an introduction to australian economy-wide modelling colin ... · an introduction to australian...
TRANSCRIPT
An Introduction to
Australian Economy-Wide Modelling
Colin Hargreaves
No. 35 - February, 1989
ISSN 0157-0188
ISBN 0 85834 807
An Introduction to Australian Economy-Wide Modellinq
by
Colin P. Hargreaves
This paper was written for a conference on the macroeconomic
models of Australia, that I arranged at the University of New
England in December, 1988. The participants at the conference
were people from the private and public sector and as such
were not academic macroeconomists. This paper was thus meant
to be a broad introduction covering the main developments in
macroeconomic modelling since its beginnings with the work of
Jan Tinbergen in the late 1930’s. After discussing some of the
earliest ideas in mathematical modelling of the economy, the
first models and their problems will be discussed leading up
to the developments at the Brookings Institution. Then some of
the major theoretical events over the last 20 years in dynamic
macroeconometric modelling will be discussed followed by the
development of other types of models such as applied or
computable general equilibrium models. I will conclude with
some comments upon the value of these models.
2
i. Early__~inninqL~ up to the 1930’s
In his posthumous and monumental work, "History of Economic
Analysis", Joseph $chumpeter points out how, thousands of
years before Christ, the Egyptians were running a planned
economy and, hundreds of years before Christ, the Chinese were
running exchange controls. Although many people were clearly
thinking about economic subjects, it is only with the Ancient
Greeks that we begin to find economic analysis along the lines
that we know today. Here we find Aristotle wrestling with the
concepts of ’value’, ~money’ and ’interest’, and Plato
discussing the Division of Labour and basically originating
the Cartal theory of money whereby money is not just another
commodity used as a medium of exchange (usually metal for
practical reasons) but an intrinsically valueless ’token’
which is socially accepted as a medium of exchange.
It is amazing how little development of economic thought there
was from this time right up until the 1600’s. Here the warring
state of affairs between England and Spain and between England
and Holland caused much discussion of the best way to increase
the wealth of England. Many Englishmen feared the wealth of
Spain in particular and its ability to buy large Armadas. Then
with the Cromwellian revolution there arose a parliament
thirsting for information upon which to make decisions about
the nation and taxes in particular.
Ioi Sir William Petty_~23-87.
The 1600’s saw a large number of great talents but one
particularly striking character was Sir William Petty. And he
only arose as a great figure because he broke his leg while a
14 year old cabin boy on an English merchantman. After being
dumped ashore in France he explained his pitiful circumstances
so well in Latin to the Jesuits that they not only cared for
him but admitted him to their college and there started his
career as one of the Renaissance’s self-made ’Universal’ men
of enormous versatility - physician, surgeon, mathematician,
inventor, engineer, member of parliament, public servant and
businessman - who refused the peerage twice before finally
accepting a third offer° One idea that he spent much time upon
was the ’Double Bottom’, a sort of catamaran; on one trial,
one of these outstripped all the boats in Dublin Harbour but
Petty was never able to obtain such good results again.
With Sir William Petty one has possibly the beginnings of
today’s mathematical economics and econometrics, with his
"Political Arithmetik, the art of reasoning by figures upon
things related to government". In this treatise he writes -
"Instead of using only comparative and
superlative Words, and intellectual Arguments, I
have taken the course ... to express in Terms of
Number, Weight or Measure; to use only Arguments of
4
Sense, and to Consider only such Causes, as have
visible foundations in Nature."
Petty is most famous in economics for his discussion of money
and the velocity of money in particular, but possibly of
greater relevance here is his analysis of National Income.
5
Table I: Sir William Petty[s Analysis of National Income
in "Verbum Sapienti", 16~
(£m) (£m)
Land: 24m acres @ 6s/8d p.acre
Houses: London - 28,000 @£15p.a. = £420,000pa => 5.04
Outer London - as many but not worthso much each, so same => 5.04
Cities and Market towns = 2 x London => 10.08
Outside cities and towns, approx same => 10.08
Houses Total say 30
Shipping: 500,000 tuns at 6d per tun 3
Cattel Stock: 1/4 of land value 36
Coined Gold and Silver: 6
Wares, Merchandises, Utensils of Plate & Furn. 31
Total of non-land 106
TotalWorth Rent
144
106 =>5.89say 7
Wealth of the Nation 250
Annual Proceed of Wealth 15
Expenditure:- 6 million people @ £6/13/4d => 40m per annum
"Then the Labour of the People must furnish the other 25"
As 250=>15, the Value of the people => 25 and so Human value => 416.67m
i.e. 420/6 = £70 each
In Table I, an attempt is made at summarising the logic used
in Perry’s ’~Verbum Sapienti" of 1665. He starts by saying that
there are 24 million acres of land worth on average about
6s/Sd per acre (a third of a pound sterling), creating a total
worth of £144m which should provide a rental income of £8m. As
to the houses on the Land, he separates these into various
categories in and outside London. In London there are about
28,000 houses with an average rental of £15p.a. giving a total
rental of £420,000p.a., so they must be worth £5.04 million.
Houses in the outer London suburbs are at least as many as
those in London but not worth so much, so, say, another £5.04
million. Houses in the other cities and market towns must be
worth twice as much as London and hence £10.08 million, and
finally all other houses must be worth about the same again
and hence another £10.08 million. All in all this makes a
housing total of, say, £30 million.
To this we must add £3million worth of shipping (500,000 tuns
at 6d per tun, 6d=£0.025), £36million of ’cartel~ stock ( say
a quarter of the land value ), say £6 million worth of coined
gold and silver, and say £31 million worth of Wares,
Merchandise~ Utensils of Plate and Furniture. This creates a
total non-land wealth of £106 million, leading to a convenient
total wealth of the nation of £250 million. Then just as the
£144m land wealth led to a rental income of £8 million, the
non-land wealth leads to, let us say, £7 million creating an
annual total proceed of wealth of a nice round £15 million.
7
Now there are 6 million people spending about £6/13/4d p.a.,
(£6o66), making £40 million p.a. Given the £15 million from
’rent’, "Then the Labour of the People must furnish the other
25". Probably for the first time, we have here none other than
an expression of the key national income-expenditure identity
that is at the core of all macroeconomic models.
He goes on to say that since £250 million of property wealth
led to £15 million worth of ’rents’, the value of the people
that creates £25 million of income must be £250,25/15 =
£416o67 million. Given that there are 6 million people, then
on average each person is worth just £70° This must also be
one of the first evaluations of the modern concept of human
wealth. He developed this argument as a reason for spending
the equivalent of a few pounds per person to save Londoners
from the ravishes of the plague; for his purposes, there was
no reason to be numerically more accurate°
Sir William Petty was the leader of a small group of
outstanding people including Gregory King, Charles Davenant
and John Graunt, the founder of demography. In Gregory King’s
"Natural and Political Observations and Conclusions upon the
State and Condition of England in 1696" we find the first
definite expression of demand equations that was not to be
improved on for over two hundred years. In discussing the
demand for wheat he comments that, if production fell by 10%
or 20% etcor be!ow normal, then the price will rise by some
8
30% and 80% above trend° This image of a demand curve was not
to be improved upon till the early 20th century with the work
of Moore. For an interesting re-appraisal of Petty, see the
book by Alessandro Roncaglia (1977, English text 1985) wherein
the author tries to escape from the linear development of
economic thought implied by Schumpeter’s historical analysis;
Roncaglia argues that the many aspects of Perry’s thought,
other than those developed by later economists, are now of
relevance today°
1.2 Adam SmithL 1723-90~ and David Ricardo~ 1772-1823o
While Petty had been a very dramatic figure of his age, a man
about whom there are many tales, not all apocryphal, the next
great watershed came with a very different type of man. In an
amazing indictment Schumpeter writes about Adam Smith, ’Few
facts and no details are needed about the man and his
sheltered and uneventful life (1723-90). It will suffice to
note: first that he was a Scotsman to the core, pure and
unadulterated; second, that his immediate family background
was the Scottish civil service .o. ; third, that he was a
professor, born and bred, ... by character indelebilis; fourth
oo. that no woman, excepting his mother, ever played a role in
his existence: in this as in other respects the glamours and
passions of life were just literature to him .... He was
conscientious, painstaking to a degree, methodical, well-
poised, honorable° He acknowledged obligation where honour
required it, but not generously° He never uncovered the
footprints of predecessors with Darwinian frankness. In
criticism he was narrow and ungenerous. He had the courage and
energy that exactly fit the scholar’s task and go well with a
good deal of circumspection°’ (op.cito p181-2).
Like many ’great’ works, Adam Smith’s Wealth of Nations,
published in 1776, was in no way innovative but rather it was
a detailed methodical statement of the ideas of the day put
into a logically consistent form. As such it publicised the
ideas, and economics became an established subject. ’But no
matter what he actually learned or failed to learn from his
predecessors, the fact is that the Wealth of Nations does not
contain a single analytic idea, principle or method that was
entirely new in 1776o~ (Samuelson p184) o
One of the major changes of this time was the relatively
important position given to the concept of labour. Value was
no longer a matter of exchange but of labour "conceived as
toil and time, the working day that divides up, and uses up,
men’s lives." (Sheridan, p66). Michel Foucault in his
magnificent analysis of human thought, ’Les mots et les
chosesW, marks this point as the move away from the Classical
Age of representation.
Foucault’s translator, Alan Sheridan, says "By making labour
the constant measure of exchange value" Adam Smith "laid the
i0
foundations for a political economy that could be based not on
exchange (representation) f but on production. But in Smith’s
analysis labour is still regarded as a commodity that can be
bought and soldo As long as representation retained its
precedence all commodities represented a certain labour and
all labour a certain quantity of commodities. The second stage
of the founding of political economy was the work of David
Ricardo .... Ricardo pointed out that labour could not be used
as a constant measure since it is ’subject to as many
fluctuations as the commodities compared with it’. This
insight led him to make the productive activity of labour
itself not the measure of value but the source of value. Value
is no longer a sign in the system of equivalences but a
product of labourf which is prior to any system of exchange.
The theory of production must now precede that of circulation.
Circular causality has been replaced by linear causality.
Ricardo has made possible the articulation of economics on
history." (Sheridan p68-9) o From here it was only a natural
progression in one respect to the historicological work of
Marx and the mathematical ’causal’ models of Popperian
positivism.
1o3 Leon Walra_~s 1834-1910.
The next major figure~ however~ of relevance to us, arising
from the organismic systemic thinking of the 19th century, is
the Frenchman, Leon Walras~ with his ’Elements d’economie
ii
politique pure’ in 1874; Schumpeter referred to his system as
’the Magna Carta of economic theory’ (op.cit. p242). Although
Isnard back in 1781 was possibly the first to try
mathematically to define and prove the concept of economic
equilibriums it is in Walras’s theory of general equilibrium
that we finally see a system that ’determines the equilibrium
values of all the economic variables, to wit: the prices of
all products and factors~ that would be bought in perfect
equilibrium and perfect competition by al! households and
firms.’ (Schumpeter~ p998-9).
Although this is the foundation of modern economics, Leon
Walras gained very little credit for it during his own
lifetime. As he himself once wrote in a letter to a friend,
"If one wants to harvest quickly, one must plant carrots and
salads; if one has the ambition to plant oaks, one must have
the sense to tell oneself: my grandchildren will owe me this
shade."
1.4 John Maynard Keynes, 1883-1946.
The thrust of development in economic thought turned from
comparative statics to analysis over time, and the business
cycle in particular° On the econometric side we have the path-
breaking work of Henry Moore and his ’Economic Cycles: Their
Law and Cause’ (1914) and ’Generating Economic Cycles’ (1923).
Joseph Schumpeter wrote his own two-volume historical book on
12
’Business Cycles’, (1939)o On the economic side we have John
Maynard Keynes and ’The General Theory of Employment, Interest
and Money~, (1936) o Although this may be in part a synthesis
of the work of Joan Robinson, RoG.Hawtrey, RoFoHarrod and
RoFoKahn, and at the same moment~ contemporaries in Sweden,
EoRoLindahl and CoFOhl~ were in the process of writing down
the same ideasr Keynes had a startling analytical flair.
Although nobody seems to read the original Keynes any more, I
would at least highly recommend reading such works as ’The
Economic Consequences of the Peace’ (1919)~ ’The Economic
Consequences of Mr Churchill’ (1925) ~ ’Can Lloyd George Do
It?’ (1929) and ’How to pay for the War’ (1940) o
Keynes~ major economic innovation was possibly what Schumpeter
beautifully describes as his stagnationist image of "an
arteriosclerotic economy whose opportunities for rejuvenating
venture decline while the old habits of saving formed in times
of plentiful opportunity persist" (op.cit. p 1171). As far as
today is concerned~ however, one must remember that Keynes’
concept of unemployment was that of cyclical unemployment, not
the Marxian structural unemployment which may seem more
relevant today.
13
2. The Development of National Macroeconomic Models.
2.1 Jan Tinbergen’s Early Models
While Henry Moore’s analysis of cycles and his estimation of
elasticities possibly led to the partial equilibrium analysis
of Allen, Wold and Stone, the general interest in cycles also
led to the pioneering work of Jan Tinbergen who may well be
called the forefather of macroeconomic models of today. As
early as 1935 he published a model of the Dutch economy, but
more famous is his 1939 model of the US economy. Though his
structures owe more possibly to common-sense considerations
than pure economic theory, they are no mean feats. The US
model has 32 stochastic equations and 18 identities with 14
exogenous variables. There are three consumption, three
investment and three liquidity preference functions and it
even includes the stockmarket. In his 1951 model of the UK
economy there are 45 equations with 9 exogenous variables. The
gold stock is modelled endogenously thus linking the monetary
sector to international trade and share prices are also
modelled endogenously.
A major aspect of Tinbergen’s work is that the statistical
analysis was used to estimate parameters and test hypotheses
resulting from theory, not the other way around. In
particular, it was not like the more statistical analysis of
Wesley Mitchell and the NBER, whose analysis aimed more at
14
establishing the facts so that the theory could be based on a
firm foundation° While Tinbergen’s approach has held sway, the
other methodology has reappeared recently with the VAR
modelling started by Christopher Sims (1980, "Macroeconomics
and Reality", Econometrica,48,1-48) who was worried about the
inaccuracies arising from imposing unproven theoretical
constrictions on the data~ We will return to this approach
later°
A major factor in the late 1930’s and the 1940’s was the
quality of the data. Firstly, for the major time series, only
annual data was available. Secondly, what with two world wars
and a depression, the economies had not been going through a
stable period from which stable statistical results could be
derived. For lack of other information, it is not~surprising
that many economists, during World War II, predicted another
depression after this war. Note also how, when estimating a
model in 1950 to analyse business cycles, Tinbergen used 44
observations from 1871 to 1914. Nobody now would use
observations over thirty years old to estimate a model to be
of relevance to today. Klein used 20 pre-war observations to
estimate his 1950 model and the Klein-Goldberger (1955) model
was estimated with 13 pre-war and 7 post-war observations!
2.2 The Age of Klein.
The post-war period saw a great explosion of macrodynamic
modelling led by the work of Lawrence Klein. One might almost
15
call this jubilant confidently optimistic period, the ’Age of
Klein’. People had such high hopes for economics; the world’s
problems would soon be solved. These grandiose claims were not
the fault of economic theory but rather forced upon it by the
social attitudes of the time. When scepticism started to creep
in, economics, and macroeconomic modelling research in
particular, were criticised more than they deserved. Now
people are beginning to be more realistic in their attitudes
and are realising that, while macroeconomic models do not
represent the economy perfectly, they are a valuable source of
information needed to make reasonable business and policy
decisions.
The first Klein model was an extraordinarily simple demand-
driven model of expenditure; note how many things are missing.
There is
i. n_~o explicit production function and so no analysis of
the supply side and technological effects,
2. n~o price or wage equations and so no analysis of
inflation,
3. n__Qo labour market and so no analysis of unemployment,
4. n_9o financial sector and so no analysis of interest
rates,
5. n__go international linkages are modelled and so no
analysis of the balance of trade, the exchange
rate etco
And so the list could go on. This is not meant as a criticism
16
of this early model, but, when you think of all the questions
attacked by modern modelsr it shows just how far such models
have developed over only thirty years.
One could say that model development really got under way with
a rush of new models when quarterly data became available.
Most models were fairly small with some 28-37 equations -
Duesenberry et al’s recursive model (1960), Klein-Popkin
(1961), Fromm (1962) and Liu (1963). On the other side one had
the idea that greater detail would lead to greater accuracy
and one has the Brookings model developed over 1961-4. This
had some 272 equationsr of which 125 were stochastic, 51 a
priori and 96 definitional. The debate over whether models
need to be highly disaggregated or not still continues today
and may never be decided; rather than asking this in general,
it may be more useful to ask such a question only in the
context of a specific planned use of the model.
2°3 ’Brookin~
The Brookings Bureau was an unparalleled phenomenon. For many
years it seemed to tower over macrodynamic theory° It showed
clearly how collecting a group of researchers together into a
Bureau can be so productive, especially in its early exciting
formative years ( two books and forty-three papers in four
years). The Bureau is probably more famous for its innovative
research than any one model°
To list just a few of the Bureau’s earliest ideas -
17
a) Detection of Structural Chanqe - they found
significant changes across 1953 and so omitted the 1948-1953Q2
data and used just 1953Q3 to 1962,
b) Use of an Input-Output Core, relating final demand
categories to industrial sectors - 7 price equations, 9 price
conversion equations relating industrial sector prices to
final demand prices, 7 equations relating industrial sector
outputs to components of GNP,
c) Use of Demographic Information with separate labour
participation rates for men and women of different ages and
even a marriages equation,
d) Modelling ’intentions’ to invest separately from
’realisations’,
e) Use of a Highly Disagqreqated Government Sector to
model federal expenditures separately from state expenditures,
and so the list could go on and on. But let us not imply that
the USA had total hegemony over this new modelling enterprise.
Many models were being developed in countries all over the
world, especially Canada, England, Holland, Japan and also
Australia (see Nerlove, 1966).
18
2~4 The Neo-Classical Input
Early Brookings and Klein were part of this optimistic period
of the 1950~s and early 1960’s, the glowing age of the welfare
state° By the late 1960’s a dramatic social change had set in.
For many people this is marked by the dramatic events in
France in May 1968, especially on the streets of the Latin
Quarter in Paris. This was possibly more a social
emanation/release of something that had been brewing for a
while, a disaffection with the state, a fear of the state and
disbelief in its benevolence and even its ability to ’do good’
if it tried. So many people at the time wanted to express
their dissatisfaction with ’the powers that be’, whatever they
might be.
Within this atmosphere arose a shift of opinion against
government interference, what one might call the ’keep your
hands off’ revolution, the so-called neo-classical revolution
in economics and the rise of monetarismo The word ’revolution’
over-emphasises the number of new innovative ideas and
probably reflects more of a social phenomenon amongst
economists° Although Milton Friedman had written his "Studies
in the Quantity Theory of Money" as early as 1956, his ideas
only really gained public notice in the late 1960’s at which
time he published ’The Optimum Quantity of Money’ (1969)o
There are three major effects upon modelling that are still
being felt today - longer time horizons, supply side modelling
and greater financial modelling. With the concept of
’permanent income’ and the longer time horizon also came the
analysis of technical change.
19
Now we automatically distinguish between models that have
shorter or longer horizons, as what you can assume to be fixed
depends on your time horizon. In the short-run the labour-
supply is fixed and there is no resolution of disequilibria.
In the long-run ’all’ can vary and so there must be a method
of resolving disequilibria. Because of the extra effort needed
to do this, long-run models tend to be smaller. Also it may
not be sensible to model the immense detail of some models in
the long run as the structure of the economy is probably not
stable enough. A difficulty with long-run models is that of
testing, in that we never experience an economy settling down
to a long-run stable equilibrium. Thus these models have to be
assessed more by consistency with theory rather than data,
raising many methodological questions.
Long-run models tend to be Classical as what we term
’Keynesian’ economics is more interested in the short run.
With the revived classical use of production functions also
came the introduction of the supply side, and with the
discussion of the velocity and supply of money came much
greater financial modelling. In the so-called ’Keynesian’
model, supply was fixed and so output was demand driven; there
was no real production function. This meant, for instance,
that one could not study the effects of short-run capacity
2O
constraints and disequilibria. Above all, these models are
useful as logically consistent expressions of a body of theory
and its implications; I will return later to the use of models
to express theory concisely and consistently.
21
3. Major Post-War Developments in Macroeconomic Theory.
A few of the major developments are probably worth mentioning
as they feature in most macroeconometric models in one way or
another. I will relate them in particular to four Australian
models -
a) the Treasury model called NIF88 (see $imes, 1988),
b) the MURPHY model (see Murphy, 1988),
c) the McKibbin and Sachs Global Mode! called MSG (see
McKibbin, 1988)
d) the computable general equilibrium model called ORANI
(see Dixon et al, 1982).
3.1 Wealth and the Consumption Function
Firstly there is the analysis of wealth and in particular the
Modigliani/Ando consumption function. A typical expression of
this function, used in both the NIF model and the Murphy model
of Australia, would be -
$Con = f(current labour income, non-human wealth (lags of),NPOPxPI ..
The left-hand-side is real per-capita consumption (consumption
divided by the population and a relevant price index). Note
that the current labour income is seen as arising from human
wealth; this distinction reminds one of Petty’s original work.
22
With wealth comes the introduction of a new stock variable,
that we still find very hard to measure° This also led to the
demand for stock-flow consistent models such as Warwick
McKibbin and Sach’s MSG model of the world that includes a
specific Australian subBmodel. With the lengthening of the
time-horizon and the emergence of new, better data, one can
now deal with stock-flow effects and, in particular, see the
work of EoPoDavis of the Bank of England.
3.2 Investment and Interest Rates°
An example of a stock effect is Tobin’s Q, the valuation
ratio. This is the ratio of the market valuation of equity to
the replacement cost of the capital stock. The market
valuation of equity should equal the discounted net present
value of all future profits. If the present and future profits
of the capital are worth more than the cost of that capital,
then it is sensible to invest in more capital. This approach
is used in Murphy, NIF and MSG. However it leads to serious
problems in deriving the ’market value of equity’
The first solution was to use Portfolio Demand Equations, a
large set of equations for different types of assets demanded
by different sectors - foreign, public, private and household.
This approach was used in an earlier version of NIF called
NIFI0o Through these equations one can determine the interest
rate, the price of equity, the exchange rate and many other
23
things. This approach has since been found to be unduly
complicated and may lead to unstable equations. A method used
now is to determine a key interest rate and from this
determine others by arbitrage conditions as is done in NIF88
(see Simes, 1988). Also in MSG, "Exchange rates, long-term
interest rates and equity prices are determined according to
intertemporal arbitrage conditions based on rational
expectations of the future path of the global economy."
(McKibbin, 1988)
3.3 Unemployment, Inflation and the NAIRU.
Another major theme has been the Phillips curve, originated in
1958. The trade-off between unemployment and inflation was not
new. The shock was the level of unemployment at which
inflation started, over 5%! This implied that if you wanted
unemployment below 5% you had to accept some degree of wage
inflation. Initially this was treated as a constraint and it
was a political nightmare; the aim of policy soon became to
shift the whole curve rather than just move along it.
This Phillips curve development was in fact a dubious one in
the first place for it was purely a statistical result. In the
late 1960’s new data led to the derivation of families of
curves and so the idea proliferated. In many if not by far the
majority of models, the rate of unemployment was used in the
causal determination of the rate of inflation whereas this
24
apparent causality was partly just a spurious effect of the
simultaneous forces in the economy. For this reason, and
partly the difficult political problems raised by the idea, it
is pleasing to see that the rate of unemployment is used less
now in determining inflation. For instance, in Ray Fair’s
Model (Fair, 1984), the labour supply is an intermediary
function of much else such as taxes, real wages, interest
rates and transfer payments, and so there is no clearly stable
Phillips curve°
With the debate upon the Phillips curve, there arose the idea
of the ’Natural Rate of Unemployment’ (NRUI in possibly the
most influential paper since Keynes, "The Role of Monetary
Policy", given by Milton Friedman in his Presidential Address
to the American Economic Association in 1968. The NRU is the
rate at which the average real wage rate is held constant and
any expansionary policy will just cause inflation in the long
run. This led to the notorious vertical long-run Phillips
curve. The form of this concept used today is the ’Non-
Accelerating-Inflation-Rate-of-Unemployment’. This is known as
the NAIRU, a name which is not much of an improvement. MSG,
NIF and Murphy all make use of this concept though the effect
is mollified by allowing some hysterisis; hysterisis is when
the particular path taken towards the long-run solution
affects that long-run solution.
3.4 Expectations
25
Friedman’s paper talked of people’s expectations of inflation.
This led to the idea that the government had to reduce
people’s inflationary expectations to shift the curve. To do
this they had publicly to espouse a long-term deflationary
policy. Friedman assured the policy-makers that this would not
take long and one could soon get back to a more publicly
acceptable policy° Out of this arose the grand experiment,
Margaret Thatcher~s "Medium term financial strategy" in which
her government announced a five-year slow reduction in the
rate of growth of money and in the PSBRo The result?
Unmitigated disaster. Not only did unemployment rise to
incredibly high levels but so did inflation! It will be many
years before Great Britain recovers from the scars (for
further comments on this tregime’, see Buiter and Miller,
1983). The more recent reduction in inflation was a result of
the reduction in global inflation levels and there are
disturbing signs of galloping inflation again in Britain. One
doubts whether the British government had read Friedman’s
later 1976 Noble Memorial lecture, in which he talks of
slumpflation, positively sloped Phillips curves and much
longer time periods.
People’s expectations were not only discussed in Friedman’s
theory but had been a standard part of the debate for years,
even in Keynest work. However, in 1960, J.F. Muth laid the
theoretica! foundations of a new development, known as
26
’rational expectations’. The name is rather misleading for in
at least one sense expectations had always been rational. The
distinguishing idea is that the expectations are forward-
looking, not adaptive and backward-looking. By this I mean
that the expected values are the actual values predicted by
the model, given all the information it has at this present
point of time. Given this, a much better name would be ’model
consistent expectations’ There is still much evidence for and
against the hypothesis but, above all, it is a very
parsimonious modelling strategy. The modeller does not have to
explain how people derive their expected values; he simply
equates them with the models expected values.
Early rational expectations models go back to Lucas (1973 and
1975), Sargent (1976), Sargent and Wallace (1976) and Barro
(1976). In England, Patrick Minford has used rational
expectations extensively in his Liverpool Model. In
Australia, they are used in both Murphy and MSG but not in
NIFo The debate for and against is at times surprisingly
heated and vitriolic. The popular middle line is to use
rational expectations not across the board but in specific
areas such as in financial markets. The expectations debate
goes on to discuss the different effects of anticipated and
unanticipated changes in government policy and even the
effectiveness altogether of government policy.
27
4. Developments in Statistical Estimations
There is not time or space in this paper to describe the many
other developments in macroeconomic theory, for developments
in other ways in modelling must be mentioned. Here I can pay
little more than lip-service to the dramatic developments in
econometrics, starting with the work of Trygve Haavelmo (for
example 1940, 1943 and 1947) and the Cowles Commission in the
1940’s and 50’s. Most estimation techniques were developed
some while ago now and it is possibly surprising to see the
lack of use today of instrumental variable methods and LIML
(if not FIML) techniques. A major criticism of econometrics is
possibly that too much effort has been put into hypothesis
testing rather than estimation. Economics is not without its
own major debates such as the procedural differences between
David Hendry, Christopher Sims and Edward Leamer (see Pagan
1987). Curiously such issues, while vitally important, are
commonly omitted from econometrics courses which concentrate
on technique. For instance, it is rarely mentioned in student
texts; it is not discussed at all in Gujarati’s popular basic
text (1988).
After going through a rather dry period full of asymptotic
theory, there have been a number of exciting developments in
econometrics in the last few years, notable amongst which are
the specification tests which have been written about by a
number of Australasians, such as David Giles, Max King, Mike
28
McAleer and Adrian Pagan (see references below). Also there
has been the development of the methodologically more dubious
diagnostic tests for heteroscedasticity, autocorrelation, non-
normality, etc. Hardly a model is estimated without extensive
use being made of these tests° Many models are still estimated
with single equation techniques, and have to be for lack of
data; this raises the risk of simultaneity bias and here the
Hausman tests are very useful~
Finally the increased computing power available has permitted
the development of computer-hungry techniques such as Jack-
knife and Bootstrap methods (see Miller 1974, Efron 1982 and
Hinckley 1988), and the movement of computer-aided design
(CAD) techniques into statistics with the development of
exciting new graphical techniques. Further information on the
development of econometrics can be gained from Epstein’s
(1987) ’History of Econometrics’, which is particularly
relevant to modelling as it concentrates especially upon
structural estimation, the question of exogeneity, rational
expectations and vector autoregressions.
29
Extensions of the Macroeconometric Model.
5.1 VAR Modellinq.
One econometric technique deserves greater mention as it has
led to a whole new field of macroeconometric modelling. The
new models are ~ector ~uto-~egressive ( ~oving-~verage )
models known as VAR(MA) models. They are essentially just
multivariate Box-Jenkins ~uto-~egressive/~oving-~verage
(AR!4A) models. As mentioned earlier, they arose because of
fears that theoretical restrictions were too constrictive on
the data, not allowing the data to ’speak’ The comments in
Christopher Sims’ critique of standard methods remind one of
the work of Wesley Mitchell.
Since his tAutoregressive index model for the U.So, 1948-75’,
published in 1981, many VAR models have been estimated
including Sims own world model, and, here in Australia, one
estimated by Trevor and Thorp of the Reserve Bank (1988). The
main trouble with these models is that in the end they need as
many arbitrary restrictions as the normal macroeconometric
models and I prefer the normal model’s theoretical rather than
the VAR modelts statistical justifications for these
restrictions (see Doan et al, 1984, and Trevor & Thorpe).
Although data-oriented, these models do not produce any better
predictions although they may improve with the development of
cointegration. A major use of VAR models might be to use them
30
as benchmarks for other theoretical models.
5°2 International Models.
Aside from the econometric developments, mention must be made
of the development of other new types of model. In the early
1600’s it was forbidden to export gold from England as it was
thought that this would reduce the wealth of England and hence
its strength to fight Spain and Holland. When an East India
Company boat sank with a large cargo of gold bullion soon
after leaving an English port, it raised a furor of public
consternation. The director of the East India Company, Thomas
Mun, sprang to its defence with his tract "England’s Treasure
by Forraign Trade" in 1630. Possibly from this arose the
mercantilist theories in economics. Given this and England’s
large entrepot trade~ it is surprising that there are few
British international models°
Now there are many types of international model with different
geographical coverage and country groupings. Another
distinction is whether they use a common model structure for
each country or combine differing sub-models of each country.
The Project Link uses 79 different models developed in their
original countries (Waelbroeck, 1976) o The Japanese EPA world
econometric model uses different models for each country,
constructed however by the EPA itself (Yoshitomi et al, 1984).
MSG uses similarly structured models for each sector of the
31
world but the Japanese labour market in particular has to be
modelled differently. These models are now used extensively by
the IMF and others to help devise international policy. This
is possibly the greatest area of development in modelling
today (for exampler see Bryant et alr 1988).
5.3 Multi-Reqional Modellinq~.
Of a similar nature in a way to international models are the
multi-regional models of a single country. Here two distinct
approaches abound. Firstly there is the top-down approach (TD)
where some national model is used to create some aggregate
forecasts which are then broken down into regional forecasts
in some consistent fashion. An example is the model of Milne,
Adams and Glickman (1980) which uses the Wharton model’s
aggregate forecasts. The alternative approach is of course the
bottom-up approach (BU) which builds up from individual sets
of equations for each region, adding the results together in
some consistent manner to obtain national aggregates. An
example is the National Regional Impact Evaluation System
(NRIES) of Ballard, Glickman and Gustely (1980). Here there
are 51 state models with approximately 230 equations each plus
a 50 equation national model; there are some 4000 behavioural
equations in some ii,000 equations all told! Given the lack of
good state-based statistical data, the TD approach is probably
more suitable to Australia.
32
6o General Equilibrium Modellinq~
6.1 The ’Tableau Economique’o
The final development that I will discuss is General
Equilibrium Modelling° Schumpeter writes "few sequences in the
history of economic analysis are so important for us to see,
to understand, and to fix in our minds, as is the sequence:
Petty-Canti!lon-Quesnay’~ (p218) o While Richard Cantillon
(1680-1734), a Parisian banker of Irish extraction, was
clearly inspired by the work of Petty, his achievement stands
out clearly in its own right. Most outstanding of all was his
understanding of the circular flow of the economic process as
the payments of each sector of society become the incomes of
other sectors. In a!l its bare essentials, Cantillon really
created the ~Tableau economique’ that was developed by the
surgeon-physician to Madame Pompadour, Francois Quesnay (1694-
1774) o
Quesnay was a member of the physiocrats who tried to find
valid criteria for judging particular policy proposals. While
the mercantilists had aimed at increasing wealth by ’forraign’
trade, the physiocrats aimed at increasing wealth by
increasing production, particularly agriculture. Quesnay’s
work lies directly in the tradition, since Petty, of
’describing the facts’o Amongst his many treatises on economic
theory, the most famous is the ’Tableau Economique’ of 1758.
He followed Cantillon’s categorisation of people into three
classes, the landowners (classe proprietaire), the farmers
(classe productive), and anyone not involved in agriculture,
the merchantmen, which he called the ’classe sterile’. The
Tableau depicts the flows through the economy of funds from
one sector (classe) to another.
33
Apart from the great simplification provided by the Tableau it
lays the way open for statistical analysis; Quesnay was fully
aware of the possibilities here° Like Petty he also tried to
estimate the national income. Schumpeter wrote "the Cantillon-
Quesnay tableau was the first method ever devised in order to
convey an explicit conception of the nature of general
equilibrium° It would seem impossible to exaggerate the
importance of this achievement if admiring disciples had not
actually succeeded in doing so". (op.cit. p242)o However this
approach has led a checkered history. It lay essentially
undeveloped until Walras took it up over a hundred years later
and even then, because of the mathematical complexity, it was
not further developed until the 1930’s and 50’s.
6.2 Input-Output Analysis°
The essential concept of interchanging flows was developed by
Wassily Leontief in his Input-Output analysis (Leontief, 1941
and 1964). This led in particular to Richard Stone’s Input-
Output model of the UK economy, which has developed into the
34
massive present-day Cambridge Growth Project model. In
Australia the first I-O tables were produced by Cameron in the
late 1950~so In 1964 the Australian Bureau of Statistics (ABS)
under Cameron produced its first table for 1958-9. While
initially such tables were infrequent, dependent on censuses,
more recently the ABS has produced tables for consecutive
years in the early 1980’s and even a state breakdown for 1985-
6o For further information on Australian Input-Output Tables,
see the article by Barbetti (1987) in the new Australian
Journal of Regional Studies; for a computer package to use
these tables~ there is the GRIMP package, described in Jensen
and West (1986) and distributed by the Department of
Immigration~ Local Government and Ethnic Affairs~
6°3 The Johnasen and Scarer A roacheso
If you add some neo-classical production functions and some
price-sensitive demand equations to the basic I-O table, you
have the essence of a highly non-linear Walrasian general
equilibrium model° This was developed by two different groups.
One group stems from the mathematical developments of
Scarf(1967~ 1973)o Scarf’s mathematical solution produced a
whole stream of Applied General Equilibrium models° In the
U.K., one has the models of Whalley (1975), Miller and Spencer
(1977) and Piggott and Whalley (1977). In the US, one has
Fullerton et al (1978) and in Australia, Piggott (1980). Much
of this work is described in the book ’New Developments in AGE
analysis’, edited by John Piggott (1985) o Note that Warwick
McKibbin describes his MSG model as Wa dynamic general
equilibrium model of a multi-regional world economyWo
35
Most general equilibrium Scarf-type models were developed to
study particular policy changes. The other group to develop
the Walrasian scheme have tended to develop more long-lasting
models of the whole economy. These stem from the pioneering
work of L. Johansen with his 1960 model of Norway° Essentially
this is based on the idea of taking logs and differentiating
to create a linear model. Strangely the Scarf and Johansen
groups appeared initially to be totally unaware of each
other’s work°
The Australian example of a Johansen type model is ORANI
developed mainly by Peter Dixon in the IMPACT Project directed
by Alan Powell. Now it is developed by Peter Dixon and Brian
Parmenter at the Institute of Applied Economic and Social
Research in Melbourne (see Dixon et al, 1982). This approach
has developed greatly over the last ten years, partly through
the mathematical expertise of Ken Pearson and the development
of the GEMPACK set of computer programs for running models°
Peter Dixon and Brian Parmenter have also now produced a
forecasting version of ORANI simply called ORANI-F. These lie
directly in the tradition stemming from Quesnay of a
macroanalytic description of the economy based on a
microanalytic theory of exchange and, as such, they are often
called, rather confusingly, microeconomic models of the
36
economy.
Such models should be contrasted with models that use a
microeconomic input-output core as a foundation for a
macroeconomic model. An Australian example of this is the
Institute Multi-Purpose model (IMP), developed largely by
Peter Brain of the National Institute for Economic and
Industrial Research. This is a highly detailed, largely
Keynesian, demand-driven econometric model containing some
6000 equations in ten modules; apart from the main module,
other modules cover demography, finance, energy transport,
industrial activity, agriculture, international trade and
telecommunications° There are two further modules, a linear
programming module to determine patterns of demand for new
energy types and a module splitting key aggregates by state.
For further information see Peter Brain (1986).
37
7. The Value of Models.
Let me conclude quickly with some brief comments on the value
of these various types of models. Firstly, modelling forces an
open expression of theory in a logically consistent way. Great
analysts may have magnificent subconscious models in their
heads but they are of little use to other people and die with
them. Models also help the analysis of the economy and the
effects of policies. And, however much criticism they have
received, they are also helpful in understanding the future.
As Samuelson wrote ’However good is the qualitative judgment
..o our judgment is kept tuned up by looking at computer
forecasts.’ One should not believe one model’s forecast
wholeheartedlyo Rather the analysis of many different models’
forecasts, and other people’s personal judgements, help create
better judgements in the on-going decision-making processes of
government and private industry.
Different models reveal different aspects of the economic
system. The aim of conferences on the models, and of modelling
Bureaux, is not to produce one grand supermodelo It would
probably be as bland as the well-known cartoon and now film
super-heroo Different models implicitly summarise the
implications of different theories and hence comparative
analysis leads to greater understanding of these theories and
inevitably to their development. One should not think of it as
the ’War of the Models’ but rather construe it as the
38
construction of a well-informed debate. And it is precisely
this debate which I hope to enhance by the establishment of an
economy-wide modelling Bureau to be called the Economic
Modelling Bureau of Australia (EMBA).
39
References.
Adams, FoGo, and N.J. Glickman (1980) editors, Modelin~ theMultire ional Economic S stem, Toronto: Lexington Books.
Amano, A., A.Marayama and M.Yoshitomio (1982) "EPA WorldEconomic Model", Vol I: Model Structure, Vol 2: SimulationResults° EPA World Econometric Model Discussion Paper ii,Tokyo, EPA, ERIo
Ando,Ao (1981), "On a Theoretical and Empirical Basis ofMacroeconometric Models’t, in "Large-Scale Macro-econometricModels", edited by J.Kmenta and J.B.Ramsey, Amsterdam: North-Holland°
Ando,A. and F.Modigliani (1963), "The Life Cycle Hypothesis ofSaving: Aggregate Implications and Tests", American EconomicReview, 55-64.
Ball, RoJo (1973) "The Economic Models of Project Link"° InBall,R.Jo,ed. The International Linkage of National EconomicModels, Amsterdam: North-Holland~ 65-107.
Ballard, K.Po, N.J. Glickman and R.Do Gustely (1980) o "A Bottom-up Approach to Multiregional Modeling", p147-160 in Adams,FoG. and N.Jo Glickman, editors, Modeling the MultiregionalEconomic S~, Toronto: Lexington Books.
Barbetti, AoJ. (1987) "Systems for the compilation of AustralianInput-Output Tables", The Australian Journal of RegionalStudies, No 2, p49-63
Barker, To and Wo Peterson (1987) The Cambridg_9 MultisectoralDynamic Model of the British Economy, Cambridge: CambridgeUniversity Press.
Barro, RoJ. (1976) "Rational Expectations and the Role ofMonetary Policy", Journal of Monetary economics, 2, pi-33.
Blaug, M. (1985) Economic Theor~ in Retros ep_~, 4th edition,Cambridge University Press.
Brain, P. (1986) The Micro-economic Structure of Australia,Longman-Cheshire.
Bryant, RoC., DeW° Henderson, G. Holtham, Po Hooper, SoA.Symansky (1988), editors. E_~pirical Macroeconomics forInterdependent Economies, Washington: Brookingso
Buiter and Miller (1983)o "Changing the Rules: EconomicConsequences of the Thatcher Regime", Brookin~ers onEconomic Activity, 2, p305-365~
4O
Cantillon, R~ (1931) Essai sur la Nature du Commerce en General,in A.Eo Monroe (1930) Opocit.
~Challen, DoWo, and A.J. Hagger (1979) Modellinq the AustralianEconomy, Melbourne: Longman Cheshire.
Doan,To, Ro Litterman, and Co Sims (1984) "Forecasting andConditional Projections Using Realistic Prior Distributions",Econometric Reviews, 3, 1 : i-i00
Dixon, PoB., BoR. Parmenter, J. Sutton and DoP. Vincent (1982),ORANI: A Multisectoral Model of the Australian Economy,Amsterdam: North-Hollando
Dornbusch, (1976) t~Expectations and Exchange Rate Dynamics",Journal of Political Economy,
Duesenberry, J., O. Eckstein and G. Fromm (1960) "A Simulationof the US Economy in Recession", Econometrica, XXVIII, 749-809.
Economic Planning Agency, World Economic Model Group (1986) "TheEPA World Economic Model: An Overview". Discussion Paper No37° Tokyo:EPA,ERI.
Efron, Bo (1982) ~’The Jackknife, the Bootstrap and otherResampling Plans’�, in Regional Conference Series in AppliedMathematics, No 38, Philadelphia: SIAM.
Epstein, R.Jo (1987) A Histor~ of Econometrics, Amsterdam:North-Holland
Fair, R.Co (1984), Specification, Estimation, and Analysis ofMacroeconometric Models, Cambridge, Mass: Harvard UniversityPress.
Foucault, M (1966) Les mots et les choses, Paris: Gallimard.Translated by Alan Sheridan as The Order of Thinq~, Tavistock:Pantheon, 1970
Friedman, Mo (1968) "The Role of Monetary Policy", PresidentialAddress to the American Economic Association.
Friedman, Mo (1969) The O_ptimal 9uantit~ of Money and otheressay_~, London: Macmillan
Friedman, Mo (1976) Nobel Memorial Lecture
Friedman, Mo (1970) "A Theoretical Framework for MonetaryAnalysis", Journal of Political Economy.
Fromm~ G~ (1962) ~IInventories, Business Cycles, and EconomicStabilization", in Inventory Fluctuations and EconomicStabilization, Part IV, Joint Economic Committee, U.S.Congress 87th Session, Washington: U.S. Government PrintingOffice, p69-89.
41
Fromm, G. and L.R.Klein (1975) editors. The Brookings Model:Perspective and Recent Develop, Amsterdam: North-Holland.
Fullerton, Do, J.B. Shoven and Jo Whalley (1978) "GeneralEquilibrium Analysis of U.S. Taxation Policy", in 1978Compendium on Tax Research, Washington: Office of TaxAnalysis, U.S. Treasury.
Gujarati, DoN. (1988) Basic Econometrics, 2nd edition, McGraw-Hill International Edition.
Haavelmo, Trygve (1940) "The inadequacy of testing dynamictheory by comparing theoretical solutions and observedcycles", Econometric___~a, 8:312-21, October.
Haavelmo, Trygve (1943) "Statistical Testing of Business CycleTheories", Review of Economics and Statistics, 25:13-18,February.
Haavelmo, Trygve (1947) "Methods of measuring the marginalpropensity to consume", Journal of the American StatisticalAssociation, v42, no 237, March, pi05-122.
Hausman, J.A. (1978) "Specification Tests in Econometrics",Econometrica, 1251-72.
Hickman, B.G., and L.R.Klein (1985) "Recent Developments inProject LINK", Social Science Research Council Items 39, 1-2,p7-11.
Hinckley, D.V. (1988) "Bootstrap Methods", Journal of the RoyalStatistical SocietlL~ Series B, 50, no 3, p321-337
Jensen, R.C., and G.R. West (1986) I__~put-Output forPractitioners: Theory and Applicationsr AGPS, Canberra.
Johansen, Lo (1960) "A Multisectoral Study of Economic Growth",Amsterdam: North-Holland
Keynes, J.M. (1919) The Economic Consequences of the Peace,London: Macmillan.
Keynes, J.M. (1925) The Economic Consequences of Mr Churchill,London: Hogarth.
Keynes, J.M. (1927) The End of Laissez-Faire, London: Hogarth.
Keynes, J.M. (1929) Can Lloyd Geor e D~?, London: Nation andAthenaeum.
Keynes, JoM. (1936) The General Theory of Em~loyment, Interestand Mone~, London: Macmillan.
Keynes, J.M. (1940) How to PaZ for the War, London: Macmillan
42
King, Gregory (1696) Natural and Political Observations andConclusions ~Don the State and Condition of England in 1696
King~ Mo, and DoEoA. Giles (1987) editors. SpecificationAnalysis in the Linear Model~ London: Routledge & Kegan Paul.
Klein, LoR0 (1950) Economic Fluctuations in the United StatesL1921.1941~ New York: John Wiley and Sons
Klein, LoRo (1986) ~’Economic Policy Formation; Theory andImplementation~’o In Handbook of Econometric , vol 3, p2068-70,edited by Zo Griliches and MoDo Intriligator, Amsterdam:North-Hollando
Klein, LoRo and AoSo Goldberger (1955) An Econometric Model ofthe United States ~ Amsterdam: North-Holland.
Klein~ LoRo and Jo Popkin (1961) "An Econometric Analysis of thePost-War Relationship between Inventory Fluctuations andChanges in Aggregate Economic Activity~’, p 69-89 in InventoryFluctuations and Economic Stabilization, Part III, JOintEconomic Committee, UoSo Congress 87th Session, Washington: SGovernment Printing office°
Kmenta~ J~ (1966) ~An Econometric Model of Australia, 1949-61",Australian Economic Pap_@_r~_, vol 16, p121-9o
Leontief~ W (1941) The Structure of the American Economy_~ 1919-i92~~ Cambridge, Mass: Harvard University Press (secondedition: 1951, 1919-1939, NY: Oxford University Press°
Leontief~ W (1966) I~_~n~ut-out~ut Economics, NY: Oxford UniversityPress°
Liu, ToCo (1962) ’~An Exploratory Quarterly Model of EffectiveDemand in the Post-War UoSo Economy", Econometrica, XXXI, 301-48.
Lucas, (1973) "Some International Evidence on Output-InflationTrade-offs~’, American Economic Review~ 63, p326-34.
Lucas, RoEo Jnro (1975) "An Equilibrium Model of the BusinessCycle~ Journal of Political Economy, 83, pli13-44.
McAleer~ Mo (1987) ~’Specification Tests for Separate Models: asurvey~’ in King and Giles~ Opocito
McKibbin, WoJo (!988), ~Policy Analysis with the MSG2 Model",Australian Economic Pa~ers, July, vol 27, p126-50.
McKibbin, Wo, and JoSachs, (1988) The McKibbin-Sachs (MSG)Global Model of the World Economy, July
Miller, RoGo,Jnr (1974) "The Jackknife: a review",Biometrika,61,1-17.
43
Miller, M.H., and J.Eo Spencer (1977) 1’The Static EconomicEffects of the U.K. joining the E.C.: A General EquilibriumApproach", The Review of Economic Studies, 38, April, p175-208.
Milne, W.J., F.G. Adams and N.J. Glickman (1980) I’A Top-DownMultiregional Model of the US Economy" p133-145 in Adams, F.G.and N.J. Glickman, editors, Modeling the MultireGionalEconomic ~, Toronto: Lexington Books.
Minford, A.P.L., S.Marwaha, K.G.P.Matthews and A.Sprague (1984)"The Liverpool Macroeconomic Model of the United Kingdom’t,
Economic ModellinG, 1,24-62.
Monroe, A.E. (1930) editor, Early Economic Thought: Selectionsfrom economic literature 9rior to Adam Smith, Cambridge:Harvard University Press.
Moore, H. (1914) Economic Cycles: Their Law and their Cause,Macmillan, NYo
Moore, H. (1923) GeneratinG EconomicS, Macmillan, NY.
Mun, Thomas (1630) W’England~s Treasure by Forraign Trade", p12-37 in Masterworks of Economics: DiGests of i0 Great Classics,edited by L.Do Abbott, New York: Doubleday (1946).
Murphy, C.W. (1988a) W’An Overview of the ~JRPHY Model",Australian Economic Pa ep~, July, 175-199.
Murphy, CoW. (1988b) "Rational Expectations in Financial Marketsand the Murphy Model~’, Australian Economic Pa ep_e_r_~, July, vol27, p61-88
Murphy,C.W., I.Ao Bright, RoJ. Brooker, W.Do Graves and B.K.Taplin (1986) "A Macroeconometric Model of the AustralianEconomy for Medium-term Policy Analysis", Office of theEconomic PlanninG and Advisory Council Technical Pa~, No 2.
Muth, J.Fo (1960) "Optimal Properties of Exponentially WeightedForecasts", Journal of the American Statistical Association,55, p299-306
Nerlove, M. (1966) "A Tabular Survey of Macro-EconometricModels’~, International Economic Review, Vol 7, No 2, p127-175
Pagan, AoR. (1987) "Three Econometric Methodologies: A CriticalAppraisal", Journal of Economic Survey_~, Vol i, No i, p 3-24.
Pagan, A.R., and A.D. Hall (1983) ~’Diagnostic Tests as ResidualAnalysis~’, Econometric Reviews, 159-218.
Petty, Sir W. (1963) The Economic Writin~l~ of Sir William Pettv,ed by C.H. Hull (1899), two vols. Cambridge: printed by J. &CoFoClay at the university press.
44
Piggott, J~ (1980) A General E~uilibrium Analysis of AustralianTaxation Polic~, PhD Thesis, University of London.
Piggott, Jo, and Jo Whalley (1977) "General EquilibriumInvestigations of UoKo Tax-Subsidy Policy", in M.J. Artis andAoRo Nobay (eds) Studies in Modern Economic Anal sy_~i~, Oxford:Black~ell.
Piggott, Jo and Jo Whalley (1985), editors. New Developments inA_p_plied Equilibrium Anal sy_~, Cambridge University Press.
Quesnay, Fo (1972) Quesna~’s Tableau Economi ugg_~, edited with newmaterial, translations and notes by Kuczynski, M., and RoL.Meek, London: Macmillan.
Ricardo, Do (195!) The Works and Correspondence of DavidRicardo, edited by Po Sraffa, in nine volumes, Cambridge:Cambridge University Press for the Royal Economic Society°
Roncaglia, Ao (1985) PETTY: The Oriqins of Political Economz,Armonk, NY: oEoSharpe, Inco
Sargent, R.Do (1976) "A Classical Macroeconometric Model of theUnited States", Journal of Political Economy, 84, pp207-38.
Sargent, ROD., and No Wallace, (1975) "Rational Expectations,the Optimal Monetary Instrument and the Optimal Money SupplyRule’~, Journal of Political Economy, 83, p241-54.
Sargent, RoDo(1982) "Beyond Demand and Supply Curves inMacroeconomics", American Economic Review, May, 72, p382-89.
Scarf, H~Eo (1967) "On the Computation of Equilibrium Prices",in WoMo Fellner et al, eds, Ten Essay_~ in Honour of IrvinqFisher, New York: Wiley, p207-230o
Scarf, HoEo (1973) (with the collaboration of T.Hansen) TheComputation of Economic Eguilibrium, New Haven: YaleUniversity Press.
Schumpeter, Jo (1939), Business ~, New York: McGraw-Hill
Schumpeter, J. (1954), The Historz of Economic Analysis,posthumously edited from manuscript by EoB.Schumpeter, OxfordUniversity Press°
Sheridan, Ao (1980) Michel Foucault: The Will to Truth, London:Tavistocko
Simes, RoMo (1988), "An Outline of the NIF88 MOdel of theAustralian Macroeconomy", mimeo for the conference: "TheAustralian Macro-economy and the NIF88 Model" March 1988.
Simes, RoMo, and Po Horn (1988) "Design of the NIF88 Model",Australian Economic Pa~, July 1988, 155-170.
45
Sims, C. (1986) "Are Forecasting Models Usable for PolicyAnalysis?" Quarterly Review i0 winter: 2-16. Minneapolis,Minnesota: Federal Reserve Bank of Minneapolis.
Sims, C. (1988) "Identifying Policy Effects". In E__~piricalMacroeconomics for Interdependent Economies, vol i, edited byBryant,R.C. et al, The Brookings Institution, 1988.
Smith, Adam (1776) The Wealth of Nations, Eo Cannan edition1904, Skinner edition in Penguin 1970 (not the full work), andCampbell, Skinner and Todd edition (1976) o
Taplin, G.B. (1973) "A Model of World TRade" In TheInternational Linkaqe of National Economic Models, edited byR.J.BalI, 177-223. Amsterdam: North-Holland.
Tinbergen, J. (1939) Business Cycles in the United States ofAmerica, 1919-1932, Part II of Statistical Testinq of BusinessCycle Theories, Geneva: League of Nations.
Tinbergen, J. (1951) Business Cycles in the United Kingdom~1870-1914, Amsterdam: North-Holland.
Treasury (1981), The NIF-10 Model of the Australian Economy,AGPS, Canberra.
Treasury (1984), Proceedings of the Conference on the NIF-10Model, AGPS, Canberra.
Trevor, R.G. and SoJo Thorp (1988) "VAR Forecasting Models ofthe Australian Economy: A Preliminary Analysis", AustralianEconomic Papers, vol 27, June, pi08-120.
Yoshitomi, Masaru, et al. (1984) "EPA World Economic Model", voli, "Model Properties and Multiplier Analysis"; vol 2, "ModelStructure"; vol 3, "Multiplier Analysis (Detailed)". EPA WorldEconometric Model Discussion Paper 16. Tokyo:EPA,ERIo
Waelbroeck,JoL.,(1976) edo The Models of Project LINK,Amsterdam: North-Holland.
Wallis, K.F. (editor), PoG. Fisher, J.A. Longbottom, D.S. Turnerand J.Do Whitley (1984,1985,1986,1987), Models of the UKEconomy: First to Fourth Review of the ESRC MacroeconomicModelling Bureau, Oxford University Press°
Walras, Leon (1874-7) Elements d’economie politi ue~, istedition, (5th definitive edition, 1926), translated intoEnglish by W. Jaffe (1954)
Whalley, (1975) "A General Equilibrium Assessment of the 1973U.K. Tax Reform", Economica, 42, May, p139-61
WORKING PAPERS IN ECONOMETRICS AND APPLIED STATISTICS
The Prior Likelihood and Best Linear Unbiased Prediction in StochasticCoefficient Linear Models, Lung-Fei Lee and William Eo Griffiths,No. 1 - March 1979.
Stability Conditions in the Use of Fixed Requirement Approach to ManpowerPlanning Models. Howard E. Doran and Rozany Ro Deen, No. 2 - March1979.
A Note on A Bayesian Estimator Zn an Autocorrelated Error Model.William Griffiths and Dan Dao~ No. 3 - April 1979.
On R2-Statistics for the General Linear Model with Nonscalar CovarianceMatrix~ GoE. Battese and WoE. Griffiths, No. 4 - April 1979.
Construction of Cost-Of-Living Index Numbers - A Unified Approach°D.S. Prasada Rao, No. 5 - April 1979o
Omission of the Weighted First Observation in an Autocorrelated RegressionModel: A Discussion of Loss of Efficiency. Howard Eo Doran, No. 6 -June 1979.
Estimation of Household Expenditure Functions: An Application of a Classof Heteroscedastic Regresston Mode~s. George E o Battese andBruce P. Bonyhady, No. 7 - Septe~Joer 1979.
The Demand for Sawn Timber: An Application of the Diewert Cost Function.Howard Eo Doran and David F~ Williams~ No. 8 - September 1979.
A New System of Log-Change Index Numbers for Multilateral Comparisons.D.S. Prasada Rao, No. 9 - October 1980.
A Comparison of Purchasing Power Parity Between the Pound Sterling andthe Australian Dollar - 1979. W.F. Shepherd and D.S. Prasada Rao,Noo i0 - October 1980.
Using Time-Series and Cross-Section Data to Estimate a Production Functionwith Positive and Negative Marginal Risks, W.E. Griffiths andJ.R. Anderson, No. ii - December 1980o
A Lack-Of-Fit Test in the Presence of Heteroscedasticity. Howard E. Doranand Jan Kmenta, No. 12 - April 1981o
On the Relative Efficiency of Estimators Which Include the InitialObservations in the Estimation of Seemingly Unrelated Regressionswith First Order Autoregressive Disturbances° H.Eo Doran andW.E. Griffiths, No. 13 - June 1981.
An Analysis of the Linkages Between the Consumer Price Index and theAverage Minimum Weekly Wage Rate. Pauline Beesley, Noo 14 - July 1981.
An Error Components Model for Prediction of County Crop Areas Using Surveyand Satellite Data. George Eo Battese and Wayne A. Fuller, No. 15 -February 1982.
Networking or Transhipment? Optimisation Alternatives for Plant Lot;atDecisions~ H.Io Tort and P.A. Cassidy, No. 16 - February 1985.
{~ht~imm!.ic Tc,~I;s fo~, the Partial Adjustment and Adal,tioeModels. HoE. Doran, No. 17 - February 1985.
A F~rther Consideration of Causal Relationships Between Wages and Prices.
JoWoB. Guiss and PoA.Ao Beesley~ No. 18 - February 1985.
A Monte Carlo Evaluation of the Power of Some Test& For Heteroscedasticity.
WoE. Griffiths and Ko Surekha, No. 19 - August 1985.
A Walrasian Exchange Equilibrium Interpretation of the Geary-KhamisInternational Prices. DoS. Prasada Rao, No. 20 - October 1985.
Using D~rbin~s h-Test to Validate the Partial-Adjustment Model.H~E~ Doran, No. 21 - November 1985.
An Investigation into the Small Sample Properties of Covariance Matrixand Pre-Test Estimators for the Probit Model. William ~. Gr±~fiths,R. Carter Hill and Peter J. Pope, No. 22 - November 1985.
A Bayesian Framework for Optimal Input Allocation with an UncertainStochastic Production Function. William Eo Griffiths, No. 23 -February 1986,
A Frontier Production Function for Panel Data: With Application to theAustralian Dairy Industry. ToJ. Coelli and G.E. Battese, No. 24 -February 1986.
Identification and Estimation of Elasticities of Substitution for Firm-Level Production Functions Using Aggregative Data. George E. Batteseand Sohail J~ Malik~ No. 25 - April 1986.
Estimation of Elasticities of Substitution for CES Production FunctionsUsing Aggregative Data on Selected Manufacturing Industries in Pakistan.George Eo Battese and Sohail J. Malik~ No.26 - April 1986o
Estimation of Elasticities of Substitution for CES and VES ProductionFunctions Using Firm-Level Data for Food-Processing Industries inPakistan. George E. Battese and Sohail J. Malik, No.27 - May 1986.
On the Prediction of Technical Efficiencies~ Given the Specifications of aGeneralized Frontier Production Function and Panel Data on Sample Firms.George Eo Battese, Noo28 - June 1986.
A General Equilibrium Approach to the Construction of Multilateral IndexNmmberso DoS. Prasada Rao and J. Salazar-Carrillo, No.29 - August1986o
Further Results on Interval Estimation in an AR(1) Error Model.W.E0 Griffiths and PoA. Beesle~ No.30 - August 1987.
Bayesian Econometrics and How to Get Rid of Those Wrong Signs.Griffiths, Noo31 - November 1987.
H.E. Doran,
William E.
Confidence Intervals for the Expected Average Marginal Products ofCobb-Douglas Factors With Applications of Estimating Shadow Pricesand Testing for Risk Aversion. Chris M. Alaouze, No. 32 -September, 1988.
Estimation of Frontier Production Functions and the Efficiencies ofIndian Farms Using Panel Data from ICRISAT’s Village Level Studies.G.E. Battese, T.Jo Coelli and ToC. Colby, No. 33 - January, 1989.
Estimation of Frontier Production Functions: A Guide to the ComputerProgram, FRONTIER. Tim J. Coelli, No. 34 - February, 1989.
An Introduction to Austratian Economy-Wide ModeZling. Colin Pc Hargreaves,No. 35 - February, 1989.