0.02
�0.59
�0.34
VY
2*
0.0330
0.0364
0.0410
0.0450
0.0500
0.0490
0.0680
0.0720
0.0610
VY
20.0410
0.0343
0.0390
0.0440
0.0500
0.0490
0.0670
0.0710
0.0610
DVY
225.00
�5.86
�4.73
�2.65
�0.57
�0.19
�0.13
�0.88
0.16
VY
3*
2.23
2.46
2.80
3.07
3.36
3.68
4.03
4.48
4.88
VY
31.900
2.463
2.800
3.070
3.360
3.680
4.030
4.480
4.890
VY
3�1
4.63
0.08
0.07
0.03
0.02
0.02
0.00
0.11
0.28
VP*
106.40
106.60
106.80
106.70
107.50
108.40
118.80
109.50
VP
106.57
106.81
106.95
106.83
107.64
108.50
118.90
109.50
DVP
0.00
0.16
0.20
0.14
0.12
0.13
0.09
0.09
0.00
4.2 National Economic Evolution Control Based on the Computable Model. . . 235
V 3;VPÞ, measured values, and deviations of computed values of the main output
variables of the model from the respective measured values. Here years 2000–2007
correspond to the period of model parametric identification; 2008 is the retrospective
prediction period; VY is GDP (� 1012 tenge in prices of year 2000);VY i is the grossvalue-added by the ith sector (� 1012 tenge in prices of year 2000);
GDP (� 1012 tenge in prices of year 2000); VP is the consumer price index in
percentage with respect to a preceding year; the symbol * corresponds to the
measured values; and the symbol D corresponds to the deviations (in percentage)
of the computed values from the respective measured values.
The mean error of the computed values of endogenous variables of the model
relative to the respective measured values in the period of retrospective prediction
is 1.04%.
The results of verification show acceptable adequacy of the CGE model of
economic branches.
4.2.2 Estimation of the Macroeconomic Theory Provisionson the Basis of the Computable Model of GeneralEquilibrium with the Knowledge Sector
On the basis of the computable model with the knowledge sector, we carried out
experimental testing of the main provisions of the macroeconomic theory aimed at
detecting possible sources of cyclic oscillations of macroeconomic indicators as a
result of shocking changes of demands in final and investment goods. The
experiments are similar to the experiments described in Sect. 4.1.2.4 for a comput-
able model of general equilibrium of economic branches; they consist of
computations of the following scenarios of changes of demands of the final and
investment products:
(a) O1 t½ � ¼ O1 t� 1½ � þ a Y t½ � � Y t� 1½ �ð Þ; (4.190)
(b) O2i t½ � ¼ O2i t� 1½ � þ b Y t� 3½ � � 2Y t� 2½ � þ Y t� 1½ �ð Þ i ¼ 1; 2; 3ð Þ; (4.191)
(c) Joint application of scenarios (a) and (b);
(d) Increase in the shares O1ðtÞ k times in comparison with the base variant;
(e) Increase in the sharesO2iðtÞ i ¼ 1; :::; 16ð Þ l times in comparison with the base
variant;
(f) Joint application of scenarios (d) and (e).
Here t ¼ 2010, . . ., 2015 is the time in years; O1 t½ � ¼ CO p 4c t½ � is the share ofthe budget of households spent for purchasing final products (exogenous function);
O2iðtÞ is the share of the budget of the ith branch spent for purchasing investment
products (i ¼ 1, 2, 3) (exogenous function); Y[t] ¼ VY[t] is GDP in prices of year
2000 (endogenous variable); a, b, l, k are some positive constants (l > 1, k > 1).
236 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
As a result of computational experiments on the basis of the model aimed at
realization of scenarios (a)–(f), one can observe cyclic oscillations of variable P(t)(consumer price index) (see Figs. 4.21, 4.22, 4.23, and 4.24). We did not reveal the
oscillatory phenomena for values of real indicators (in particular, Y[t]). While
applying all proposed scenarios, one can observe a decrease in the GDP indicator
Y[t] in comparison to the base period.
The experimental results presented in Figs. 4.21, 4.22, 4.23, and 4.24 prove
respective provisions of the macroeconomic theory [40], just as for the case of the
computable model of economic branches considered above.
Base Scenario
Fig. 4.21 Values of the consumer price index with application of scenario (a) with a ¼ 3 � 10�13
and for the base variant
Base Scenario
Fig. 4.22 Values of the consumer price index with application of scenario (b) with b ¼ 10�13 and
for the base variant
4.2 National Economic Evolution Control Based on the Computable Model. . . 237
4.2.3 Finding Optimal Parametric Control Laws Basedon the CGE Model with the Knowledge Sector
4.2.3.1 Attenuation of Cyclic Oscillations of Macroeconomic Indicatorsby Parametric Control Methods
In computational experiments, we consider the problem of attenuating cyclic
oscillations in the consumer price level arising from application of scenario (c) to
the economic system development while taking into account its linear dependence
between the volume of consumer expenses and its current income as well as the
linear dependence between investments and income growth. This problem is
Base Scenario
Fig. 4.23 Values of the consumer price index with application of scenario (c) with a ¼ 3 � 10�13;b ¼ 1013 and for the base variant
Base Scenario
Fig. 4.24 Values of the consumer price index with application of scenario (d) with and for the
base variant
238 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
similar to the problem considered in Sect. 4.1.3.1 for the case of a computable
model of economic branches.
On the basis of the computable model of economic branches and application of
scenario (c) (see Sect. 4.2.2, where a ¼ 3 � 10�13; b ¼ 10�13), find the values of
the taxation ratesCT vad t½ �; CT pr t½ �; CT pod t½ �; andCT esn t½ � from 2010 to 2015,
which provide the lower bound of the following criterion Kp; which characterizes
the deviations of computed values of the consumer price index VP½t� from their
respective desired values of P[t]:
Kp ¼X2015t¼2010
VP t½ � � P t½ �P t½ �
� �2:
Here we use the computed base values of the model consumer price index
without parametric control as the desired values of P[t].The constraints on the adjusted parameters are as follows: 0:104 � CT vad ½t�
� 0:156; 0:24 � CT pr½t� � 0: 36; 0:04 � CT pod½t� � 0:06; 0:16 � CT esn½t�� 0:24:
The constraints on the growth of the macroeconomic indicator:Yr t½ � 0:95Yc t:½ �Here Yc[t] are the computed scenario GDPs without parametric control; Yr[t] arethe computed values of the GDP with parametric control.
The value of criterion KP without parametric control is Kp ¼ 0.424. The optimal
values of the criterion KP with application of the parametric control law appears to
be Kp ¼ 0.000844.
The base values of macroeconomic indicators Y(t) and P(t), values obtained withapplication of scenario (c), and values obtained with application of the optimal law
of parametric control are presented in Figs. 4.25 and 4.26.
AttenuationBase Scenario
Fig. 4.25 Computed values of GDP
4.2 National Economic Evolution Control Based on the Computable Model. . . 239
Analysis of the results of computation experiments presented in Figs. 4.25 and
4.26 shows that application of the derived optimal parametric control law to the
consumer price index in 2011 appears to be less than the scenario index and
practically coincides with the desired values starting from year 2012. Also, the
values of the GDP practically coincide with the initial values of the scenario
variant.
4.2.3.2 Finding Optimal Parametric Control Laws on the Basisof the Stochastic CGE Model with the Knowledge Sector
The stochastic computable model of economic branches was derived from the
respective deterministic model (with the estimates of values of the exogenous
parameters found from solving the parametric identification problem) via adding
discrete Gaussian noise with independent constituents to the right-hand sides of all
the model dynamic equations. These equations include the following for computing
the endogenous variables:
– Values added VY 1; VY 2; VY 3ð Þ in three producing sectors by means of the
respective production functions;
– Values added VK 1; VK 2; VK 3ð Þ in three producing sectors;
– Annual budgets VB 1; VB 2; . . . ; VB 5ð Þ of sectors 1–5.The noise added to the expressions with respect to VY__i can initiate respective
cyclic oscillations caused by quick shifts (shocks) in the development of techno-
logical progress and random changes of the population growth rate. The additive
noise in the expressions with respect to VK i also characterizes random changes to
the shares of budgets of the producing agents spent for purchasing investment
AttenuationBase Scenario
Fig. 4.26 Computed values of the consumer price index
240 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
goods and random character of the coefficients of retirement funds. The additive
noise in the expression with respect to VB i describes the random character of the
incomes obtained by the sector within the current period.
In this work the estimations of mean square deviations of the generated Gaussian
random values defining the mentioned noise were obtained by using statistical data
on economic development of the Kazakhstan Republic from 2000 to 2008 as
follows:
For each time series of measured values of the aforementioned variables, we
computed some selected mean square deviations of differences between the
measured values and trends of those values. The values obtained were accepted
to be the estimated mean square deviations of the components of the generated
discrete Gaussian noise xðtÞ added to the right-hand sides of the 11 dynamic
equations indicated above.
In computational experiments with the stochastic computable model of the
economic branch, we use the following optimization criterion:
Ks ¼ E1
6
X2015t¼2010
VY t½ �( )
! max (4.192)
Here Ks is the expectation value of the average GDP of the country in prices of
year 2000 in the period 2010–2015.
In experiments with optimization criterion (4.192), we applied additional
constraints on the growth of the consumer prices of the following form:
E VPr t½ �ð Þ � 1; 09E VP t½ �ð Þ; t ¼ 2010; . . . ; 2015 (4.193)
Here VP t½ � is the computed level of the consumer prices in the model without
parametric control, VPr t½ � is the consumer price level with parametric control.
In the computational experiments, we realize regulation of 18 adjusted
parameters Oi t½ � ¼ CO s1 i ½t� t ¼ 2010; . . . ; 2015; i ¼ 1; 2; 3ð Þ; shares of the
consolidated budget spent for the financing of three producing agents in
2010–2015. A natural limitation must hold; namely, the sum of all eight shares of
the consolidated budget used in the model (including the shares Oi t½ �; i ¼ 1; 2; 3,mentioned above) must not exceed 1:
X8
i¼1Oi t½ � � 1; t ¼ 2010; . . . ; 2015: (4.194)
We considered the following problem of finding optimal values of the adjusted
parameters. On the basis of the stochastic computable model with the knowledge
sector, find the values of the shares Oi ½t� of the consolidated budgets spent for
financing three producing agents, which provide the upper bound of the criterion Ks
with additional constraints (4.194) on these shares.
This optimization problem was solved using the Nelder–Mead algorithm.
After application of parametric control to the shares of the stochastic model
4.2 National Economic Evolution Control Based on the Computable Model. . . 241
budgets, the value of the criterion appears to be KS ¼ 7:234 � 1012; its value
increases by 24,93% KS ¼ 5:79 � 1012: in comparison to the base variant.
A similar parametric control problem with the respective constraints was also
solved on the basis of the deterministic CGE model with the knowledge sector with
use of the deterministic analog of criterion (4.192):
Kd ¼ 1
6
X2015t¼2010
YðtÞ:
After application of parametric control to the shares of budgets of the producing
agents, the value of the criterion for the deterministic model appears to be equal to
Kd ¼ 9:23 � 1012: The criterion value increases by 33.14% in comparison with the
base variant.
4.3 National Economic Evolution Control Basedon the Computable Model of General Equilibriumwith the Shady Sector
4.3.1 Model Description, Parametric Identification,and Retrospective Prediction
As we know, the national economies of many countries [24] work with the shady
sector. Their activities negatively influence the national economy’s evolution.
Using a mathematical model of a national economy for the estimation of and search
for effective economic measures to decrease the influence of the shady sector
activity on the national economic system’s evolution seems to be proper direction.
4.3.1.1 Economic Agents of the Model
The considered model [27] describes the behavior and interaction of 10 product
markets and 3 labor markets in the following 7 economic agents: The first three are
producing agents.
Economic agent no. 1 is the state sector of the economy. This includes enterprises
with more than a 50% share of the state.
Economic agent no. 2 is the market sector consisting of the legally existing
enterprises and organizations with private and mixed ownership.
242 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The state and market sectors produce products distributed among the following
four directions:
1. The final product for households including consumer nondurable products
(foodstuffs, etc.), durable products (house equipment, motor vehicles, etc.), as
well as services;
2. Final products for economic agent no. 5, including
(a) The final product for public institutions, including free services for
inhabitants rendered by the enterprises and organizations in the field of public
health, education, and culture; services satisfying the needs of the society as a
entirety, i.e., general public administration, maintaining law and order,
national defense, nonmarket science, housing, and communal servicing, etc.;
(b) The final product for nonprofit organizations servicing the households,
including the free services of a social character;
3. Investment products, namely, the costs of creation of capital assets. This type of
product does not include the state (or governmental) investment since it is taken
into account in the preceding type of products. The capital assets are considered
a separate type of product in this model;
4. The exported products. Since the imported products form one of the components
of the products considered above, only the net export is included in exported
products.
Besides produced products, the state and market sectors trade capital assets
represented by the capital products in the model.
Economic agent no. 3 is the shady sector.
There are various types of shady economics [27], namely,
– White-collar shady economics is the unofficial economic activity of employees
in the registered economy concerned with their official professional activity.
This includes the economics of informal ties (i.e., offstage performance of the
ordinary production programs), upward distortion economics (presenting ficti-
tious results as real), and bribe economics (abuse of official status of the public
officers for achieving private goals).
– So-called gray (informal) shady economics is the lawful economic activity that
is not accounted for by the official statistics. This sector of shady economics
produces mainly the ordinary products and services (just as in legal economics),
but the producers avoid official taxes, not wishing to pay additional costs
concerned with the discharge of taxes, etc.
– So-called black (underground) shady economics is the statute-prohibited eco-
nomic activity concerned with the production and selling of prohibited products
and services (selling drugs, racketeering, etc.).
As for the shady sector of the considered model, it includes “gray” shady
economics as well as white-collar shady economics represented by the production
of the final goods for households by the market sector of the economy.
4.3 National Economic Evolution Control Based on the Computable Model. . . 243
The shady sector sells only one type of product, namely, the final product for
households. This economics agent does not pay taxes and receive subsidies. The
shady sector realizes the following actions:
– By distribution of its budget, it pays the services of the labor force and
determines the share of the retained budget;
– By distribution of produced products, it determines the share of final products for
selling in the market of final products for the households at the shady price.
Economic agent no. 4 is the aggregated consumer representing all households of the
country. Moreover, within the frameworks of this sector, the supplies of the labor
force for the state, market, and shady sectors are determined.
Economic agent no. 5 is the government represented by the aggregate of the central,
regional, and local governments as well as the off-budget funds. In addition, this
sector includes nonprofit organizations servicing households (political parties, trade
unions, public associations, etc.).
Economic agent no. 5 establishes taxation rates and defines the sum of the
subsidies to the producing agents and social transfer and spends its budget for
purchasing final products produced by the state and market sectors.
Economic agent no. 6 is the banking sector, including the central and commercial
banks.
The banking sector establishes interest rates for the attracted deposits and issues
money.
Economic agent no. 7 is the outer world.
4.3.1.2 Exogenous Parameters of the Model
The exogenous parameters of the model include the following:
– Coefficients of the production functions of the sectors;
– Various shares of the sectors’ budgets;
– Shares of production for selling in various markets;
– Depreciation rates for capital assets;
– Deposit interest rates;
– Issuance of money;
– Various taxation rates;
– Shares of the consolidated budget spent for purchasing final goods, backing the
state and market sectors, as well as for social transfers;
– Export prices of final goods for the outer world.
The list of the exogenous parameters of the model is given in Table 4.19.
244 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table 4.19 Exogenous variables of the computable model of economic branches with the shady
sector
Economic agent 1: State sector of the economy
O_1k_P2 The share of the state budget spent for purchasing the capital products at the price of
P_2k
O_1i_P2 The share of the state budget spent for purchasing investment products at the price of
P_2i
E_1c_P2 The share of the state sector products for selling in the markets of final products at the
price of P_2c
E_1g_P1 The share of the state sector products for selling in the markets of final products for
economic agent no. 5 at the price of P_1g
E_1g_P2 The share of the state sector products for selling in the markets of final products for
economic agent no. 5 at the price of P_2g
E_1i_P1 The share of the state sector products for selling in the markets of investment
products at the price of P_1i
E_1i_P2 The share of the state sector products for selling in the markets of investment
products at the price of P_2i
E_1k_P1 The share of the capital assets of the state sector for selling in the markets of capital
products at the price of P_1k
E_1k_P2 The share of the capital assets of the state sector for selling in the markets of capital
products at the price of P_2k
E_1ex_Pex The share of the capital assets of the state sector for selling in the markets of capital
products in the outer world countries at the price of P_ex
A_1_r The empirically determined coefficient of dimension of the state sector
A_1_k The coefficient of the state sector capital
A_1_l The coefficient of the state sector labor
A_1_n The depreciation rate for the capital assets of the state sector
Economic agent 2: Market sector
O_2l_P2 The share of the budget of the market sector spent for purchasing the labor force at the
price of P_2l
O_2k_P2 The share of the budget of the market sector spent for purchasing the capital products
at the price of P_2k
O_2i_P2 The share of the budget of the market sector spent for purchasing investment products
at the price of P_2i
E_2c_P2 The share of the market sector products for selling in the markets of final products at
the price of P_2c
E_2c_P3 The share of the market sector products for selling in the markets of final products at
the price of P_3c
E_2g_P2 The share of the market sector products for selling in the markets of final products for
economic agent no. 5 at the price of P_2g
E_2i_P2 The share of the market sector products for selling in the markets of investment
products at the price of P_2i
E_2k_P2 The share of the market sector products for selling in the markets of capital products
at the price of P_2k
E_2ex_Pex The share of the market sector products for selling in the markets of final products in
the outer world countries at the price of P_ex
A_2_r The empirically determined coefficient of the dimension of the market sector
A_2_k The coefficient of the market sector capital
A_2_l The coefficient of the market sector labor
A_2_n The depreciation rate for the capital assets of the market sector
(continued)
4.3 National Economic Evolution Control Based on the Computable Model. . . 245
Table 4.19 (continued)
Economic agent 3: Shady sector
O_3l_P3 The share of the budget of the shady sector spent for purchasing the labor force at the
price of P_3l
E_3c_P3 The share of the shady sector products for selling in the markets of final products at
the price of P_3c
A_3_r The empirically determined coefficient of dimension of the shady sector
A_3_k The coefficient of the shady sector capital
A_3_l The coefficient of the shady sector labor
Economic agent 4: The aggregate consumer
L_1_a The part of the employees entering the state sector (e.g., starting their working
activity in the state sector)
L_1_r The part of the employees withdrawing the state sector (for example, retired
employees)
L_2_a The part of the employees entering the market sector (e.g., starting their working
activity in the market sector)
L_2_r The part of the employees withdrawing from the market sector (for example, the
retired employees)
O_4c_P1 The share of the budget of the households for purchasing final products at the price of
P_1c
O_4c_P2 The share of the budget of the households for purchasing final products at the price of
P_2c
O_4c_P3 The share of the budget of the households for purchasing final products at the price of
P_3c
O_4_$ The share of the budget of the households for purchasing the foreign currency
O_4_b The share of the budget of the households for saving (as banking deposits)
L_1_2 The part of the state sector employees leaving for the market sector
L_2_1 The part of the market sector employees leaving for the state sector
L_12_3 The part of the state and market sector employees partially employed in the shady
market
Economic agent 5: Government
T vad The value-added tax
T pr The profit tax for organizations
T prop The property tax
T pod The income tax for physical bodies
T esn The single social tax
O 5g P2 The share of the consolidated budget for purchasing final products at the price of
P_2g
O_5_s1 The share of the consolidated budget for backing the state sector
O_5_s2 The share of the consolidated budget for backing the market sector
O_5_tr The share of the consolidated budget for payment of the social transfers
O_5_f1 The share of expenditures of the off-budget funds spent for the state sector
O_5_f2 The share of expenditures of the off-budget funds spent for the market sector
O_5_f4 The share of expenditures of the off-budget funds spent for the households
B_5_Other The sum of tax proceeds (not included into consideration), nontaxable incomes, and
other incomes of the consolidated budget
(continued)
246 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.3.1.3 Endogenous Variables of the Model
The endogenous variables include the following:
– The budgets of the sectors and their various shares;
– The shares of the produced products for selling in various markets;
– The remainders of the agents’ budgets;
– The produced values-added of the producing sectors;
– Supply of and demand for various products and services;
– The gains of the sectors;
– The capital assets of the producing sectors;
– The parts of the employees withdrawing from each of the producing sectors;
– The parts of the employees entering each of the producing sectors;
– The wages of employees;
– The various types of expenditures of the consolidated budget;
– The various types of prices of products, services, and the labor force;
– The subsidies to the producing sectors;
– The social transfers to inhabitants;
– The gross production of goods and services;
– The GDP.
The list of the endogenous variables of the model is given in Table 4.20.
Table 4.19 (continued)
Economic agent 6: Banking sector
M_1 The issuance of money of the state sector
M_2 The issuance of money of the market sector
P_b The deposit interest rate for the enterprises
P_b_h The deposit interest rate for the physical bodies
General parts of the model
P_1l The state prices of labor force
P_1C The state prices of final products for households
P_1g The state prices of final products for economic agent no. 5
P_1i The state prices of investment products
P_1k The state prices of the capital products
P_1ex The state prices of investment products for the outer world
Model constants
etta The constant used for correction of the shares of budgets of the agents while leveling
the aggregate supply and demand in the markets with the state prices
Q The iteration step
C The iteration constant used for changing the velocity of computations of the
equilibrium state of the CGE model
4.3 National Economic Evolution Control Based on the Computable Model. . . 247
Table 4.20 Exogenous variables of the computable model of the economic branches with the
shady sector
Economic agent 1: State sector of the economy
O_1l_P1 The share of the state sector budget for purchasing the labor force at the price of P_1l
O_1k_P1 The share of the state sector budget spent for purchasing capital products at the price
of P_1k
O_1i_P1 The share of the state sector budget spent for purchasing investment products at the
price of P_1i
O_1_t The share of the state sector budget for discharging taxes to the consolidated budget
O_1_f The share of the state sector budget for discharging taxes to off-budget funds
O_1_s The share of the retained budget of the state sector
E_1c_P1 The share of the state sector products for selling in the markets of final products at the
price of P_1c
Y_1 The value-added of the state sector (in the prices of the base period)
S_1c_P1 The supply of final products by the state sector at the price of P_1c
S_1c_P2 The supply of final products by the state sector at the price of P_2c
S_1g_P1 The supply of final products by the state sector for economic agent no. 5 at the price of
P_1g
S_1g_P2 The supply of final products by the state sector for economic agent no. 5 at the price of
P_2g
S_1i_P1 The supply of investment products by the state sector at the price of P_1i
S_1i_P2 The supply of investment products by the state sector at the price of P_2i
S_1k_P1 The supply of the capital products by the state sector at the price of P_1k
S_1k_P2 The supply of the capital products by the state sector at the price of P_2k
S_1ex_Pex The supply of the exported products by the state sector at the price of P_ex
D_1l_P1 The demand for the labor force in the state sector at the price of P_1l
D_1k_P1 The demand for the capital products in the state sector at the price of P_1k
D_1k_P2 The demand for the capital products in the state sector at the price of P_2k
D_1i_P1 The demand for investment products in the state sector at the price of P_1i
D_1i_P2 The demand for investment products in the state sector at the price of P_2i
Y_1_p The gain of the state sector in current prices
B_1 The state sector budget
B_1_b The balance of the banking accounts of the state sector
K_1 The capital assets of the state sector
Economic agent 2: Market sector of the economy
O_2k_P1 The share of the market sector budget spent for purchasing capital products at the
price of P_1k
O_2i_P1 The share of the market sector budget spent for purchasing capital products at the
price of P_1i
O_2_t The share of the market sector budget for discharging the taxes to the consolidated
budget
O_2_f The share of the market sector budget for discharging the taxes to the off-budget
funds
O_2_s The share of the retained budget of the market sector
Y_2 The value-added of the market sector (in the prices of the base period)
S_2c_P2 The supply of final products by the market sector at the price of P_2c
S_2c_P3 The supply of final products by the market sector at the price of P_3c
S_2g_P2 The supply of final products by the market sector for economic agent no. 5 at the price
of P_2g
(continued)
248 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table 4.20 (continued)
S_2i_P2 The supply of investment products by the market sector at the price of P_2i
S_2k_P2 The supply of the capital products by the market sector at the price of P_2k
S_2ex_Pex The supply of the exported products by the market sector at the price of P_ex
D_2l_P2 The demand for the labor force in the market sector at the price of P_2l
D_2k_P1 The demand for capital products in the market sector at the price of P_1k
D_2k_P2 The demand for capital products in the market sector at the price of P_2k
D_2i_P1 The demand for investment products in the market sector at the price of P_1i
D_2i_P2 The demand for investment products in the market sector at the price of P_2i
Y_p The gain of the market sector in the current prices
B_2 The market sector budget
B_2_b The balance of banking accounts in the market sector
K_2 The capital assets of the market sector
Economic agent 3: Shady sector
O_3_s The share of the retained budget of the shady sector
Y_3 The value-added of the shady sector (in the prices of the base period)
S_3c_P3 The supply of final products by the shady sector at the price of P_3c
D_3l_P3 The demand for the labor force in the shady sector at the price of P_3l
Y_3_p The gain of the shady sector in the current prices
B_3 The shady sector budget
B_3_b The balance of the banking accounts of the shady sector
K_3 The capital assets of the shady sector
Economic agent 4: The aggregate consumer
M_4 The issuance of money to households
O_4_tax The share of the household budget for discharging taxes to the consolidated budget
O_4_s The share of the retained budget of the households
L_1 The supply of the labor force to the state sector
L_2 The supply of the labor force to the market sector
L_3 The supply of the labor force to the shady sector
D_4c_P1 The household demand for final products at the price of P_1c
D_4c_P2 The household demand for final products at the price of P_2c
D_4c_P3 The household demand for final products at the price of P_3c
W_1 The wages of employees of the state sector
W_2 The wages of employees of the market sector
W_3 The wages of employees of the shady sector
B_4 The budget of households
B_4_b The balance of banking accounts
Economic agent 5: Government
O_5g_P1 The share of the consolidated budget spent for purchasing final products at the price
of P_1g
O_5_s The share of the retained consolidated budget
O_5f_s The share of the retained off-budget funds
D_5g_P1 The demand for final products at the price of P_1g
D_5g_P2 The demand for final products at the price of P_2g
G_1_s The expenditures of the consolidated budget aimed at backing the state sector
G_2_s The expenditures of the consolidated budget aimed at backing the market sector
G_4_tr
(continued)
4.3 National Economic Evolution Control Based on the Computable Model. . . 249
Table 4.20 (continued)
The social transfers to the inhabitants formed from the funds of the consolidated
budget
G_1_f Off-budget funds assigned to the state sector
G_2_f Off-budget funds assigned to the market sector
G_4_f Off-budget funds assigned to the inhabitants
B_5 The consolidated budget
B_5_b The balance of banking accounts of the consolidated budget
F_5 The money assets of the off-budget funds
F_5_b The balance of banking accounts of the off-budget funds
General part of the model
P_2l The market prices of the labor force
P_2c The market prices of final products for households
P_2g The market prices of final products for economic agent no. 5
P_2i The market prices of investment products
P_2k The market prices of capital products
P_3l The market prices of the labor force
P_3c The market prices of final products for households
I_l The deficiency indicator for the labor force market
I_c The deficiency indicator for the market of final products for the households
I_g The deficiency indicator for the market of final products for economic agent no. 5
I_i The deficiency indicator for the market of investment products
I_k The deficiency indicator for the market of capital products
D_sl_P1 The total demand for the labor force at the price of P_1l
D_sl_P2 The total demand for the labor force at the price of P_2l
D_sl_P3 The total demand for the labor force at the price of P_3l
S_sl_P1 The total supply of the labor force at the price of P_1l
S_sl_P2 The total supply of the labor force at the price of P_2l
S_sl_P3 The total supply of the labor force at the price of P_3l
D_sc_P1 The total demand for final products at the price of P_1c
D_sc_P2 The total demand for final products at the price of P_2c
D_sc_P3 The total demand for final products at the price of P_3c
S_sc_P1 The total supply of final products at the price of P_1c.
S_sc_P2 The total supply of final products at the price of P_2c
S_sc_P3 The total supply of final products at the price of P_3c
D_sg_P1 The total demand for final products at the price of P_1g
D_sg_P2 The total demand for final products at the price of P_2g
S_sg_P1 The total supply of final products at the price of P_1g
S_sg_P2 The total supply of final products at the price of P_2g
D_si_P1 The total demand for investment products at the price of P_1i
D_si_P2 The total demand for investment products at the price of P_2i
S_si_P1 The total supply of investment products at the price of P_1i
S_si_P2 The total supply of investment products at the price of P_2i
D_sk_P1 The total demand for capital products at the price of P_1k
D_sk_P2 The total demand for capital products at the price of P_2k
S_sk_P1 The total supply of capital products at the price of P_1k
S_sk_P2 The total supply of capital products at the price of P_2k
(continued)
250 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
4.3.1.4 Model Markets
The equilibrium prices are formed in 13 markets of the model as a result of
equalization of the supply and demand of various types of products, services, and
labor force:
– The markets of final products for households with the governmental, market, and
shady prices;
– The markets of final products for economic agent no. 5 with the governmental
and market prices;
– The markets of the capital products with the governmental and market prices;
– The markets of investment products with the governmental and market prices;
– The markets of the labor force with the governmental, market, and shady prices;
– The markets of exported products.
For each market, we determine the total supply and demand equalized during
iterative recalculation. The formulas determining the deficiency indicators for the
markets with governmental prices used in the model are presented below.
The labor force market:
I l t½ � ¼ S sl P1 t½ �=D sl P1 t½ �; (4.195)
The market for final products for households:
I c t½ � ¼ S sc P1 t½ �=D sc P1 t½ �; (4.196)
The market for final products for economic agent no. 5:
I g t½ � ¼ S sg P1 t½ �=D sg P1 t½ �; (4.197)
The market for investment products:
I i t½ � ¼ S si P1 t½ �=D si P1 t½ �; (4.198)
The market for capital products:
I k t½ � ¼ S sk P1 t½ �=D sk P1 t½ �: (4.199)
Table 4.20 (continued)
Y The GDP (in the prices of the base period)
Y_p The GDP (in current prices)
P Inflation of consumer prices
L The number of people employed in the economy
K The capital assets
4.3 National Economic Evolution Control Based on the Computable Model. . . 251
As is obvious, the deficiency indicator is the ratio of the product supply to its
demand.
Let’s now present the model formulas reflecting the market process of changing
the labor force prices:
P 2l Qþ 1½ � ¼ P 2l Q½ � � C 2l;C 2l ¼ Abs D sl P2 Q½ �=S sl P2 Q½ �ð Þ=C; (4.200)
The prices of final products for households:
P 2c Qþ 1½ � ¼ P 2c Q½ � � C 2c;C 2c¼ Abs D sc P2 Q½ �=S sc P2 Q½ �ð Þ=C; (4.201)
The prices of final products for economic agent no. 5:
P 2g Qþ 1½ � ¼ P 2g Q½ � � C 2g;C 2g¼ Abs D sg P2 Q½ �=S sg P2 Q½ �ð Þ=C; (4.202)
The prices of investment products:
P 2i Qþ 1½ � ¼ P 2i Q½ � � C 2i;C 2i ¼ Abs D si P2 Q½ �=S si P1 Q½ �ð Þ=C; (4.203)
The prices of capital products:
P 2k Qþ 1½ � ¼ P 2k Q½ � � C 2k;C 2k¼ Abs D sk P2 Q½ �=S sk P2 Q½ �ð Þ=C; (4.204)
The equilibrium price in the shady markets is the same as the market price. The
respective formulas are presented below:
The labor market price:
P 3l Qþ 1½ � ¼ P 3l t½ � � C 3l;C 3l ¼ Abs D sl P3 Q½ �=S sl P3 Q½ �ð Þ=C; (4.205)
The price of final products for households:
P 3c Qþ 1½ � ¼ P 3c t½ � � C 3c;C 3c ¼ Abs D sc P3 Q½ � S sc P3 Q½ �=ð Þ C= : (4.206)
The formulas describing the total supply of and demand for products for each
price used in this model are discussed below. The final formulas determining the
supply of and demand for each economic agent are also given ahead.
The total demand for the labor force at the governmental, market, and shady prices:
D sl P1 t½ � ¼ D 1l P1 t½ �; (4.207)
252 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
D sl P2 t½ � ¼ D 2l P2 t½ �; (4.208)
D sl P3 t½ � ¼ D 3l P3 t½ �: (4.209)
The total supply of the labor force at governmental, market, and shady prices:
S sl P1 t½ � ¼ L 1 t½ �; (4.210)
S sl P2 t½ � ¼ L 2 t½ �; (4.211)
S sl P3 t½ � ¼ L 3 t½ �: (4.212)
The total demand for final products at governmental, market, and shady prices:
D sc P1 t½ � ¼ D 4c P1 t½ �; (4.213)
D sc P2 t½ � ¼ D 4c P2 t½ �; (4.214)
D sc P3 t½ � ¼ D 4c P3 t½ �: (4.215)
The total supply of final products at governmental, market, and shady prices:
S sc P1 t½ � ¼ S 1c P1 t½ �; (4.216)
S sc P2 t½ � ¼ S 1c P2 t½ � þ S 2c P2 t½ �; (4.217)
S sc P3 t½ � ¼ S 2c P3 t½ � þ S 3c P3 t½ �: (4.218)
The total demand for final products for economic agent no. 5 at governmental and
market prices:
D sg P1 t½ � ¼ D 5g P1 t½ �; (4.219)
D sg P2 t½ � ¼ D 5g P2 t½ �: (4.220)
The total supply of final products for economic agent no. 5 at governmental and
market prices:
S sg P1 t½ � ¼ S 1g P1 t½ �; (4.221)
S sg P2 t½ � ¼ S 1g P2 t½ � þ S 2g P2: (4.222)
The total demand for investment products at governmental and market prices:
D si P1 t½ � ¼ D 1i P1 t½ � þ D 2i P1 t½ �; (4.223)
4.3 National Economic Evolution Control Based on the Computable Model. . . 253
D si P2 t½ � ¼ D 1i P2 t½ � þ D 2i P2 t½ �: (4.224)
The total supply of investment products at governmental and market prices:
S si P1 t½ � ¼ S 1i P1 t½ �; (4.225)
S si P1 t½ � ¼ S 1i P2 t½ � þ S 2i P2 t½ �: (4.226)
The total demand for capital products at governmental and market prices:
D sk P1 t½ � ¼ D 1k P1 t½ � þ D 2k P1 t½ �; (4.227)
D sk P2 t½ � ¼ D 1k P2 t½ � þ D 2k P2 t½ �: (4.228)
The total supply of the capital products at governmental and market prices:
S sk P1 t½ � ¼ S 1k P1 t½ �; (4.229)
S sk P2 t½ � ¼ S 1k P2 t½ � þ S 2k P2 t½ �: (4.230)
So, we have 24 formulas for determining the total supply of and demand for
products considered in this model.
4.3.1.5 Economic Agent No. 1: The State Sector
In the markets with governmental pricing, equalization of the total supply and
demand occurs via the correction of budget shares and the share of the finished
product. This process is described by the formulas
O 1l P1 Qþ 1½ � ¼ O 1l P1 Q½ � � etta 1lþ O 1l P1 Q½ � � I l t½ �� 1� etta 1lð Þ; (4.231)
O 1i P1 Qþ 1½ � ¼ O 1i P1 Q½ � � etta 1iþ O 1i P1 Q½ � � I i t½ �� 1� etta 1ið Þ; (4.232)
O 1k P1 Qþ 1½ � ¼ O 1k P1 Q½ � � etta 1k þ O 1k P1 Q½ � � I k t½ �� 1� etta 1kð Þ; (4.233)
E 1c P1 Qþ 1½ � ¼ E 1c P1 Q½ � � etta 1cþ E 1c P1 Q½ � � I c t½ �� 1� etta 1cð Þ: (4.234)
254 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
HereQ is the iteration step and 0 < etta_1l, etta_1i, etta_1k, etta_1c < 1 are the
model constants. With its increase, the process of attaining equilibrium is slower.
Nevertheless, the equation system becomes more stable.
Let’s s now proceed to the formulas determining the behavior of the state sector.
The production function equation is given by
Y 1 tþ 1½ � ¼A 1 r � power K 1 t½ � þ K 1 tþ 1½ �ð Þ=2ð Þ;A 1 kð Þ� powerðD 1l P1 t½ �;A 1 lÞ: ð4:235Þ
Here power(X, Y) corresponds to XY; A_1_r, A_1_k, and A_1_l are the
parameters of the production function.
The following formulas determine the demand for the production factors within
the state sector.
The demand for the labor force at governmental prices:
D 1l P1 t½ � ¼ O 1l P1 t½ � � B 1 t½ �ð Þ=P 1l: (4.236)
The demand for capital products:
At governmental prices:
D 1k P1 t½ � ¼ O 1k P1 t½ ��B 1 t½ �ð Þ=P 1; (4.237)
At market prices:
D 1k P2 t½ � ¼ O 1k P2� B 1 t½ �ð Þ=P 2k t½ �: (4.238)
The demand for investment products:
At governmental prices:
D 1i P1 t½ � ¼ O 1i P1 t½ � � B 1 t½ �ð Þ=P 1i; (4.239)
At market prices:
D 1i P2 t½ � ¼ O 1i P2� B 1 t½ �ð Þ=P 2i t½ �: (4.240)
The following formulas determine the supply of products of the state sector.
The supply of final products for households:
At governmental prices:
S 1c P1 t½ � ¼ E 1c P1 t½ � � Y 1 t½ �; (4.241)
4.3 National Economic Evolution Control Based on the Computable Model. . . 255
S 1c P2 t½ � ¼ E 1c P2� Y 1 t½ �: (4.242)
The supply of final products for economic agent no. 5:
At governmental prices:
S 1g P1 t½ � ¼ E 1g P1� Y 1 t½ �; (4.243)
At market prices:
S 1g P2 t½ � ¼ E 1g P2� Y 1 t½ �: (4.244)
The supply of investment products:
At governmental prices:
S 1i P1 t½ � ¼ E 1i P1� Y 1 t½ �; (4.245)
At market prices:
S 1i P2 t½ � ¼ E 1i P2� Y 1 t½ �: (4.246)
The supply of capital products:
At governmental prices:
S 1k P1 t½ � ¼ E 1k P1� K 1 t½ �; (4.247)
At market prices:
S 1k P2 t½ � ¼ E 1k P2� K 1 t½ �: (4.248)
The supply of exported products:
S 1ex Pex t½ � ¼ E 1ex Pex� Y 1 t½ �: (4.249)
The following formula calculates the gain of the state sector:
Y 1 p t½ � ¼ S 1c P1 t½ � � P 1cþ S 1c P2 t½ � � P 2c t½ � þ S 1g P1 t½ � � P 1g
þ S 1g P2 t½ � � P 2g t½ � þ S 1i P1 t½ � � P 1iþ S 1i P2 t½ � � P 2i t½ �þ S 1k P1 t½ � � P 1k þ S 1k P2 t½ � � P 2k t½ � þ S 1ex Pex t½ �� P ex: ð4:250Þ
256 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
As is obvious, the gain consists of the gain from selling final products and
rendering services for households and economic agent no. 5, investment, capital,
as well as exported products.
The budget of the state sector is determined as follows:
B 1 t½ � ¼ B 1 b t½ � � 1þ CP b t� 1½ �ð Þ þ Y 1 p t½ � þ G 1 s t� 1½ � þ G 1 f� t� 1½ � þM 1: (4.251)
The agent’s budget is formed from the following:
1. Funds in banking accounts (subject to the interests on deposits);
2. Gain received within the current period;
3. Bounties received from the consolidated budget G 1 s;4. Part of the off-budget funds G 1 f ;5. Emission of money M_1;
The dynamics of the banking account balance of the state sector is as follows:
B 1 b tþ 1½ � ¼ O 1 s t½ � � B 1 t½ �: (4.252)
The capital assets’ dynamics is determined by
K 1 tþ 1½ � ¼ K 1 t½ � � 1� E 1k P1� E 1k P2ð Þ � 1� A 1 nð Þ þ D 1k P1 t½ �þ D 1k P2 t½ � þ D 1i P1 t½ � þ D 1i P2 t½ �:
(4.253)
This formula calculates the volume of capital assets, taking into account their
selling and wear and tear. The purchased assets and investments to the capital assets
enter the formula with a plus sign.
The share of the budget of the state sector for discharging taxes to the
consolidated budget is given by
O 1 t t½ � ¼ Y 1 p t½ � � T vadð Þ=B 1 t½ � þ Y 1 p t½ � �W 1 t½ � � K 1 t½ � � P 1k � A 1 nð Þð� T prÞ=B 1 t½ � þ K 1 t½ � � P 1kð Þ � T propð Þ=B 1 t½ �:
(4.254)
This formula takes into consideration the value-added tax (VAT), profit tax, and
property tax. While calculating the share of the budget for discharging profit tax, the
gain is subtracted from the costs of labor forceW_1, as well as depreciation chargesK_1[t] � P_1k � A_1_n.
The share of the budget for discharging the single social tax to the off-budget
funds is described as
O 1 f t½ � ¼ W 1 t½ � � T esnð Þ B 1 t½ �= : (4.255)
4.3 National Economic Evolution Control Based on the Computable Model. . . 257
The remainder of the budget of the state sector of the economy is given by
O 1 s t½ � ¼ 1� O 1l P1 t½ � � O 1k P1 t½ � � O 1k P2� O 1i P1 t½ �� O 1i P2� O 1 t t½ � � O 1 f t½ �: ð4:256Þ
4.3.1.6 Economic Agent No. 2: The Market Sector
Since the behavior of the market sector is similar to that of the state sector, we
shorten the description composed by analogy with agent no. 1 in some places.
The market sector corrects the shares of its budget O 2k P1 and O 2k P1 for
purchasing capital and investment products with governmental prices.
This process is described by the formulas
O 2k P1 Q½ � ¼ O 2k P1 Q½ � � etta 2k þ O 2k P1 Q½ � � I k t½ �� 1� etta 2kð Þ; (4.257)
O 2i P1 Q½ � ¼ O 2i P1 Q½ � � etta 2iþ O 2i P1 Q½ � � I i t½ �� 1� etta 2ið Þ; (4.258)
where Q is iteration step and 0 < etta_2k, etta_2i < 1 are model constants.
Let’s proceed to the formulas determining the behavior of the market sector.
The production function equation is given by
Y 2 tþ 1½ � ¼ A 2 r � power K 2 t½ � þ K 2 tþ 1½ �ð Þ=2ð Þ;A 2 kð Þ� power D 2l P2 t½ �;A 2 lð Þ: ð4:259Þ
Here A_2_r, A_2_k, and A_2_l are the parameters of the production function.
The following formulas determine the demand for the production factors in the
market sector.
The demand for the labor force at market prices:
D 2l P2 t½ � ¼ O 2l P2� B 2 t½ �ð Þ P 2l t½ �= : (4.260)
The demand for capital products:
At governmental prices:
D 2k P1 t½ � ¼ O 2k P1 t½ � � B 2 t½ �ð Þ P 1k= ; (4.261)
At market prices:
D 2k P2 t½ � ¼ O 2k P2� B 2 t½ �ð Þ P 2k t½ �= : (4.262)
258 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The demand for investment products:
At governmental prices:
D 2i P1 t½ � ¼ O 2i P1 t½ � � B 2 t½ �ð Þ P 1i= ; (4.263)
At market prices:
D 2i P2 t½ � ¼ O 2i P2� B 2 t½ �ð Þ P 2i t½ �= : (4.264)
The following formulas determine the supply of products of the market sector.
The supply of final products for households:
At market prices:
S 2c P2 t½ � ¼ E 2c P2� Y 2 t½ �; (4.265)
At shady prices:
S 2c P3 t½ � ¼ E 2c P3� Y 2 t½ �: (4.266)
The supply of final products for economic agent no. 5 at market prices:
S 2g P2 t½ � ¼ E 2g P2� Y 2 t½ �: (4.267)
The supply of investment products at market prices:
S 2i P2 t½ � ¼ E 2i P2� Y 2 t½ �: (4.268)
The supply of capital products at market prices:
S 2k P2 t½ � ¼ E 2k P2� K 2 t½ �: (4.269)
The supply of exported products:
S 2ex Pex t½ � ¼ E 2ex Pex� Y 2 t½ �: (4.270)
The following formula calculates the gain of the market sector:
Y 2 p ¼ S 2c P2 t½ � � P 2c t½ � þ S 2g P2� P 2g t½ � þ S 2i P2 t½ � � P 2i t½ �þ S 2k P2 t½ � � P 2k t½ � þ S 2ex Pex t½ � � P ex: ð4:271Þ
4.3 National Economic Evolution Control Based on the Computable Model. . . 259
As is obvious, the gain consists of the gain from selling final products and
rendering services for households and economic agent no. 5, investment, capital,
as well as exported products. As presented above, the gain from selling final
products and services for households at the shady prices is not accounted for here.
The budget of the market sector is determined as follows:
B 2 t½ � ¼ B 2 b t½ � � 1þ CP b t� 1½ �ð Þ þ Y 2 pþ G 2 S t� 1½ � þ G 2 f� t� 1½ � þM 2: (4.272)
The agent’s budget is formed from the following:
1. Funds in banking accounts (subject to the interests on deposits);
2. Gain received in the current period;
3. Subsidies received from the consolidated budget G 2 S2;4. Part of the off-budget funds G 2 f ;5. Emission of money M_2.
The dynamics of the banking account balance of the market sector is as follows:
B 2 b tþ 1½ � ¼ O 2 s t½ � � B 2 t½ �: (4.273)
The capital assets’ dynamics is determined by
K 2 tþ 1½ � ¼ K 2 t½ � � 1� E 2k P2ð Þ � 1� A 2 nð Þ þ D 2k P1 t½ � þ D 2k P2 t½ �þ D 2i P1 t½ � þ D 2i P2 t½ �: ð4:274Þ
This formula calculates the volume of the capital assets, taking into account their
selling and wear and tear. The purchased assets and investments to the capital assets
enter into the formula with a plus sign.
The share of the budget of the market sector for discharging taxes to the
consolidated budget is given by
O 2 t t½ � ¼ Y 2 p� T vadð Þ=B 2 t½ � þ Y 2 p�W 2 t½ � � K 2 t½ � � P 2k t½ � � A 2 nð Þð� T prÞ=B 2 t½ � þ K 2 t½ � � P 2k t½ �ð Þ � T propð Þ=B 2 t½ �:
(4.275)
This formula takes into consideration the VAT, the profit tax, and the property
tax. While calculating the share of budget for discharging the profit tax, the gain is
subtracted by the costs of the labor force W_2 as well as the depreciation charges
K_2[t] � P_2k[t] � A_2_n.The share of the budget for discharging the single social tax to the off-budget
funds is described as
O 2 f t½ � ¼ W 2 t½ � � T esnð Þ B 2 t½ �= : (4.276)
260 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The remainder of the budget of the market sector of the economy is given by
O 2 s t½ � ¼1� O 2l P2� O 2k P1 t½ � � O 2k P2� O 2i P1 t½ � � O 2i P2
� O 2 t t½ � � O 2 f t½ �: ð4:277Þ
4.3.1.7 Economic Agent No. 3: The Shady Sector
Let’s compose the formulas determining the behavior of the shady sector.
The production function equation is given by
Y 3 tþ 1½ � ¼ A 3 r � power K 3 t½ � þ K 3 tþ 1½ �ð Þ 2=ð Þ;A 3 kð Þ� power D 3l P3 t½ �;A 3 lð Þ: ð4:278Þ
Here A_3_r, A_3_k, and A_3_l are the parameters of the production function. The
production function equation is similar to that of the state and market sectors, but
one of its arguments (the capital assets) is calculated another way.
The shady sector does not have its own capital assets. The same can be seen in
real life, where the representatives of “white-collar” and “gray” economies use
capital assets of the state and market sectors. Therefore, the capital assets of the
shady sector are formed as follows:
K 3 t½ � ¼ gamma� K 1 t½ � þ K 2 t½ �ð Þ; (4.279)
where gamma is the share of capital assets of the state and market sectors used in
shady economics.
The demand for the labor force at shady prices is calculated similarly to that of
the other sectors:
D 3l P3 t½ � ¼ O 3l P3� B 3 t½ �ð Þ P 3l t½ �= : (4.280)
Now, we’ll calculate the supply of final products for households at shady prices:
S 3c P3 t½ � ¼ E 3c P3� Y 3 t½ �: (4.281)
The following formula calculates the gain of the shady sector:
Y 3 p t½ � ¼ S 2c P3 t½ � þ S 3c P3 t½ �ð Þ � P 3c t½ �: (4.282)
This formula takes into account the final goods produced by “white-collar” and
“gray” shady economics.
The budget of the shady sector is determined as follows:
B 3 t½ � ¼ B 3 b t½ � � 1þ CP b t� 1½ �ð Þ þ Y 3 p t½ �: (4.283)
4.3 National Economic Evolution Control Based on the Computable Model. . . 261
The agent’s budget is formed from the following:
1. Funds in banking accounts (subject to the interests on deposits);
2. Gain received in the current period.
The dynamics of the banking account balance of the shady sector is as follows:
B 3 b tþ 1½ � ¼ O 3 s t½ � � B 3 t½ �: (4.284)
The remainder of the budget of the shady sector of the economy is given by
O 3 s t½ � ¼ 1� O 3l P3 t½ � (4.285)
4.3.1.8 Economic Agent No. 4: The Aggregate Consumer (Households)
Let’s proceed to the formulas determining the behavior of the aggregate consumer.
The household demand for final products:
At governmental prices:
D 4c P1 t½ � ¼ O 4c P1� B 4 t½ �ð Þ P 1c= ; (4.286)
At market prices:
D 4c P2 t½ � ¼ O 4c P2� B 4 t½ �ð Þ P 2c t½ �= ; (4.287)
At shady prices:
D 4c P3 t½ � ¼ O 4c P3� B 4 t½ �ð Þ P 3c t½ �= : (4.288)
The movement of the labor force:
In the state sector:
L 1 t½ � ¼ L 1 t� 1½ � � ð1� L 1 2 t� 1½ � þ L 1 a t� 1½ �� L 1 r t� 1½ �Þ þ L 2 t� 1½ � � L 2 1 t� 1½ �; ð4:289Þ
In the market sector:
L 2 t½ � ¼ L 2 t� 1½ � � 1� L 2 1 t� 1½ � þ L 2 a t� 1½ � � L 2 r t� 1½ �ð Þþ L 1 t� 1½ � � L 1 2 t� 1½ �; ð4:290Þ
In the shady sector:
L 3 t½ � ¼ L 1 t½ � þ L 2 t½ �ð Þ � L 12 3: (4.291)
262 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The number of employees in the shady sector is determined as the share of the
number of employees in the state and market sectors.
The wages of the employees:
In the state sector:
W 1 t½ � ¼ D 1l P1 t½ � � P 1l; (4.292)
In the market sector:
W 2 t½ � ¼ D 2l P2 t½ � � P 2l t½ �; (4.293)
In the shady sector:
W 3 t½ � ¼ D 3l P3 t½ � � P 3l t½ �: (4.294)
The budget of households is determined as follows:
B 4 t½ � ¼ B 4 b t½ � � 1þ CP b h t� 1½ �ð Þ þ VB 4 t� 1½ � � VO 4 s t� 1½ � þW 1 t½ �þW 2 t½ � þW 3 t½ � þ G 4 tr t� 1½ � þ G 4 f t� 1½ � þM 4 t½ �: ð4:295Þ
The agent’s budget is formed from the following:
1. Funds in banking accounts;
2. Retained money in cash remaining from the preceding period;
3. Wages received in the state, market, and shady sectors;
4. Pensions, welfare payments, and subsidies received from the off-budget funds;
part of the off-budget funds G_1_f;5. Emission of money M_4;6. Income from property, commercial activity, and other incomes. This constituent
part of the budget enters the model exogenously to complete the budget to the
values of official statistics.
The dynamics of the banking account balance of households is as follows:
B 4 b tþ 1½ � ¼ B 4 t½ � � O 4 b: (4.296)
The share of the budget for discharging income tax is as follows:
O 4 tax t½ � ¼ W 1 t½ � þW 2 t½ �ð Þ � T podð Þ B 4 t½ �= : (4.297)
The remainder of the money in cash is
O 4 s t½ � ¼ 1� O 4c P1� O 4c P2� O 4c P3� O 4 tax t½ �� O 4 b� O 4 buck: (4.298)
4.3 National Economic Evolution Control Based on the Computable Model. . . 263
4.3.1.9 Economic Agent No. 5: Government
Economic agent no. 5 corrects the share of the budget for purchasing final products
at governmental prices. This process is described by the following formula:
O 5g P1 Q½ � ¼ O 5g P1 Q½ � � etta 5gþ O 5g P1 Q½ � � I g t½ �� 1� etta 5gð Þ; (4.299)
where Q is the iteration step, and 0 < etta_5g < 1 is the model constant.
Let’s now proceed to the formulas determining the behavior of economic
agent no. 5.
The consolidated budget obeys the relationship
B 5 t½ � ¼ O 1 t t½ � � B 1 t½ � þ O 2 t t½ � � B 2 t½ � þ O 4 tax t½ � � B 4 t½ � þ B 5 other
þ B 5 b t½ � � 1þ CP b t� 1½ �ð Þ: ð4:300Þ
This formula sums up the money collected as the taxes from the state and market
sectors as well as from inhabitants. The value B_5_other entering the model
exogenously is the sum of other taxes (not included in the list of taxes considered
in the model), nontaxable income, and other income of the consolidated budget.
The obtained sum is supplemented by the funds in banking accounts (subject to the
deposit interests).
The dynamics of the banking account balance of the consolidated budget is
described by
B 5 b tþ 1½ � ¼ O 5b s t½ � � B 5 t½ �ð Þ; (4.301)
f 5 t½ � ¼ O 1 f t½ � � B 1 t½ � þ O 2 f t½ � � B 2 t½ � þ F 5 b t½ �� 1þ CP b t� 1½ �ð Þ: (4.302)
This formula calculates the sum collected from the state and market sectors in
the form of the single social tax entering the accounts of the following off-budget
funds:
– Pension fund;
– Social insurance fund;
– Federal and territorial funds of obligatory medical insurance.
The derived sum is supplemented by the funds in banking accounts (subject to
the interest on deposits).
The dynamics of the banking account balance of the off-budget funds is deter-
mined by
F 5 b tþ 1½ � ¼ O 5f s t½ � � F 5 t½ �: (4.303)
264 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
The demand for final products:
At governmental prices:
D 5g P1 t½ � ¼ O 5g P1 t½ � � B 5 t½ �ð Þ P 1g= ; (4.304)
At market prices:
D 5g P2 t½ � ¼ O 5g p2� B 5 t½ �ð Þ P 2g t½ �= : (4.305)
Subsidies to the producing sectors are as follows:
The state sector:
G 1 s t½ � ¼ O 5 s1� B 5 t½ �; (4.306)
The market sector:
G 2 s t½ � ¼ O 5 s2� B 5 t½ �: (4.307)
Social transfers to inhabitants:
G 4 tr t½ � ¼ O 5 tr � B 5 t½ �: (4.308)
The assets of the off-budget funds made available for
The state sector:
G 1 f t½ � ¼ O 5 f1� F 5 t½ �; (4.309)
The market sector:
G 2 f t½ � ¼ O 5 f2� F 5 t½ �: (4.310)
The assets of the off-budget funds made available to inhabitants:
G 4 f t½ � ¼ O 5 f4� F 5 t½ �: (4.311)
This includes the assets of the pension fund and social insurance fund for paying
out pensions and welfare payments.
4.3.1.10 Economic Agent No. 6: The Banking Sector
The banking sector of this model includes the central and commercial banks. This
economic agent implements the following functions:
1. It realizes the emission of money, M_1, M_2, M_4;
2. It establishes the deposit interest rate for enterprises and physical bodies.
4.3 National Economic Evolution Control Based on the Computable Model. . . 265
4.3.1.11 Economic Agent No. 7: The Outer World
In this version of the model, all the economic indices of the outer world are
specified exogenously. This means that the domestic producers cannot export
more products than the outer world requires.
4.3.1.12 Integral Indices of the Model
In this subsection we present the formulas for calculating some integral indices of
the economy.
The GDP (in prices of the base period) is
Y t½ � ¼ Y 1 t½ � þ Y 2 t½ �: (4.312)
The GDP (in current prices) is
Y p t½ � ¼ Y 1 p t½ � þ Y 2 p t½ �: (4.313)
The inflation of consumer prices is
P t½ � ¼ 100� P 2c t½ � P 2c t� 1½ �=ð Þ: (4.314)
The number of people employed within the economy is
L t½ � ¼ L 1 t½ � þ L 2 t½ � þ L 3 t½ �: (4.315)
Capital assets are
K t½ � ¼ K 1 t½ � þ K 2 t½ �: (4.316)
The considered CGE model with the shady sector is presented in the context of
relation (1.12) by 11 expressions n1 ¼ 11ð Þ; in the context of relation (1.13) by 98
expressions n2 ¼ 98ð Þ; in the context of relation (1.14) by 14 expressions
n3 ¼ 14ð Þ:The analyzed model includes 144 exogenous parameters (whose values are
required to be estimated by solving the parametric identification problem) and
123 endogenous variables.
4.3.1.13 Parametric Identification of the CGE Model with the Shady Sector
The problem of the identification (calibration) of the exogenous model parameters
was solved by the methods applied to the parametric identification of the
266 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
computable model of economic sectors and computable model with the knowledge
sector (see Sect. 4.1).
To estimate the quality of the retrospective prediction on the basis of economic
data from the Republic of Kazakhstan for the years 2000–2004, for some starting
point o1 2 O , we solve the problem (Problem A) of estimation of the model
parameters and initial conditions for the difference equations by searching for the
minimum of criterion KIA:
K2IA ¼ 1
10
X2004t¼2000
Y�½t� � Y½t�Y�½t�
� �2
þ P�½t� � P½t�P�½t�
� �2" #
: (4.317)
Here t is the number of years;
Y[t] is the calculated GDP in billions of tenge in the prices of year 2000;
P[t] is the calculated level of consumer prices.
Here and below, the * symbol corresponds to the measured values of their
respective variables. The problem of the model parametric identification is consid-
ered to be solved if such a pointo0KIA
2 O exists thatKIAðo0KIA
Þ< e for a sufficientlysmall e.
In parallel with Problem A for the point o1 , we also solve a similar problem
(problem B) with use of the extended criterion KIB instead of criterion KIA:
K2IB ¼ 1
12:15
( X2004t¼2000
Y�½t� � Y½t�Y�½t�
� �2
þ P�½t� � P½t�P�½t�
� �2" #
þ 0:1X2004
t¼2000
L� 1½t� � L 1½t�L� 1½t�
� �2
þ L� 2½t� � L 2½t�L� 2½t�
� �2" #
þ 0:1X2004
t¼2000
K� 1½t� � K 1½t�K� 1½t�
� �2
þ K� 2½t� � K 2½t�K� 2½t�
� �2" #
þ 0:01X2004
t¼2000
Y� 1½t� � Y 1½t�Y� 1½t�
� �2
þ Y� 2½t� � Y 2½t�Y� 2½t�
� �2
þ Y� 3½t� � Y 3½t�Y� 3½t�
� �2" #)
:
(4.318)
Here:
L_1[t] is the number of employees in the state sector;
L_2[t] is the number of employees in the market sector;
K_1[t] are capital assets of the state sector;K_2[t] are capital assets of the market sector;
Y_1[t] is the state sector GVA;Y_2[t] is the market sector GVA;
Y_3[t] is the shady sector GVA.
4.3 National Economic Evolution Control Based on the Computable Model. . . 267
As a result of simultaneously solving problems A and B by the described
algorithm, we obtain the values KIA ¼ 0:0025 and KIB ¼ 0:12. The relative value
of the deviation of the calculated values of variables used in criterion (4.318) from
the respective measured values is less than 0.25%.
The results of the retrospective prediction of the model for the years 2005–2008
presented in Table 4.21 demonstrate the calculated values and measured values, as
well as the deviations of the calculated values of the output variables of the model
from the respective actual values.
4.3.2 Finding the Optimal Values of the Adjusted Parameterson the Basis of the CGE Model in the Shady Sector
In the context of analysis of the connection between some processes of shady
economics and the basic macroeconomic indices of a country (the GDP and
consumer price index), a number of computational experiments described below
(the simulation of the scenarios specifying some possible negative effects within
the country’s economy) are carried out. These simulations are similar to
experiments in [27].
We consider the following six scenarios:
1. Simulation of the process of cash withdrawals (10%, 20%, 30%) from a
consolidated budget of a country and reassigning this cash to households from
year 2005 (scenarios 1, 2, and 3). We also simulate the process of direct stealing
or the quite legal process of development of budgetary funds (the kickback
process).
2. Simulation of the process of cash withdrawal (10%, 20%, 30%) from producers
and reassigning this cash to households from year 2005 (scenarios 4, 5, and 6). In
this case we simulate the process of giving (from producers) and taking (by
households) bribes.
The results of applying the enumerated six scenarios of the economic develop-
ment of the country with a negative effect of shady economics in comparison with
the base variant of evolution are presented in Tables 4.22 and 4.23.
The analysis of Tables 4.22 and 4.23 also shows that the analyzed scenarios
insignificantly affect the country’s GDP. At the same time, the consumer price
Table 4.21 Results of
the model’s retrospective
prediction
Year 2005 2006 2007 2008
Y� t½ � 4258.03 4715.65 5136.54 5303.27
Y[t] 4221.69 4586.33 5004.12 5478.31
Error (%) �0.861 �2.820 �2.646 3.195
P� t½ � 107.6 108.4 118.8 109.5
P[t] 108.4 109.5 112.6 112.0
Error (%) 0.706 1.017 �5.528 2.240
268 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
index increases significantly during the first year of applying scenarios 1–6. In the
subsequent year, their effect on the price index becomes weaker.
Note that the considered aspects of shady economics, namely, stealing from
budget and bribes, results in pronounced negative consequences for the country’s
economy. In both cases, the demand for consumer products grows, which leads to a
natural rise in consumer prices. Additionally, the producer often includes the
expenditures on bribes in the price of its production, which also leads to a rise in
prices. In any case, finally, this hurts inhabitants who are not related to partitioning
the budgetary funds and not accepting bribes and kickbacks.
The next part of computational experiments is aimed at reducing the negative
effect from each of the considered scenarios to the main macroeconomic indexes,
namely, the level of prices by the methods of parametric control.
Within the framework of applying the parametric control approach, the problem
statement is as follows: Find the optimal values of 76 parameters (uli ; i ¼ 1; :::; 19 isthe parameter number, l ¼ 2005; :::; 2008 is the year number) regulated by the
government for 2005–2008 for each of the considered scenarios. The regulated
parameters are the following:
– Various taxation rates;
– Shares of the consolidated budget for financing the state and market sectors of
the economy as well as for purchasing final products;
– Shares of the state sector budgets for purchasing various types of products;
Table 4.22 Values of the
GDP (in billions of tenge
in the prices of the year 2000)
for the base variant and
for scenarios 1–6
GDP
Year 2005 2006 2007 2008
Base variant 4,300,103 4,618,653 4,963,707 5,337,048
Scenario 1 4,301,026 4,623,221 4,975,060 5,357,813
Scenario 2 4,301,887 4,627,487 4,985,442 5,376,495
Scenario 3 4,302,752 4,631,527 4,994,972 5,393,343
Scenario 4 4,298,244 4,612,732 4,953,878 5,324,870
Scenario 5 4,296,520 4,607,483 4,945,717 5,315,665
Scenario 6 4,294,935 4,602,927 4,939,176 5,309,146
Table 4.23 Values of the
consumer price index (in %
with respect to preceding
year) for the base variant
and for scenarios 1–6
Consumer price index
Year 2005 2006 2007 2008
Base variant 107.624 108.602 109.334 108.816
Scenario 1 115.575 109.706 109.986 108.989
Scenario 2 123.530 109.761 110.470 109.044
Scenario 3 131.481 108.962 111.001 109.006
Scenario 4 138.576 118.506 113.760 111.462
Scenario 5 171.450 123.439 115.029 111.904
Scenario 6 206.522 125.441 114.879 111.508
4.3 National Economic Evolution Control Based on the Computable Model. . . 269
– Shares of various types of products produced by the state sector of the economy
for selling in the various markets.
The level of consumer prices of the country for 2008 in comparison to 2004 with
the use of the jth (j =1, . . . ,6) scenario is used as the minimized criterion K:
K ¼ P 2c½2008�=P 2c½2004�:
The following constraints on the GDP of the country are used among the
constraints of the solved variational problem:
Yj½t� �Yj½t�; j ¼ 1; :::; 6:
Here �Yj½t� is the value of the GDP with the use of the jth scenario without
parametric control; Yj½t� is the value of the GDP with the use of the jth scenario andthe values of the controlled parameters optimal in a sense of criterion K.
The constraints on the controlled parameters uli are presented in Table 4.24.
We consider a problem of finding optimal values of economic parameters uli. Onthe basis of model (4.195), (4.196), (4.197), (4.198), (4.199), (4.200), (4.201),
(4.202), (4.203), (4.204), (4.205), (4.206), (4.207), (4.208), (4.209), (4.210),
(4.211), (4.212), (4.213), (4.214), (4.215), (4.216), (4.217), (4.218), (4.219),
(4.220), (4.221), (4.222), (4.223), (4.224), (4.225), (4.226), (4.227), (4.228),
(4.229), (4.230), (4.231), (4.232), (4.233), (4.234), (4.235), (4.236), (4.237),
(4.238), (4.239), (4.240), (4.241), (4.242), (4.243), (4.244), (4.245), (4.246),
(4.247), (4.248), (4.249), (4.250), (4.251), (4.252), (4.253), (4.254), (4.255),
(4.256), (4.257), (4.258), (4.259), (4.260), (4.261), (4.262), (4.263), (4.264),
(4.265), (4.266), (4.267), (4.268), (4.269), (4.270), (4.271), (4.272), (4.273),
(4.274), (4.275), (4.276), (4.277), (4.278), (4.279), (4.280), (4.281), (4.282),
(4.283), (4.284), (4.285), (4.286), (4.287), (4.288), (4.289), (4.290), (4.291),
(4.292), (4.293), (4.294), (4.295), (4.296), (4.297), (4.298), (4.299), (4.300),
(4.301), (4.302), (4.303), (4.304), (4.305), (4.306), (4.307), (4.308), (4.309),
(4.310), (4.311), (4.312), (4.313), (4.314), (4.315), and (4.316), determine values
of the economic parameters uli ; i ¼ 1; :::; 19, l ¼ 2005; :::; 2008, which are optimal
in a sense of criterion K under the above constraints.
The stated problem of finding the minimum value of criterion K as the function
in 76 variables (and the respective values of the controlled parameters uli ¼ arg
minKÞ for each of the considered six scenarios under the aforementioned
constraints is solved using the Nelder–Mead algorithm. The results of the stated
problem solution are presented in Table 4.25.
Analysis of Table 4.23 shows that in the case of the considered scenarios, the
parametric control approach allows both reduction of the level of prices by
9.4–13.2% and increase of the GDP of the country by 1.3–2.44% for year 2008 in
comparison to the case without use of the parametric control approach.
270 4 Macroeconomic Analysis and Parametric Control of Economic Growth. . .
Table 4.24 Controlled model parameters and the constraints imposed on them
No. Controlled parameter ui
Interval of admissible
values of controlled
parameter
1 Rate of the VAT [0.135; 0.165]
2 Income tax rate for organizations [0.27; 0.33]
3 Property tax rate [0.009; 0.011]
4 Income tax rate for physical bodies [0.135; 0.165]
5 Rate of the single social tax [0.099; 0.121]
6 Share of the consolidated budget for purchasing final products [0.117; 0.143]
7 Share of the consolidated budget for backing the state sector [0.325; 0.398]
8 Share of the consolidated budget for backing the market sector [0.028; 0.034]
9 Share of the consolidated budget for social transferring [0.320; 0.391]
10 Share of the consolidated budget for purchasing the capital products [0.129; 0.158]
11 Share of the consolidated budget for purchasing investment products [0.068; 0.083]
12 Share of products of the state sector for selling in the markets of final
products for the market sector
[0.101; 0.123]
13 Share of products of the state sector for selling in the markets of
final products for economic agent no. 5 at the exogenous prices
[0.039; 0.048]
14 Share of products of the state sector for selling in the markets
of final products for economic agent no. 5 at market prices
[0.039; 0.048]
15 Share of products of the state sector for selling in the markets
of investment products at exogenous prices
[0.107; 0.131]
16 Share of products of the state sector for selling in the markets
of investment products at market prices
[0.107; 0.131]
17 Share of products of the state sector for selling in the markets
of the capital products at exogenous prices
[0.200; 0.244]
18 Share of products of the state sector for selling in the markets
of the capital products at market prices
[0.200; 0.244]
19 Share of products of the state sector for selling in the markets of final
products in foreign countries
[0.230; 0.281]
Table 4.25 Results of the application of the parametric control approach
Year
Criterion K without
parametric control
Criterion Kcorresponding to
the obtained optimal
values of parameters
Yj � ½2008�in billions tenge
Yj½2008�in billions tenge
Scenario 1 1.52 1.32 5.35 5.47
Scenario 2 1.63 1.41 5.38 5.45
Scenario 3 1.73 1.50 5.40 5.50
Scenario 4 2.08 1.87 5.32 5.44
Scenario 5 2.72 2.47 5.31 5.44
Scenario 6 3.32 3.04 5.31 5.44
4.3 National Economic Evolution Control Based on the Computable Model. . . 271
Chapter 5
Conclusion
This book describes the theory of parametric regulation of economic growth on the
basis of the deterministic continuous-time dynamical models, deterministic
discrete-time dynamical models, and stochastic discrete-time dynamical models
with additive noise. The presented theory consists of eight constituent parts. In the
context of these parts, we
• Proposed applied methods of estimation of structural stability and stability of the
mathematical models.
• Formulated and proved theorems on existence conditions for solutions of varia-
tional calculus problems of synthesis and choice of the optimal parametric
control laws.
• Proposed the method for analyzing dependence of the results of solving the
considered variational calculus problems on the values of uncontrolled
parameters. In particular, we gave a definition of the bifurcation points of
extremals of the variational calculus problem of choosing the optimal parametric
control laws and formulated and proved the theorem on a bifurcation point
existence conditions.
The efficiency of the parametric control theory is illustrated by a number of
examples, including the following: the mathematical model of the neoclassical
theory of optimal growth; the mathematical model of the national economic system
subject to influence of the share of public expenses and interest rate of governmen-
tal loans on economic growth; the mathematical model of the national economic
system that considers the influence of the international trade and currency exchange
on economic growth; Turnovsky’s monetary model; Jones’s endogenous model; the
computable model of general equilibrium of economic branches; Forrester’s math-
ematical model of global economy, etc.
Based on the dependence of a solution to a system of algebraic equations on their
coefficient, we proposed an approach to parametric control of the static equilibrium
of a national economy. The efficiency of this approach to parametric control
of static equilibrium is shown on the basis of the Keynes’s mathematical model
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2_5,# Springer Science+Business Media New York 2013
273
of general economic equilibrium and mathematical model of an open economy of a
small country.
This book illustrates the instrumental capabilities of a number of mathematical
models, such as the IS, LM, IS-LM models, Keynes’s model of general economic
equilibrium, model of an open economy of a small country, many models of market
cycles, a computable model of general equilibrium of economic branches, etc.,
respectively, in the following fields: macroeconomic analysis of the macroeconomic
markets’ conditions; estimation of characteristics of the market cycles in form of
amplitudes and periods of the respective business cycles; retrospective analysis
within the identification period; verification of the mathematical model and perspec-
tive analysis within the period of middle-term prediction of the indicators of the
economic agents and national economy, as well as their instrumental capabilities for
estimating the elasticity of the respective endogenous variables, sources of economic
growth, and estimation of several provisions of the macroeconomic theory on the
market cycles on the basis of the respective mathematical models.
The materials of this book have an important applied focus illustrating capabilities
of working out recommendations on economic policy. Hence, the diagrams of
dependencies of the optimal values of criterions of optimization problems formulated
on the basis of Keynes’s mathematical model of general economic equilibrium
and the mathematical model of the open economy in a small country (with the
coefficients estimated by the statistical data of the national economy) allow us to
consider the respective optimal values of economic instruments being a part of the
optimal solution of the formulated optimization problem as the recommendations on
economic policy.
The respective optimal parametric control laws can be considered the recom-
mendations on the economic policy for regulation of the business cycle within the
framework of the (selected by a decision maker) criteria for regulation of market
cycles’ evolution.
In the context of the computable models of general equilibrium, the respective
(agreed with a decision maker) optimal laws of parametric control of the criteria of
economic growth effectiveness can be considered the recommendations on economic
policy. Thus, the results of research presented in this book represent the elements
of the modern paradigm of macroanalysis and working out recommendations on
economic policy.
274 5 Conclusion
References
1. Aivazyan SA, Mhitaryan VS. Applied statistics and foundations of econometric: A course for
institutes of higher education. Moscow, UNITY, 1998; 1022 p. (in Russian).
2. Anosov DV. Coarse systems. Trans Institute Mathematics of USSR Acad Sci. 1985;169:59–93
(in Russian).
3. Arnold VI. Geometric methods in the theory of ordinary differential equations. Moscow:
MTsNMO; 2002 (in Russian).
4. Ashimov AA, Borovskiy YuV, Volobueva OI, Ashimov AsA. On choosing effective laws of
parametric control of market economic mechanisms. Avtomatika i telemekhanika. 2005;
No. 3:105–112 (in Russian).
5. Ashimov AA, Borovskiy YuV, Nizhagorodtsev RM, Novikov DA, Sultanov BT. Parametric
control of the characteristics of market cycles by the example of the Goodwin model. In:
Nizhagorodtsev RM, editors. Informational economics: the institutional problems.
Proceedings of the Ninth Drucker Readings; 2009; Moscow: Dobroe slovo, pp. 53–66
(in Russian).
6. Ashimov AA, Borovskiy YuV, Novikov DA, Nizhagorodtsev RM, Sultanov BT. Structural
stability and parametric control by the example of the models of cyclic dynamics of macro-
economic systems. Problemi upravleniya. 2010;No. 1:12–7 (in Russian).
7. Ashimov AA, Borovskiy YV, Sultanov BT, Iskakov NA, Ashimov AA. Elements of the theory
of parametric control of the evolution of the national economic system. Moscow: Publishing
House of Physical and Mathematical Literature; 2009 (in Russian).
8. Ashimov AA, Sagadiev KA, Borovskiy YuV, Iskakov NA, Ashimov AsA. On the theory of
parametric control of the development of market economics. Upravlenie bol’shimi sistemami.
2007; No. 17:3–27 (in Russian).
9. Ashimov AA, Sagadiev KA, Borovskiy YuV, Iskakov NA, Ashimov AsA. The development
and application of the theory of parametric control of the evolution of the economic system on
the basis of some neoclassic model of optimal growth. Avtomatika i telemekhanika. 2008;
No. 8:113–19 (in Russian).
10. Ashimov AA, Sultanov BT, Adilov ZhM, Borovskiy YuV, Borovskiy NYu, Ashimov AsA.
On the application of the parametric control theory to the computable models of common
equilibrium. Economika i matematicheskie metodi. 2010; No. 3. (in Russian).
11. Ashimov AA, Sultanov BT, Adilov ZM, Borovskiy YV, Novikov DA, Nizhegorodtsev RM,
Ashimov AA. Macroeconomic analysis and economic policy based on parametric control.
Moscow: Fizmatlit; 2010 (in Russian).
12. Bautin NN, Leontovich EA. Methods and techniques for qualitative analysis of the dynamic
systems on the plane. Moscow: Nauka; 1990 (in Russian).
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2,# Springer Science+Business Media New York 2013
275
13. Belen’kaya OI. Analysis of the influence of instruments of credit and monetary policies of the
Bank of Russia to the parameters of the real investment [www.optim.ru/fin/20012/rbelenkaya/
asp]. (in Russian).
14. Bobylyov NA, Emelyanov SV, Korovin SK. Geometric methods in variational calculus
problems. Moscow: Magistr; 1998, 658 p. (in Russian).
15. Stolyarov II (editors). State regulation of market economics. Moscow: Delo; 2001
(in Russian).
16. Hookenheimer J, Holmes F. Nonlinear oscillations, dynamic systems and the bifurcations of
vector fields. Moscow/Izhevsk: Institute of the Computer Research; 2002 (in Russian).
17. Dubovskiy SV. The Kondratiev cycle as a simulation object. Matematicheskoe modelirovanie.
1995;7(6):65–74 (in Russian).
18. Evtushenko YuG, Malkova VU, Stanevichus AA. The parallel search of global extremum of
multivariable function. Zhurnal vychislitel’noi matematiki i matematicheskoi fiziki. 2008;49
(2 C):255–69 (in Russian).
19. Ioffe AD, Tihomirov VM. The theory of extremal problems. Moscow: Nauka; 1974.
20. Kolemaev VA. Mathematical economics. Moscow: UNITY; 2002.
21. Koplyk IV, Polennitsa PV, Ostapova OP. Searching the Global extremum of function specified
by a simulation model. Vestnik SumGU, series “Technical sciences.” 2009; No. 2:105–12
(in Russian).
22. Kryukov RV. Governmental regulation of national economics. Moscow: Priorizdat; 2005
(in Russian).
23. Larichev OI. Theory and practice of decision making. Moscow: Logos; 2002 (in Russian).
24. Latov YV. Economics outside the law: Studies of the theory and history of shady economics.
Moscow: Moscow Public Scientific Fund; 2001 (in Russian).
25. Lotov AV, Pospelova II. Multi-criterion problems of decision making. Moscow: Publishing
Department of Faculty VMiK MGU, MAKS Press; 2008 (in Russian).
26. Magnitskiy NA, Sidorov SV. New methods of chaotic dynamics. Moscow: Editorial URSS;
2004 (in Russian).
27. Makarov VL, Bahtizin AR, Sulakshin SS. The application of computable models in state
administration. Moscow: Nauchniy ekspert. 2007; 304 p. (in Russian).
28. McConell KR, Brew SL. Economics. Principles, Problems Politics. 1992: 1, 2. Moscow:
Respublika (in Russian).
29. Matrosov VM, Izmodenova-Matrosova KV. Doctrine of Noosphere, global modelling, and
stable development. Moscow: Akademiya; 2005 (in Russian).
30. Matrosov VM, Khrustalyov MM, Arnautov OV, Krotov VF. On highly aggregated model of
the development of Russia. Analysis of development instability on the basis of mathematical
modelling. Preprints of the Second International Workshop. Moscow; 1993 (in Russian).
31. Orlov AI. Econometrics. A textbook. Moscow: Publishing House "Ekzamen"; 2002
(in Russian).
32. Petrenko EI. The development and realization of the algorithms for constructing the symbolic
set. Differentsialniye uravneniya i processi upravleniya (electronic journal). 2006; No.
3:55–96 (in Russian).
33. Petrov AA, Pospelov IG, Shananin AA. The experience of the mathematical modelling of
economics. Moscow: Energoatomizdat; 1996 (in Russian).
34. Pontryagin AS. Ordinary differential equations. Moscow: Nauka; 1970 (in Russian).
35. Sax DD, Larren FB. Macroeconomics. A global approach. Moscow: Delo; 1996 (in Russian).
36. Krasovskiy AA (editor). Handbook on the automatic control theory. Moscow: Nauka; 1987
(in Russian).
37. Abdiev KS (editor) The statistical yearbook of Kazakhstan. Almaty: Statistical Agency of the
Republic of Kazakhstan; 2001–2009 (in Russian).
38. Surinov AE (editor) The statistical yearbook of Russia. Moscow: Rosstat; 2010 (in Russian).
39. Tarasevich LS, Grebennikov PI, Leusskiy AI. Macroeconomics. Moscow: Vysshee
obrazovanie; 2006 (in Russian).
276 References
40. Tumanova ES, Shagas NA. Macroeconomics. Moscow: Infra-M; 2004 (in Russian).
41. Ulam S. Unsolved mathematical problems. Moscow: Nauka; 1964 (in Russian).
42. Friedman M. The quantitative theory of money. Moscow: El’f Press; 1996. 77 p (in Russian).
43. Holodniok M, Klich A, Kubicek M, Marek M. Methods of the analysis of nonlinear mathe-
matical models (Transl. from Czech.) Moscow: Mir; 1991 (in Russian).
44. Chernik DG, Morozov VP, et al. An introduction to the economical and mathematical models
of taxation. Moscow: Finansi i statistika; 2001 (in Russian).
45. Sharaev YV. Theory of economic growth. Moscow: Publishing House of SU HSE; 2006
(in Russian).
46. Yanovskiy LP. Controlling chaos in models of economic growth. Ekonomika i matematicheskie
metody. 2002;38(1):16–23 (in Russian).
47. Ashimov AA, Borovsky YuV, Ashimov AsA. Parametrical regulation of market economy
mechanisms. Proceedings of 18th International Conference on Systems Engineering ICSEng;
2005 Aug 16–18; Las Vegas, Nevada, pp. 189–93.
48. Ashimov AA, Sagadiev KA, Borovsky YuV, Ashimov AsA. On the bifurcation of extremals of
one class of variational calculus tasks at the choice of the optimum law of a dynamic systems
parametric regulation. Proceedings of XVIII International Conference on Systems Engineering
(ICSE 2006); 2006; Coventry University, UK, pp. 15–20.
49. Ashimov AA, Sultanov BT, Adilov ZhM, Borovskiy YuV, Borovskiy NYu, Ashimov AsA.
The development of parametrical regulation theory on the basis of one class computable
general equilibrium models. Proceedings of 12th International Conference on Intelligent
Systems and Control; 2009; Cambridge, MA, pp. 212–17.
50. Ashimov А, Sagadiyev КА, Borovskiy YuV, Iskakov NA, Ashimov AsА. The development
of the market economy evolution parametrical regulation theory on the basis of the growth
Model. 27th IASTED International Conference on Modelling, Identification and Control; 2008
Feb 11–13; Innsbruck, Austria, pp. 83–6.
51. Ashimov АА, Iskakov NA, Borovskiy YuV, Sultanov BT, Ashimov AsА. On the develop-
ment and usage of the market economy parametrical regulation theory on the basis of one-class
mathematical models. Proceedings of 19th International Conference on Systems Engineering
ICSEng 2008; 2008 Aug 19–21; Las Vegas, Nevada, USA, pp. 43–8.
52. Ashimov АА, Iskakov NA, Borovskiy YuV, Sultanov BT, Ashimov AsА. Parametrical
regulation of economic growth on the basis of one-class mathematical models. Syst Sci
2009;35(1):57–63.
53. Ashimov АА, Sagadiyev КА, Borovskiy YuV, Iskakov NA, Ashimov AsА. Elements of the
market economy development parametrical regulation theory. Proceedings of the Ninth
IASTED International Conference on Control and Applications; 2007 May 30–June 1;
Montreal, Quebec, Canada, pp. 296–301.
54. Ashimov АА, Sagadiyev КА, Borovskiy YuV, Iskakov NA, Ashimov AsА. On the market
economy development parametrical regulation theory. Proceedings of the 16th International
Conference on Systems Science; 2007 Sept 4–6; Wroclaw, Poland, pp. 493–502.
55. Ashimov АА, Sagadiyev КА, Borovskiy YuV, Iskakov NA, Ashimov AsА. On the market
economy development parametrical regulation theory, kybernetes. The International Journal
of Cybernetics, Systems and Management Sciences. UK, Bingley: Emerald. 2008;37
(5):623–36.
56. Ashimov АА, Sagadiyev КА, Borovskiy YuV, Iskakov NA, Ashimov AsА. The develop-
ment and usage of the market economy parametrical regulation theory on the basis of one-class
mathematical models. 14th International Congress of Cybernetics and Systems of WOSC.
Book of Abstracts; 2008 Sept 9–12; Wroclaw, Poland. Book of Abstracts, p. 14.
57. Cass D. Optimum growth in an aggregative model of capital accumulation. Rev Economic
Studies.. 1965;32:233–40.
58. CIA World Factbook – https://www.cia.gov/library/publications/the-world-factbook/index.
html.
References 277
59. Diamond P. National debt in a neoclassical growth model. Amer Economic Rev.
1965;55:1126–50.
60. Grossman G, Helpman E. Innovation and growth in the global economy. Cambridge, MA:
MIT Press; 1991.
61. Jones ChI. Was an industrial revolution inevitable? Economic growth over the very long run.
Adv Macroeconom. 2001;1(2, Article 1):49 p.
62. Koopmans T. Capital accumulation and economic growth. Pontificae Academiae Scientiarum
Scripta Varia.. 1965;28:225–300.
63. Leontief W. Essays in economics: theories and theorizing. London: Oxford University Press;
1966.
64. Lukas R. Making a miracle. Econometrica.. 1993;61(2):251–71.
65. Lukas R. On the mechanics of economic development. J Monetary Economics. 1998; No. 2.
66. Nelder JA, Mead R. A simplex method for function minimization. Comput J. 1965; №7:308–13.
67. Robinson C. Structural stability on manifolds with boundary. J Differential Eq. 1980; No. 37:
1–11
68. Schneider F, Enste D. Shadow economies: Size, causes and consequences. J Economic Liter.
2000;8(1):100.
69. Solow RA. Contribution to the theory of economic growth. Quart J Econom.. 1956;70:65–94.
70. Strongin RG, Sergeyev YD. Global optimization with non-convex constraints. Sequential and
parallel algorithms. Dordrecht/Boston/London: Kluwer Academic; 2000.
71. Swan T. Economic growth and capital accumulation. Economic Record. 1956;32(2):334–61.
72. Turnovsky S. Macroeconomic dynamics and growth in a monetary economy: a synthesis.
J Money, Credit Banking. 1978;10(1):1–26.
73. Turnovsky SJ. Methods of macroeconomic dynamics. Cambridge: MIT Press; 2000.
278 References
About the Authors
Ashimov Abdykappar Ashimovich is an academician at the National Academy of
Sciences of the Republic of Kazakhstan, doctor of technical sciences, and professor
at Kazakh National Technical University named after K.I. Satpaev, e-mail:
Borovskiy Yuriy Vyacheslavovich is a candidate of physical and mathematical
sciences and an assistant professor at Kazakh National Technical University
named after K.I. Satpaev, e-mail: [email protected].
Sultanov Bahyt Turlykhanovich is an adviser of the state scientific and technical
program at Kazakh National Technical University named after K.I. Satpaev, e-mail:
Abdilov Zheksenbek Makeevich is a doctor of economic sciences and a professor at
Kazakh National Technical University named after K.I. Satpaev, e-mail: [email protected].
NovikovDmitriy Alexandrovich is a correspondingmember of theRussianAcademy
of Sciences, doctor of technical sciences, and a professor at the Trapeznikov Institute
of Control Sciences of Russian Academy of Sciences, e-mail: [email protected].
Alshanov Rakhman Alshanovich is a doctor of economic sciences and a professor
at Kazakh National Technical University named after K.I. Satpaev, e-mail:
Ashimov Askar Abdykapparovich is a researcher of the state scientific
and technical program at Kazakh National Technical University named after
K.I. Satpaev, e-mail: [email protected].
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2,# Springer Science+Business Media New York 2013
279
Index
AAnalysis of national economy.
See Macroeconomic analysis and
parametric control
BBanach–Alaoglu’s theorem, 105–106
CChain-recurrent set, 7–8
Computable general equilibrium (CGE) model
economic branches
aggregate consumer, economic agent
17, 171
cyclic oscillations, 205–208
discretionary public policy
determination, 175
elasticity coefficients, 199–200
endogenous variables, 163–165
exogenous variables, 162–163
governmental prices, algorithm,
159–160
government, economic agent 18,
171–173
growth sources analysis, 200–203
integral indexes, 173
macroeconomic theory, 204–205
model agents, 160–161
model markets, 165–167
optimal parametric control laws,
208–211
parametric identification and
retrospective prediction, 173–175
producing products and services,
economic agent 1-16, 167–170
prospective analysis, endogenous
variables, 186–199
retrospective analysis, exogenous
variables, 176–186
statistical data, Republic of
Kazakhstan, 176
subsystems, 158–159
interbranch connections, 158
knowledge sector
aggregate consumer (households),
economic agent 4, 230–231
consumer price indexes,
237–238
cyclic oscillations, 238–240
economy branches, economic agent 3,
227–229
endogenous variables,
213–216
exogenous parameters, 213
general equilibrium, economic
branches, 236–237
government, economic agent 5,
231–233
innovation sector, economic agent 2,
225–227
integral indices, 233–236
macroeconomic indicators, cyclic
oscillations, 236
model agents, 211–213
model markets, 216–222
optimal parametric control laws,
240–242
science and education sector,
economic agent 1, 222–224
mathematical models, 157
phenomenological models, 157
regression equation, 157
A.A. Ashimov et al., Macroeconomic Analysis and Parametric Controlof a National Economy, DOI 10.1007/978-1-4614-4460-2,# Springer Science+Business Media New York 2013
281
Computable general equilibrium (CGE) model
(cont.)shady sector
aggregate consumer (households),
economic agent 4, 262–263
banking sector, economic agent 6, 265
economic agent no. 3, 261–262
economic agents, 242–244
endogenous variables, 247–251
exogenous parameters, 244–247
government, economic agent 5,
264–265
integral indices, 266
market sector, economic agent 2,
258–261
mathematical model, 242
model markets, 251–254
optimal values, 268–271
outer world, economic agent 7, 266
parametric identification, 266–268
state sector, economic agent 1,
254–258
Continuous-time deterministic dynamical
system
parametric control laws for, 9–11
solvability conditions, 11–12
Cycle stability, Kondratiev cycle, 147
Cyclic market dynamics
Goodwin mathematical model
description, 147–149
optimal parametric control law,
150–152, 156
structural stability analysis, 149–150,
153–156
Kondratiev cycle, mathematical model
criterion K optimal value, variational
calculus problem, 147, 148
description, 141–142
economic systems evolution, 144–146
estimation, robustness, 143
structural stability estimation, 147
DDecision makers (DM)
IDSS (see Information decision support
system (IDSS) and DM)
interaction vs. IDSS, 24Deterministic systems. See Discrete-time
deterministic dynamical system
Discrete-time deterministic dynamical system
parametric control laws, 13–14
solvability condition, 15
Discrete-time stochastic dynamical system
parametric control laws, 16–18
solvability conditions, 18–19
EEconomic branches, CGE model
aggregate consumer, economic
agent 17, 171
branch demand, labor force, 183, 185
budget, branch 1, 181, 182
capital productivity, 183, 185
cyclic oscillations, 205–208
discretionary public policy
determination, 175
elasticity coefficients, 199–200
endogenous variables, 163–165
exogenous variables, 162–163
fixed assets, branch, 183, 184
governmental prices, algorithm,
159–160
government, economic agent 18,
171–173
gross value added, branch 4, 181, 182
growth sources analysis
coefficients factors, 201–203
GVA, 200–201
retrospective data, 200
indicators, 178–183, 187–193
integral indexes, 173
investment products, economic agent 1,
176, 177
macroanalysis, retrospective, 194–196
macroeconomic theory, 204–207
model agents
constants and variables, 161
description, 160
economic sectors, 160–161
model markets, 165–167
optimal parametric control laws, 208–211
parametric identification and retrospective
prediction
algorithm, 174–175
macroeconomic mathematical
model, 173
measured and computed values, output
variables and deviations, 175
producing products and services, economic
agent 1-16, 167–170
production, investment products, 185–187
prospective analysis, 196–199
retrospective analysis, exogenous variables,
176–177
282 Index
statistical data, Republic of
Kazakhstan, 176
subsystems, 158–159
Economic equilibrium models, 125–127
Economic growth, one-sector Solow model
analysis, structural stability, 43, 45
choosing optimal laws, 43–45
dependence, optimal value, 45
estimation, model parameters, 42–43
model description, 42
Economic instruments
equilibrium solutions and payment balance
states, 136–139
equilibrium state macroestimation
equilibrium values, endogenous
parameters, 127
Keynesian mathematical model,
125–126
labor supply price and production
function, 126
IS model and analysis, 118–121
wealth and money markets
joint equilibrium and actual
values, 124
macroestimation, joint equilibrium
state, 124
plots, IS and LM models, 124, 125
Estimating defense cost, Richardson model
analysis, structural stability,
47, 48–49
choosing optimal laws, 47–48
dependence, optimal value, 49–50
estimation, model parameters, 46–47
model description, 46
FForrester’s mathematical model
analysis, structural stability, 86–87, 90
choosing optimal laws, 87–90
finding bifurcation points, 90–91
model description, 83–86
GGeneral equilibrium in economy. See National
economic evolution control based
CGE model
Global and national economy, Forrester’s
mathematical model
analysis, structural stability,
86–87, 90
choosing optimal laws, 87–90
finding bifurcation points, 90–91
model description, 83–86
Goodwin mathematical model
description, 147–149
optimal parametric control laws
computational experiments, 16, 152
economic parameter k, 151
market cycle, 152
relations set, 150
solving problems, stages, 151
structural stability analysis
with parametric control, 153–156
without parametric control,
149–150
IInformation decision support system
(IDSS) and DM
application of parametric control theory,
23–24
effective public economic policy based
aggregate scheme, 25
“choice of one or several econometric
models”, 27–28
“economic instruments of the economic
policy implementation in the chosen
direction”, 34–35
“evaluation of the national economic
conditions”, 26–27
“statement and solution of the
problem(s) of estimation(s)”, 28–34
Invertibility test, 8
Investment-savings (IS) model. See IS model
and analysis, economic instruments
IS model and analysis, economic instruments
macroeconomic theory, 119, 120
macroestimation, 118
public expenses and taxation, 120
statistical characteristics, 118–119
JJones’s model
analysis, structural stability, 100–101
dependence of optimal values,
104–105
finding optimal values, 102–104
model description, 98–99
parameters evaluation and retrospective
forecast, 100
parametric sensitivity estimation,
101–102
Index 283
KKeynesian model
economic instruments
equilibrium values, endogenous
parameters, 127
labor supply price and production
function, 126
parametric control, 127–128
Knowledge sector, CGE model
aggregate consumer (households),
economic agent 4, 230–231
consumer price indexes, 237–238
cyclic oscillations, 238–240
economy branches, economic agent 3,
227–229
endogenous variables, 213, 216–218
exogenous parameters, 213–216
general equilibrium, economic branches,
236–237
government, economic agent 5, 231–233
innovation sector, economic agent 2,
225–227
integral indices
economy, Russian Federation, 233–234
model and identification process, 234
Nelder-Mead algorithm, 234
output variables and respective
deviations, 234, 235
relations, 234
macroeconomic indicators, cyclic
oscillations, 236
model agents
economic agent 1, 211–212
economic agent 2, 212
economic agent 3, 212–213
economic agent 4-7, 213
model markets, 216, 219–222
optimal parametric control laws
computational experiments, 241
endogenous variables, 240
Nelder-Mead algorithm, 241–242
parametric control problem, 242
science and education sector, economic
agent 1, 222–224
Kondratiev cycle, mathematical model
criterion K optimal value, variational
calculus problem, 147, 148
description, 141–142
parametric control, economic system
evolution
capital productivity ratio, 145, 146
coefficients and criteria,
optimal laws, 145
efficiency, innovations, 146
optimal laws, relations, 144
optimal values, criteria, 144
robustness estimation, without parametric
control, 143
structural stability estimation, 147
MMacroeconomic analysis. See Computable
general equilibrium (CGE) model
Macroeconomic analysis and parametric
control
equilibrium solutions and payment balance
states
common economic equilibrium,
136–137
dependence, optimal values, 138, 139
IS-LM-ZBO, 138
money supply and public expenses, 236
national economy in 2007, 136, 137
open economy model, small country and
estimation, 129–136
small country model
dependencies, optimal values,
139, 140
external exogenous parameters, 139
unemployment, 129–130
Market economic development.
See Parametric control theory
Mathematical model, national economic
system
discrete-time dynamical system
(see Robinson approach)
international trade and currency exchange
analysis, structural stability, 73, 79–81
choosing optimal laws, 73–78
finding bifurcation points, 81–83
model description, 70–72
public expense and interest rate,
government loans
analysis, structural stability,
53, 62–65
choosing optimal laws, 53–59
finding bifurcation points, 66–69
model description, 50–52
parametric control, market economic
development, 59–62
stability indicators evaluation
absolute stability indicator, 5
CGE model, economic branches, 6
maximal absolute stability indicator, 5
Monte Carlo method, 4–5
284 Index
normalized input data vectors, 3–4
Orlov’s definition, 3
Mathematical models of cycles
Goodwin (see Goodwin mathematical
model)
Kondratiev cycle (see Kondratiev cycle,
mathematical model)
Money markets
equilibrium conditions
estimation, money velocity, 121
Fisher equation, 121
LM model, 121, 123
property demand, econometric
estimation, 122
statistical characteristics, 123
values, money supply and
aggregate, 122
values of multipliers, 122
and wealth, 124–125
NNational economic evolution control based
CGE model
economic branches
aggregate consumer, economic agent
17, 171
cyclic oscillations, 205–208
discretionary public policy
determination, 175
elasticity coefficients, 199–200
endogenous variables, 163–165
exogenous variables, 162–163
governmental prices, algorithm,
159–160
government, economic agent 18,
171–173
growth sources analysis, 200–203
integral indexes, 173
macroeconomic theory, 204–205
model agents, 160–161
model markets, 165–167
optimal parametric control laws,
208–211
parametric identification and
retrospective prediction,
173–175
producing products and services,
economic agent 1-16, 167–170
prospective analysis, endogenous
variables, 186–199
retrospective analysis, exogenous
variables, 176–186
statistical data, Republic of
Kazakhstan, 176
subsystems, 158–159
interbranch connections, 158
knowledge sector
aggregate consumer (households),
economic agent 4, 230–231
consumer price indexes, 237–2238
cyclic oscillations, 238–240
economy branches, economic agent 3,
227–229
endogenous variables, 213–216
exogenous parameters, 213
general equilibrium, economic
branches, 236–237
government, economic agent 5,
231–233
innovation sector, economic agent 2,
225–227
integral indices, 233–236
macroeconomic indicators, cyclic
oscillations, 236
model agents, 211–213
model markets, 216–222
optimal parametric control laws,
240–242
science and education sector,
economic agent 1, 222–224
mathematical models, 157
phenomenological models, 157
regression equation, 157
shady sector
aggregate consumer (households),
economic agent 4, 262–263
banking sector, economic
agent 6, 265
economic agent no. 3, 261–262
economic agents, 242–244
endogenous variables, 247–251
exogenous parameters, 244–247
government, economic agent 5,
264–265
integral indices, 266
market sector, economic agent 2,
258–261
mathematical model, 242
model markets, 251–254
optimal values, 268–271
outer world, economic
agent 7, 266
parametric identification, 266–268
state sector, economic agent 1,
254–258
Index 285
Neoclassical theory
analysis, structural stability, 37–38, 40–41
choosing optimal laws, 38–39
finding bifurcation points, 41
model description, 36–37
OOne-sector Solow model
analysis, structural stability, 43, 45
choosing optimal laws, 43–45
dependence, optimal value, 45
estimation, model parameters, 42–43
model description, 42
Open economy of small country
balance of payments, 134
consumption, 131
domestic commercial interest rate, 133
double balance in 2007, 135
economic indices, 129
equilibrium and actual values in 2007
and 2008, 135–136
investment model, 132
labor supply price, 131
money velocity, 130
net capital export model, 132
solving system, 134, 135
wealth export model, 133
Optimal control and uncontrolled
parameters, 156
Optimal control problems
cyclic oscillations
economic branches, 205–208
knowledge sector, 238–242
parameters, CGE model, 268–271
Optimal growth, neoclassical theory
analysis, structural stability,
37–38, 40–41
choosing optimal laws, 38–39
finding bifurcation points, 41
model description, 36–37
Optimal laws, parametric control theory
continuous-time deterministic
dynamical system
parametric control laws for, 9–11
solvability conditions, 11–12
discrete-time deterministic dynamical
system
parametric control laws, 13–14
solvability condition, 15
discrete-time stochastic dynamical system
parametric control laws, 16–18
solvability conditions, 18–19
uncontrolled parameter variations
bifurcation points, 21–22
continuous dependence of optimal
values, 20–21
description, 19
Optimal parametric control
Goodwin mathematical model, 150–152
Orlov’s definition, 3
PParametric control laws
continuous-time deterministic dynamical
system, 9–11
discrete-time deterministic dynamical
system, 13–14
discrete-time stochastic dynamical system,
16–18
Parametric control theory
algorithm of application
definition and implementation, 23–24
DM with IDSS, 24
applications
endogenous Jones’s model, 98–105
Forrester’s mathematical model, global
economy, 83–91
mathematical model, national economic
system, 50–83
neoclassical theory, optimal growth,
36–41
one-sector Solow model, economic
growth, 42–45
Richardson model, 46–50
Turnovsky’s monetary model, 91–98
components, 2–3
linear constraints, 1–2
optimal laws (see Optimal laws, parametric
control theory)
proof of statements
Banach–Alaoglu’s theorem, 105–106
Jacobian matrix, 114
Lipschitz constant of function,
106–108, 111–112
Robinson’s Theorem A, 115
Weierstrass’s theorem, 109–110, 112
rules of interaction between DM and IDSS,
24–35
stability and structural stability,
mathematical model, 3–9
Parametric identification, 173–175, 211, 242,
266–268
Problem solutions, 270
Public economic policy, 24–35
286 Index
RRetrospective prediction
knowledge sector
economic agent 4, aggregate consumer
(households), 230–231
economic agent 3, economy branches,
227–229
economic agent 5, government,
231–233
economic agent 2, innovation sector,
225–227
economic agent 1, science and
education sector, 222–224
endogenous variables, 213–216
exogenous parameters, 213
integral indices, 233–236
model agents, 211–213
model markets, 216–222
and parametric identification,
173–175
shady sector
economic agent 3, 261–262
economic agent 4, aggregate consumer
(households), 262–263
economic agent 6, banking
sector, 265
economic agent 5, government,
264–265
economic agent 2, market sector,
258–261
economic agent 7, outer world, 266
economic agents, 242–244
economic agent 1, state sector,
254–258
endogenous variables, 247–251
exogenous parameters, 244–247
integral indices, 266
model markets, 251–254
parametric identification,
266–268
Richardson model
analysis, structural stability, 47,
48–49
choosing optimal laws, 47–48
dependence, optimal value, 49–50
estimation, model parameters, 46–47
model description, 46
Robinson approach
analysis, structural stability, 6–7
chain-recurrent set, 7–8
construction, computational algorithm, 7
invertibility test, 8
weak structural stability, 9
Robinson’s theorem, 115, 143
SShady sector, CGE model
aggregate consumer (households),
economic agent 4, 262–263
banking sector, economic agent 6, 265
economic agent 3, 261–262
economic agents, 242–244
endogenous variables, 247–251
exogenous parameters, 244–247
government, economic agent 5, 264–265
integral indices, 266
market sector, economic agent 2, 258–261
mathematical model, 242
model markets
economic agent 5, 253–254
equilibrium prices, 251
governmental prices, 251–252
labor prices, 252–253
optimal values
GDP, 270
parameters, 270
parametric control approach, 270, 271
simulation process, 268
outer world, economic agent 7, 266
parametric identification, 266–268
state sector, economic agent 1
budgets, 256–257
capital assets’ dynamics, 257
economic agent 5, 256
exported products, 256
governmental prices and capital
products, 255
households, 255–256
investment products, 255
process, 254–255
VAT, profit tax and property tax,
257–258
Small country model
dependencies, optimal values, 139, 140
external exogenous parameters, 139
Solvability conditions
continuous-time deterministic dynamical
system, 11–12
discrete-time deterministic dynamical
system, 15
discrete-time stochastic dynamical system,
18–19
Stability indicators
absolute stability indicator, 5
CGE model, economic branches, 6
maximal absolute stability indicator, 5
Monte Carlo method, 4–5
normalized input data vectors, 3–4
Orlov’s definition, 3
Index 287
Stability of mathematical models
discrete-time dynamical system
(see Robinson approach)
stability indicators evaluation
absolute stability indicator, 5
CGE model, economic branches, 6
maximal absolute stability
indicator, 5
Monte Carlo method, 4–5
normalized input data vectors, 3–4
Orlov’s definition, 3
Stochastic systems, 18
TTheory of optimal growth, 36–41
Turnovsky’s monetary model
analysis, structural stability, 95
dependence of optimal values, 97–98
estimation, parametric sensitivity, 95–96
finding optimal values, 96–97
model description, 91–93
parameters estimation and retrospective
forecast, 93–94
UUncontrolled parameter variations
bifurcation points, 21–22
continuous dependence of optimal values,
20–21
description, 19
VVariational calculus problem
continuous-time deterministic
dynamical system
parametric control laws for,
9–11
solvability conditions, 11–12
discrete-time deterministic dynamical
system
parametric control laws, 13–14
solvability condition, 15
discrete-time stochastic dynamical system
parametric control laws,
16–18
solvability conditions, 18–19
uncontrolled parameter variations
bifurcation points, 21–22
continuous dependence of optimal
values, 20–21
description, 19
WWeak structural stability, discrete-time
dynamical system
analysis, structural stability, 6–7
chain-recurrent set, 7–8
construction, computational
algorithm, 7
invertibility test, 8
Weierstrass’s theorem,
109–110, 112
288 Index