A Model of Technological Capacity

To Support Survivors of Nuclear Attack, ~~~~~~~~~~ R INGH' - - ' ...... .US-F-

- i ...2.1

RAC-TP-313

Summary

A MODEL OF TECPNOLOGICAL CAPACITY

TO SUPPORT SURVIVORS OF

NUCLEAR ATTACK

Bernard Sobin

September 1968

This report was prepared for the Officeof Civil Defense, under OCS Contract

DAHC20-67-C-0137OCD Work Unit 3534B

This report has been reviewed in the Office of CivilDefense and approved for publication. Approval doesnot signify that the contents necessarily reflect theviews and policies of the Office of Civl Defense.

RESEARCH ANALYSIS CORPORATIONMcLean, Virginia 22101

U.'.

SU MMARY

Problem

To describe an economic model that was developed for estimating theability of surviving productive capacities to support human survivors of directweapons effects after a major nuclear attack on the US. 1t covers the princi-ples on which the model is based and enough information on the origins of thenumerical estimates to give a reade-" some idea of the value rf its use.

The model is limited to technological possibilities without regard towhether the economy would actually be managed well encugh to realize suchphysical potential. The economy analyzed by the model is as of some date.-fter decontamination and completion of any minor repairs that can restoreservice of szmic daniaged capaciti, s.

Background

Development of the model was part of the work under a contract to in-vestigate the economic-viabilit) imolications of light population casualties butheavy damage to nonhuman productive resources. Results of applications ofthe model to specific damage assumptions appear in classified documents.

Discussion

Details of the model struct.,,re are based on the assumption that a marginbetween the number of pe-ple actu-1I1y needing support by the economy and thenumber that potentially could be supported indefinitely is a good index of via-bility of the economy. A larger favorable margin suppor's greater confidencethat the population will iurvive despite either errors in the model or ineffi-ciency of the postattack system for managing the economy. A larger marginalso implies immediate availability of more slack 'n the economy for applica-tion to growth-producing investments.

The model allows for no relief of bottlenecks by imports. To the extentthat imports are available, particularly without having to be paid for 1mme-diately by exports, the survival problem will be easier than indicated by themodel.

The model is a linear-programming analysis with activities that consistof alternative production processes, support of population, and support of

governmental activities. The amount of the last is a fixed charge on the econ-omy. The model chooses combinations of 'evels of the various productionprocesses that maximize the number of people who can be supported withoutviolation of any of the capacity constraints or other feasibility conditions ofthe model. The capacity constraints are capacities surviving an attack (afterrepair of light damage). In principle the model could be modified to includemilitary and investment activities, and the objective of maximizing the numberof people that can be supported could be modified.

Although a ,number of dynamic elements could be introduced to the model,i is important to note that the model is presently static. It is static in thesense that the calculated rates of delivery of end products of the economy togoverriments and households are rates that can be maintained for an indefiniteperiod, The rates of delivery do not depend on depletion of inventories ofeither end products, materials, or goods-in-process. The absence of invest-nient activities as the model is now formulated also implies that the model isstatic in the sense that there is no provisi, for changes in the capacity of theeconomy to make deLliveries of end products. As noted earlier, a good marginbetween requirements and potental for survival production implies the exist-ence of slack for investment in growth of the cconomy'b ability to deliver endproducts.

It is conceivable that, after an actual attack, investment would be neededvery soon to halt depletion of inventories of commodities crucial to survival.Here the investment would be to preserve an existing rate of delivery of endproducts rather than to expand the rate. The main part of this paper includessome discussion of how the present mode! might be modified to include someprovision fui' investment activities needed to maintain feasibility of a givenrate of delivery of end products.

This report has been reviewed in the Office of

Civil Defense and appvoved for puoDicotion. Ap-

proval does not signify that the contents nec$-

ECONOMICS AND COSTING DEP RTMENT sorily reflect the views and policies of the Officeof Civil Defense.

TECHNICAL PAPER RAC-TP-313 Contract DANC20-67-C-0137

Published September 1968 OCD Work Unit 3534B

A Model of Technological Capacity

To Support Survivors of Nuclear Attack

by

Bernard Sobin

DISTRIBUTION STATEMENT

This document has been approved for public

release and sale; its distributlon Is unlimited.

RESEARCH ANALYSIS CORPORATION

McLEAN, VIRGINIA

FOREWORD

This is one of two RAC papers prepared as part of a study of the capa-bility of the US economy to continue to support the immediate survivors of anuclear attack in which the damage to industrial capacity is large in relationto the population lost to direct weapon effects. The other paper has beenissued as Bernard Sobin and David F. Gates, "Economic Implications of HighPopulation and Low Property Survival in Nuclear Attack on the United States(U)," SErRET, RAC-TP-317, August 1968.

The present paper describes a r.;odel that was used as one of the ingotsto thp elassified paper. The model is designed to calculate the capability ofthe economy to support survivors after hypothesized attacks, and it can dealwith attacks other than the ones hypothesized for the classified paper.

The model is a static one, with no provision for postattack investmentto relieve any capacity bottlenecks that remain after decontaminaton and re-pair of light damage. A later publication will describe an augmentation ofthis model to include a period of investment before the steady state.

Arnold ProschonHead, Economics and Costing Department

iii

L

9 CONTENTS

Foreword

Summary1iProijaem-Background -Discuss iuti

Introduction 5

General Characteristics of the Model 6

Activities-Balance Constraints

Derivation of Model Inputs 13

Overview of Data Requirements-Food -Production Submodel-Industrial -Technology, Submodel -Per Capita Requirements -Governm, -it Requirement s-Co. stant Terms

Uses of the Model 3

Potential Improvements in the Model 3

Appendix

A. Teceinology Matrix kir Linear -Programming Computations 39

References 6 5

Tables

1. General Form of the Mtodel2. 1958 Er'TAndiure I~emi and Factors for Conxersion to

Postattack Con(Ntions 2 93. Nonfood Consumption Expenditures by Industry 3l4

v

Problem

To describe an economic model that was developed for estimating theability of surviving productive capacities to support human survivors of directweapons effects after a major nuclear attack on the US. It covers the princi-ples on which the model is based and enough information on the origins of thenumerical estimates to give a reader some idea of the value of its use.

The model is l iitec to technological possibilities without regTar6'whether t1he economy would actually be managed well enough to realize suchphysical potential. The economy anlyzed by the model is as of some dateafter decontamination and completion of any minor repairs that can restoreservice of some damaged capacities,

Background

Development of the miodel was part of the work under a contract to in-vestigate th.. economic- viabili'v inmplications A light population, casualties butheavy damage to nonhumian productive resources. Results of applications ofthe miodel to specific damage assunipf ns appear in classified documents.

Discussion

lXetai Is of the miode I structure are based on the a ssu mpt ion that a nmargi,between the rinmbr of people ac tua lly needing suoport bv the ecoinmy alI.I thenumber that potent ially could bc supported indefinitely is a g~o ine ovia-bilitv of the ovonoiY. A larger fa vorabit margzin suppo rts greateir con! idenctthat the txpulation will survive despite either u'r rors in the miode I or ne ffi -c iencV oif the postattack svstemi for managing th, vconm niv. A IL rk:er m11.1-4i1also impnlies innied iate av~ai labilIity of mo re slack in the evononlyv for appli1 ca -tion to growth-prodiucing icie st me tits.

The aiode I allows for no rc Iief of bott lent ck. b%. i lilports. To~ the ex~tentthit imports are ava ilable,. particular lv without h'Aving to be paid for innicdiAtely lv exports,. the sar vi cAl probit in will be easier that indi cated by the

The rrodel is a linear progrArniinv analv--s w ith activities that consistof altrrnative production processes. supptirt of population, and support iof

governmental activities. The amount of the last is a fixed charge on the econ-omy. The model chooses combinations of levels of the various productionprocesses that maximize the number of people who can be supported withoutviolation of any of the capacity constraints or other feasibility conditions ofthe model. The capacity constraints are capacities surviving an attack (afterrepair of light damage). In principle the model could be modified to includemilitary and investment activities, and the objective of maximizing the numberof people that can be supported could be modified.

Although a number of dynamic elements could be introduced to the model,it is important to note that the model is presently static. It is static in thesense that the calculated rates of delivery ot end products of the economy togovernments and households are rates that can be maintained for an indefiniteperiod. The rates of delivery do not depend on depletion of inventories ofeither end products, materials, or goods-in-press. The absence of invest-ment activities as the model is now formulated also implies that the model isstatic in the sense that there is no provision for changes in the capacity of theeconomy to make deliverieF of end products. As noted earlier, a good marginbetween requirements and potential for survival production implies the exist-ence of slack for investmen!t in growth of the economy's ability to deliver endproducts.

It is conceivable that, after an actual attack, investment would be neededvery soon to halt depletion of inventories of commodities crucial to survival.Here the investment would be to preserve an existing rate of delivery ,f endproducts rather than to expand the rate. rhe main part of this paper includessome discussion of how the present model might be modified to include someprovision for investmnvnt activities needed to maintain feasibility of a givenrate of delivery of end products.

2

A Model of Technological Capacity

To Support Survivors of Nuclear Attack

I-

INTRODUCTION

Developrnnt of the model described here was part of a project to deter-mine whether civil deiense measures directed at saving lives but nct propertyfrom direct weapons effects have am, significant ikelihood of being defeatedultimately by inability of surviving production capacity to support all the sur-viving popul-tion. Another RAC paper' includes an application of the model toa particular nuclear attack.

The basic problem can be subdivided according to a number of orinciplesof classification. One possible subdivision is by time period dealt with. Othersubdivisions can b physical problems when physical resources are usedin optimal fashion or management problems (monetary and liscal policy, pricecontrols, matcriai controls, etc) for optimizing the use of resources. Themodel described here deals with phys'cal capabilities (with management as-sumed perfect) of resources that exist after a period of ,Aezr'ng rubble, de-contamination, and completion of repairs that do nc. cause significant drain onproductive resources. The model is used to determine whether the fixed plantand equipment plus land and labor resources are then sufficient, without furtheraugmentation by inv,?stment or iy imports in excess of normal, to provide forminimum levels of governmental activity and indeiinite support of the popula-tion in good health.

The model goes into gTeat detail on food production and supporting activi-ties as central to survival. The detail is in tiie form of highly disaggregatedfood-proauction activities and provision of flexibility in mixes of lood-produc.,tion activities to meet survival requirem nts. A distinguishing feature of themodel is concentration o" detaii in Ihe sectors considered mc,'-t impol'Tant forthe principal model appiicati3n. More detail in other sectors would te neededfor useful application to otner kinds of problems.

i5

GENERAL CHARACTERISTICS OF THE MODEL

The model is an application of linear-programming mathcniatics. Theproblem is to find a combination of levels of available types of activities thatmaximizes a linear function of those activities subject to a set of linear in-equalities in which the activity levels are the variables.*

Mathematically the problem may be stated as

maximize

subject to, \h,

\-//

where X is a variable level of the ith activity rate (' , the given contributionto the objective Z per unit of the ith activity rate; a, the amount of the ithitem used or made available per unit of the ith activity. The model has 149activity rates (annual) into which all resource-using activity for survival hasbeen divided, and there are 82 coastraining rows.

The a,, are, by convention, positive for inputs to activities and negative

for outputs. The expression Z a, J for an i representing goods and services

currently produced may be considered an excess ot c'urrent input requi-_nmentsover current output. When this sum equals zero, requirements equal supply,and the usual row constrains the sum to be less than or equal to a b, of zero.Occasionally there is a row with a constraint that the sum be greater than orequal to a b, of zero. Such a row occurs, for example, in certain nutritioncases (to be discussed later) where a constraint is needed to make sure thatwhat can healthfully be consumed be at least as much as is supplied to the diet.

For an i representing fixed capacities, the expression al I\J consists entirely

of requirements, and this sum is constrained to be less than or equal to a b,that represents postattack capacity. Finally, in a few cases, b is preceded by

*A theorcm in linear programming ztates that if there is more than one combina-tion that meets the requirement there are infinitely many.

6

an equality sign to represent a stipulated level of one activity or a stipulatedlevel of use of a fixed capacity.

Appendix A has the entire matrix of a,, . Table 1 classifies e matrixelements by column (activity), row (constraint), and sign (noted earlier as in-dicating outputs if negative anti inputs if positive). The following commentsdescribe the matrix in terms of both activities and constraints.

ACTIVITIES

Persons Supoorted

Persons supported is a class consisting of only one activity, the numberof millions of people supported each year. It is the only variable with a non-zero (,. This is convenient becausc the absence of other nonzero terms of theobjective function implies that the selection of activities that rrmaximize the ob-jective is identical for any positive maimitude that may be assigned to the valueof human liie,

Each million persons supported requires minimum quantities of caloriesand particular nutrients. There is also a maximum number of calories that iscompatible with health and a maximum weight of food that can be eaten. Theplus entries in the box for human nutrients represent a, describin, amountsof these minima and maxim. per million persons supported.

The industrial capacities indicated per million persons in the -personssupported" column are capacities that are required dirPctlv and indirectly tosupport delivery cf nonfood commodities that the aver age person is consideredto need for his support. For example, footwear requirements imply need forfootwear-production capacity directly plus leather tanning and finishing, elec-tric power, fuel, and many other kinds of industrial capacity ;ndirectly. Eachrow of the group "industrial capacities" has the sum of all direct and indirectdemands on that capacity for all nonfood requirements of one million people.

Pursuing the footwear illustration further, it may be noted that an indirectinput to footwear production is hides and skins. The production process thatproduces hides and skins happens to be defined in terms of the amount of meatproduced, and any capacity constraints appear in the model as constraints onmeat production. Regardless of how production processes and constraints aredefined, however-it would be just as correct to call hides and skins the princi-pal product aad meat the byproduct-the requirement for hides and skins permillion persons must be provided for.

The byproducts covered by the model are the hides and skins just men-tioned and cotton fiber. The latter is a byproduct of cottonseed (a food andanimal-feed product) production. Here the byproduct is, at least in peacetime,more valuable than the principal product. It illustrates the principle that thedetermination of which of joint products is the byproduct is arbitrary. Appli-cations of the present model are not affected by which of the joint products aredefined as the byproducts.

Labor requirements are similar in concept to industrial capacities andbyproducts. The production processes that directly or indirectly produce thenonfood requirements per million persons all use labor. The entry in a "labor"row is the sum of all these labor requirements.

7

0 Im

0 -

2i 0~u ~ C

o go C E

cu 0I CL u C

00 0 C

CL.

The foregoing listing of items in the "persons supported" column has bothdirect and indirect inputs to nonfood requirements but only direct inputs to foodrequirements. This reflects the fact that the indirect inputs to nonfood require-ments per million people are the same for all model applications, whereas theindirect inpiLs to food requirements depend on the solution mix of foods andways of producing them. The model does, of course, account for indirect in-puts to nutrition but does so as inputs to separate activities for each way ofproducing each food.

Production

Maximizing the number of persons that can be supported involves decisionsabout three classes of production variables: crops, livestock products, and in-dustrial inputs to agriculture. The decisions for the last group are really im-plied by those of the first two groups and could have been omitted from the listof decision variables. The way they were retained, however, avoids any over-determination of the model and has some conveniences in analysis of solutions.

The crop-production activities pr- Iuce both food and animal-feed crops.The negative entries in the human-nutrient and livestock-nutrient rows referto amounts of nuL,'ients yielded in later processing (including byproduct produL-tion) per billion pounds of each crop after any crop deductions for seed.

The plus entries in the next two classes of rows represent amounts oftypes of land ano industrial inputs to agriculture required per billion pounds ofeach crop.

A single row for soybean-grain mix has soybean-meal equivalents forbread baking per billion pounds of soybeans and of breadgrains. The sign isplus for the soybean crop and minus for the breadgrains. The reason for thisrow is discussed in the section "Balance Constraints," subsection "SpecialConstraints."

The industrial-capacity rows are industrial capacities required directlyas inputs to agriculture (fertilizer, pesticides, and petroleum products) andrequired both directly and indirectly as inputs to the processing and distribu-tion of agricultural products. The indirect inputs to agriculture are accountedfor in the columns for industrial inputs to agriculture. As suggested earlier,it would have been possible to consolidate the columns for industrial inputs toagriculture with the columns for the agricultural activities.

The negative entry in the byproducts section is production (negative re-quirement) of cotton fiber per billion pounds of cottonseed. The remaining e,-tries in the crops columns are self-expianatory.

The labor entries are for labor employed in agriculture per billion poundsof each crop plus labor required directly or indirectly to process and distributeeach billion pounds of crop.

The columns for livestock products (beef, milk, pork, lamb, poultry meat,ind eggs) can be explained in much the same way as the columns for crops.The negative requirement in the byproduct section is hides and skins yieldedby beef production.

The industrial inputs to agriculture are columns each of which refers toa row of the group with the same class name. Each column has a negative en-try in the corresponding row giving the aniuwil of the inaustrial output (nega-tive requirement) in units used for the row per unit used to measure the levelof the column production activity.

9

The entries for industrial capacity and labor rows are amounts of eachindustrial capacity or unit of labor that is required directly or indirectly foreach unit of delivery to agriculture. In the case of a capacity required directly(e.g., petroleum-refining capacity per unit of petroleum-product deliveries toagriculture), the entry is unity for the d4"ect requirement plus any feedback ofadditional capacity of the same type required indirectly (e.g., additional petro-leum-refining capacity required indirectly to support output of the industriesthat supply inputs to petroleum refining). The government-operations class ofactivity has only one column in the present model although a breakdown by typeof operation would be possible. The industrial capacity and b product erariesare direct and indirect requirements to support sales of goods and services toFederal, state, and local governmems. The labor row consolidates the laborfor purchased goods and services with labor employed directly.

The entry in the row for level of government is a factor to convert unitsof the column activity to units of the row constraint. The model happens to de-fine both units identically; hence the scaling factor is unity.

BALANCE CONSTRAINTS

The matrix elements whose signs appear in Table 1 have already beenstated to be coefficients of variables in a linear inequality or equation. Thetypes of constraining inequalities and equations are discussed in the followingparagraphs.

Human Nutrients

A human-nutrient row for a lower bound per rIlion persons states thatthe requirements of the nutrient to support surviving persons, minus thenutrients produced by crops, minus the nutrients produced by livestock, mustbe less than or equal to zero.

A row for an upper bound states that the maximum consumption for allsurviiors must be no less than the quantities yielded by crops and livestock,i.e., production must not exceed what can healthfully be eaten. There are upperbounds for calories, fats, and weight of ration.

This second kind of balance constraint would not be necessary if foodweight, calories, and nutrients 1.ere not yielded in fixed proportions by cropsand livestock products. An optimal production program would not produce moreof any item than could healthfully be consumed. Constraints may be needed,howeve-7, to prevent the satisfaction of requirements of some nutrients by typesof food with such small percentages of those nutrients and such large percent-ages of calories, fats, or weight as to make the intake of food excessive n oneor more of calories, fats, and weight.*

Livestock Nutrients

In the cases of ILwer bounds, a typical livestock-nutrient balance con-straint states that what is required to produce the solution quantities of live-

*At present it appears that a typical situation is one in which diets that maximizethe number of persons supported are well below maximum per capita levels of calories,fats, and weight. The constraints on maximum production have therefor, not been neededyet.

10

stock products minus what is yielded by crops (including pasture and byproductsof human-food crops) must be less than or equal to zero. That is, requirementsmust be less than or equal to supply. A row for upper bounds (on weight of ra-tion) states that the maximum consumption of nutrients in production of thesolution quantities of livestock products mvrt be no less than the quantitiesyielded by the cru, s.

Land Classes

The typical land constraint deals with the maximum acreage that can bedevoted to a crup or class of crops. Thus there are maximum acreages foreach crop, for grains as a class, for all crops other than pasture, and for allcrops including pasture. Each such constraint for a class of land states thatthe combined requirements for all crops must be less than or equal to theamount of the class of land available. Convertibility of land to alternative usesis allowed for by the rough device of having the sum of subclass maxima of aclass exceed the class maximum. For example, the sum of maximum acreagesfor each grain exceeds the grainland maximum, which implies that maximumuse of grainland for some grains is not consistent with maximum use of grain-land for other grains.*

In addition to the maximum acreages, there are three minima. The amountof land devoted to fruits and vegetables must be at least as great as in recentpeacetime years. The principal reason is that the iiodel omits specific account-ing for vitamin C and other nutrients for which fruits and vegetables might beimportant. In addition the small amount of land invo lved would not add much tothe output of crops higher in calorie and protein yields, and in the case of fruitsthe conversion to otl:,r production would mean loss of trees well beyond theend of the food emergency. The third minimum acreage constraint relates towheat. At least one-half the land suitable for wheat must be devoted to wheat.At least that much wheatland is considered not usable for other crops with nor-mal inputs and yields.

Industrial Inputs to Agri 'ilture

The model accounts fur three industrial inputs to agriculture: fertilizers.pesticides, and petroleum products. A row for each states that the agregaterequirements for crop production minus the amount produced by industry mustbe no greater than zero.

Special Coritraints

The constraint with respect to mix of soybeans and grains is designed tolimit the proportion of soybean meal in bread flours and as a meat extealder.Soybean meal is far superior to wheat flour, rye flour, and cornmeal as asource of protein and is also superior to meat, but there are limits to the per-centage of soybean meal that can be included in bread acceptable to consumersor that can be used to 'adulterate- n-wat. The existence of any such limitsunder emergency conditions is perhaps debatable, but the model provides forsuch limits well above what is accepted now.

The correct way in principle to handt land concrtibilitV is to cla,.sify land byproductivity in alternative use and to have a separati alto rnative way of prolucing eachcrop for each class of land that niav be used. This may tx, attempt( in :1 lat'r v%,rsionof the model,

11

The way the constraint works is as follows. Soybean production has threekinds of food columns: one in which the soybean meal byproduct of oil produc-tion is fed to livestock; one in which the meal is used for human food, primarilyas a substitute for breadgrains; and one in which the meal is used as a meatextender. The second and third types of soybean production have positive co-efficients in the special-constraint row. These positive coefficients are eQ',alto the quantities of meal yielded per billion pounds of soybean production. Eachof the breadgrain columns has a negative coefficient equai to that yield multi-plied by the maximum ratio of soybean meal for flour to total flour. Each ofthe meat columns has a negative coefficient equal to the product if the soybean-meal yield of soybean production and the maximum ratio oi proportion of ex-tender in total meat. The constraining inequality states that the total amountof soybean meal produced minus the maximum used in bread illours and meatmust not exceed zero.*

The constraint with respect to level of government activity states howmany units of the 1958 mix of direct and indirect requirements generated bygovernment operations are stipulated for the problem. The a,, here is unity,and the stipulation is a percentage of the 1958 level multiplied by 0.001 to con-vert millions of dollars of the activity vector to billions of dollars for theactivity variable. The scaling by 0.001 could also have been in the a,, ratherthan in the stipulation constraint. t

Industrial Capacities, Livestock Capacities, and Labor

Each industrial-capacity, livestock-capacity, or labor row states thattotal requirements by solution activities must not exceed the postattack capac-ity available. In each case the postattack capacity available is an output of adamage assessment. The industrial and livestock capacities are defined inunits of the outputs they support. The labor capacity unit is millions of workers.

Byproducts

Each of the two byproduct rows (hides and skins: cotton fiber) states thattotal requirements by using activities minus total supply provided by producingactivities must be nogreater than zero. This is in effecta stipulation of no stock-pile depletion. For a situation in which there is considered to be a largv stock-pile of a byproduct (e.g., survival of a large peacetime stock of cotton or salvageof large quantities of hides and skins from livestock dying or slaughtered soonafter the attack), the constant term may be any maximum inventory-depletionrate that is considered proper.

*This is a palatability constra!nt for which the Justiflcation iP. emer-pricy ig qr;-tionable. However, if the concessions to palatability are accepted. it ought to be vith aseparate constraining inequality for each use rather than with only 4 single aggregateconstraint. The procedure used so ar overstates capabilities slightly by permitting thecomputer to use the meat-extender activity (which requires no ultimate baking Inputs)for production of soybean meal both as a meat extender and as a bread-flor subotitute.

* Use of an activity vector with requirements measured in millions tnstead of bil-lons of 1958 dollars is a vestige 4f early-stage k'omputations with direct requirementain which use of billions of dollars would have required t,-o many decimal places for thecoaputer program.

12

DERIVATION OF MODEL INPUTS

Evaluation of a model may use one or both of two kinds of criteria: (a)theplausibility of answers yielded by the model and (b) the apparent soundness ofthe estir.;ates of inputs to the model. The purpose of this section is to providebases ior judgments on the second kind of criterion.

Full achievement of this purpose here would involve providing worg sheetsfor complete derivation of every number in the model. However, that wouldmake the paper ii-practically long and would add little to ach.eving the objec-tive beyond what ma , be yielded by more aggregated descriptions of the deriva-tions of most numbers. The latter level of detail of the derivation of estimatesis used here.

The discussion is in terms of the following major categories: (a) an over-view of data-requirements sources, (b) the food-production submodel, (c) theremaining industry submodel, (d) industrial capacities, (e) land capacities,(f) per capita consumption requirements, and (g) government-operation require-ments.

All the final estimates. except those for capacities, appear in App A. Theland capacities are in a text table, and the industrial capacities cannot be pre-sented because they vary with the: problem.

OVERVIEW OF DATA REQUIREMENTS

The sectural classification system of the model and the selection of sourcesare tailored iA-gelv to the intended use of the model in analysis of the capabil-ity of the economy to provide minimum subsistence to survivors of weaponseffects. Since physiological standards for nutrition are much less compressibl-than customary standards of other kinds of household collsumption, the mainemphasis f the model is on resources and drcisions ieded for nutrition, akey physiological requirement. (Survival standards are discusstd in more de-tail in the subsection *Nutrient Requirements.') The emphasis on nutritiontakes two forms. One is the use of considerably more dirtail for the agricul-tural sectors than appears in any pxblished model of the U.' econonv Theother is the use of altcrnative input-output combinations for each agriculturaicrop and infinitely many alternative human diets, with the mathematlcal procc-dure selccting the cumbinations that make the most effctIvU Ut.e of available;'esources. Outside the food-production area the momc' .- rather highly aggr,-gated, and there are no process alternatives. The mode! could probably beimproved by disaggregation and provision for alternative prouesse outside the

13

i

food area, but it is felt that there is a reasonable balance at the margin between

efforts at accuracy in food and nonfood areas.

FOOD-PRODUCTION SUBMODE L

The food-production submodel consi.ts of the columns for food-production

activities and rows for nutrients, land and industrial inputs to agriculture, the

soybean- grain- mix constraint, and part of each labor input to food production.

The food-production activities cover production and subsequent processing

of the following crops and livestock products: wheat, corn, oats, rice, rye,peanuts, cottonseed, soybeans, edible dry beans, fruits, potatoes, sweet potatoes,

beet sugar, cane sugar, other vegetables, grain sorghum. barley, hay, pasture,

beef, pork and lard, lamb, poultry, eggs, and milk. Omitted are a number of

minor food and feed crops, fish, and exotic foods. The fish are omitted because

of difficulty of measurement of a postattack fishing and fish-preservation capac-

ity that would probably be too small to affect the diet significantly.*

Each of the crops has at least four different aLtivity columns. The four

columns correspond to production with normal fertilizer and pesticide inputs,

without lertilizer, u 'hout pesticides. -mnd with neither fertilizer nor pesticides.

In addition there are further columns in some cases for alternative ways of

processing crops. For example, corn may provide nutrients exclusively for

livestock or mainly for humans with byproducts for livestock. When corn is

intended for livestock, a distinction is made between the activities of pr,'ducing

for poultry and for other livestock because the digestible-nutrient con'ents are

not the same for the two classes. Appendix A has a cumplete list of activities.

Human Nutrients

The hun in-nutrient rows deal with minimai for calories, proteins, fats.iron, and calcium, and they deal with maxima for calories, fats, and weight of

ration. The ritionalv for this list is discussed below in the discussion of percapita requirements.

All the foo-processing activities are measured in billions of pounds of

crop or livesttok products. The human-nutrient yield is estimated through twofactors, the weight of edible food per unit of the activity and the amount of each

nutrient per unit of foAx (with calories considered as a nutrient). Factors for

converting crop weights to food weights are readily available from an intro-ductory table of Agricultural Statistics- and from stch food engintering texts

as Parker, Harvey. and Statler. rhv, nutrient contents of foods ar, fronm Watt

and Merrill.' A third factor that might be considered is the percentage of

nutritive value of each crop or livestock product that is lost before reaching a

human stomach. In the absence of data with respect to individual crops, the

midel, as noted elsewhere, applies a unifornm 15 percent margin on per capita

nutrient requirements.

*IrI 1!1' fish at :owied Ior oi 7 1'r ccnt of !hr tota! I > ,or U1?jA1W t~ , Ii A.1weijht Of tr.eat, lault!% aiA fir h.4 arml ,(if ;t~: ) rts, And fi. .l,g %cs,,rt-l xhould

reduce thr t al icit%,

14

It may be noted that the nutrient value of wheat is based on conversion towhole-grain flour. There is now a substantial use of wheat in this way to makewhole wheat bread: but by far the bulk of wheat is now used to produce a morehighly refined bread flour and a rather high-protein byproduct for animal feed.On the assumption that protein for human food would be scarcer than proteinfor livestock feeds in the postattack situation, only the processing to wholewheat flour was provided for in the model despite the general consumer prefer-ence for bread made with more highly refined flours. It is not clear, however,that the assumption of greater shortage of human-food protein is correct forall damage patterns: it may therefore be worthwhile to add an activity for thepeacetime flour-extractioil rate as an alternative to manufacture of whole-grainflour.

The model similarly increases the nutritive value of rice for humans atthe expense of animal-feed byproducts by omitting processing stages that nor-mally convert parts of the grain with high-protein values to animal feeds.

Livestock Nutrients

The livestock nutrient rows are classified by type of livestock and by typeof nutrient. The classification by type of livestock serves to assist in distin-guishir,6 among the nutritive values of identical feeds for different livestockclasses. The classification by nutrient helps constrain the livestock.

For any one of the two kinds of livestock nutrients, there are four rowsin which entries for nutrient outputs of a crop may appear: total, excludingpoultry- cattle and sheep; hogs: and poultry. A crop intended to be fed to live-stock other than poultry has an entry for one or both of cattle and sheep and/orhogs. There is an identical entry for both if the feed is suitable for both classesof livestock (e.g., corn is suitable for both, but hay is suitable only for cattleand sheep). Having a unit of activity recorded as supplying a unit of output toeach of two classes of animals leaves open the possibility, even though only oneunit of nutrient is produced, of a total animal feeding that exceeds production.However, the first row, for total excluding poultry, balances the total nutrientinput requirements of all livestock products other than poultry against the once-counted production by all crop activities. This adds an upper bound on totalnutrient availability to the overlapping (and therefore double-counted) avail-abilities for individual classes of livestock. Outputs of a poultry nutrient havenot only a separate row (the fourth), but also separate crop-production activi-ties. In the case of a type of feed that can actually be fed either to poultry orto other livestock (e.g., corn), the model defines two distinct producing activi-ties even though the products are identical in appearance and input requirements.The reason for the separate treatment of poultry-feed-producing activities isthat the amount of digestible nutrient per unit of some feeds when fed to poultryis too different from the amount of digestible nutrient of the same feeds whenfed to other livestock. It would not be possible to have an unambiguous nutritivevalue of a feed without specification as to whether the feed would be fed to poul-try or to other livestock.

The model has two classes of livestock nutrient. The first class is a USDepartment of Agriculture measure of nutrient quantity called "feed units." Afeed unit is the weight of feed that is equivalent in nutritive value to a s-mdardquantity of corn grain when fed to a particular class of livestock for ma-.ginal

15

changes in a balanced ration. The unit is not based on any chemical analysisof the content of the feed but rather on experimental substitutions: hence nothingrigorous can be said about the content of a feed unit in proteins, fats, etc. Thedefinition in terms of marginal changes in a balanced diet, however, suggeststhat the feed-unit value of a pound of feed is primarily a matter of the energyvalue of the digestible nutrients, which may be measured gross or, sometimes,net of energy consumed in eating and digesting the feed.

Since the model is concerned with formulation of complete rations forlivestock, not just with marginal changes in a balanced ration, it is necessaryto consider requirements other than for food units. The model assumes, how-ever, that accounting for only the single most important of such other require-ments, proteins, is necessary for reasonable assurance of approximate balancewhen common types of feed are used. Of course, to the extent that the assump-tion is wrong, the model tends to overstate the amount of flexibility in the ration.

The feed-unit and protein values per unit of feed weight for the modelcome generally from a Department of Agriculture publication by Hodges. 5 In afew cases, the ratios are weighted (by production) averages of more detaileddata in the source (e.g., the model average for hay instead of the source's dataon individual types of hay). There was also some supplementing of the Hodgesdata with data from appendix tables and text of Morrison.6

The weights of byproduct feeds produced in processing of human-foodcrops come from the same types of food engineering texts used to estimate thefood portion.

Minimum requirements of feed units per unit of livestock produced areaverage ratios for 1960-1965 in which the numerators are feed units consumedas reported by the Department of Agriculture in "Statistical Bulletin 337." anddenominators are livestock products produced as reported in Agricultural Sta-tistics.

2

Minimuir irequirements of protein are based on leed input and productoutput experience for 1959 only. The denominator of livestock product outputis from the same source as the denominator for feed-unit requirements; butthe numerator is calculated from the Hodges ' data cited in connection with di-gestible-protein values of feed classes and from data of the same source onthe quantities of each kind of feed estimated to have been consumed by eachclass of livestock in 1959. The cumulative product of percentage of proteinano amount of each kind of feed represented an aggregate protein consumption.

Inadequate data on green pasture presented special dfficulties both forestimating nutritive value per unit of weight consumed and for calculating thepasture )ortions of the estimates of nutrient inputs per unit of livestock product.rhe Hodges data do not include feed-unit values or protein values per unit ofpasture, but they do include a table (Table 22) that compares the total digestible-nutrient percentages of individual grasses and hays. The model uses --n averageof these ratios (0.30) to convert feed-unit and protein values of hay to corre-sponding values for grass. The weight of grass consumed in 1959 by each classof livestock is derived from the estimate of feed-unit value per weight unit ofpasture grass and estimates in the previously cited "Statistical Bulletin 337"of the number of feed units each class of livestock must have obtained frompasture in addition to the directly measured feeds to produce the livestockproducts listed. The fepd units from pasture were estimated by the Department

16

of Agriculture as a residual because there is no direct way of measuring whatan animal consuaies in grazing. The limal estimates of the model are subjectto errors in the Department of Agriculture procedure for estimating pasturefeed units as a difference between tIwo roughly estimated quantities and to sam-pling errors in arriving at the ratio between~ nutritive values of hay and pasture.Actual pastures may include grasses and weeds that are inferior to the grassesused in the sample.

Land Inputs to Agriculture

Land requirements per unit of output in the real world vary continuouslyas a function of quantities of fertilizers and pesticides of various kinds, thequality of seed, the quality of land, the quality of management, and the amountof effort expended in preparation. of the land and rare of the crop. The model'sapproximation to reality uses average quantities and qualities of all these inputsfor one possibility with respect to each crop, and then adds three other discretepossibilities: normal inputs of all kinds except for complete absence of fertil-izers, normal inputs of all kinds except for complete absence of pesticides.and normal inputs of all kinds except for complete absence oi both fertili7ersand pesticides. Absence of one or both of fertilizers and pesticides impliesan increased requirement of land per unit of output.

Although the activities of the model are restricted to those with full orzero avail,, -ility of one or both of fertilizers and pesticides, the solution to anyproblem using the model can approximate the effects of partial availability byhaving an appropriate mix of activities with both zero and full availability offertilizer and pesticides. For example, availability of sufficient fertilizer fora 20-percent-of-normal application to the land for a crop is treated as full avail-ability on an activity using 20 percent of the land and zero availability on theactivity using the other 80 percent of the land devoted to a crop. This approxi-mation to reality is wrong to the extent that the true relation between land pro-ductivity and inputs of fertilizers and pesticides departs from linearity. Thereare also errors in the assumption that there is no need to distinguish amongdifferent kinds of fertilizers and different kinds of pesticides.

The land requirements per unit of average current application of fertilizerand pesticide come from 1963 experience as reported in appropriate tablesof Agricltra Statistics. The iequirements in the absence of fertili7zer comealmost entirely from a Department of Agriculture publication by Thach andLinberg. 8 They studied differences in yield per acre for farms using and notusing fertilizers. Their data on the farms not using fertilizers in 1954 areassumed for the present model to be applicable to current farming where thereare normal inputs of all other kinds.

This undoubtedly overestimates the effectiveness of fertilizers in increas-ing yields per acre. It is quite likely that the 1954 farmers who used no fertil-izer were also less- than- average users of pesticides and in general less pro-gressive in their farming techniques. In addition, some of the increase inyields between 1954 and 1963 is attributable to factors other than increasedapplication of fertilizers. Furthermore, it would take more than 1 year ofabsence of normal fertilizer use to geL the full loss of productivity; although,in the other direction, it would take more than 1 year of iormal fprtili~er useto restore any productivity lost in a prolonged period of absence of normal

17

/LL

applications of fertilizer. There are other, more random errors in addition tothese biases in favor ot overe iin.,a5 , i, I e importance of fertilizer, and furtherresearch on the influence of fertilizers would be worthwhile in any further workon the model.

The data on effects of dispensing with pesticides are even more sketchy.A Stanford Research Institute study by Moll, Cline, and Marr" has some lightlydocumented estimates of effects of pesticide absence in the first year after anattack for wheat, potatoes, sugar beets, corn, and alfalfa. Presumably the ef-fects could grow after the first year. For the present model, therefore, theStanford Research Institute estimates of crop yields without pesticides are re-duced slightly. The wheat and corn estimates are also applied to other grains,and a yield of 85 percent is used. Sweet-potato yields are made the same asyields of potatoes, with an estimate of 70 percent, and sugarca is combinedwith sugar beets for a yield of 85 percent. On the basis of scattered qualitativeinformation, it was judged that fruits and vegetables depend on pesticides muchmore than the foregoing crops, and a yield of 60 percent was arbitrarily as-signed. In the absence of any information abcut peanuts, cottonseed, and soy-beans, they were assigned intermediate yields at 75 percent of normal.

Following the precedent of the Stanford Research Institute study, the caseof absence of both fertilizers and pesticidep i- as~igni - a perc-ntage-of-1,nimayield that i-; the product of tne pereentages fc. ferili, . aa pesticide ab n acesseparately. This is based on the plausible but unproved assumption that nsects,rodents, and weeds take the same percentage of a crop regardless of the amountof fertilizer used. Another plausible assumption might be that they take thesame absolute amount, in which case the aggregat" loss from absence of bothfertilizers and pe '"cides would be greater than indicated by the product of thetwo separate absence packages.

Yields per acre for any given crop depend nct only on how much fertilizerand pesticides are applied but also on which land is used. The model classifiesagricultural acreage of the US by the types of crops for which the acreages arewell suited, and all requirements of land per unit of crop output are require-ments of wel-suited land.

Soine lands are well suited for more than one crop or class of crops, sothat the total amount of land of all types is less than the sum of potential well-suited lands for each crop. The model makes sure that total computed use ofland in any class or aggregate of classes does not exceed the amount of landavailable by requiring balance of requirements and supply at every level ofaggregation. For example, a given wheat production places an equal require-ment on each of wheatland, grainland, cropland, and agricultural land. Thisprevents land with alternative uses from being calculated as being used morethan once.

The classification system for land and the estimates of amounts of landin each class are taken almost unchanged from informal estimates of the Depart-ment of Agriculture"' Of two alternative estimates from the Department, onewith relatively large and the other with relatively small flexibility of land use,the one with lesser flexibility was taken, and additional restrictions were in-serted on the amounts of land normally used for wheat, fruits, and vegetablesthat could be diverted to other uses,

18

L4

Industrial Inputs to Agriculture

[he model accounts for three kinds of industrial inputs to agriculture:the fertilizers and pesticides just referred to in the discussion of land require-ments and petroleum products. The units oi measurement used by the modelare as follows: for fertilizers the unit is aggregate weight of nitrogen, phos-phoric acid, and potash; for pesticides anC petroleum products the units arevalues, respectively, for Industries 29 and 31 of "The 1958 Interindustry Rela-tions Study" " of the Office of Business Economics (OBE) of the US Depprtmentof Commerce. (The study is discussed later.)

The fertilizer inputs are 1963 totals distributed araong 1963 crop acreagesin proportion to 1959 use per ace as reported 'Ll a Department of Agriculturepublication by Ibach, Adams, and Fox.12 the pesticides and petroleum inputsper acre are those of 1958. based on distributions according to a slightly moreaggregated classificdton of crops in an unpublished Department of Agriculturcstudy"l in support of the 1958 interindustry model of the O13h ot the Departmentu: Commerce. That sturi, d.tributed 1958 use of the inputs among crops. Datafrom Aricultural Statistics2 permitted substitution of quantities of land forcrop values: then inputs of pesticides and petroleum products per unit of landwere multiplied by 1963-experience land requirements per unit of agriculturalproduct to yield inputs of pesticides and petroleum product per unit of agri-cultural product. It was not practical to blow up the 1958 inputs to exhaust a1963 total consumption, hence, unlike the case for fertilizers, the coefficientshere have a probable downward bias for reflection of 1963 technology.

For activities in which either fertilizers or pesticides or both are absent,the model has the other industrial requirements per unit of output rise in pro-portion to the i ?quirement for land. Thi, probably overstates input require-ments a little. For example, the complete loss of a harve.. in one area becauseof an insect invasion made possible by absence of pesticides should eliminatethe petroleum- product requirements associated with harvesting.

The cumulative product of each row of negative input coefficients and crop-production-activity levels provides total requirements for the industrial inrutconcerntd. A negative coefficient in the column for the industrial activity thatproduces the product generates a balancing output of the product in the model.The industrial activities are all OBE industry levels measured in dollar value.In the cases of pesticides and petroleum products, the units of measurementfor row and colun are identical: hence the coefficient is unity. In the case offertilizers, the row is weight of nutrient and the column is value: hence thecoefficient is a factor for converting column units to row units.

Capacity-constraint rows for production of fertilizer and pesticides, re-spectively, have coefficients of unity in corresponding industrial-activitycolumns. In the final solution the product of unity and the activity level (i.e.,the activity level) must not exceed the stipulated capacity. Petroleum-productcapacity, which has uses other than for agriculture, is dealt with in a row tobe discussed later.

It may be noted that the OBE industry for pesticides also includes fertil-izers. The model here, however, has separate ro . "' columns to permituse of data on separate capacities.*

*Aggregation errors of the OBE classification are preserved despite the provision

for separate output capacities in that there is no provision for differences between fertil-

izer and pesticides in the mixes of flow inputs and indirect capacity requirerment .

19

Livestock Capacities

Levels of livestock-product production need to be constrained by theamount of livestock available. The livestock-production activities are mea-sured in terms of weight of output per year, and for convenience the livestock-capacity constraints are measured in the same units. Using the same unitsmakes it possible to make the coefficients all equal to 1. With row units ofnumbers of animals or their liveweight, the coefficients would have been theconversion factors that are now used to convert livestock numbers to productcapacity. The cumulative product of a row of coefficients and solution levelsof livestock producL is the aggregate amount of the row capacity needed for thesolution activity levels.

INDUSTRIAL-TECHNOLOG Y SUBMODE L

All rows other than those already considered and the special constraintrows belong to the industrial-technology submodel. Each coefficient describesan amount of a row industrial capacity needed directly or indirectly to supporta unit of a column activity.

Industrial- Capacity Requirements

The industrial capacities considered by the model are capacities for eachof fertilizers and pesticides and capacities for certain OBE industries and anaggregation of OBE industries. The fertilizer and pesticide rows have entriesof 1 in the corresponding fertilizer and pesticide activity columns. The co-efficients for other rows are scaled elements of data from the OBE interindustrymodel.

The basic OBE model has the economy divided into 83 producing sectorsand has a table of direct and indirect requirements for each of 82 sectors (thescrap sector being excluded) per unit of final demand (delivery for purposesother than as input to current production). In addition, data are available forproducing such a table with more detailed breakdowns of the interindustrysectors. OBE has described its project in a number of published articles ofthe Survey of Current Business14- 1 7 and in mimeographed material that it willsupply on request. -

The 1 PC model uses as inputs a direct- and indirect-requirements tablecomputed with data published by OBE' 7 for disaggregating three of the sectors:Industry 14 (food and kindred products), Industry 38 (nonferrous metals), andIndustry 65 (electricity, gas, and water and sanitary services).

The general procedure for cdIculating a table of direct and indirect re-quirements from a table of interiniustry flows in a base year is as follows.Let X, be the base-year flow from Industry i to Industry i; and let X, be thetotal output of Industry i (including both interindustry flows and deliveries tofinal demand). The first step is to calculate the matrix A. consisting of ele-ments aI, eich of which is calculated as

2 \ \

20

The table of direct r,2 zc ket rcquiremclts is ( A) 1, where i is an identitymatrix of proper dimensions.

Both OBE and the RAC model use this general procedure; but there is adifference in definition of an industry output. The OBE procedure defines anindustry output as including the value of scrap and byproducts. The RAC pro-cedure treats these as negative inputs instead of output.

This difference in definition leads to a more fundamental difference thanmerely a lower nominal value of X, and a higher value of each a,, in the RACapproach. In c. der to prevent a requirement for output of the ith industry fromgenerating a requirement for a normal mix of outputs of the ith industry by wayof the formula

.1 - a \1

when the base year X, has a flow largely of byproducts of the principal produc-tion of the ith industry, OBE transfers any part of \, that represents byproductsfrom X, to X, before it uses X, in calculation of a,,,. OBE considers it im-proper, for example, for a model to generate a dollar's worth of meat-industryoutput of standard mix in an attempt to supply the leather-tanning indus'ry witha dollar's worth of hide and skin byproduct. They prefer to have meat-in',istryoutput of standard mix determined entirely by the requirements for food, witha side calculation determining whether this implies sufficient hides and skinsoutput to supply needs of the leather-tanning and -finishing industry.

The RAC procedure has byproduct requirements calculated as part of themain model rather than as a side calculation but takes care of the mix problemby having different rows for the principal products and the byproducts of an in-dustry. If a byproduct is a principal product of some other industry of the OBEmodel, the RAC procedure estimates all requirements on the row for the in-dustry producing it as a principal product. If there is no industry producingthe item as a principal product, as in the case of hides and skins, the distribu-tion row is a brand new one for a new industry defined as inventory depletionof the item. All industries that use the item have positive coefficients for Lhatrow, and any industry or industries producing it as a byproduct have negativecoefficients. A solution level of the inventory-depletion activit3 is the excessof requirements for the byproduct over supply. For a problem in which the in-ventory-depletion rate is to bc constrained (perhaps even to zero), there maybe no feasible solution unless the model permits production in excess of re-quirements for the principal product of an industry producing the byproduct.The RAC model has this kind of flexibility. The activity level for any industryproducing products in fixed proportions is generated by requirements for theproduct that needs the highest level, and surpluses of the other products areproduced. In the RA - model it is immaterial which of joint products unique toan industry is designated as the principal product.

The RAC version of the OBE A\ matrix has seven byproduct rows (identi-fications of byproduct element of \, supplied by OBE on magnetic tape) withno industries producing them as (defined) principal products. Two of theseappear in th;, final RAC model as byproducts, the supply of which must be ac-counted for and shown sufficient in any feasible solution. The two byproductsare hides and skins and cotton fiber. Hides and skins output is a byproduct ofthe activity that produces beef. Cotton fiber is a byproduct of the activity that

21

I

produces cottonseed. It may be noted that in peacetime the demand for cottonfiber has much more effect on production of cotton and cottonseed than the de-mand for cottonseed. As has been noted, however, the choice of which of twojoint products should be defined the byproduct in the RAC model does not affectthe results.

In addition to defining industry outputs in the special way just described,the preparation of the OBE data for use in the RAC model also involved a num-ber of changes in the base year X, before calculation of the aj. All rows fordistribution of products of OBE's two agricultural industries were reduced tozero since the RAC food submodel was replacing them. Those elements thatreferred to textile-fiber requirements, however, were transferred to the rowfor cotton byproduct of cottonseed production. The OBE food-processing in-dustries were assigned both additions to and subtractions from their inputs.The additions were values of transportation and warehousing plus wholesaleand retail margins needed to move outputs of the food-processing industries toconsumers. The data for this came from OBE work-sheet records of marginsbetween producer prices and purchaser prices of foods. The subtractions wereflows of food among food-processing industries and some of the packaging andprinting inputs. The reductions of packaging and printing were as emergencyausterity measures. The elimination of the interindustry flows of foods wasconsistent with the treatment of food-processing activities as minimum amountsof activity outside agriculture needed to release the nutrients in agriculturalproduction. For example, release of the nutrients in oilseeds requires plantsthat extract the oil from the seeds and make shortenings; release of nutrientsin grain requires milling and baking. However, the use of shortening in bakingis not required to release the nutrients in either oilseeds or grain, thereforethe RAC model suppresses the OBE requirement of oilseed-null output per unitof bakery output.

Once all X had been adjusted as needed according to the above rules,the a , and (I A) I were calculated in the standard way. This provided a tableof direct and indirect requirements for each row industry per unit of final de-mand for the principal product (as defined) of the column industry. The amountof each kind of food processing required per unit of each of the RAC model'sfood-production activities per unit of agricultural food product may then beconsidered a scaling factor to apply against the corresponding food-processingindustry's direct and indirect requirements per unit of its final deliveries."Final" here refers to requirements generated from outside the sectors definedin the (I A) ( table, not to the more commonly denoted demand categories ofhousehold consumption, exports, investment, and government purchases ofgoods and services.

The RAC model preparation used such a scaling to obtain the direct andindirect inputs to food processing and distribution associated with each agri-cultural food product but first eliminated some rows of (I A) 1 and aggregatedsome others. The eliminations were of rows for service and other industrieshaving very flexible capacities or which were considered to be too unimportantunder emergency conditions to be permitted to affect estimates of viability ofthe economy. For an example of the latter it was felt that any shortage ofhousehold-furniture capacity could be made good either by doing without thenormal output of that capacity or by production of the needed items outside the

22

household-furniture industry. The row aggregations covered all the metalwork-ing industries. All metalworking industries were combined into a single indus-try providing parts for repair and maintenance of production machinery andtransportation and communication equipment. Of course the aggregation wouldbe improper for production of major pieces of equipment. A locomotive cannotbe built in a textile-machinery plant, but it is quite likely that a textile-machineryplant could use some of its tools to make certain locomotive parts. The elimi-nations and aggregations left only 13 capacities of modified OBE industries (in-cluding inventory-depletion industries) that are accounted for in the RAC model.

The 13 industries with their OBE and RAC codes are listed in the accom-panying tabulation.

OBE code RAC code Industry

16 BE16 Broad and narrow, iabrics: yarn and thread mills24 BE24 Paper and allied products except containers27 BE27 Chemicals and selected chemical products29 BE29 Drugs: cleaning and toilet preparations31 BE31 Petroleum refining and related industries37 BE37 Primary iron and steel manufacturing39 BE39 Metal containers

40-63 BEMET Metalworking industriesE BErI Transportation and warehousing

3 .1 BE93 Copper manufacturing:38.2 BE 94 Aluminum ianufacturing68.1 BE96 Electric utilit ies14.3 BE86 Canning :and preserving of fruits, vegetables, and seafoods

It may be noted that no separate railroad-transportation industry, eithertotal or by any geographic classification, appears specifically in the model.The flow inputs to all transportation are accounted for in use of the matrix(0 A) 1 in full detail for calculating direct and indirect requirements for thelisted class of capacities, but there is no pr ovision for a possible shortage ofany particular class of transportation capacity.

This is a major weakness of the model should it ever be applixvd to aproblem requiring heavy-traffic movements. It is considered of no great im-portance, though, for a minimum-survival model. It is not easy to destroy alarge enough pe:'centage of rolling stock and iles of route to make the aggre-gate capacity for train or truck movenwnts insufficient to handle a bare-survivalvolume of traffic. Nuclear- bombing can create bottlenecks by destruyit classi-fication yards and traffic ct nters such as Chicago and St. Louis, but it should1w possitle to break the resulting botth, necks sufficientlv for light traffic tomove thruugh or around them. Military-type bridges, transshipment o cargos,and laying of new track for short distances are the kinds of expedients that areadequate for light trafixc. Our experience in inte rdiction of North Vietnamesesupply routos is an .. uOi.at io, of how quickly emerg n('y relpi rs ,an bx, made.

Each row oif the subiiodel for these industrial-capac lv requirvimentsglives an .'mnourA of the corresponding industrial conint olity that is requireddirectly and indirectly per unit of each activitv of the model, including the ac-tivities for nonfood support of population and governments and the provision ofindustrial inPuts to agriculture as xeil as the activities that are defined as foodproducing. The cumulative product of the coefficients along a row and th,

23

associated activity levels of a solution is a total (net of any negative terms forbyproduct output) requirement for capacity to produce the product (or supply itfrom inventories in the case of a byproduct).

Labor Requirements

The model has a single row of labor requirements that has one or two orthree components for each column activity. One component consists ofagricultural labor in food and feed production. The second consists of laboremployed in the i'dustrial sectors of the economy. It is the only component forthe persons- supported activity and for the three activities supplying direct in-dustrial inputs to agriculture, and it is added to agricultural labor for agricul-tural outputs that require industrial processing. The third component consistsof labor employed directly by the government sector. The government-opera-tions activity has this component plus an industrial-sector component forgovernment purchases from industry. The first two components are discussedin the following para,.-aphs. The third is dealt with in the section, "GovernmentRequirements."

The agricultural-labor component with normal inputs of fertilizer andpesticides consists of 1965 ratios between inputs of labor and outputs of agri-cultural products. The labor inputs are based on man-hours statistics as re-ported by the Department of Agriculture for individual crops and livestockproducts. 4 Man-hours are converted to employment on the rather arbitraryassumption of 2000 man-hours/year for an agricultural worker. The seasonalnature of agricultural-labor requirements implies probably a lower number ofman-hours per year for workers employed entirely in agriculture, but it isassumed that, under conditions of labor shortage, useful nonfarm work couldbe found for some agricultural workers. Outputs of agricultural products forthe denominator of each ratio of labor inputs to product output come from Agri-cultural Statistics.)

The industrial component is calculated as

'here [ is a row vector of direct and indirect requiremnts for labor per unitof final demand for output of each industry. N is a corresponding vector ofdirect requirements, and I and A have the meanings previously assigned. Ineffect, this procedure estimates the direct and indirect labor requirement perunit of a particular industry's final demand as a cumulative product of the directand indirect requirement for each industry's output per unit of this final demandand the amount of labor required directly per unit of each industry output. Theratios \ would be 1965 experience for OBE industries as estimated by Faucettif the labor row were needed for a current problem in which a labor shortagewas possible. Since the only problem worked on with a potential labor shortageib a problem for 1975. the ratios \ have been proiected to 1975 before the pro-cessing described here. The source of productivity changes is a study by theEngineer Strategic Studies Group' of productivity changes in industry. Themodel has no provision for increases of labor production in agriculture.

The labor row before the projection of productivity changes to 1975 nas anumber o biases. A general downward bias is associated with loss of labor

24

productivity to be expected when both farmers and industrial entrepreneursundertake unfamiliar kinds of production with facilities and equipment that areoptimized for other kinds of production. On the other hand, hours of workcould be increased over peacctime levels. Sociological and psychological fac-tors in labor-productivity changes have potential importance and direction ofeffect that have not been investigated.

PER CAPITA REQUIREMENTS

Basic Concepts

Any estimate of per capita requirerne its Afor survival is likely to sufferfrom two classes of ambiguities. The first is the representativeness of thecalculated average. The second is the standard of austerity used.

The representativeness problem arises from the fact that people differ intheir survival requiirements. For example, a person with diabetes ma',. need asteady supply of insulin to survive. But estimating a per capita requirementfor insulin as total insulin requirements divided by total population, both dia-betic and nondiabetic, implies in a mathematical model that the percentage ofthe population that can survive is no greater than the insulin capacity as a per-centage of total requirements.

The correct way in principle for dealing with differences among pIoplein their survival needs is to classify the population into groups that are honto-,eneous with respect to their requirements. This was n t practical for theRAC model. To avoid major errors of assigning to the entire population survivalrequirements that are significant only to small segments of the population, theRAC model attempts to deal only with the types of things that are widely needed.These are food, clothing.and shelter ior protection against the environment.and the medical and saritary supplies and services that may be needed to pre-vent the spread of epidemic disease.

The austerity problem is more complicated than the representativenessproblem. Factors that affect the desired degree of average austerity include(a) the length of time the austerity is to last. (b) factors other than living stan-dards affecting the psychological state of the population during the ixriod ofausteritv, (c) the consequunes for health and productivity of marginal changesin any tentativelv pro,- ,sed average level of living, atd di the value s'tet

for trading off increments in these consequences against increments in anycosts of changui the living standard.,. In general, authors do not try to ,pXiaintheir survival lists in such terms. ano no one who tried )&ould ever bw able tocommunicatt with complete success all four classes oif factors in his decision.

Nevertheless a brief attempt will be made along these lines for the sur-vival-requirem,,rts vector of the present model. The length of timne involvedis I to 2 years. The model is organized as a steady-state model, but therecould b some slight improvevn-ret in ..tandards during the p'riod. It is assumedthat the populati.3n is psychologicalMv prepared to assun., heavy burdens foryears in the hope of building towaro e tttr life. Mo rale and producti ity o flabor are as ;ured to require no higher le\v'e!; of living than are netded to main-tain physical health for both prtx+' t!ve :,nd nonproductive members of the. popu-lation. Peacetime poverty standards are of no use as a guide here because the

25

aspirations of the poor in peacetime -. e assumed to be c'-ditioned in largepart by living levels achieved by the more successful elements of the popula-tion. With uniform levels of living for the entire population, the standard neednot be much more than physical health. The value system of the present authoris not communicable tuilv, but it is of a type that would permit virtually no in-crement in living levels for survivors that would significantly reduce the num-ber who could survive. If the trade-off in survivor living standards should beagainst postattack foreign aid, against peacetime costs of pro-iding for higherlevels of postattack living, or against other conceivable costs, somewhat higherlevels of living might well appear lo be more proper standards.

The -nodel distinguishes betweer food and all other survival requirements.Comparatively little compression of food standards is compatible with healthover a period exceeding a year when the standards are measured in terms ofrequirements for specific nutritive elements of food. One may eliminate steakfrom a diet but not the requirements for proteins, iron, and other nutrients thatare supplied by steak. On the other hand, use ot many items of clothing andtoilet articles can be reduced or eliminated without sigrificant loss of health.The food-nutrient requirements of the model are therefore approximately equalto peacetime standards, but other requirements are slashed heavily.

Nutrient Requirements

The model has minimum per capita requirements for calories, proteins,fats, iron, and calcium. There are of course many other nutrients and vitaminsthat are necessary for health. It was judged in preparation of the model, how-ever, that a diet of natural foods that was clearly adequate with respect to theitems listed was unlikely to be seriously inadequate with respect to others.For example, it \.)uld be practically impossible to meet requirements for thelisted items without heavy use of grain. potatoes. and sugar that would provideall the carbohydrates required. Meeting calcium requirements, together witha model restriction on use of fruit lands for purposes other than fruit production.probably takes care of any vitamm C requirements. The rather generous al-lowance of proteins, plus the impossibility of optimizing without use of livestock-product capacity, virtually eliminatrs any prospect that the protein supplywould not be properly balanced with respect to particular essential amim acids.Nevertheless, some shortages of particular nutrients or vitanins are cnrce lv-able. and the model could easily be modified to add a row for any new nutrientthat may be considered v'orthv.

The mrinimu m per capita nutrient rettvirenients for foxi actuallv eatenare mostly recommendations (with an U111noWn safety mtarcin for it lls otherthan calorit 0 of the National |esarch ('uncil !",r age and sex classesweighted by Bureau of the (Census IN)pulatiOn statlstcs.' There are two ,x-ceptions. Althtough th, National Rese.arch ('ounc i has no standard for fat, itrecognizets that some unkmn amount is impxortant. A RAND Corp)orati,,n dNt u-meni by Pogrund" rettommends that fat provide at least 40 percent of all (alo-ries. and this require mer' was built into early vvrmons of the mde l. Thepresent version uses 30 percent of minimum (caloriv6 , which is closer to anofficial Armv recommendation- that 25 to 30 percent of actual c.lotrivs isclearly suffik lent. The other change is a reductlin of tw National ResearrhCouncil recommendation of 0149 mg day of calcium to 744 mg.

The rtvason for the reduction is asfollows. The Army source 3ust citd (, n-curs in the National H :earchCouni o that 800 mg day is desirable for an

I-'

adult male but states that 700 mg is sufficientand cites an Army regulation thatrequires only 700 mgfor a soldier even though diets actuallv served at Armymesses appear to include about 900 mg. The same Army source suggests that400 mg,. day is sufficient for an indefinite period under emergency conditions. Thedecision to reduce the model's initial requirement by one-eighth came after trialruns in which the caicium requirement was of major importance in determiningthe size of population that could be supported. It was then decided that the averagerequirement would be reduced in proportion to the difference between NationalResearchCouncil recommendations and Army regulations for an adult male.

The maximum intakes per capita presented greater problems. The modelhas three kinds of upper limits on calories, fats, and total weight of food thatcan be eaten

The thir(O upper limit is not a limitation that can be used very accurately,because there are no experimental data and because the weight depends )n thecomposition of the food. The consumption is ordinarily about 3.9 lb of retailweight per capita. The miodel temporarily has a arbitrarily high limit designedto pr,_,vide no effective constraint in any run.

With respect to calories,, nformal advice from the Department of Agriculturesuggested that the maximum should not be more than 15 percent above the mini-mum. The model, however, has a maximum about 10 percent above what nowappears to be eatenwhich is itself well above 15 percent in excess of the mini-mum. Itwas felt that in an emergency people could overeat at least as much asthey do now. In addition, the presvnt national average represents many who eatconsiderably less than the average aswvell as many who eat c(,nsiderably more.

The Army source cited earlier for minimum requirements recommends

that not more than 0 percent of calories be supplied by fat and reports that ArmyField Ration A has 40 to 45 percent fat calories. Th,, average civilian diet for1965 as reported by the Department of Agriculture' has a number of gram,; ofretail weight of fat that may be caculated as equal to about 41 percent of thetotal calories in retail purchases: m.ny people proablv consume considerahivlaiger percentage. than the average. The model has a maximum amou,,t ,,f fitthat would be 60 percent of mini mum calories a.nd 47 p rcent of nr.,xinin calo-ries. For cases where maximum fat niav be associated wit,, inininuni calories,the upper limit on fat is probably tco high. (oinpute r , 'put> Ahould always 6cinspected tor desirabilitv of rvcomL)utati) With .i t0owt r mIt.

The average dail; requirements us#'A" In t h.1i' l SU1 .ue sum.nlari ed iII tht,accomp.t,,;ying tahulatiin.

I i . 1z

(' Aliorie i 2 7 T I, Ii.

IronL. K ~.i2. -

t'h'~l , , '4-i -

% , ight oaLalI lb

Thtr arv no data on thc percvn-i4. )I h.- ibri.s :e, tz.i!i; .'.un'-I '0..! lv .

accvxn trd for by thi' Lft In jxirch svd f(f.

The r -)del scales these amounts upward to an annual basis and for anaverage wastage of 15 percent of the nutrient content of all ioods. Wastage inpeacetime can occur in distribution to consumers, in kitchens, and in failureto eat all that is prepared. The 15 percent for the model assumes negligiblewastage at the table, a kitchen loss at the low end of a normal 4 to 21 percentrange found in experiment in Army kitchens, and the remainder lost to rodentdamage and other food spoilage in the distribution process. No data were foundon peacetime average losses in distribution before purchase at retail of about3150 cal per capita. Some of these losses would undoubtedly result from short-ages of proper transportation and warehousing capacity, but otners could proba-bly be re&dced by application of more than customary care.

Nonfood Requirements

The direct nonfood requirements are bared on personal judgment of whatmight be dispensed with in recorded 1958 personal-consumption expenditures.The judgmer.,s were applied to details of those expenditures as published bythe Department of Commerce in the October 1965 issue of the Survey of CurrentBusiness." The ciassification there is not only by producing industry of the1958 interindustry stud-, which is not very irformative for some highly aggre-gated industries, but also by type of household purpose. For example, Industry72 is "hotels and lodgipg places and persbrial and repair services except auto."But a further classification of one expenditure states that it is for "shoe clean-ing and repair." The judgment tnen is on how much of industry 72 output isneeded for shoe repair.*

After judgments were made on eacb combination of consumer purpose andindustry number. the expenditures for each industry number were aggregatedto yield a total consumption exuenditure for each industry. Division of thesetotals by 1958 population yielded a per capita estimate.

Table 2 has the published Department of Commerce estimates for 1958with major classification by consumption purpose and minor classification byindustry. The table i3 shortened as compared with the published one in thatthe industry detail is omitted for functional categories thai weit judged not tobe worthy of inclusion on the final list. The judgments added to the table appearin a column of percentages for the expenditure items 'hat were judged worthyof reter.tion in whole or in part.

There is no point in trying to prove here any necessity for the particularpercentages noted in the last column. A few notes on the assumptions may,however, add to the plausibility of the stated percentages..

All food items are omitted because, as has been noted, food requirements"re considered on the basis of nutrients rather than dollars.

Nearly all clothing is omitted on the assumption that, although displacedpersons may provide a one-time burden on the economy, the absence of stylechanges, the possibility of great reductions in most wardrobes without damageto health, and the sharing of wardrobes would permit the economy to go forseveral years without much need for ne-' clothing.

*Quite another problem, of course, is the suitability of the Department of Com-inerce input pattern for Industry 72 as an indicator of what it takes to repair shoes.

28

TABLE 2

1958 Expenditure Itetni and Factors for Conversion to Postattack Conditions

I 1958 amout, .

miullionis of doflirs PretoExpeorure lass 0 emergency

Prhsr eopendit-res

triduca 1 lowedroles P e

'"" ~ aI' il lw,e- rage, 7,913 1 ; .:1,21F ood furn -. hed ga,% ernm.'nt int lu in. m l:,t, and . ,n n':r-

c ial rntp;,-, 8P.1 1.2-1 _b

V .i4 nsuret-d and f'r.)du cd on farnis I. l fljt - 11

Iba( . prodju :, 1..2 I 3.982 (

shtes an ther fit;tvrr rdqal 2. " 7 1.07,

12' rubber danti nnis, ,- ilan e ,u. pda.ti i p ;tu rs t'l - 1 1 1

11 f-r seair andi ,thrr :eaher pr f'.2.itr, 1~t -

inor ts1

>[-o- lealinui dand repaitr, t)ta 219 219).2-. ho.tel n kd ii , plac, e'" , pvrr a n c are~t

219tlt't" 211) 20V

I .aundering in -tablinlyent.'. total 977 772-. noivt . n5 b.f,, is.zng f

t .ttC... ittre,).lj and re .1, rL !

es,!p auto 944 909 80.mediuti dnd educti tondi ser'.ice antl,.ip,~

e stablisi ent. 28 2 S281- )od produced and L msumned itn farms% 1.410 i.110 _b

tlornies and childrien's Iothing and a( cessories except foot-

sear. t.,at 7912 13.3I 6. bro, d and nar-tn f abr! ts,. darn and t hread i i 1 06 6H' 0

7.- miscellaneoos textile goo~ds and flo~or % ering% 14 3oi 018, apparel -. 049 11 .596 21)19, miscellanefous fabri _aicd textile prodUctS4170

2 '. t..-r ant~ aiiied pr;.ducts excewt coftainterS 3. 7110 0:32, rubber and ii;scelianeous piastit products 9 1 7' 0i1. t oot 'tear and ot her lee ther products :1 0 7: _15 0(A. Hid !c~smnUiacluring (16 101 080. gross triports 'If goodts anid ser itoj's 129 260 0R~1, scrap, used. ted secondhand go-ds - It -3c6 0

Mlen's and his' ciltithin-i artd at cessortes. totai 4,.1 7. 116, broad and rlarros% fabric s, Narn and thread Millst 818, apparel 10.6 6.713 2019, miscellaneous fabricated texti Itprtducts o t

32, rubber ,tnt IM~ieldneous plast.(c products I0.31, fototwear arnd tlhr ipather vrtducts 1621- 298 0W*, griss rr's 'if go)ods ant1 servii es 18 I 18 0

M1, strap. used, andl sec,ttlhad goods 117 o

Standard , ltthing issue(] to mlilitari ptersonnell 57 0(ledn ing. ds e ng, press ing, alIteration, itotrtge, and repair of

garments. to idtittlng furs tin shqs. 1,71)7 1.7917 0it'welr% antd %at, '1,s Q214 .S0( 013th -r cttiting. ac, f-sries. ornd vtr 172 172 0

Ftilei art istit ad repatrations 1,39 1 , 3tX) 0I

kiLrstts eat% ~roirrs. atnd iaths -2. 111 2 01I; 0

()wnrrocctupted ntnfatrm tkln, space-reniI %alue, t.)lt) 2ttlg19 2 6. 8)1

1, rettl estate andl rental 26 ,.01 1t 26, 809 to

Venalt-,)t upiel nttnfarnt tlstlltngs .tnd othe'r htuwsitg, totatl 12.1t,) 12 .1.')7

reat e-t ate andi re~ntal 1 .2~ 1.. 12.Q10

29

TABLE 2 (coitinued)

1958 amount,

milons of dollars Percent ofExpeditue clss'emergen'cyExpoditue clss0

expenditu~es,Producer Purchaser allowed

72, hnt ekI and l odg Ing p lat e. rersona I and rpa IT se r% e M,'Xcept auto i1 VI1. to7.me-dical and educati-inl ser% ice and nonrip-f!te staIi sh me nt s 82 82 t

Rental %alue )f farmhou.se%, total 1.861 1,9o171, real fStiAte and rental 1'86 1 1.R61 to

Furniture, includintg miattresse, anid t.'1spriigs 2,141 1,1t6 0Kitchen ad other household a-)pliarites 2.712 1.-i12 0China, glassware, zable" are, and utensils 911 1.697 0Other du~able house furrishing 1.958 3.728 0v-idurable houe furnishings 1,3 11 -2.521 0

Cleaning aind polishing prepartions and misve!:aneous house-1h,-ld supplies and paper products. totol 2,08,- 3.1:W)8

9, stone ard cla.N mining anid quarrying 12 "1 0't, paper and allied produdis. exc.lu ding . rnain-s and

boxes 57-,0 97)8 (6)2S, paperboard containers and boxes 38 tt 026, printing and publishiav 3 027, chernicais and selected cherocal products 104 145S 6029, drugs, cleaning and toilet preparations 1,1)76 1.S571t 6030, paints and allied products 18 .1 (-0t

:36, store and clay products 27 50 -1042, other fabricated met.! products t8 8tt 105 3. electrical industrial equipment an-1 ppacatus 5 t055. electric lighting and wiring equipment lit 161058, miscellaneous electrical equipment. niachinerN, and

supplies 2939 40641, miscellaneous manufacturing .44 t 1080, imports I I t0

Stationers and writing supis506 9:38 0

VlectricitN, total 1,381 t.381%6, electric ity 1,381 4,381 30

Gas, total 2,68.5 2,68597, gas 2,685 2,685 50

Water and other sanitary sersices, total 1,0_18 I ,04898, water Q90 980 10079. state and local government enterprises 68 68 100

Other fuel and ice, total 2,351 t1,153

7coal mining 261 t. I-: 4020, lumber and wood prodlucts, excluding containers :8 r76 41027, chemicals and selected chemical products 10 76 t

31, petroleum refining and related products 2,008 3,462 1037. prmary iron and steel manufacturing 10 Il 068, electricity, .ns, water, and other sanitarN services 14 It 0

Telephone and :elegraph :3,892 3,892) 0Domestic servive :1,503 '1, 503 0

Other household operation, total 1,708 1 ,7, -,65, transportation and warehousing 286 28670, finarce and insurance 117

30

TABLE 2 (continued)

____ __-1958 amount, Preto

Expenditure class' [ -mergency

ProCe Pucsr expenditures

.2, cl 2212'In 122I liiig )2h1225. 321 rs.w2,2i l nd ip112

ser\ it l , ex,2 -I tauto 7 1 .1.1 4o

73, buie., o-r%\ 22 vs (22 62 60

78, 3-eliral zi.,,ernym22in! vntirpris '170 -70 6 0

D~rug 3rip.Iratlm 2 and s12 undries, till.- 1,01,12 31Q5231, vall'i~ and3 alied jiril 3 td ', ,\ ilvp 2 ,mnenrs 3' 233

29, diutz. ICe'liiiiig 12223 1221121 rvi(2lrail 21 2 *687 20

12. rulbrin(!in s,22 21252eI 1222iIMp last511 it ii u t s 1. 790 it

'! ec-tri( lighing and i' 2ring ei pment22fl 3 0

6i2. .i ct fi and .2(1 inti,1iig instirumenti s 757.. 39 0Opht hI ic C 1 ro2)GUcL s :hpqi ~ ap~pIiatn, i s I'll 613 60)PI l ,i 21.1225 3..571 3.,7 60Iterlt st, 1 .876 1,.870 60() her professioinal setu it v 8.32 83 2 6(0IPriu~itek ,intr,dlid hl)53,135 anfd Msln,w2ria 1.2(02 1.202 100I I a it Ii itis ura 9 ( I ,1 I( 10 i.130 (0lirlikerage 2 liargvs andii in~iestme~nt t-2lllfseiing 881 883 0Bank sC2 cri chargc'.. trust ,er% ic es. .it I 1 ep.'it-b., rintil 810 83e, 0Sctrvjies furnished %,,ithmlt pa~nmen! bN financial iflteri liales

f'\ci32t insurance 22(o Ifies 1,0731 1.071 C

V 3i~ o ~ ndl in, Ir ifnsurImc 11:.21(0 .3 ,210 01 aI li (5 3.5. SI _31 1.531 0

FufelC and burial c\3 2cni - tuta) 1.:108 1.328Q, stone1 ulni lax ruin ing and quarr!, rig 5 12 0

16i. stlife anti la% products 18 60 (0

71, real3 estalte and rental 23V2 23 Q0

72,- i22te .121a2,1( Iolg iritg p323.221v2, pecr s in aI anid re pa ir se r i Le s.

evvt ituto112 1.01.3 1'0 13 20

(tier personual blisiness, total 93 1 9:3 166, i luwai,2n iat ions, e\, Iud ing r.:dio n ,122 Iv,1 is io

1radi list ig 17 17 207.1, btU5121O5 s1lrv2 '-SI 137 1 1 0

76. aniuseinent s 6 6077 i2cdkul and eIUcio((al seric and noniproifit

1staubi ishmients 729 72() 50-8, FederalI golvrnmniit ent erpirises 62 6)2 0\tt , ars a1221 3iu2'5r' 223 o used airs .2i 13,2 58 0

-outiai,t ti-125 tilbvs. at2 l-ssles. pats repFair, greasinug.l2.lshiig. parking, stiralge. and rentl) , total1 .,181 62,507

19, niisi-t.larliius falurialted tex.ile priduits 30 7 527, (heuiils and22 5lvIv(t(2l hlii A2iuil3U ts "32 76 5

N~. drugs, Lleanling aniii toiilvt Firvpa~rlti. 22 .38 .5:32, rubbir ,lnd imi2cIlanvouis plasti1c pr2 -dwl2-' 80( 1 1,372 5

32, ithier fabric ated iieal iriidu2 ts 32 23031. . V iLV ii)11il u ia.irm 222 2nels 16 :15 5

553 , 'cltrii lighting and %\iring equilnill 15 J30 52

56u, r ii t , t v I(e ui l i. an2 d iit i i i in i i i t 22222 i cu i i i2t 3 2 53 5

58, 12i2is (v Ili viwu s vI i-tr i-1 2222 hina ti n vrN , v quti1)ime n t ,11and

.4 11i 11I i v SI103 28 tH

59t, 22222! (ir 1 h i( Ills an1d2 eq-6 ieii t 81 11 3575,~i~~u2~ repair 1iii sivr\ ; 13387 1,387 5

3:3, s( rap,2 1151'.I 21222 sI, o iai1I123 n 53 20 .

31

TABLE 2 (continued)

1958 amo-int, Percent of

millions of dollars

Expenditure class, emergency

:hsrexpendituresProducer Purchaser alloendue

prices prices allowed

Gasoline iad oil, total 5,251 5,251

31, petroleum refining and related products 5,251 5.251 5

Bridge, tunnel, ferry, and road tolls 250 250 0Automobile insurance premiums less claims paid 1,606 1.606 0

Street and electric railway and local bus, total 1,219 1,219

65, transportation and warehousing 1.219 1,219 100

Taxicabs 574 574 0Railway (commutation) 121 124 0Railway (excluding commutation) and sleeping and parlor car 3:38 338 0

Intercity bus 296 296 0Airline 479 .179 0

Othet intercity transportatioi 32 32 0Books and maps 632 1.022 0

Magazines. eiewspapers, and sheet music, total 1,439 2,061

26, priming bind publishing 1,468 2,090 583, scrap, used and secondhand goods -29 - 20 5

Nondurable toys and sport supplies 1,188 2,115 0Wheel goods. durable toys, sport equipment, boats, and

pleasure aircraft 1,080 1,845 0Radio and television receivers, records, and musical

instruments 1,644 2,836 0

Radio and television repair, total 681 68172, hotels and lodging places, personal and repair

services, except auto 681 681 5

Flowers, seeds, and potted plants 338 544 0Admissions to motion-picture theaters 992 992 0

Admissions to legitimate theaters, opera, and entertainmentsof nonprofit institutions (except athletics) 297 297 0

Admissions to spectator sports 249 2.19 0

Clubs and fraternal organizations except insurance 692 692 0

Commercial participant amusements 848 848 0

Pari-mutuel net receipts 451 454 0

Other recreation 1,174 1.181 0

Private higher education 1,282 1,282 0

Private elementary and secondary schools 1,006 1,006 0

Other private education and research K.52 852 0

Religious and welfare activities, total t,178 4.17877, medical and educational service and nonprofit

establishments 4,178 4,178 100

Foreign travel by IS residents 1,900 1,900 0Expenditurs abroad by I'S government personnel (military

and civilian) 892 1,077 0Fxpenditures in the I'S by foreigners - 1,046 --1,0.16 0

Personal remittances in kind to foreigners - 107 107 0

Total personal consumption expenditures 'L4.032 290,069

Durable commodities M,2 62 37.881Nondurable commodities 88.765 140,15 2

Services 112.005 112.036

a'l'he major classes4 are standard personal consumption expenditure classes of the N itionai Ir.tome

Division of the Office of Business IEconomic-s. '[he classes with numerical codes are sectors if the

195H input-output table of the National Economic Division, Office of Business c,)nomics.

bConsidered elsewhere under another classification system.

32

Some production of footwear is permitted to replace that which can nolonger be repaired and may be needed for essential travel to and from work.It is assumed that the number of workers who walk to work instead of ridingwill greatly increase. With the sharp reduction in new-shoe purchases, adoubling of shoe repairs is allowed.

For most inputs to household operation that are accepted at all, the factoris about 40 percent per capita. This reflects both an expected increase inhousing density and such austerity as reduction in the use of electricity forappliances and lighting.

The automotive-expenditure judgments are based on the assumption thatnobody will drive a personal car. A small allowance for automotive expendi-tures is intended to represent some incremental input to public automotivetransportation, where the requirements for gasoline, tires, batteries, etc, percapita are lower than for personal autos. Those who do not use public trans-portation will be expected to walk or stay at home in the hypothesized circum-stances of a large labor supply in relation to capital resources. Dormitoryaccommodations near Places of work are conceivable for workers who wouldotherwise have to commute by auto.

Such an approach yields levels of living that are probably well above levelsordinarily achieved in large areas of the world right now but well below whatis considered extreme poverty in the US. However, as noted earlier, povertyis a relative matter. What is properly considered an intolerable level of livingfor a person in the present environment may be quite bearable under conditionsof austerity for all after a nuclear attack.

Table 3 has the results of applying the factors and aggregating to the levelsof "1958 Interindustry Study" industries.*

The direct and indirect requirements for nonfood items that appear in thePERCA column of the model are calculated starting with Lhe matrix-vectormultiplication

I- \) I(

for the nonlabor elements and LC for the labor elements, where C is the vectorof direct requirements from Table 3 and A and L have the meanings previouslyassigned. After these matrix-vector multiplications, there are deletions andmetalworking industry aggregations to conform to the industry classificationof the industrial submodel.

GOVERNMENT REQUIREMENTS

The nature of governmental activity at Federal, state, and local levels isbound to change drastically after a heavy nuclear attack. How it may changedepends on such things as whether military action continues, whether tihere areresources in excess of what is needed for survival, and what those faced withoperation after a nuclear attack consider important. Some kinds of governmental

*An example of a particularly regrettable aggregation is the inclusion of shoe-re-

pair expenditures by households in Industry 72 (hotels and lodging places, Irsonal andrepair services, except auto).

33

activity can also depend on the size of the surviving population. A large popula-tior needs more of many kinds of government services than a small population."r

In general, however, except for construction and military activities ofgovernments, it is probably accurate enough to consider that all governmental

TABLE 3

Nonfood Consumption Expenditures by Industry

(1958 popuatilon and producer prices)

Expenditures

Code Industry Total, PeCaiamillions Per capit

of dollars dllr

7 Coal mining 104.10o 0.59718 Apparel 2,226.80 12.73320 Lumber and wood products, except containers 19.20 0.1102t Paper and allieu products, except containers and boxes 168.50 2.6~7926 Printing and publishing 73.10 0. t20

2-1 Chemicals and selected cheniical products 69.00 0.39's

29 Drugs. cleaning and toilet preparations 9:35.10 5.3 t-

31 P'etroleumu refining 764.55 t.3 72312 Rubber and miscellaneous plast'c products 62,75 0.35934 Footwear and other leather products 4i3.210 2.,36312 Other fabricated mnetal products 17".4 to .099

52 Ser~ ice industr', machines 0.80 0.0055:3 Eliectrical industrial equipment and appiaratus 2.00 0.011

55 Eectric lighting and niring equipment 12.3,7 0.011.SO Radio, television, and communcadtion equipmerit 3.60 C.021

58 Miscellaneous electrikcal equipmet ind supplies 19.7-5 0.113Motor vehicles and equipment 4. 1.,

64 Miscellaneous manufacturing 25.60 0.116Iransport at ion and "arehousing 1,219. 00 6.()70

I1 Real estate and rental 7360.0 HO 112.090

72 Hotelsa and lodging platec s. personal aind repair se rv ices.except auto) 1.784.25 10.202

73 llusivess services 3 7. 20 0. -13.

75 A.utonviiile repair and er~icrs 219.3s 1. 25 177 Medical andt edut ational serx ice and nonprofit establish'

mern'% 13.131.50 7510578 Fede~ al go~ ,rnment enterprises 12.01.9-)6

79 taer and local go~ ernuient enterprises 68.00 0.18"~

91 Scrap 0.4V) 0.0(12

%6 Hect rit, it s 1.11t. 10 7.5 '71 -

G7(as 1.3 12.t0 7.67798 ater and "%itar% ser% i e 980.00 -1.004

activities have labor as their principal direct inputs, with other direct inputsconsisting primarily of supplies and maintenance of office buildings. Newconstruction has no place in a steady-state survival model, and the model makesno provision for military requirements. It is therefore assumed for the currentmodel that the 1958 mix of direct and indirect requirements, excluding require-ments generated by new construction and military expenditures. is satisfactory.The level can be tailored to the size of the surviving population.

*The PARM project of the National Planning Association, under Office of Ymer-gency Planiring (OEP) contract, has develope~d a model that generates requirements forvarious classes of governmental services as functions of population.

34

The mix of direct requirements consists of state and local expendituresother than for construction in 1958 plus an estimate of the nonmilitary expendi-tures other than for construction by Federal government. The state and localdata are as published by OBE in the Survey of Current Business."" The nonmili-tary expenditures of the Federal government are from unpublished work sheetsmade available by OBE.*

The model can, of course, use other levels and mixes of government re-quirements. However, if mixes are used that r- iaire substantial amounts ofoutput from the ordnance industry or other particular metalworking industries,it will become necessary to disaggregate the present aggregation of all metal-working industries.

CONSTANT TERMQ

The column of constant terms changes with each problem to which theproblem is applieu. Depending on whether it is preceded by -- , - , or =, eachelement of the column is an upper bound, a lower bound, or both on the cumu-lative product of a row of coefficients and corresponding solution magnitudes.In the cases of the nutrient rows, the constant terms are necessarily zero forall problems since neither humans nor livestock can live for very long on nu-trients stored in their bodies. The other zeros of the constant-terms columnof Table 1 could be rates of desired inventory depletion for any problem inwhich that should appear to be a useful concept. A zero implies that require-ments may not exceed supply from current production. The remaining constantterms preceded by inequality signs are fixed capacities of kinds that are notused up in current production. In the case of land, the postattack capacity isthe same as preattack. For industrial capacities and labor supply the percent-age of preattack capacity is an output of assessment of physical damage fromthe attack. The stipulated level of govtrnment activity tells how much is de-sired of the government-operations activity when measured in the units fixedfor level of government.

"he most difficult data problems are associated with the determination ofpreattack capacities. A practical definition of capacity is an amount of outputthat would not be exceeded under any reasonably possible circumstances of in-

put that did not include more fixed plant and equipment. If it is further assumedthat industry capacities never declined in peacetime, the highest output actuallyachieved in recent years is a lower bound on existing capacity. The amount ofunderstatement depends on the percentage of capacity that was idle at the timeof this historical peak, on the possibility of increasing output by use of extrashifts or more workers per shift, and on any new investment or technologicalchange that might increase production potential.

Despite the foregoing elements of understatement, lack of research timeand a desire to err if at all on the siae of understatement of economic capacityof the country has led to use of historical production peaks for most kinds ofcapacity. Faucett " is the source of data on OBE industry outputs in 1958 prices,the prices of the OBE model. The only adjustment of Faucett data needed forthe present model was a technical one of adding what OBE calls secondary

*The work-sheet data have not been releast'd by OBE for general use; hence it isnot possible to givt. the table here. The RAC mo(del processes the data in a way thatprevents its accurate reconstruction from information appwaring in this paper.

35

L

transfers to the output of each industry.* For some more detailed capacitiesto process agricultural crops, the historical experience came from AgriculturalStatistics2 data on the amounts of crops ,o processed in recent years.

An exception for industrial capacities is the use of current data 2" on per-centage of capacity of wheat grain mills being used. Together with AgriculturalStatistics data on amount of wheat being processed this yields a capacity mea-sured in amount of wheat processed. The Bureau of the Census data are basedon operation of a maximum number of shifts 5 days/week. Informal advicefrom the industry is that operation 7 days/week would be practical for extendedperiods of time. Negligible amounts of downtime are needed for maintenance.Accordingly the calculated capacities for operation 5 days, week are increasedby 40 percent before being used in the model.

Data on livestock-product production peaks are from Agricultural Statis-tics2 and on the labor force from current issues of the Survey of Current Busi-

2 7ness .

The livestock capacities involve a certain amount of ambiguity for a steady-state model that purports to describe rates that may last fo" more than a year.Capacities of some livestock can be increased at little investment cost and ina relatively short time compared with the length of the period studied. This isparticularly true of poultry flocks in the production of poultry meat and eggs.A hen reaches peak egg production in about 7 months and will lay hund.( .Is ofeggs within a year. A minor diversion of eggs from breakfast tables to hatcher-ies, plus a diversion of female chicks from butcher shops to egg laying a fewmonths later, has a biological potential of greatly increasing poultry-flockcapacities in a short time. There are technological problems of differences inplant and equipment of poultry pro'..ro... - .. . meat, and breedingflocks, but it is reasonable to assume that a great deal of these kinds of plantand equipment could be improvised. The potential for increasing poultry flocksduring the period under consideration is enormous. The model now contains noprovision for the reduction in liv'stock output per umIt if input that would in-evitably be associatec with use of improvised facilities and hence overstatesthe capability of the economy. However, the model uses what is considered avery conservative steady-state equivalent of 1.33 and 3.00 times the flockcapacity for meat and eggs respectively that exist immediatelN after the lossesfrom nuclear-weapon effects* There are no such adjustments for other classesof livestock, although an argument might well be made for a small adjustmentto be applied to swine herds, which also have rap&( reproduction cycles.

Projections of preatta-k capacity are based on the previously cited Engi-neer Strategic Study Group report,^' which, in turn. is based on a NationalPlanning Association study." ' Minor adjustments using other data were neededto convert the source's estimates of percentage change between 1960 and 1975to a percentage change between 1965 and 1975.

*Each industry produces the products that define the scope of the industry plussome prr'ducts of kinds that define the scopes of other industries. These ar, called pri-mary and secondary products. rtspectively. Ihe OBE definition of an industr "s output.whicn is also the RAC definition, includes all the primary and secondary productior, ofthat industry plus the production in all ot'er industries of prodicts that are primary !othat one. The Faucett data did not include this akitional production, and tht, RAC adjust-ment of Faucett data added such production in the proporti. n added by OBE for 1959.

*These nearly arbitrary numbers can b, of great importance in some problemsbecause of the low input requirements and the excellent balance of human nutrients ineus. When dairy products are scarce. tgs can be a principal source of calcium.

36

USES OF THE MODEL

The model is designed primarily to indicate whether a postattaci economyis physically viable without substantial investment in reconstruction. It doesthis by estimating a maximum number of weapons-effects survivors that canbe supported for I to 2 years without depletion of inventories that would leadto subsequent collapse. A large margin between the number that could be sup-

ported and the number that needs to be supported suggests a requirement of nomore than a relatively low standard of management efficiency and a likelihoodof potentially rapid economic recovery as resources not needed for survivalare diverted to reconstruction.

A narrow margin or a deficit of economic capability suggests a need forefficient management, of course, to use surviving resources as effectively aspossible. The model also suggests some of the kinds of things management,both preattack and postattack, might consider to increase the number of peoplesupportable. Standard output of the model includes not only the number of peoplewho can be supported under the conditions postulated but also shadow pricesof the constant terms of the model. A -shadow price" is an amount of increasein the objective (millions of persons supported) per unit of change in the con-stant term (capacity, allowable depletion rate, or stipulated level of -overnmentoperations). Additional options in most linear- programming routines (includingthe one used by RAC) allow testing of the sensitivity of various aspects of theanswers to stipulations and parameters of the model. It can be used to measurethe worth of preattack preparations such as provisions of stockpiles and oi -'ost-attack measures for conservation of resources and for highest priority recon-struction after an itLick. In the case of reconstruction measures, the modeltells where marginal changes in capacity are most needed, but it does not sayanything about what it takes in time and sacrifices of ofiwr uJes of the inputsto achieve any given change in capacity. This w .mld rtquire a more dynamicmodel.

37

Ir

POTENTIAL IMPROVEMEN £S IN THE MODEL

Potential improvements in the model may be classified as (a) increasedaccuracy in the estimate of particular matrix coefficients and preattack capaci-ties, (b) provision of grea.er detail for greater realism of the model structurein applications to the steady-state problems for which the model is designed,and (c) introauction of capability to handle problems with dynamic elements(circumstances that change during the time horizon of tile model).

Potential improvements of the first kind have in many cases been indicatedand are in other cases inferable from earlier discussion. Earlier discussionhas 1lso suggested directly or indirectly points at which the model structuremight be made more suitable for practical problems. Included among the pos-sibilities are (a) additional rows for human and livestock nutrients. (b) additionalactivities for producing human and tivestock nutrients, and kc) more detail inthe industrial sectors of the economy that support food processing and inputsto agriculture (primarily chemical industries). Potential modification to includedynamic elements is discussed in the following paragraphs.

A fully dynamic model, such as PARM - " or IRAM." has each input to aproduction activity identified with respect to lead time before output as well aswith respect to type of input, and it has investment activitits with similarlytime-phased input requirements. Such a model has vastly expanded data andcomputation requirements and would be impractical ft. a linear-programmingmodel with the optional food-production and supporting-industry activity detailconsidered necessary here.

A possible compromis, would te to consider the model as covering twosteady-state periods. The second period would be a submodel similar to thepresent model. having no investment activities. The first period would bw longenough to satisfy the following cor :ttions as well as pvossible (a) reconstructionactivities can be completed or nearly t ,,ipleted during ZNe perlkAd, but the re-sulting capacity incrrases are not available to any significant extent until thenext period and b) the depletion of g ods in process and working inventories(as distinguished from deliberately prepared stockpiles) carnot be a majorsource of inputs to production. The strucluxe of the submto.(el lur th., firstperiod would differ from that of the second peritod in that the optior-il activitieswould include investment activities of kinds that previous exercise of the statikmodel might show to be potentially usciul. -he kind of problem to be run onthe model would be a maximimation of the number of lrvivors in both periods.taking a,-count of the possibilities of depleting cert~a stockpiles in the firstperiod provided that investments in new capacity make further depletions in thesecond period unnecessary.

38

Appendix A

TECHNOLOGY MATRIX FOR LINFAR-

PROGRAMN'NG COMPUTATIONS

TablesAl. lcftniti0' to, (oristrjan tg Rows of th1 U cchnolmg- Matixt 40A2. Defini mns lor Column Ac tics oft ht Tcchi-olop- Matir ix 42A.-\: Tec hno bogx Matri 4 6

Tables Al and A2 prc sent definitions of the constraints and activities ofthe linear- programming model with the constraints and activities classified asrows and columns, iespectively. Table A3 contains a tabulation of the techhol-ogy matri..

TACI E Al

Definitions for Constraining Rows of tke Technology Matrix

RLW nomea Type of constraint Ur. It

2ALMI Food calories (ininimm) Billions

* HPRO Food proteir, Millions ot kF*F'kT Food fat (mininrim) Millions of kg

*IRON F'id iron Millions of kg~C AI.C F '04 (alcmumf Millions of kg

+ NUTX U'nused

SNUT Y i nused-.HWT Food weight Billions of lb-FATM4 Food fat (maxi.-um) Mllions of *r,

CALMA Food calco ia~.imum) Millions

F I TOT Feed unit all livestock Billions of lb+VI'R (M Feed units for uattle and sheep. FUHO(; Feed units for hogs

PHOTO Feed protein for all livestockI

PHOHV Feed protein for cattle and sheepPROHO Feed protein for hogs

IPHOPO Feed protein for poultrN- WIT lerd weight Billions of lb

1,AV Land for wheat maximum) Millions of acres

-WI.MIN Land for wheat (minimum)* ACOR Land f or corn1* lA 0 A Land for oatsI ,4A1 IC L and f or rice

1,YI AR)1 Land for rveLA PE A A Land fir veanut

* IACtT L and for cottonI 'ASOY L and f or Roy beans1. k HF: A .and foi edible beans

0O1A4rRI Land for fruits

L.IA PO'1, Land f, ,iitatoies

i LASWI' Land for sweN potatoes

LAI-IK I aini f or sugar bvets.I AC AN La.~nd fior sagani aneIA\ F KG Land for vegetables

IASNOH i and f or sorghumi grain osoare

40

TABLE Al (comtinued1)

Row riarne" Type of constraint Unit

* IAiI'tt 1 -4 fine haN %ItIlonq 'J acres* 1--V lrid for pasturelA(.I t~i ,rainland

I Il . A Tota d rI aIrrt urdi land Milons of a(cresVKH I F rtilvzer lnutrients, in t. It basis) Iilliions 4f lb

[ IF!,I Prst ik de IHinPions 4f 195)1 donillars'* IFHI PezrinJeum produ i s Hillions of 1958 dollarsC Al"PH ('apai for iert ie BC illions of lb*C A PF> 1 Capnaiti fors pestic ide.. Billioins of1 19,58 dollars*l PIt A ( apd~IN t for grain mili~ - r.p.u! measurre Bijilons .f ib

SPO'll CdpaL .t ffoi ,'real preparatioinslnpu! me auueI

*SPIIIC Capa it forw ricn. millinz lin put nwasure;

- SO~l(.,pan it% fi iolserd piro-essing (input measure'- 'lI Uapai t[N for ..ikir-beet nli ) ig nput measure'

P SC' 11 Cpa, ;t% f r SugarcanC' milling ;i;nput measure)I

s PSOt V ooJ sot bean ruea! FBils oft lb-HF!6 Caparitt for fabrit tarn and thread Hiilhonrw 0 f 1958 dollars

*H2 aat for paper and allied products exceptLontaine.rs

* F2T Capacit% fr chemicals atid scelcted c~hemicalpt oduk ts

BEF19 (:aPait'l for drugs. cieaningininud toiletpreparat ions

*HIQI 1 Ca.pa( it% foe petroleum refining and related

ploUO

-HF3 ( apacatN for prt".art iron and teei-HF39 f apac it% r or mnetal wincers

* FCr(apat- fo t ir talAorking-H~u (. o.C-pa ( itt. for transportation and mNareh using

13 K 86 ( apur itt fo. in-nning and pre.,erv ing fruits and%egetable-.

H ['93 ( apa - it% for copper rmanufact!uring

B1,I91 ( spait ti f tr aluiminum madnufacturing8f95 ( ips inl for other nonfinrrons-nietals

mranulai turinpB V96 ( sputiaN for elec tric pomer1W Bl97 (apavitin for gas-ut ilitN ser icr

H 1IC96 ( upaiti for %ater and sianitartN sent ice

HFIt? oto illis 4 98dollars

I. AHOR L abonr %li IlIions f.r ikers

( , N~ll A f;klt~t for 11i lk produtiton 0li\ estock I Iio n s of lbV P~~IOF Ca pac, tv fIor pork produoct inni ( it e st w kM1A H FV (i it i z ftir beef hprluction iiesiocL)

ki WGG CloiacitiN f or e gg lindut r (Ii % es tti ck I

%I IA1'0l 1 Cit a c it % for lilu iitr -nmra t I ridnui v i Ili% ies t i\! MA I( N apai it~ fu1 r lamb production (liinst,ik)

IFfil ; I-resh vegetables (ainiislm), [I Fl1 Fresh fruit (niaxinium) Hi,1lions it 11)

(;VI ILevel of goivernien- acItiviti (1951 packLag~ 10l00)? units

ali'ecedittg at ron% namie, , 0, or t tellIs %%het her t he cumu lat in tiproidiorI of at rim% 's velevme nt s and m

iisponnding ciuntn-activitN levelS is constrained to be I* or ireSpie (t ite % v a ito iant term.

41

N TABLE A2

Definitions for Column Activities of the Technology Matrix

Column name Type of acttity' Unit

PYRC Populat ion SUI*Vival Mlillions of people% i I Food wheat Billions o~f lb

ticol Food corn with b,%proau-t feed for livestock

other than poultrIICO2 Food corn with b ' prooduct feed for poultr

IIOA I Food oatmealHIi Food rice with bvproduct feed for livestock

other than poultry

111112 Food ri- e with bN product feed for pouitr.

HRY I Food rye wit b product feed for livestockother than poul-ry%

1111)2 Food rye with b'prodac: feed for posultr%1l1~LlFood peanuts

IWSlI Food cottonseed oil1ISYi Food soybean oii tnd food meal for use in

bread

IISY2 Food soy bean oil and feed for livestockother than poultryv

IIS13 Food soy bean oil and feed meal for poultryv

IIS14 Food scybean meal for use in bread

IIS15 Foio4 "ybean oil and food meal for use as

meat extender

HI 6 Foiid meal for use as meat extender

HBNI Food edible be!ans

IIFI Fo fresh fruit

HFR2 Food processed fruit

lni Food pot at oesI-SWI Food sweet potatoesHBTI Fond sugar beets

tic Xl F'oodl -iigarcane

It\ F I Fo(od fresh vegetables

M VF2 Food proL essed vegetab~es

I1101 Feed corn fed to livestock other than poultry

ILC02 Fe ed corn Ifed to poultryI SG I Sorghumn grain fed to livestock other than

pouti I c

LSG2Sorghum grain fed to pioultr~

I ,0Al Feed oats fed to livestock other than piiiitrN

LOA2 Feed oats fed to poultry

LH IFeed barlev fed to livestoc k other than

PoUlt rv

I HA2 Feed h arle y fed to poultrN,I AI Feed hay

IPA] Ieed grossII[tRVIItI, Food "heat

iC 0l1 F Food ciirn with byproduct feed fur I i~estm 1,

iither than .ioultrN

11('02F Food corn witth hvproduct -ed for lioultrk

IIOA I F Food oatmealIIHIiF Food rice witij bN product fee,' ',)r livesto, k

other than poultr11111121 Food rice with byproduct feed for poultr.. Billliiiti if lb

42

TABLE A2 (contnued)

Column name Type of ,,ivtyc' Unt

IH)' IFV Food r\,-- with hr orodu~t feed for fixestock Bil lions of IbIit her than p',ultr%

Ifl 1t21, Food rne it h hr prod(tut fe--d for poultrNIpiFll !*ood peanutsH(:SIV F ood cottoinseed ilINI IF F'od so )bean oiI and f,.od meal for use in

breadIISI 2V F'ood sor bean oil and feed for livestock

ot her than poultr%IS1~31j: Food s0ob 'ipan oil and feed meal for pou Itrv\1151 Wh Vood soy\bean meal fo r use in breadB

FisYSI 5[od sor bean il and food ;n'a I for use c

meal -xtenderIS) 6f- Food meal fo.r use as meat extender

IIBNIl- hod edible beanis111 lj F' o (I( fresh fruitIIF'R2F' Food pre.cssed fruit111111 F Food potatoes

IN; I I F'oodJ sweet potatoesHFIT 1F- F'ood sugar beetsliICNiF' Food sugarcane

FIIFF'ood fresh IregetablesiYfkK21- Food processed \egetablesI. 0(.IF' F'eed corn fed to livestock other than poultrNI.C021- F'eed corn fed to poultr),I S(; FI Sorghum grain fed to livestock other than

poultryI'SG 2F. Sorghurn grain fed to peultrLOXN IF F eed oats fed , livestock othe, than p ,uitr,LOA 21- Feed outs fed to poultry%

1.13A IF F eed barle; fed to livestock other thanp(-u ltfN

1,13A2F' Feed hurle% fed to poult.,IiAIF VF-eed haN

11WHIP hood IheatIICOI P1 hood corn with byproduct feed fir livestock

other than poultrxhFI021) F'ood corn w ithI byproduOct feed fin po lt r)I FOA I1 P ood oatmealFIRI IP F1 ood rice with b vproduct feed forlistk

other than jiouitrN

IIF12F F'ood rite with byproduct feed for poultr%IM Y 1 ) F'ood rye with byprodat t feed for livestock

other than poultrc,IR 121) hood rye with bYproduct feed foi poulircIIN 11) hood pea nul s

lIfC5 IP hood cottonseed oilFFS't II P hood soybean oil and food meal for te

ini breadIMS21 P Food soybean oil and feedl for livestoick

other than poultrFS't :11) hood sirebean il arid feed ineil for poultry

FISN 01l Foo0d soybean ineal for use itn bredl

HISY 51, Fioil sovbean oil and] food nical for use asmneat extenider Bi Ii irs if lb)

TABLE A2 (continued)

Column name Type of activity' Unit

"SI'6P Food meal for use as meat extender Billions of lb

IIHNIP Food edible beansHFRIP Food fresh fruitHFR2P Food processed fruit1PTIP Food potatoesHSWIP Food sweet potati.esHBT1P Food sugar beetsHCNIP Food sugarcaneHVEIP Food fresh vegetablesiIVE2P Fo.d protessed vegetablesICO IP Feed corn fed to livestock other than

poultryICO2P Feed corn fed to poultryISG IP Sorghum grain fed to livestock other than

poultryLSG2P Sorghum gra n fed to poaltry

LOAIP 1-eed oats f-d to livestock other than poultryLOA2P Feed oats ied to poultryLBAIP Feed barle. fed to livestock other than

poultryIBA2P Feed barley fed to poultryLIIAIP Feed hayit%. IN Food wheatIICOIN Food corn with byproduct feed for tivestock

other than poultryHCO2N Food corn with byprodict feed for poultryItOAIN Food oatmealHRIIN Food rice with byproduct feed for livestock

other than poultrv1111I2N Food rice with byproduct feed for poultryHRYIN Food rye with byproduct feed for livestock

other than poultry

HRY2N Food rye A.Lh byproduct feed for poultryHPNIN Food peanutsHCSIN Food cottonseed oilHSYIN Food soybean oil and food meal for use in

breadIISY2N Food soybean oil and feed for livestock

other than poultryHSY3N Food soybean ol and feed meal foreoultry

SYI tN Food soybean meal for use in breadtlSY5N Food soybean oil and food meal for use as

meal extenderHSI 6N Food meal for use as meat extenderHHNlN Food edible beanstIFRIN Food fresh fruitHFR ZN Food processed fruitHI1FI N Food pot at oe st|SWIN Food sweet potatoesIIITIN Fool sugar beetsHCN IN Food sugarcaneIVE IN Food fresh vegetables1IE2N Food processed vegetablesILCO IN Feed corn fed ti livestock other than poultrNL.COZN Feed corn fed to poultry Billions of lb

44

TABLE A2 (continLed)

Column name Type of activity' Unit

1.GIN Sorghumn grain fed to livestock other than Billions of lbpoultryN

I N; 2N -Sorghumn grain fed to poultry

LOA IN Feed oats fed to livestock other thanI

LOA2N Feed oats fed to pouli

LBA IN Feed barle fed to livestock other thdnpoultry

LBX2 Feed barley fed to puoltr.

LHIN Feed hayMYF Food beef (dressed)

EGGS Food eggs110lL food pouiltrY meat (dressed)POR K food pork (dressed) andf lard

J POR K Food porkb (dressed)

MILK Food rniilk Billios(fl

BC27F Fertiliizers Billions of 1956 dollarsBC 27P Pest ic ides Billions of 1958 dollarsBC3I Petroleum products Billions of 1958 dollarsSTGOY Govcrnmert operations (1958 package) One unit

aF, P, or N ait the end of a crop .. irne denotes a process without the normal inputs pf fertilizer, pesti-

cides, or both, respectiveli .b1-0o reduce intake of fats and or calories, the oil or fat is not used for food.

45

TABLE A3

Technology MatriK

C9LA 42 1

A .* s44 210 00 -4.4030 -1.4 -43:4"022 2492(0

I .. N0( 0. 01W243 -3.aO1100 912143 -. Icr 9 40 01 . ('-j'(.

CALL~t 0.(0400 -. '00 -0., 10 0 O.113 -2 9 1' -9 03

& 4 NUT.

, - sT 9000.O31000 -0.9.200 -. '3 'O'' -. 3,

.4 IM - 4144 0.91))U0 1. 003030 -1 1 .'700 -11. 8C30 10 -13., S2 .'21"

IA 0 (FU 1 .9310c

Ii * M4343.tiL2-.l)3

19 L.1"4 01

20 L A-F3.732

24 * 4001 2/ 21

29 L 4(

40 L4I A UII I 14(,31

AS.[14

L4 ?

04 0 44.4

L6 0 1 4

* l 474 '1242 ... ' 0.1(24. I1 .~>~. .

s.s03 A,' 13 0. '' 1' .. .9 t ,

4 90 4 0 ( (I4 I 1 . ' ' .

611- 1 (4

42~ 0.,C,

A4 00" II)4

4% I (414 '.331I3 .. ') 40 '0'1 3.,.49~~~~~~ ~~~ I 14 ,3(36 '1104 ''17 301' .2j.0 > 1 ' . 1~

10 * v I07 .. 311 .1?4 '02 .'6. . 9 1 4. . ' . 1

49 (0( 01141 .1(114 .114 .. 1664 1~..I . '4e

TABLE A3 (continued)

68 44P32- "loll 1 2 1$2 I68 81 844

4 7 Ij 76 Q -0.00600 -o0.3 3900 -0:044400 '.4 4'- '"I 7 L ' .6 -0.96000 -0.2 26 00 -. 904000 -. i 't""' -3

7 NU 4 r7- T78 -0.76000 0.70.0 -G.15)000 -0.436000 -D.170330 0O. 73 '.87" 1'.17202

12 7 U Um.3 -0.473000 4, 800U,,0923 -1.080000o

]476 -0.1&000-07922

27 * .38 -. 63003 o~.30

F8 * l" D836 -6.22000 -2.4"20-3C. 3)U

29 ' 8 -6.00 o 7000 -. $303 .o0

2 - LA 129

.4 A 8 7834. 7700

24 7 ,8.', 0-0233

27 T *.f~ 2863 1.144324 Il "A28,62

0 2oO,0 9 '' ,', 403

68A6A 07600

39 7 8 13 o 9 0 0 1 '

37 ST 2.7 oD 73871

"6 7 t8 0c 7700

-2 , % 86 .403 0473 .401 39'0 ,8,. ., 0 442 7 A 1872 08 00 0680 .403 06'0 9

,3 -. ' 4 '? ' ")

43 7 6687 0.02 000 .170 00 0 2970 2 0. 266 >2 -03 13 , 2 o ' ~ '

46 7 6871

69 A 4F26014?

n9 2 7 j o . 1 .2 . ( 0 3 .1 .0 2 I. , 0 0 1. 2 ' I I T%' 0

92 * 4747

TABLE A3 (continued)

1f 1& 70 20 2. 1

88336 HON7I .47 33A2 WTI 57 7 98

I~~~~~~~~~~9 o"68 78000 -.. 000 9.0000 -M93 00 -2V.203 .'''.lC

2 p U S78 -126000Oa -70.40000GO '.9lOO A M10002:000 -2.00O -. 030 3.'

At8 .. 000 -6.400000 -,..6000on -1.00031r -2.3o300 2~)

806 0.0440 0.020 -n.006. -0 .:00200 -n .0 3 -0. 031 . '7

S CAL( -0.914000 -0.631000 - z.0 200 _0_.0000 -0 ,010 -C . III', ... ' 0

9 4 67% -1.0000.0 -. 00o0 -1.600000 -1.40000 -2.300003 -. 2'

I1 6 UTOT.3

'

34 * 308

16 3 P*OU

20 * 18.7

26 LA-IC

2a, 1.6878

I8 a 110 0.820

29 * .1661 0.083000

8 2 6 18!.6

3

336 16310OA

60 D 108

43 * .L 78 0.60 00 0800 0 .01 670 0 .0337oo 0 0 .0897C. 7 39 30)1

44 : 3117 0.1 000 3 0) 01 0 1.62 1.600 20) .0' '3 ' .w '

43 : 38760 k.170000 *. 0000 033 0 ~ 0

96 C A16'7,

SOW * 301 1. 000.100

14 13307 0. 1,11000

33 * 6116 0.103 . ' 1'a,

'A') O*ll r 882 0.02 6..'

CA , 87900002, ' 1

30 *61.0 72

* 11 1 683.001 660003~

61 * 73372.01037 . '8

TAP' 'E AJ (contlnued)

3! * 3T - I'. )o .,' -.. , I .003g

4l (t )c -<,11t2303 - 3.132-0 00

$' - 4*7 --.'r. 7-3.333.'- -. ,3. .7-3..3).33 .

.1 - 7444 - .)223 -

II * 3,7 )7 -3.32,'''? - ..3..>

13 * 3j4. ), -1 2 >"r. 3 . A.14 * t3) I2 -^.-q.'3'

IA * 43.,'0 7. -

20 * tls"

26 * t34+27 t 44c37

33 * LAP)?

36+ * 12.1 . v ~ ? '1 3 ) . ' ) O . 2' '.4i ') .41< .',37 4 43t4

4$ 4 14>..34

31 * 1 I3444),!

+.3 I4., i.263,'' 3 264' . <') .

41 14<4% ,4'1 ') .24 332 2 . ' - '. 6 - ' 41 '- '442 * .4714 '. 44 ,fl 6'3 .2 4 ' 3. 6 '.4 . > -3

4.. .. -t ,4.,

64443...'44

43 * 3'? '''249

644 ;

TABLE A3 'continued)

449) 44*4' 14*) t'*: 441 I. I' ' 9*1.

* faLul ii'*4

- -, .9 *1 4 4.4.

* #*9 ., 1 - ,. -. ~. -

9 * 9*9 9* 1' -. ''.4 '- -1'.~* 9494 . 1*1.7 .*,1.44 .* 4 tat. -. 149 - >' - -* 4.49'9 * 4q.j99

* - 9499 ..4

4.,i *,'44 - 44

4 - *Ta* 9. - *1. -Ii - )*(4* 1'4. -I.. - I -

LI * 9,919 -C..*92023 -4. .V74

' .. . 4-I.) * 9-S4.44 '493,30) 4441 - -. 4)4.- -

99 * *.~t.. - .49.13217 . 4 -4 9 -W.. ''1.94.4703

-. 9.7 . . II **M.)~ 4 -'.. 449-7 -' - - . - -

99 * ,b.i -7. 949430) --- - - -'IS * 49)4) 342'j4 949 -'44.., -7. 1422)) ''7 ' - .. - -'.1 .1

.9 - t41. 4444

234441*9 .4

2* 9 **4 A09 * *ai9

9' ta94,99.44*'!

49.9

'4 * 9*4*. 4.449'. .44957

94 4 9*999-P. - ' ' '

.3 4 1994 >'~'A.' 94., .

499,4 .44,4 .4~49'' 94.4 ,'44~. 9.49.

1.49 .'' .49~. 9,' .92, .4. . .4.* *'tI,! ).4 '1 9.44' ~49' 4 . .4

9* * .4999.9 * apt...4 * ,.

4.7 4 9fl9

4. 4 \~>4 - 4. . . .

44 94999 . -'4 . .

4494944 .9. .-. *

............. , 9

<<"9 .499 1)4 .9.4

50

I

TABLE A3 (continuod)

• 4'41 -,4 1 + - .t +-1 .4 .. *t- P. 4. - 4.m. q , .I, I* 4 i u ) +. 4'..)3 -.2- ) ,2+ ,4 .4 ,l - ..* .. ...... ..44... ' C 24:) ,+ l -'.4.. ....-A+ 4 A+ -44,4': OIZ," 3. ,"J •. , +

4-4 * 4. . - ).2%4. ,;. - . .'' .41 4 4.#. 2. 2. .2.' -.. .-

.

-. * *4*I -. , - .44 , - t ' .t ,

4, * .4.44,

*+ 4 4 1 4 4.. * 4 4 ., ' + - ' " .4L,.+ . . . ... . . .... . 4 , ,- . . . . . .

''4444 . '1.

.* " -.- ' - +''444 /

I 4 4 4: o .. "= • 2 ...

'

.. . .*. . +4*$ - 4'¢" .........

'1l +* .++:+. +.

-4 :+I 1: ..

1 5'i+ •"

TABL-E A3 (continued)

0 0 3 5340 ?O j

I9 3 CA94 oo'1o -,5 04

5 ". 0P00 0.'

0lc _0

2 * s~6o -152 100 -827.%13000 -11000 -I.0 3 ), 1", -

A. 69 36 1.00000 -3.80000 -47.00000 -I D.i000 -o9.430031 6"'

* iu -0.751003 0900 0.030 -. 900l 0924

11 Z'0911

19 * :3:6-0 %7000

20 L P606

IS - L 160 -. 90

2 1 1*006

2t 169827 + ,*10

2: * , S'i, 1.011003 1.010000 1.01)000 .100

29 LABIA I

640 111

36 L A03.8

.0 LAG.&6

41 0 1.R 1.010000 1000 .100 i)C' .032 .10 1

42 0 .1 L1 1.01t0300 1.010000 10)0 .lC"3.f 0 U 9' 'I .9

45 * FS 980 0.6,000 .400 7.0 )0 .,00 1.003-. -

41 * 1" CA00P.0ES1 C 7 01

46 , A906.

49 4 66

SI 0 SPOIL 1.030000 1.00000o 6.003000, 1.0000

52 SPOT6

86 * 6900 0. 786000 0.76TARO30 0.70230D

5* 62. 0.00260% 3063 .24 .0,5 ".9

57 FI0 880 0.9416 0.005103 0031 .)010

5 E 0 969 0 00458 0.001 105 0.)05 C11.050-

"I*63 0054 0005 0.0 4 o.ocoo .0 2

40 of663 0.0029; 0.005104 '.0.... 00..3. ; 0

61 SE669 0.0035, 0.0<6 0391 0.00314 701 2

60 " (8 .00491 0.01123 0.095 00,3510T"

64 Sta6b.599

68 * 6893 0.0 10181 o.00626 0.338 .011'0 0

66 *6694 0.000250 0.501210 9.027 0.000700""

67 68% 00419 0 .00 1099 0020416 0.000.i8

610 11;1<0579 6 00173 0.0006 9 .00179 0.000

71 R 6606 0.10o00 0.00. . 05 0.0000041 2 "661 0.0000S2 0.0000116 0.02)0? 0.00,00520.''

7) LA1*BOR 0.01620 0.035278 0.0126M0 ;.o106.70 0.070190 207< .41. 7o3

76 6661(

60 6 1696

at F 6666 I . 0010".

52

TABLE A3 (continued)

47 I "1 i -. 11.

I4 7 A7 7,( 7 v i 7 7 7 ~ . 2 7

I C A J O ' - 3 7 O s '2.7 3 3 - 5 . , 7 - . 7 2 2 7 -

2 ~ ~ ~ ~ ~ ~ ) J 113 -. 032-70?

7 *~1 - 601~00 17.02'

* . 0-.- -(342 2 o322 27

S ~ ~ ~ ~ )D *IL?*.j0 -. 433 '1-372 -. 2

1 4 7jAR

-263303 0.'20 '7.2

9 -FATMA -. 62001-.0')J..O

10 (447 475077h0 ,971'70 .7,777 -s. . 7~.X2

I7 A 7.422 -27730 -t1i' -7

6 3703 70f( .2 712~

II A 707A. 103 - 29A- .' . 2)

74 t A A y

07 0

24 * LI.

S7 I 10022;

24 * 7 * 7

s2 * E L , 0...7S30

33 * I Z*37 0. 30'0009 0O 1

60 * , 1*7.2"7

43 R I47 77 T'

7. 071. 1

47 * (4 0c7 72 0 0 0L

"1 :* 7 71 1

I3 * 5P. I I00~

7 4 10111

7% 737770.23374 3.027072.. .53

'TABLE A3 (continied)

.SGF GI 66 96 A, 66 BA It LI-2 7'A

C * AL M I o

3 6 FAT -.- 34'>

9 * tALC4 6 m SOTxI NU l01

19 6 FjO 3. 1000 -. 100-'

1 4 6 FOFO -0.810000 -'C .95c00- .63))0

16 * m 866 0000 ~ 0.19601'.02

id 6 P -00.0100 66 006303I8 A POO oi~o -0.940003 -0. 1,-1 -1 . J.

19 - LA -0.910300 -C.1940000 '.403 '.90" c.6C) - .(C-

2i 0 1606 1.093000 1.C90010

75 LARYE26 L 616E6

L1 * AC OT

3 0 1-0.i!

31 L 106

31 * 6660.570 1.(73

41 * 1601 1.01100 10900 .C33 0.94> 290 ) '.0 : .o. -

36 Lto -Y' 0!>

663 SPt. I44 I S T:'~ 0:021"0 ...45 PE 1 500P00612010 l

6A 6 661

0. 56 * 662

+1 FF21.

49 6 6629

4 59 663 '.-'F' ' "Q'104

60 Spat1.'0, .Cl

63 * 66 '204 .199

ST6 665t'1 .2

61 60291196 11 6613 ~0 '0(4 0~~04'

00 13 * 6604 002245 0.02419 0.30416 ,3)26! 1.0261 00)9)6 4j1- .092

14 * 3116

74 CM I 6666

o 06 0 MAPO

'1 A LUUPS

:1 I.I1 IlI5

-~ -- -- - ~ - -~"-- ,;Z-----------

TABLE A3 (con) nised)

44 1'049 76 3

A SiN 0.39.3 -o. 30)00 -G .C37 33-226)io' - .6'

3 s 01(003 -0.390303~j -703 -3. -. 6C1 73'0 -o3 7. '1

aI *n. 60- -3 0I3 -"06 1 312 * 64)0 -0.17300 -3.3530 3 4 3

IA 7 6931. -Cu'0 663C

1 5 )0 0. '133I.372

21 - *18)22 * ALUS 0311636

2 L A T* .)40,13

79 L A( -13 3

29 * 4 1*3In LA " U

34 ILtA

36 L ASUQ37 * A*

3 ' 1*0*)43 A 1*.* 0 . 5I)30 3.3)400o.967' ..

42 1 tAL 056300 o.533 2.I4 )o r.3 .,^ .967 31) 3 -o' 74'

67 * AP!

41 R L .96 0.20-01

'F CP PI L o.03330 I' 30

1 2 * PM 3957

S4 5 506(1 -2.140003. - .3

6 A 564 o. 039)3 L .3)2 3.3)3 1.)7 o . 11 . 3 00

+0 *83 0.330 .033675 6.oo37i 0.0'3 2.'72 .'7''4'

A2 G 55 .. )4 0.014 o0)5 0.')9 - '7)6 -'7)4 "I'+7* - o 965. '017 0.309405 6 0.3'16 0)1)6 .I52 ./6?'32

'S A f 553 0.002 1) 6.0 *7o ) 0.7~03 1.0'6 3'.S o2 ) o3 3

63l q 95 0.318 o33) n.7'3 0.L33 I.412 '.)A -'576

1 5 * 1 7 0 0 , )9o 0. 0 0 6 7 2 0 . 91 0 9 9 ) 5 0 . 0 1 n 6 .1 ' 6 5 3 ' . 9 6 ). 6 7 9.7l

76 * H3 18*1*olj ~t? 5 4 Or11o " c o115)5l n0

78 * L

55

TABLE A3 {coarfnuCdl

0618 1 -1*7..0000'3 -69.032000 -696.030 7,200 -'7 Co -77.'"NO788 -1 52.600600 -152I.603lCC( -1~2627 1 1150.60' .

* 87 -82.500000 -78.800000 -763J~ *?0iC 02.03030 I - :~~ - .*.

6 * ""0388 - a9W! -OSO3 0173000 -C .j73

00 -(0.930(00 - .94000 -. 93o-oR02

I I FU6080? -1.28000012 * 60808 - .2"62000

L] , ONO~7 -1.280000o14 * 61* '.750000I s P80W -'.251000Ia6 7 fok *808S -A3 000

18 ' "806 0.610N

21 - 11

24 k 1687L

26 LA 18EA

a7 7 1800728 4 180 0.907000 0.907000 0.907003 0.907000 c.407)00 '40"

30 1 68 LAP124

3 2 * LASWE

33 *86

38 LA t jAV

48I , t8P A

43 R *3 06068c 9 c

35 * 816 0.5006a58 0.0003890 0.000383 0.000356 0.000l81 o.'0L-

4 * 1 864 00I3T.040 .)45 00833 '020 .'2'

47 , 86TR7 0.005303 0.00361 0.00411 0.05303 0.30D31 20 1'A, 7 869 0010 .045t.248 00105 0"25 .705

AI 7 63 00001 00540 0086 .008 .254 05840 * 81 00528 00215 0021 .050 .021A.~0

49 * 88 00146 00358 00388 00596 00316 .'5'

47 855 a0.0030oog6 o.00089 0.03013o# 0.OI01 56 063 .C"l

70 * fz 869 0.086 0007 003413 0:"3'036 ) 0.'373-.1?

74 * 0811

s9 7 2 C160*68

42 0(03.

E 0 ao: ':oo91j9 o.QoS~b "56dI ~l~ os6 aoI o47 it ~ r4*022~O~t..BERET~76'~* .0'lsi .6N00- .s. - -s

TABLC A3 'c antmuedr

""PN73.Un3

- ... '1<

0 ~ ~ ~ ~ ~ ~ ~ ~ o * nlj 2 3202 2210 -8f'22-4-

F4 ,TOT 022 -- 92)2 '0)2 I

4 up-9 27/00 0320 '2 22 -0 2 .,S -a 142 it 40)00000 <.12 0X30 -724 - 7

3 PROv

- " -0900 08000 -RO))* 4

19 L* ' _r I --C. c4 ' -0. 0

?. L8471 -C

L9 Al U88 -T--.

V1 AS( T88 039038 L ABE A

23 O T '880*?

.0 * 9*10?A

30 1 CA T U 0 3200 r o 5l o L 0 20 c 1 1

32 1 101 'o21233 %FO o 38 0*81 0.2 3"o r_ 1 U

36 S 9490

38 I 8 L

526 f18t 0.325000 c.100 1210 .0000" o2991 .4' '17- '9~7

4 4 1 1 3 7I3 I 18 0.000 .220 .000 0092" 38'< 234 11 . OX(4.OL"I'141982L o1

92 * f8t.00000

54 A 99 : "0 (j3

55!61 L, Sf9 0003 0 1?4 1 .20.135. E 9 866 0 .0096811'1 0.00 L.tl .41o

Mot0 * 999 0.0308s42 2 .0CC 39" I I2

8 lo * 8139 0.0 9 3 0 5 ?9 * G 0. 0 1e ..0 - 0U 1 s4 - 4.'o9 V14

42 61 319131 .03602023 700

"AL 1!F5 1098 .992 .090.3"86 8 F41 9915 -

85 G99o.0(7".0' " 40 79' '7,

44 * 3994 0.000185 4.0708 00.9 ''9

TAB -E A3 o'~e

99 F. -'.

Is *

it :46. (.0 100

b Go 0 nsfo1

19 - W :0"00301 nl 1009*O0 0 .,

2; - C 'I)'25 L00A-.103 -. 1 T- 12 I.l -c.acc-.8W2% - .9' - .I -

I25 LPK0 .9103 -it * 00 0OlOOP05130-c.'

16 * 65* 00911u.O92

20 LI a85U21 - LAI.

015 L190I6724 184 C .114000 C.9I420,

26 5 LASOy

40 LA*C 00

T ~ ~ ~ 3 *z LASI G:,,500 0

FEA 0 1*S0* 0 .480 100 0 .49,5, 0 0-?0c

C37200 1** . 901 .1,0 .J0

39 CA 9*E*I40 1*06 0350 .800 .50)11..9~l 48'

445 1:1 85

98 F I I*

90 OWS91

a4 1940

691 s 882

60 11,1:07

12 bt8858

43i 8440

84 1 (11

4*858

ALA

A N 3

* 37

,

1- 0 0 ".9 )0 93 -..

99 * I- 2.

..-. - -

92 9 9 ,.' 99. -...

1 ''

99DO.) 233.

l~~w~~ 'o...Q33.

at is 10114 ~93 9

. 'A, 4 to 2 0..91.2 )O33

-*t :191 93

-... .11) Q

All, 9j2. 9 3~9

4191 A9 9-29Qj9 .,933 t3

4 -

99 - i '. 9 5 9

ITABLE A3 (continued)

* .0.2 -@7. ACQ-0 i' ','fit, i-

Pti~a -A~' a. - 4 4 ,. . . . --. 4

4 * i&t -1;;'0 1 - . . .

up A t 111

*O - All -tft7"II100.jKi))oOo -- ~,It *

34b A 5407 -C 3J-

A P4AI tiA.0'111

39 - L~f -5.-)q r445 -- 7-n-

70 A rLgOA

it -

7II L COf Q

115 1 ItA .. ',,. . ',' 31 ,. ".

74 A

10 L AGA A

33 . 94,J000 .11, .1t

14 11

"!4111

II A LP744

4 W A A LA

A th~s 1940000 .04f0)0 1.13' Ovloi' * JI ,

4, * (a is;

1i t 4p1t

06

Ii6 441 .000 1'-'' 3 33" iS ) I

4 A 44194 at44

al Atoit3490 5 '

79 A Si .370 '.jEt .2'4-'3' . 7, -'

14 A SEJ* j<304ft~ ,'.33A71 .. C5>.0 . '1 .5'1 ---.-

47 Ar 499 5.0.0 YiiWVA444007@* 5'77"

44 ft 160

TABLE A) (coot nued1

I * it-I -S,6,,. V >)++

/ .. ),- -2,%, I' ,,Y. -'.++ , . 4. ... .. "-4

* 4.14. Q.224,3) -2. )'2A. A.' ++ t + - I -

iP - .,tIT .,.,t),.2 '

I * *j'%t.

*, aar

1% * lPi

*,+ P a Jo..

I' * l''4"

2'>

* t idit

sIl L tak l

*+ lit,,+Ii

II II *.1.

... ii el•:

.4 * l .6" .. • -

64 * 1 +

61 t I44

- I 4 I *ata

I~ +

TABLE A3 (coninued)

64I2 t01 1(0. 01 "02 I -, I

CALAI -16 0:000000

D60 -. 0000

O 4 * 140 -00)903S * CAIC -0o.112a O0

I 0 u It 900

- 64(9 -1.120000

10 - CA L.A -1 .3 02I, L1011-~~

19 *U 44 ( -0 37000p .9 39

74 MI * 040.4)1000 0."000

2A 8 600-..630)>3)0

2) 5 LAGA) . I2 0

24I L 64

? * . 0.3260 0810C011

30 012"662 z

12 1 LA0t

4s 0 t IRO1 0.320,00 1630 - )n -

39 Ill,6

41 * 167. 0.6)000 .103 .103 .6'C .813 ?0

S4 I 915)1 340 * 96692 1.4000 .100 Ir30 .),0 l>' I ' ~ 7'

4? 8 k 0.0009..0

50 . 09810. *l SPOI

93 .* 09869

'199 * 0900?

16 0 0 2 000601

1? * 12? 0 .0023019 8091 0.010121

60 8 0.010*C * 063 0.20705?

11 010611 0. 19935)1

165 * 69 0.0063 s0 0 ? SS 3 J 2

14 I 64 .00

00 * )9 00019

6911 9? .0.0

69 MA A60

706 *F 806117? NFA

62

0

IABLE A3 (COfl-nueA1

L3 1 1

4LaARE-f

1 4 8 . !

L9..-;

* 01 *-L *. -o -. ,9--

9 A * . 3

9 F3~

0''J1 E L .

-- . - :

~~7.390.229- F :

400152

3 *j6E

0 01315

'24 *

RE jiR29 * 3343

0100 J

a9 * I c 00no f A 1.330'-

33 '4L" 0 .4 21.2 * i33

TABLE A3 (coninued)

I 416 09.00000

* so *0 .0 1.000300

O AT -0: sjOOO

6 "107 .0.0002000

P&*(4. .1 : 000O

C, - 660 310 0000

it Ulu pj1 . '00001. , "'U .34.0.0

11 POOO 1.460000

10 * AW4(21 W1 14

*0 * 1.A

so L1 AP

*0 * LACOT

2 LASUE

10 LAISO

02I ? .0

I0 * tai0 0 * . 0 ( 4 6

3$ * 16900CI

46 * A'O 1.0000"

39 * AP*S

so I .40651 * . L.

"*I POT

40 6 0:.,1 '04TAO -1 .01001 4 4. 6 0 4 .6 * 0 0 06 9 3 - 4 6 ? 6 1 : 0

S? :eZ? 0.000461 0.0.0-16000'1130

s Q 0185049 *:0Og 0.1,0PC6678 .67

.:21 4o 0.0POIL0

60 : 1 1 . 1 .

19 6 1 0.001234 0.3049* 0.00412 0.00104 P1 0000

IA 6*4 .0117 0.00199 0.0091" 0:300 021~0

96 *6629 0.004 0.010 0.11404 .714 1100

63 60 00606 00004 6 0.0916 1.074 s 491.70i0

64 46166

Als 64) 0.000060 0.0014 0.001047 0.O0l412 76.1920001

46 I 619 0:000*6 0.002411 0:0, 1"s 0:0214 ,.000

44 I :t90 0.0004906 0.024160 0.024161 0.016946 *,i.000300

10 : 610? 0.000 0.0100 000100 0.000196 .i:040

7**660 000040 0.004 0004 0.00016 14.60000

70 : 160 0.10 0011 0.l1... 0. 0 00094 6 0 000

74- 0* 06 .0..0,0010000

it Oto 060400600"41 .064 l

T4 * KI a6l.0079 1 it.4

1 64

F: PIK* .....~o~a90* 4 .m

REFERENCES

1. Bernard Sobin and David F. Gates, 'Economic Implications of High Population andLow Property Survival in Nuclear Attack on the United States (U)," RAC-TP-317,Research Analysis Corporation, Jul 68. SECRET

2. Dept of Agriculture, Agricultural Statistics, 1966, US Govt Printing Office, Wash-ington, D. C., 1966.

3. M. E. Parker, E. H. Harvey, and E. S. Stateler, Element of Food Engineering,Reinhold Publishing Corp., New York, 1952.

4. Bernice K. Watt and Annabel L. Merrill, "Composition of Foods-Raw. Processed,Prepared," Agricultural Handbook 8, Dept of Agriculture, US Govt Printing Office,Jun 50.

5. E. F. Hodges, "Consumption of Feed by L.vestock," Production Research Report 79,Dept of Agriculture, Economic Research Service, undated.

6. F. B. Morrison, Feeds and Feeding, Morrison Publishing Co., Ithaca, N. Y., 1956,22d ed.

7. G. Allen, 'Supplement for 1966 to Livestock Feed Rela onshins 1909-1965," 1966supplement to Statistical Bulletin 337, Dept of Agriculture, Nov 66.

8. D. B. 1bach and R. C. Linberg, "The Economic Position of Fertilizer Use in theUnited States," Agricultural Information Bulletin 202, Dept of Agriculture, EconomicResearch Service, Nov 58.

9. K. D. Moll, J. H. Cline, and P. D. Marr, &Postattack Farm Problems, Part I: TheInfluence of Major Inputs on Farm Production," Stanford Research Institute, MenloPark, Calif., Dec 60.

10. Dept of Agriculture, informal advice, 28 Jun 67.11. Dept of Commerce, Office of Business Economics, 'The 1958 Interindustry Relations

Study," Nov 64.12. D. B. Ibach, J. R. Adams, and E. I. Fox, "Commercial Fertilizers Used in Crops

and Pasture in the United States-1959 Estimates," Statistical Bulletin 348, Dept ofAgriculture, Economic Research Service, 1964.

13. Dept of Agriculture, unpublished material provided by the Farm Income Branch,Economic and Statistical Analysis rivision, Econon.ic Research Service.

14. M. R. Goldman, M. L. Marimont, and Beatrice N. Vacara, "The InterindustryStructure of the United States: A report on the 1958 II put-Output Study," Dept ofCommerce, Office of Business Economics, Survey of Current Business, Nov 64.

15. National Economics Division Staff, "The Transactions Table of the 1958 Input-OutputStudy and Revised Direct and Total Requirements Data," Dept of Commerce. Surveyof Current Business, Sep 65.

16. "Personal Consumption Expenditures in the 1958 Input-Output Study," Dept of Com-merce, Office of Business Economics, Suivey of Current Business, Oct 65.

17. 'Additional Industry Detail for the 1958 Input-Output Study," Dept of Commerce,Office of Business Economics, Survey of Current Business, Apr 66.

18. Dept of Commerce, Office of Business Economics, "1958 Input-Output Study of theUnited States: Data on Cards/Tape," undated.

19. Dept of Agr culture, "Economic Tables," from Farm Production Economics, NationalResource Economics, and Economic Development Divisions of the Economic Re-search Serm ice.

20. J. L. Preston, et al, "Output and Employment Estimates for 1-0 Sectors, 1958-1966,by Quarters," Jack Faucett Associates, Inc., Jun 67.

65

21. Dept of Army, Office of the Chief of Engineers, Engineer Strategic Studies Group,OUnited States Labor Force 1975," Jun 66.

22. National Research Council of the National Academy of Sciences. Food and NutritionBoard, 'Recommended Dietary Allowances," 1963.

23. Dept of Commerce, Bureau of the Census, 'Population Estimates,' Cmurpnt Popula-

tion Reports, Series P-25, 352, 18 Nov 66.24. R. S. Pogrun, RNutrition in the Postattack Environment," The RAND Corporation,

RM-5052-TAB, Dec 66.25. Dept of Army, 'Nutrition," TM 8-501, 28 Apr 61.26. G. J. Novak, 'American Industry in 1976 and 1985: Projections of Ootput Employ-

ment, and Productivity," National Planning Associates, National Economic Pro-jection Series, Report 64-1, Feb 64.

27. Statistical Section, current issues of Survey of Current Business.28. M. K. Wood, 'PARM-An Economic Programming Model," Management Science,

11:619-80 (May 65).29. H. E. Fassberg, D. J. Igo, and I. Moder, 'Industrial Readiness Army Mobilization

Model," RAC-T-438, Research Analysis Corporation, Dec 64.

6i

66

I *

UNCLASSI FEDSecudty Classification

DOCUMENT CONTROL DATA - R&Df9,.i..y classification .1 fist*. booy of abstrect and indoxind annowtation. must be entSedi h0 overall report to ecae.Ified)

IO I@ItNATIN ACTIVITY (Corporate author) 2 REPORT SECUR TY C LASSIFCATION

Research Analysis Corporation UNC LASSI FIEr,McLean, Virginia 22101

, i. RP130 111 T, T.A

A MODEL OF TECHNOLOGICAL CAPACITY TO SUPPORT SURVIVORS OFNUCLEAR ATTACK

4 OESCRIPTIWE NOYZ (?irpe .1 ,*por em Inclusive deft&)

Technical paper11 AUTHO S() (Firai aPmie, mindde initiof, lou nane)

Bernard Sobin

4 RtPORT DATE 7. TOTAL NO. OF PAGe[ 7b. NO. Or

RiesSeptember 1968 75 29

to. CONTRACT OR GRANT IhO, Do. ORIGINATORS RpoRT NUMSeanS)

DA*C20-67-C-0137PR*oJec€, waL RAC -. P-313

OCD Task 3534BC. Sb. OTHER R SPORT NO(S) (Any alA., niumbers Met m~ay be e.1,dne

107.101 fi

d.

Io. DISTRIBUT;ON STATEMENT

This document has been approved for public release and sale; its distribu-tion is unlimited.

11. SUPPLIMSNTARI' NOVIES 1. SPONSORING MILITARY ACTIVITY

Office of Civil DefenseWashington, D. C.

IS. ADSTRACT

The paper presents a linear programming model of production activities ofthe US economy after nuclear attack. The model has many alternative combina-tions of agricultural production for meeting specific nutritive element requirementsof the population and one production process for each nonagricultural output. Th -

processes for nonagricultural output are those of the 1958 Interindustry Model ofthe US Office of Business Economics; the agricultural activities draw on manyother sources of data. Rows of the model generally state that the cumulative prod-uct of variables and their coefficients must be less than or equal to either zero orsome stipulated capacity. In the former case the typical row states that inputsrequired minus production of the item must not be greater than zero. In the lattercase the requirements refer to capacities, which cannot be produced in a staticmodel, and total requirements must not exceed the stipulated capacity remainingafter the nuclear attack.

DD NOV0.1473 UNCLASSI FIEDSecurity Clausiflcation

U NC LASSI FIEDMcurity aari.o OE U

KK 690 LINK A LINK is LINK r1OIZ" ROL9 U? T OE *

input-output models for US postattack survivalinteriiduastry models for US postattac' survivallinear programming for US postattack survivalpostattack capacity of

US agricultureUS economyUS transportation

UNC LASSI FIEDIsemwty 0488ifteetwe