Created: 10/1/1978

OCR scan of the original document, errors are possible

USSR:econciliation Of Marxist and Western Measures of National Income -





repared for Ike use of US. Government offiefaib. "The fonn-l, covnw and content* of the publication are designed to meet theequirement* of those users. US. Government officials may obtain additional codes of ilris document directly or through liaison channel) from the Central Inli-lllwnce Agency..

overnment user* may obtain this along with similar CIA publicationsubscription basts by addressing Inquiries to:

i * '*

I Document Eapedltlns (DOCEX) Protect !; Each*nee and Gift Division Library of Congress ! i! Washington.

rrvrrmrvent users not Interested In the DOCEX may purchase reproduction) of specific publication) on an Individual basis from:

c4odupllcahon :Library of; T" lilillllll

USSR:econciliation of Marxist and Western Measures of National Income

[ CMrW Intelligence Agency

National Foreign Assessment Outer

The Soviet Union and other Communistemploy the Marxian concept of national income rather than the Western concept of gross national product (GNP) for measuring national economic activity. The Marxian measure has been thoroughly analyzed and rejected byeconomists.

The USSR began publishing national income data in the, and this in turn stimulatedconceptual andthe relationship between Soviet national income (SNI) and GNP. Among the earliest products of this researchargely theoretical article by Nove' The United Nationset of GNP accounts for the USSR5 and attempted to reconcile its estimates with the scattered Soviet references to national income available at thatove and Zaubcrmanetailed Western estimate of Soviet national income9 uiing the Marxist constnict.'

The UN haseader for many years in the development of national economic accounting and haseries of manuals onstandards for GNPhe UN has also furthered the development of the Soviet system of accounts and recentlyechnical manual originally developed by the Council for Mutual Economicn integral part of this work has been efforts to develop the theoretical and.empirical linksthe Marxist national income and Western GNP systems. During thehe UNeries of meetings and working papers comparing the two systems. Among the issuances the most notable were apaper by Seton and an empiricalexercise by Ruggles and Ruggles using the8ecently, the UN haseexamination of this area and hasomparative accounts methodological manual to accompany the latest revision of its GNPThe UN is planning to publish anvolume with empirical results for several nations.

There has alsoistory of estimates of Soviet GNP produced by Western economists. Because these estimates are prepared from data whose coverage and accuracy are oftenWestern investigators attempt to develop checks. One such procedure is to derive the

' UN. Basic Princip'.et ofike System o/ Balances of ike Notional Economy, New

' F. Seton. Comparisonroduction Concepts, published by thecexf. Eur.ovemberRichard and Nancyemmtrtson of National Aemmts Data far Ike Untied Stem Classified According to Ike Concept' of ike United Nations System of NmtPtnaJ Accounts and the Materiel Product System, mer-KjeineJum prepared for Ihe US Barcae of the Budget for wbrntaatoa to the UN,

' UN, Comparisonhe Systemational Accounts and Ike System of Baltneet cf Ike National Economy. Pan One. Con-eeptttaJ Relailensklps. New

estimate of Soviet national income implicit in the synthetically derived GNP accounts, thenit with the officially publishedeiding authority in this field, extended his workaper on Soviet national income practices that included methodological estimates of the relation between Soviet and Western systems.*

"I In this paper we will summarize theprinciples involved in comparing tbe Marxist and Wc-tern concepts and apply them to CIA's estimate of Soviet GNP7 Earlier work is extended in two areas:iscusses the foreign trade accounting problem in some detail, andemonstrates the possible use of input-output tables for connecting the twoll of the data limitations of previous Western GNP estimates still apply to this empirical reconciliation, and! manyproblems still remain.

Another area of great uncertainty in Soviet economic accounting is the treatment of defense expenditures. With the recent publication, by the CIA, of independent estimates of Soviet defense expenditures it is useful to set out tbe relation of these estimates to the GNPhe final section of this report is an initial attempt in this direction. Again it does notefinitive effort, but indicates that thereeneral consistency between the two estimates and points out several areas that need further research.


Before discussing the differences between gross national product and Soviet nationala few basic definitions should be set out.

Gross national product (GNP) is the total value of the final sales of goods and services at iT&rkct prices produced by the laboriven nationiven time period, It is gross product rather than net because nois made for depreciation and other capital consumption allowances. GNP less capitalallowances equals net national productther goods and servics charged to current expense are deducted, however, so that GNP includes only final sales.

The principal components of GNP arc:consumption expendituresurchases of goods and servicesross investment in fixed capital and inventoriesnd net exports of goods and services.

Personal consumption expenditures include the goods and services purchased by individuals, the value of goods and services rccc:ved in kind by individuals, and the operating expenses ofinstitutions. The United Stales divides PCE into purchases of durable goodsfurniture, radios,ondura-blcs (food, clothing,nd services (rent, medical, financial, educational, transportation).

Goverrunent purchases of goods and services include wages and other goods and services purchased by government institutions. Thisvaluer, government services at theircost and makes no allowance foror profits. The United Slates also includes government purchases for investment purposes.

Gross Investment Includes the value of gross additions to fixed capital and the change in the physical volume of inventories. The United States includes private purchases of dwellings and investment by nonprofit institutions, but excludes government investment.

Net exports is the exports of goods and servicesomestic activity) less imports of goods and services (included in PCE, G, or I, butomestic activity).

Reflecting Western economic theory, GNProduct- or demand-oriented concept. It can also be viewed as an income or supply concept. In Western accounting this consists of two parts: national income, or payments to the factors of production (wages, profits, rent, and interest) and rwnfactor payments (depreciation, indirecttaxes, and subsidies).

Soviet national Income (SNI) is also aconcept. It represents the netin the production of material goods andthe limited number of services thatrequired to bring the material goodsfinal sales point. These services arecommunications servingof material production, agriculturaland the wholesale and retailIt is net rather than gross valuecapital consumption allowances areThe main branches of materialare industry (including mining andagriculture, construction, and acalled "other" branches! ofThese branchesheabove make up the productive sphere.branches form the: j; :"

' SNI is computedesidual. First, the gross value of output of each branch is computed from its production data. Then tbe purchases by these productive brancheshe material services listed above,apital jconsumption allowances (essentiallyare subtracted. The remainder is SNI. |Included are wages and other employeeprofits, turnover and other indirect taxes, ^subsidiesegative value, purchases of non-material services, interest on bank! loans, social insurance contributions, and other smallubtracting deoreciation makes SNI moreto NNP than GNP, but only in the productive sphere.

'. ; ;

Depreciationpecial type of material Input. SNI is intended to measure newly created value and must be net of the goods completely consumed in the process of production. Capital assets, by definition, participate rcpetv.edly in the productive process without physically beingThis gradual transfer of value isby calculated depreciation allowances.apital asset finally is retired, the accu-mulatcd depreciation allowances are compared with the original costookkeeping entry, plus or minus, is made to equate them. This additional allowance is also considered ainput for the year of retirement1*

SNI is then distributed directly orby the Soviet financial system until the various end users purchase the material products for final use. The domestic final use of material products is published as Soviet national Income utilized for consumption and accumulationt does not include exports, but does include imports of material products. Capital losses also are not included in SNIU, presumably because they were notosses include damage to fixed capital due to natural causes such as earthquakes; abandonment of unfinished construction; and spoilage of agriculturalTherefore SNIU is less than SNI by the sum of net exports of material products plus capital losses.

SNIU includes personal consumption,purchases of institutions serving thematerial purchases of science andthe growth of fixed capital, and the growth of working capital and reserves. The sum of the first three categories equals totaland the sum of the last two equals accumulation,

Personal consumption (PC) is roughly equal to consumption of durable and nondurable goods in the US GNP accounts, with some differences. For example, utilities areervice in PCE,aterial good in PC. The most significant difference Is the inclusion ofof housing in PC. This reflects the same reasoning that leads the Soviets to treatin the productive sphereaterial

This bookkeeplni entry Is nor mil lysubstantial po*itive value because Soviet depreciation allowances arc based on optimistic estimate* o' service lives.

t purchase. Depreciation represents the valuehe housing consumed in the current yearence is part of PC. As in the productivehe depreciation is calculated rather thanonetary expenditures.

;Thc nonproductive sphere is divided intoroups. The first is often labeled consumer sen-ices in the West, and consists of the following principal branches: housing, communal services, and public utilities; education, culture, and art; tand health, social security, and physical culture. The second, sometimes labeled state senlces, consists of science and scientific services;cicdit, and insurance; generaland other branches of the nonproductive sphere. The published data on SNIU include separate totals for the material expenditures of these two groups. As in PC, depreciation is included in both totals.

The growth of fixed capital is calculated as gross investment plus capital repair expenditures less depreciation allowances, the growth ofconstruction, and losses. Depreciation is the sum of the depreciation in the productive sphere that was subtracted in computing SNI and the depreciation included in the consumptionof SNIU.

The growth of working capital and resents represents the net growth of inventories, the growth of unfinished construction, andstrategic reserves. The latter are thought to

include military as well as civilian goods.

The Difference Between GNP and SN!

With the accounting definitions in mind, we can examine the differences between GNP and SNI moreirst, on the'product or expenditure side, start with SNIU and work toward GNP. Adjust Soviet personalto PCE by subtracting depreciation onand adding personal expenditures forSecond, delete the material expenditures of

"The theory and rational* behind the vartouiLccounUnfwill not be developed In anyealth of material hai already been publlihed on the lubjeet. For cum pic iceystem of Vattanal Accounts, Newor dbcuulcra of Weatern ce-nccpu and B. P. Plyihevikly'islon-jconcept!.

the services that sell their output. These include passenger transportation, personalhousing, utilities, and miscellaneousservices. The remaining services arc joined with the material expenditures of science and administration to form the current purchases of the government sector. To these material expenditures, wages and purchases of services must be added, and depreciation must beThe sales of services are now included in PCE and G.

Having adjusted PCE plus government outlays for goods and services, we now must adjust accumulation to equal gross investment. Asearlier, this entails adding depreciation and losses. The result mil be different from Soviet published investment data. The largest difference is that capital repair expenditures are included in the Western concept of grossbut not in Soviet published investmentmaller difference is that Sovietdata include all current expenditures on machinery and equipment, construction, and other capital outlays, while the SNIU category, the growth of fixed capital, includes the value of completed investment projects, some of which began before the current year.

Finally, we must make adjustments for foreign trade. Soviet foreign trade is an accountant's nightmare because it is computed in two sets of prices. The regularly published data are intradehich represent thevalue in foreign currency converted to rubles at the official exchange rite. These prices often differ greatly from domestic prices, hence net exports will generally be very differenton which prices are used. Appendix Athis problem at greater length. SNIprocedures appear to use foreign trade prices while Western estimates of Soviet GNP usually use domestic prices,et exports in foreign trade pricesillion rubles whileomestic prices they areatillion rubles.

What is the net result of all the aboveFirst, SNIU has been increased by the amount of productive depreciation. Thefor nonproductive depreciation and losses

only rearrange the data. Second, there has been an increase in the value of services. To sec what this increase is, divide the production of services into three components: those sold to enterprises and other intermediate consumers; those sold to individuals, the government, and other finaland those provided by the government for free. SNIU includes the materialfor all three categories. GNP does not include any of the first category except as the value added in these services is reflected in the valueood or service included in GNP by end use. They arc part of intermediateThus, GNP is less than SNIU by the amount of material expenditures involved in their production (includingor the second category, GNP incudes personal and government expenditures on services at market prices. Therefore, GNP is greater than SNIU by the value added of services sold to individuals and governmentshe value of any other services that are purchased in order to produce these services. The third category of services is valued at cost in GNP. Thus GNP exceeds SNIU by the wages of government employees. Depreciation of government-owned capital stock is subtracted to be consistent with US accounting standards. In practice, services sold to enterprises are probably very small relative to those sold to individuals and governments plus those produced by the government at cost. Therefore, thethat GNP is greater than SNIU by the value added of the service sector is not greatly in error. In fact, subtracting the materialof the first group is equal tc adding its value added and subtracting its gross output. Therefore the net adjustmenthe service sectors is equal to adding the total value added, subtracting services sold to the productive sphere and subtracting depreciation on governmentstock.

In sum GNF is greater than SNIU by (I) the value of productivehe net adjustment for the value added of, services as described above,he change in the valuation of net exports (normally negative).

The same result can be obtained bythe problem from the income side. SNI is the difference between gross output and material purchases (includingnd therefore includes any purchases of services by thesphere. To calculate GNP in the productive sphere, wc add productive depreciation to SNI and subtract purchases of services by productive sectors. Next we must add the GNP originating in the nonproductive sector. As above, this is the value added of the services which are solddepreciation) pluses ofemployees. We therefore arrive at the same net increase in the value added of services. Finally, SNI includes profits received by the government from foreign trade operations. When net exports are revalued in domestic prices, these earnings must be subtracted from SNI. Thus, we arrive at the same set of adjustments regardless of whether wc work from the income or the product side of the accounts. The preceding discussion can be illustrated mathematically by using an input-output table. This exercise is shown in appendix B.

In practice, the comparison of the GNP and SNI of the Soviet Union is not easily carried out. In the CIA estimates0 Soviet GNP, attempts were made to reconcile total GNP with total SNI, and PCE with Soviet personalWhen total GNP was adjusted to the SNI concept, it1 billion rubles higher than actual SNI. Consumption was considerably closer: the adjusted PCE value exceeded the published personal consumption valueillion rubles. The following shows the specific steps involved in0 SNI to GNP using the method described above.

Remove housing depreciationillhn rubles."

Add PCE for services (seen

Subtract the mateiial expenditures bythat sell theirbillionThis datum is not available in CIAs

USSR, Panonol Consumption Eapeodlhjras (PCE) for

SatMcot. Bflhoe Robfa


Trade union and obW tl

- M

Pmowal i. - 7i

Repair and ptnonal earn

Recreation and


accounts. It Is estimated here on the basis of data from two Soviethe first monograph provides data on totalexpenditures (including depreciation) for several nonproductive branches. The second provides data to estimate the share of depreciation in the total expenditures. Paid services were assumed to becommunications, housing andeconomy, entertainment, andservices. Total material expenditures were estimatedillion rubles and depreciation was estimatedillion rubles.

the wages and purchases of servicesgovernmentbillionwages and social security deductions

illion rubles) were taken fromIAo this was added anate of services purchased byillion rubles. Tbe latter had to be estimated from data on the value edded In these services and purchases by the private sector. This estimate is thef this reconciliation.

depreciation in thebillion

M. Ruiiaytsr. Rtiuny ftvtltjm ntprottvoditvtnnor ifrry, Mravcow.nd I. M. ShneMarmin. StallMk* aiiltf,

GJVP. Tab liatn hi Ika urn of wage* and social security oVaaetioaa la education, art. col'ore, health, pfcrvcal cahara, science, it*itChilianadmiakstratlvt organs of aodal orgaabattona, mvaktoal nrr-tcca, and tha military.

op.. 1ST.

Add total depreciationillion rubles. This is the sum ofbove, plus productive depreciation and the residual book value of the retired productive capital as calculated in the CIA accounts,

Add losses to accumulation and delete this category froffrbillion rubles. This was calculated as the total difference between SNI and4 billionless net exports in foreign irade9 billionhis does not affect the value of GNP, only the estimate of investment.

Convert net exports from foreign trade prices lo domestic7 billion rubles as calculated inthe CIA accounts.

The results are summarized in tablehe comparisons do not agree precisely with those published in the CIA accounts because of slight changes made in those accounts since theirand because of the new estimates made in adjustments I,bove.

Two possible sources of error could account for the discrepancies in the comparisons ofand government expenditures. First, it Is widely believed that significant expenditures for military hardware are included in the SNIU accumulation category. These should befrom investment to government. Since the amount of these expenditures is unknown, we will assume here that they are just enough to make the estimate of SNIU adjusted investment equM to that estimated here. Second, some of the change in unfinished construction may be double counted in the CIA accounts. The details are discussed below (see. Here, we will only note that If this Is true then CIA's estimate of investment might be reduced by as muchillion rubles. Combining these two factors would reduce CIA investmentillion rubles, require the transfer4 billion rubles from SNI investment to SNI government, and raise CIAs estimate of government expendituresillion rubles5 billion rubles. The SNI government total would2 billionTotal GNP estimated by CIA would not


ussr. reconciliation of gnp and sniu0

billion rubin



material eipendhum of nld mfvtetll.-tl waate end purchuee of hevtcee by tovem-


demwirtki of


nh -to

got dsetved free sni

gnp eetineted by

change because government expenditures are defined to be equal to government incomes in the CIA accounts.eduction in onecauses an increase in the residualcategory, outlays not elsewhere classifiedore research on these issues is needed, however,inal assessment can be made.

The large difference between total GNPfrom SNI and that directly estimated byillion rubier) is the resultarge category of unallocatedillionin the CIA accounts. Depending on how much of this unallocated income arisej in the nonproductive sphere, most of the gap between the GNP estimates can be closed.

Inrequent use of the relationship between GNP and SNI is 'o make comparisons of Soviet GNP with the GNP of WesternThisarge number of practical problems, althoughew will be mentioned here. One such problem is the definition of production boundaries: what is intermediateond what is final use? For example, many government activities are really services provided to business for free or at nominal cost and do not represent their final use. In the Soviet Union, where so many services are provided by the government, this question is more significant. The Soviets consider the science sector to be nonproductive. In comparing Soviet and US






70s 1

GNP, therefore, one must decide whether to include science expenditures in Soviet GNP or toarge share to account for private research undertaken by US private business and included in the price of their products. Questions of this type are often decided arbitrarily.

The sharply different institutional structures of the Soviet and US economies also create problems. For example, the Soviets publish data on amortization deductions. These data represent all money actually set aside for depreciation in the productive and nonproductiveost Western constructions of Soviet GNP use this value for their estimate of total depreciation income, reasoning that since the US accounts make no allowance for depreciation ofcapital, no allowance should be made in the Soviet cose. This means, however, that theof government capital stock is notThe government sectorarger share of tho economy in the Soviet Union than in the United States.epreciation allowance be computed because it is relatively morein the Soviet Union, or should depreciation be omitted, thus understating Soviet GNP in comparison with the United States? This is especially important if comparisons are to be made In the same monetary values, such as US

" thin i* ml quite accurate. collective farmi set money aside for depreeuilon. thisot normally reported, but must be estimated.

and Soviet GNP 'n dollars. The size of the ruble-dollar ratio used lo convcr* Soviet GNP to dollars depends on ihc coverage of GNP in both nations.

Olher significant problems for comparisons of GNP include:

Government Investment. In the United States, government investment in fixedis lumped together with currentexpenditures. Since the USowns relatively small amounts of capital, this iserious divergence from theory. In the USSR tlv same procedure obviously cannot be followed. How then does oneomparison? The practice has been to estimate US government investment and alter the US accounts.

Tax laws. US business often charges its capital repair expendituresurrentbecause this reduces current taxable income. The US GNP accounts reflect this practice. In the Soviet Union they are capital expenditures. Since there is no practical way

, to estimate US capital repairpractice is to reduce Soviet capitalby SO percent, an amountto be included in current expensesUS. j

and subsidies.umber of imputations are necessary in both the US and Soviet accounts, valuation procedures willirect effect cm any comparison. Agricultural income-in-kindrimeMany food products are heavilyin the USSR. Should householdof these products be valued at the subsidized retail price or at; theroducer would have received if there were no subsidy? The imputation for the rental value of private housing In theimilar problem. The United States makes imputations for financial service charges. The Soviet banking system, however, Isdifferent from the US system. Should the same type of an imputation be made as in the United States? Currently no financial service charges are imputed for the Soviet Union.

The problems discussed above illustrate just some of the difficulties that arise in comparing Soviet and USifferent, but equally long lisl could probably be compiled withto comparisons of Soviet GNP with the GNP of any other Western nation.

Defense Expenditures and GNP

One of the most important aspects ofSoviet GNP accounts is determining the relationship of defense expenditures to total GNP. There are two methods of examining this. First, an independent estimate of defensecould be explicitly inserted into the accounts and tested for consistency. Second, defense expenditures can be estimated as adifference between total expenditures and the sum of all identified outlays. The pitfalls of residual estimates arc well known: any error in any estimate will create an error in the the residual, which will always contain other small components beside that being measured.arge error range must be given to any residual estimate unless high confidence can be given to each other entry in the calculation. CIA has recently published independent estimates ofexpenditures0illionhe question addressed here is whether this is consistent with CIA's GNP estimates.

US GNP accounts consist of five accounts: current income and outlay accounts for thesector, for government, and for the rest of theonsolidated savings and Investment account;otal GNP account. For the Soviet Union, It has not been possible toaccounts in such detail. CIA hasprivate income and outlay accounts. The Soviet Union, however, does not publish enough data to separate the business sector from the government sector. Therefore, CIAonsolidated public sector income and outlay account. Savings and investment data arein the income and outlay accounts,eparate account is not formed. The total GNP account is the third and final CIA account.

"CIA, Estimated Soviet DefenseSpending In,

These accounts are shown inhe defense columns inreublished as part of the CIA accounts, but are developed in this paper.

Defense expenditures are found only in the public sector outlay account, and the income used for these expenditures are found only in the public sector income account. There are two apparent exceptions. Household incomes include military wages and military subsistence(tableinesnd 3b) and household outlays include military subsistence expenditures (tableinehe reasoning is as follows. First, military wages represent personal income at the disposal of individuals to be spent as they please for consumer products or services. These same wages are included in public sector outlays.

"io art taken from CIA, OKPp. W> and B. There haw-ew imatl ohangoi alnce thai publication wai baaed.e fleet the accounti currently usee.

The only defense-related concern with theincomes account is whether total incomes equal total outlays and if not, is ll because military wages were estimated incorrectly. In this case, there is an unidentified income residual9 billion rubles (tableine 4b) with military wages identified2 billion rubles (tableinehus, it does not seem that any reasonable increase in military wages couldthe residual although there is room for some increase.

The second defense-related entry in thesector accounts i; military subsistence, an imputed income and expenditure. It represents an estimate of how much the Soviet Ministry of Defense would have to increase military wages to permit members of the military to pay for the food, clothing, and other items now received free. If this were the case, then militar* wages would be increased by the amount of estimatedand personal expenditures for food and


b Military aufctlitnica

money Income cttrrentl* earned ami ttrtKfcal

a Private monrv Ineorm ctnrvnlly

b Unidentified money Inerana and tttfitleal

net rant , , fl

& Imputed vahaa of ownat^aoppoed building- ,

atal taeiaaiearned

NtadamatW algajgawa


* Inlaw at tacaaaa036

d Nat new bank lear. to

otal Incoaaa

tobu 4



rrlill sales of acods for cor*imptloo

eoopeeauve, and commission sain .

(arm es-vtfiesje mart* sales .consumer services

union and other

) caa imi"

imputed net

c other



epair and personal can

' eckettofl and




a. farm consurnptlon-ln-klndmilitarytotal outlays


in* eetruent-

otal outlays for eoo wrapt loo and


s. net

b. dtrect

a other payments to the start




clothing would be increased by themount. Thus, the amount of military sub* sistenoe expenditures affects total household in-comes .and outlays but not their balance. Like military wages, defense expenditures forproducts are recorded in public sector outlays, and like wages, military subsistenceivilian equivalent in income imputed for food and clothing received by nonmilitary employees because of the nature of their job. Thefor nude's uniforms, for example, wouldart of health expenditures.

Public sector outlays encompass Soviet defense expenditures, but the particular method used in CIAS accounts requires an examination ofsector Incomes as well. The major portion of defense expenditures should be contained In the categorytablehis value is determined as the difference between total public sector incomes and identified public sector outlays. Thus, any error or omission in any entry of the income or the outlay account will directly affectnd hence theavailable for defense expenditures.

Defense expenditures are not wholly contained inc Due to the manner in which the Soviets report their economic data, defenseare almost certainly included in other public sector outlays. Expenditures for research and development surely include expenditures onimilarly,expenditures may be included ineducation, health, physical culture, and administration. Defense expenditures arenot In art. general agricultural programs, forestry, municipal services, or capital repair. There is relatively little information on which to base estimates of defense expenditures contained

lUSSHi Public Sector


Income retained by

. Retained Income of collective (anna

profit* of

profit* of cotuuinee cooperatives

profits of other eeaanisaitona

to rconomJe enterprises for special fond*

Insurance end loclal aeeurltv


eiee end other payments lo the

on Incurs* of coBeettm hrew -

on tnconw of consumer ccopemnves and other

c Deductions (men profltt of state

buwmcc for sobatdlard keaea.

total chargee asjalaat curreai product, net of,

Depn >

Consolidated total charm agaleat current


> Net savings of Souk

payments to Ihe state

1 Consolidated set Income

these other categories. The following areestimates. I

There areillion people in the Soviet armed forces outopulationillion peopleercent If wc assume that they receive the same health care as the rest of the population, then defense would accountercent of totalillion rubles. The Soviet Union claims that almostillion people received education of one sort or anotherthe Soviets do have an extensive military education system, it seems unlikely ;hat much moreercent of toul outlaysillion rubles, could be military.

Expenditures for state admlnlsiration probably include substantial defense expenditures. It is likely that most expenditures for wages and materials for the current operation !of theof Defense headquarters and of majorheadquarters arc In this category. Of the total expendituresercent,illion rubles, could easily be defense.

Physical cultureery small category,illion rubles (tableineilitary expenditures are certainly lessillion rubles and are ignored for this exercise.

Research and development is heavily military. Total expenditures are estimated toillion rubles (tablestimates ofange from SO toercent of. The Soviet published sciencevalue, on whichillion rubles is based, is thought to be understated by aboutercent. Thus, if we increase this total byercent and take an intermediate share ofercent,ould berabies, p ;

|;: ;

Finally, there is evidence that some defense expenditures may be included in outlays on new fixed investment and inventory change. Defenseonsiderable volume of itemsused in civilian activities. These items, which would be considered part of investment in the civilian sphere, include trucks, cars, cranes, forklifts, transport ships and aircraft, and

j i6

USSR. Public Sactorj

Cwnuul Serrtoo*




d I'hukili Culture General admWatratlv* end i. CaanonJ apfciahwil preapaoai _

t- form ccoaoaay .

e. State adminMraiw* (appaxaO-

d Municipal and Mated awvtoai



- i;


AdrnlnMratlve onaaa of aortal

Cron Invntment a. Plied capital b-

Rneerch and dgaafapmaal (civilian and military)

Outlay, avax. (cWenaa.oru. and unkWndlkdnd atatlatkal

dlacrrpancy ;

Consolidated total aaVat of food, and am Ice. eichadn of itlr. to


Traaalc* outlay.

a. ivrn'iii and-

b Stipend.

c Interest paymeM*et naw bana loaaa CooaoUdated total oatlayi

Billion Ruble.

i aaa


equipment. Direct cost estimates of these purchases0 rangeillionillion rubles. Soviet publications dealing with investment, though somewhat ambiguous, seem to imply that such: expenditures areas investment. The midpoint of theillion rubles, will be used here. !

Finally, defense expenditures for construction me almost certainly included in investmentfor construction. Estimates of the defense portion rangeillionillion rubles. We shall use theillion rubles, in this exercise. Thus far we have allowed for defense expenditures as shown indding the total to, ofillion rubles (tablee now have potential defense expenditures9 billion rubles. Next, we must exclude ihc nondefense expenditures included in. Asin CIA's GN.I* accounts, ihesc arc net exports, outlays on the militarized KGB, and changes in strategic reserves. CIA estimates the latter two categories to totalillion rubles."

* CIA, ONPAKowtlt.

When estimated in domestic prices, netare betweenandbillion rubles. This raises the amount available for defenseillion rubles, or just into the lower end of the direct cost rangeillion rubles. The interpretation of foreign trade accounting is important because itevere limit on sources of additional financing for defense. As mentioned in CIA's GNP accounts, therearge budget income residual, almostillion rubles. CIA takesercent of this residual as current Income, but states that up toercent

USSR. Gross notional Product in EitoWlibeo Prices by End0

Billion Rubles


Goods _


Soft foods

Durables j

Trade union and otbtr


ersonal conuni

Repair and personal care

Recreation, art, and physical.

Education ,



achinery and

onstruction and other capital

et addition to livestock taventorku ..


c- Imreritorka

public sector

a. Ceneral admlnlstrailve and miscellaneous servicer

Research and development (civilian and military)

c Outlays rte.c. (defense, net eaports. and unidentified oullsys) and


national product

USSRt Eiltmcrred Dnfam* Expenditure* In the Civilian0

Blllkm Rubles



This requires that net exports be entered Into public sector expenditures in foreign tradesurplusillion rubleshe net effect of this argument is that the funds available for defense cannot be increased by increasing the share of the budget incomethat is included in GNP.

could have been used. In fact, this increase is limited by the interpretation of foreign trade accounting. The basic hypothesis is that the profit accruing to the government from foreign trade operations is an income to the budget. If true, then this profit canurrent income only If net exports are valued in foreign trade prices. If net exports are in domestic prices,he foreign trade profits cannot alsourrent income. This point is developed in appendix A. Given the size of the budget income residual and foreign trade earnings, only aboutillion rubles of the budget income residual could be Included in GNP, even with no allowance for noncurrcnt income. CIA currently usesillion rubles, whichinimal allowance for noncurrent budget income, Conversely if one takesercent or more of the budget income residual as public sector income, then theforeign trade earnings implicitly are in-

Before looking for additional sources of defense financing, an alternate possibility for testing the GNP accounts will beses an input-output) table to describe the difference between GNP and can also be used to test defense expenditures. For this we look primarily at the final demand quadrant ofn appendix B. Here we have columns for consumption,government, and defense. Until now we have looked only at the total values for these columns in our attempt to isolate defenseThat is, of total investment how much could be defense related? It is also possible lo look at the rows. Given knowledge of the gross outputector, say the food industry, and knowledge of its interindustry sales, then we have the sales of the food industry to final demand. If we can then determine how much was sold to private consumption and how much to thecomponents of government we then have an implicit value for the sales of the food industry to defense. How docs this value compare to the direct cost estimate of defense purchases from the food industry? Unfortunately, CIA estimates of defense expenditures arc0 prices, and we do notable for the Soviet Union. CIA hasilot study of this nature, for the machinery industry only. This is probably the most important single sector to examine to determine if the order of magnitude of defense expenditures is correct, but the entire economy needs to be tested to narrow the range of uncertainty around total defense expenditures.

In order to test the machinery sector, CIA collected several estimates of0 gross output of machinery and put themikely rangeillion rubles. It was thenthat the ratios of interindustry sales to gross output in6ables were the same. It was assumed that0 ratio wasarrow range around this ratio. The resulting estimate of the sales of machinery to final demandillion rubles. This range includes some sales by the sectcr,, to final demand other

r il 1

than consumption. These sales were thought likely to be militaryhe second step was to estimate machinery sales to personal consumption, civilian governmentinvestment, foreign trade, and capital repair. Subtracting these estimates produced aresidual that conceptual included onlymachinery. It docs not include all military machinery. For example, the residual does not include purchases of military machinery tothe current activity of military schools, hospitals, clubs, and so on;f the Ministry of Defense; capital repair of any military equipment, includingand exports. It does include the domestic production of: large hardware items such as planes, ships, tanks, missiles, space hardware, and nuclear weapons; small arms; ammunition; radar and other electronic equipment;equipment such as typewriters, pots and pans, andans; and machinery used in the current repair of all mili'ary equipment.

For each of the civilian end uses CIAa range of values, the width of which depended upon the source of the basic data and our knowledge of Soviet data. The total value of these four civilian categoriesillion rubles,est estimate ofillion rubles. Subtracting this from total deliveries ofto final demandange for the machinery residualillion rublesest estimate5 billion rubles.

The components of the direct cost estimates of defense expenditures were then adjusted toan estimate of defense-related machinery purchases. With allowance for uncertainty, this estimate ranged betweenillion andillion rublesest estimate ofillion.

While the ranges computed from these two independent approaches are quite broad, they

"Since Ihti itudy wn completed, additional research onabia hai vindlcaied this position. The irata output and (he nonconsumotion final demand of the machinery sector have been Increased by amounts consistent with ihe Quantity usedhis money came lareely from iheector, as hypothesized in the ttudy.

can be said to beoincident in their absolute values and their best estimates.esult, CIA feels reasonably confident that the order of magnitude of the direct cost defense expenditures for machinery is correct.

Returning to other sources of defensethere are several possible sources of error in the residual method of calculating defense expenditures from GNP accounts. As already stated, the funds available for defensein0 GNP accounts depends on the values for every other entry In the public sector income and outlay accounts. The amount available for defense will rise if any income-is understated or if any outlay is overstated. The estimated data in these accounts are reasonably well founded on published Soviet data, withxception of the income from the budget residual discussed above. As seen at that time, CIA has already taken virtually the maximum possible from this source, and therefore it Isossibility for increasing the funds available for defense expenditures. The following arc some other possible sources.

Services sold lo intermediate consumption need to be excluded. The estimates ofuch as health, education, and physicalre derived from Soviet data on totalfor these services from the budget and other sources of finance. If any of these services are purchasedroductive sector foruse, then that amount should not be counted as final use. For example, if total health expenditures include any made by an industrial enterprise and included in its production costs, then that amount should be excluded from GNP.

The;total value of subsidies Is enteredegative public sector Income {table S,oviet literature indicates that some of thesubsidies may be netted against either profits or turnover taxes. If reported profits or turnover taxes have been reduced on thisthen gross subsidies should not bebut only the amount which has not been netted out. If this is occurring, it couldubstantial amount,illion rubles.

Investment0 prices Is assumed lo equal investment9 estimate prices. To the extent that prices for construction and producerincreasednvestment would be understated, and the money available for defensewould be overstated. Inflation in one year should not be great, but investment couldillion rubles.

Expenditures for state administration arefrom employment data. It is possibleubstantial portion of these people arc employed in productive enterprises or ministriesart of production costs. In this case, their earnings would be included in the prices of the goods produced, and hence should befrom GNP by end use. One billion rubles seems plausible.

The CIA accountsillion rubles for civilian police. This is computed on the assumptionarge share of the employment category "other branches of materialconsists of civilian police. However, it Is possible that, since they are in the sphere of material production, these employees maybe security personnel at productiveIn this case, their wages and otherwould be included in the cost of production, andillion rubles should not be included in GNP.

The total value and accounting cf science expenditures Is uncertain. Soviet data indicate that toul science expenditures are considerably higher than budgeury science expenditures. The source of these additional funds is not clear. CIA addsercent of these funds to public sector incomesharge to individual enterprises for special funds (tableinet is possible that the entire amount should be added, which would have the effect ofillion rubles to the residual available for defense. In addition, many Western students of Soviet science believe that the Soviet value for total science expenditures is understated by aboutercent. Depending on how the additional expenditures are accounted for, end use expenditures for science may

crease and defense expenditures 'might beor left unchanged.

The estimates of public sector expenditures for servicestableines I,nday be overstated by the value of depreciation. The CIA accounts conceptually do not include depreciation in these expenditures. Recentindicates that the published Sovietdata on which the CIA estimates were based, probably do include depreciation. In this cose, the estimated expenditures on serviceshould be reduced byillion rubles. This would add the same amount to defense expenditures.

There is some uncertainty about Sovietfor unfinished cons'ructlon.for the net increase in this variableillion rublesre certainly inSome may also be in Soviet data on working capital. If so, then this quantity should be deletcJ from the GNP category, inventories, which in turn would raise the amount available for defense. In the reconciliation of GNP with SNI, this change would lower investment and raise current government expenditures, but would not change total GNP. Decker discussed this question4 and concluded that Soviet data on working capital do not includeconstruct7 the Sovietsa statistical handbook that showsconstruction in the category "material workinghe other data in this table, however, do not correspond with the annually published data on the "commodity-materialportion of workinghe few references to the subject in Soviet accounting

f Becker. Abraham Sorter Uilitmry OuiUyihe RAND Ceirporauon, Santa trew toveiov raei, Suibtlka,

For example, Sarodnoye kkotytyitvo3 Gotfa. Sutlttlka..

books imply that the procedure is as Becker describes it. Thus, it is prcmalurc to assert that the entire value of unfinished construction is contained in working capital, but theieossibility that some portion of it is.

Depending on how many of these possible errors have in fact occurred, the residualfor defense could exceed the high end in the rangeillion rubles. For illustrative purposes, assume that the following revisions are made to the GNP accounts:

services sold to

productiveillion rubles

some subsidies as

netted against profitillion rubles

Inflation inndillion rubles

Administrative expensesin production costs illion rubles

Depreciation deducted from expenditures on services.. illion rubles

The resulting hypothetical GNP account, in an abbreviated form, is shown below with thecomponent now equal8 billion rubles:

Comments and queries on this paper areand may be directed to theDirector for Public* Affairs, Central Intelligence Agency, Washington, D.; area. For Information on obtaining additional copies, see the Inside of the front cover.



Because of the inconvertible cuirency used by the Soviet Union the treatment of foreign trade in SNI isimple matter. The Soviet Union hasetwork of specialized foreign trade organizationsach such organization is responsible for buying certain types of domestic productsranging for their export or for purchasing foreign goods and selling them to domestic enterprises. Because the exchange value of the ruble is artificial and relative prices in the USSR are quite different from those on the world market, the domestic prices within the USSR are often quite different from the foreign prices at which goods can actually be sold or must be purchased. In order to insulate the domestic economy from International prices, the FTOs purchase goods from domestic manufacturers at the domestic wholesale prices, less any turnover taxes, and sell them to foreign buyers at some agreed price. This price, when converted to rubles at the official exchange rate, may differ greatly from the domestic price and therefore the FTO mayarge profit orarge loss in the transaction. Similarly, profits and losses may occur when the FTO sells imported goods rt domestic wholesale prices, including turnover taxes.

ic profits or losses sustained by the FTOs are unrelated to their economic performance, the state budget is used either to skim off the excess profits or to make up losses. The question here is how foreign trade should be accounted for in SNI and GNP. The Soviets are noticeably reticent in discussing their accounting procedures. From the little evidence available, it appears the value of net exports in SNIU is computed in foreign trade prices, data for which are published in the annual statistical handbooks. These are not the prices for which the exports were sold by the domestic producers or imports were purchased by domestic consumers. Assuming that net profits accrue to the state from foreign trade, then these profitsust be enteredegative value on the product side, orositive value on the income side, of the accounts in order for them to balance. The Soviets appear to enter these profits as income produced in the trade sector.1

1 Tha Soviet method of aocouailnf. for forelin trad* In national Incomeuncovmd and developed by Vladimir Treml.omplete dlteoaaion ten, Vladimir O. Treml. DimltH M. Oalllk.osUnsky. and Kurt W. Kraftr. Th*comomy, NoSO.

i r

! '

The result of the foreign trade accounting is (I) toormally large profit due to price differences of internationally traded goods as part of SNIo enter what normallyery small net export value in SNIU. The profits, often called special earnings of foreign trade, can be considered as the profits accruingonopoly trade organization or as the tariffs collected on imported goods.

The Soviet practice generally has not been followed by Westernin constructing Soviet GNP accounts. In most cases the net export value entered into the product account is computed in domestic prices. Since large profits normally accrue due to the price differences, the net exports computed in domestic prices often have large negative values.

When the Western practice of including net exports at domestic prices is adhered to, one must be careful not to include as an income the special foreign trade earnings. If this happens, incomes are overstated relative to expenditures and the residual of incomes less expenditures is incorrect, fhis is particularly important since the residual is often taken to be an indicator of defense expenditures.

Htlzman extended Trend's work to GNPhere are two basicNP accountant can take to the foreign trade problem. First, one can consider the price differences toype of indirect business tax (in the caserofit to theubsidy (in the caseoss) and include the profits or losses with other nonfactor charges against GNP. In this case an imported good is considered to have entered the country at the low foreign trade priceomestic tax has been added to it to reach the domestic price. In order to remove the influence of imports from GNP so that GNP reflects only domestic activity, one should subtract only the value in foreign trade prices because the remainder is the result of domestic activity. Similarly, if exports areomestic subsidy then they should be treated like any other subsidized product. The sale is entered in the GNP account at the subsidized (foreign trade) price,egative value is entered in the income account soalance is maintained. With this approach, the net gains or losses to the government from taxing or subsidizing foreign trade are entered in the income account, and exports and imports are entered in the GNP account in foreign trade prices.

The second general approach is to assert that GNP should measure expenditures at market prices for domestic production. Subtracting imports at foreign trade prices does not achieve this objective; it must be done at market prices. One can consider the difference to be an import tariff which does not represent any domestic production. Similarly, exports must be valued at the domestic market price. In this case, exports and imports are entered in the GNP account at domestic prices and no entry is made in the income account.

D. Heton-r. Foreign Trade Under Central Flannlng. Unbent* Frew..

To further clarify these approaches, we can construct an input-output table. Let us divide the economy into three sectors: all domestic productionn import sector which purchases all imports and sells them at domestic pricesnd an export sector which purchases all domestic production intended for export and sells it abroadhows the transactions among the sectors. In the figure: d, m,epresent the intermediate purchases made by the three sectors in order to conduct their

figure 1

USSRi Domntlc Production cod orticM Trod*



Wen Id


: i; !

) represent the normal value added of the three sectors in the form of wages,axes, and)) are exports and imports in domestic)) are exports and imports in foreign trade prices; FD is the domestic sales to final demand (consumption, investment, and government) of goods and)) are the special earnings (losses) of FTOs due to the price differences. In other words, the special earnings of the export (import) sector are equal to what it sells its product foress the cost of acquiring its product)he value of purchases from other sectors; and its own value added.

In the figure all;imports are shown as being sold to intermediate consumption. Thisimplification because many imported goods are consumer or investment goods that go directly to their final purchaser. In fact, however, virtually every imported good has some value added to it by some part of the domestic economy, ifransportation charge from the port of entry to the delivery point. Thus, we can show all imports as sold to some domestic production sector which then uses the product or sells it to final demand. The purpose of subtracting imports from GNP is to remove the foreign content of the domestically purchased goods. This is not the price paid by the final purchaser, but the price paid by the importer. The difference is domestic value added. Thus, the FD variable in the figure includes the value

of imports and the* the subtraction of that value.

j s shown in; the)) are calculated as,


The figure as shown, illustrates the first approach described above. GNP is equal to domestic expenditures for consumption, investment, andplus net exports In foreign trade prices, orOn the income side,o convert to the second approach (that is, ONP measured at the market prices of domestic production) we would eliminate) term from the

.? i i


income side and also subtract it from final demand in order toalance. This is not quite the same as putting foreign tiade in domestic prices, as can be seen from the earlier equations)hese two quantities equal the profit (loss) due to price differences less thediate purchases and value added of the FTOs. Thus, after converting we have:


-) ' )))

Thus GNP equals consumption. Investment, and government, plus netdomestic prices, plus the intermediate expenditures of FTOs, plus theof the FTOs. Since thererOs, and they arebe khozraschei organizations, the last two terms should not beTheir sum, however, isillion rubles would probablygenerous estimate.|



Input-output) tables frequently can illustrate what would otherwise be obscure accounting concepts, such as the relation between GNP and SNI. This section reviews the territory already covered In the main body of this paper, but in terms ofable.howsable in the Western style in which the final demand quadrant equals GNP and the value added quadrant equals the charges against GNP.arge quantity of mathemaiical symbols are required. Tbe figure shows an economyroductive sectors in the Soviet sense, plus two service sectors. One service sector (xp) Is an aggregate of those services which are sold. The other (xf) represents the services provided free by the government. Subscripts are placed in parentheses. Thatj) represents the sales of the ith sector to the jth sector. Thus, thereroducing sectors in the Western sense. The final demand quadrant consists of sales ofectors to personal consumptionnvestmentivilian governmentnd militaryhe sum of these four! sectors is GNP. Net exports are ignored in this example. The value added quadrant consists of just two rows, depreciationnd other charges against GNP (V, hereafter referred to as valuehereouple of peculiarities in the figure which should be noted. Firstn accordance with the US practice of not including depreciation in the cost of government services. The cost of such services is simply the materials plus wages. Second, although xf is the sector of free services, it does appear to) to governmento the military.

This is just the technical way of inserting wages into the final demand columns. The xf sector buys only the wages of government employees and sells these wages to theigovernment and military. In this manner the column sum ofector will equal the current expenditures for these services and the column sum ofector will equal defense expenditures.

In figureheable is convertedable, This Is not the same as SNI. Statistically, It is impossible to distinguish between the current sales to Investment that represent the accumulation of new assets and those that replace worn outumerical total can be computed using depreciation rates, but not the detail thatable requires. Therefore,ractical master, the Soviets show depreciationeparate row between the interindustry quadrant and the value added quadrant. The final demand quadrant includes total investment. The sum of all entries in the final demand quadrant.equals the sum of the value added quadrant plus the depreciation row.

The major adjustment in going fromos the shift of the two service sectors out of the first quadrant. The value added quadrant (including depreciation) has decreased by the value added ofnd by depreciation in the service sector,n the other hand, it has been increased by the value of sales of services to the productive sphere, the sum of xp(i).

That the final demand quadrant still equals the value addeddepreciation can be seen from the following: in an input-output tabletoul is always equal to Its row total Thus, for thesector the gross output, xf equals its columnndhus, when the row and column are shifted out ofquadrant, lvalue added is reduced) and finalherefore valuefinal demand are reduced by the same' *


The analysis of Ihe paid services sector is somewhat more complicated. By the equality of rows and columns wc know that,


Moving the xp row into the value added quadrant increases its total byof xp(l) throughut decreases it by xd(p)in the equation above wc see that the net change in the value

-) -

Since the right hand side of this equation is exactly the net change in the final demand matrix, equality is retained. The equality of row and column sums ensures that the value of the final demand and value added quadrants will always be equal, but it is sometimes valuable to trace the algebra of the situation to improve one's understanding of the problem.

Theormat is only halfway to SNI. The other step is to move the depreciation row into the interindustry quadrant so that the value added quadrant equals SNI. Moving the row is accomplished easily enough. The problem comes in trying to move the equivalent column. There is none. As mentioned above, it is statistically impossible to determine which products are equivalent to depreciation. If machines were always replaced by identical machines then this might be possible. But old machines are replaced by technologically improved machines ci perhaps not replaced at all. In this case we cannot identify the currently produced machine which is being used for replacement of depreciated capital stock. Algebraically we can perform this transformation. We) as the production of the ith branch of the productive sphere which onsets depreciation. The column sum is toul depreciation, not just productive depreciation. The result of creating this column and moving the row and column into the interindustry quadrant is shown in

The first point to notice ins that the investment column has been transformed into accumulotion. This occurred when the depreciation column was created, and each entry in this column is definedecond, we have switched from US to Soviet accounting practices regarding depredation of government capital stock. The entries now labelednd xd(c) previously were defined to be zero. Finally, and most Important, the final demand and value added quadrants now equal SNI. The final demand quadrant does not quite resemble SNIU. The column sums of the final demand columns nearly equal Soviet published dau on SNIU. The

remier SorfctiMJi icedrfit, M. R. Eroer*en. thb point at tome knell. Inbook. MrtkawtrvorC^ktitMtnma*ie. StatlMlka.he Sovei weetical toJulion wai to leave the depreciation column In final demand and to pal Ihe deoreciaiion tc-bet-een ihe Interindwirr and national Income qwadr.nua flaunt J.

joint sums of the xp, xf, g,olumns equal the material purchases of the services, science, and administration.olumn sum equals private consumption. The sum ofolumn is overstated by the amount ofew column should be created for losses, but total losses are so small that this can be ignored here. Finally, there is no export or import column. Again, introducing foreign trade would confuse the illustration without clarifying any important points.

Comparingithe can easily see that GNP equals SNI)) plus xd lessn the product side, GNP equals SNInd xd,lso explains why the Soviets include nonproductivein consumption. Many Western observers have criticized theseestimates as arbitrary, andepending upon how they arc calculated they may indeed be arbitrary and misleading, but they are necessary. If their value is incorrectly calculated, the total value ofwill be biased In the opposite direction by an identical amount, since depreciation plus accumulation must equal investment.

or example, toe Abraham S. Becker, Soviet ttolfonalnlvcrtliy of California Preta, Berkeley.

Original document.

Comment about this article, ask questions, or add new information about this topic: