Created: 9/1/1997

OCR scan of the original document, errors are possible

Intelligence Report

Estimating Soviet Spending for Military Research and Development



CENTRAL INTELLIGENCE AGENCY Directorate of Intelligence1


Estimating Soviet Spending for Military Research and Development


Unlike the United States, the Soviet Union publishes no breakdown of its military budget or expenditures. Soviet outlays for military research and development are fundedariety of budget accounts along with allocations for other items, inashion as tothe amounts. The measurement of Sovietfor military research and development thereforeomplex problem of estimation, requiring the evaluation of direct, and indirect data from Soviet budget announcements and related literature as well as other evidence on Soviet military and scientific activities available to the West.

The costing of military RDT&F. (research, development, testing, and evaluation) is further complicated by the nature of these activities. To develop its estimates of Soviet expenditures for deployed forces the Office of Strategic Researchechnique of direct costing. That is, judgments based chiefly on what is observable as to the numbers of Soviet weapons and forces in being are multiplied by estimates of what these weapons and forces would cost in rubles and in dollars. Byarge part of therocess--alroost all basic researchreat deal of appliednot observable directly and cannot be associated with afinished product. As the direct costing technique

Hate:Thin report wan prepared by the Office ofm and a I. <


is therefore inadequate for costing military RDT&E, otherin this intelligencemust be used.

One purpose of the report is to present expenditure estimates for the0 An equallypurpose is toescription of the basic data and the estimative process so that the reader can appreciate the nature of these estimates and their inherent limitations.

The reader should also noteistinction must be drawn between two conceptual problems in the assessment of overall Soviet militaryctivities. One is the problem of estimating the amount of resources expressed in rubles and dollaron these activites. The other treats the effectiveness of these activities in accomplishing Soviet militarybjectives. This report is concerned entirely with problems of the firstSoviet militaryxpenditures and dollar valuations ofctivities.

A summary of this report begins on page 5.




Estimative Methodology

Sources of Funds




Reliability of the Ruble Estimates .

Ruble-Dollar Conversions

Trends in


Dollar Valuations


Estimated Sources and Disposition of Soviet Funds


Estimatednd Space Expenditures .


I, Announced Soviet "Science" Expenditures and

Related Indicators




Official Soviet announcements on annual expenditures for "science" are believed to cover essentially all outlays for military RDT&E,, and space The Soviet announcements do not provide direct information on the utilization of these funds, however, and the overall spending figures must beusing earlier, more detailed Soviet breakdowns of "science" activities together with direct costing of space programs and some fragmentary information on other aspects ofderive estimates of expenditure series for military RDT&E,, and space programs.

These estimatesigorous Sovietprogram during the past decade. Total outlays increased on the average by aboutercent per year, fromillionillion)0 to aboutillion rublesillion)

Aboutercent of the totalnd space effort0 went for military RDT&Ebut by the late Sixties the military share had declined to aboutercent. The main reason for this shift was the increasing financial importance of the soviet space program, ledgling venture the Soviet space effort then used Onlyercent of totalunds. By the late Sixties this figure had risen toercent of the total. outlays forepresented ercent of total "science" funding The share of outlays foreclined slowly to ercentnd has since remained at about that level.

Although militaryunding has declinedhare of total "science" outlays, it has steadily increased in relation to total defense spending,0 militaryxpenditures of


billion rublesillion) were aboutercent of total military outlays. By0 they hadtoillionomeercent of all Soviet military costs.

These estimates of Soviet spending for militaryust be viewed with some caution. First, the basic data come from Soviet publications. The validity of the resulting estimates therefore depends both upon the accuracy of Soviet financial accountingorrect interpretation of the published The use of data pertaining to the Fiftiesrimary basis for splitting subsequentffort into its components is an additional source of uncertainty. Finally, the transition from rubles to dollars and dollars to rublesumber of theoretical complexities as well as practical problems. The estimates of militaryre more uncertain for recent years. specially, they should be regarded as preliminary and subject to change- The large increase in funding indicated is not yetby any direct observations.

To test Lhe reasonableness and internalof the estimates derived from financial data published by the Soviets, they have been correlated with other indicators ofctivity such as scientific employment, numbers of scientific research institutes,alculated scientific wage bill. These statistical chocks all provide general support to the estimated trends in totalpending.

In spite of uncertainties and potential for error, it istho cross-checks tend to SUpporL this belief--that on balance the estimates presented in this report are reasonably accurate representations of the trends and levels of Soviet efforts, although less confidence can be attached to the detailed breakdown amonq military RDT&E,, and space than to the total.

There is, of course, no clear lino of demarcationoen research and developmente preparatory stages of series production, nor between innovative design in new product development and productiondirected at modifications of existing products. In these greyccounting is as often as not doneatter of convention and convenience rather than according to strict accountingis true in the US as well as in the USSR. For all these reasons, international comparisonsxpenditures are subject to ovon greater uncertainty than exists for measures of other economic activity.

Finally, it is important to note that theand limitations of hardware, the relativepositions of the USSR and the US, and the future threat to tho US of the Soviet militaryesult of their militaryffort cannot besolely on the basis of expenditure estimates. Such estimates provide atood appreciation of the level of effort and its change over time, but they cannot in themselves directly relate this effort to relative strength or weakness.



Estimative Methodology Sources of Funds

Because of the secrecy surrounding much of Soviet scientific activity and the inherent complexity of making economic assessments, it is difficult for intelligence analysts to judge precisely the level of Soviet scientific activity or its distribution among military, civil, and space projects.

The Soviets provide data identified as "totalfor science from the [state] budget- and other sources." Official Soviet writings indicate that this figure includes funding for civilctivity as well as for tho space program.

The Soviet state budget is composed of six major elements:

the National Economy

(basically investment funds)

Measures -- Defense


Loan Services

Residual and Reserve Funds of the Councils

of Ministers.

The Social-Cultural Measures account is further broken down into allocations for Enlightenment (Education and the Arts), nd other items (culture, health, and social welfare). Finally, the "Science" account is divided between Ail-Union expenditures {spending at the national, level) and Union Republic expenditures (spending at the republic or local levels) . Tlae Soviets have notreakdown of the "Science" account between All-Union and the Union Republic expenditures, however, The distribution5 between these two accounts has been used to allocate their share since then.

The largest single source of funds nd space is the Ail-Union "science" category of the state budget, which accounts for about one-half of the total. Capital investmentnd space facilities-funded from another category of the statefor aboutercent of total "science" expenditures. Other funds for "science" are provided by individualunive"ities, and ministries, includingthe Ministry of Defense. Expenditures by these other sources are not reported but rather appearesidual termed "funds of enterprises and other Estimates of expenditures by universities and enterprises have been made by extrapolation fromavailable for certain years. This procedureizable share of "science" funds unattributed to asource. This unattributable category may largely represent funds from the Defense budget.


H The Soviets do notreakdown of their science" expenditures among civil, militaryresearch, development, testing, andnd tneir space program. owever, they dida detailed resourceexample, wages and salaries, instruments, books,f expenditures foreriod under the All-union and the union Republic "science" budqets. The budgets of the Unionare devoted to developing "local industry" (essentiallytemized expenditures about equal to their announced totals.

The sum of the detailed cost for the Ail-Union budget, however, did not equal the announced budget figure. It loft unexplained a large annual residual that grew steadily over7 period. Analysis on the official expenditure data and pertinentliterature suggests that this residual is the major source of funds for those activities considered sensitive by theRDT&E, nuclear, and the space program. Soviet publications, for example, have identified the All-Union budgetfor "science"ource of funds for "work of national importance" and associate "science" budgetswith ballistic missile development and the space program. The Office of Strategic Research there-


fore considers this residual as militarynd space spending, or "classified" expenditures.

The Soviets have not published an itemized All-Union "science" budget For this reason the Office of Strategic Research has used the trend of thes the basis for estimating "classified" budget expenditures for "science" Both the res ual and the itemized portion increased0ut the residualarger share. 0 the residual accounted for aboutercent of the total and7 it had increased to almostercent.

To estimate the "classified" residual for the yearshe growing share oferiod was extrapolated but onlyhen it reached aboutercent. Somectivity that is of "nationalnd therefore funded by the All-Union "science" budget, almost certainly occurs. For the yearshe methodology used maintains the estimated share for the "classified" residual atercent of the All-Union "science" budget.

In addition to allocating all of the "classified" residual to militarynd space spending, the estimative methodology allocates one-third of the remaining portion of the All-Union "science" budget to militarynd space spending. This allocation is based on the judgment that so-ne of the research financed by theseexample, development work of anature on electronicto military and space urograms.

The estimative methodology allocates the other types ofunds between civil programs on the one hand and militarynd space on the other as follows:

-- Enterprise "science" expenditures are allocated on the basis of the relative shares of civilian and military production in the total, output of machinery and equipment. During the Sixtiesj percent of output was allocated to defense andercent to civilian uses. It can be assumedimilar portion of the enterprise funds for "science" were provided by the industries contributing to this pattern of produc tionpercent forandercent for militarynd the space program.

" Ail university funds devotedctivity in the USSR are believed to be for civil purposes. What littlehese institutions mayis probably done under contract and hence the financing would probably originate from some other funds.

Capital investment funds forand unattributable funds presumably support They are consequently apportioned on the same general basis as the civil-military splitfor "science" budget funds, enterprise funds, and university funds considered as ais, one-quarter civil and three-quarters military and space.

esult of this procedure, it is estimated that aboutercent of total Soviet "science" outlays goes tondercent to militarynd space for the. These shares should bo regarded only as approximations because of thein the estimative procedure.

Space Expenditures

The next step in the process ofpending is to estimate expendituresSoviet space program. In this case theapproach is used. Enough detailedthe Soviet space effort is available fromto permit an assessment of what eachSoviet space program would cost if itin the US. The sum of these programallowances based on US analoqy for suchelements as administration and advanced ot n total Soviet spar-

expenditures. This estimate of space expenditures conceptually covers the full range of Soviet space activities.

There is no information which suggests that the Sovietsfunding or institutionalbetween space programs related to military requirements

- 12


and those related to civil objectives. The estimate of total Soviet space spending, however, has beenbetween military and civil for comparison with US programs. Soviet programs that correspond to those funded and conducted in the US by NASA are considered to be civil and those which correspond to programs funded by the Department of Defense are considered to be military. The estimated outlays for these civil and military spacethe exception of those for thehases of military spacesummed, converted to rubles, and then subtracted from the total for militarynd space. The resultuble estimate of Soviet spending for militaryhich includespending for military space systems, but excludes spending for operating military space systems.

Price Adjustments

The next step in the methodology is to adjust the ruble dataonstant price base so that estimated annual changes in expenditure levels will reflect only changes in the underlying level of activity and not any changes in prices.

The Sovietsajor price revision infirst overall change in the Soviet wholesale price structure The Office of Strategic Research estimates, on the basis ofreceived to date, that this price revision and subsequent adjustments had the effect ofosts byercent TheSoviet "science" expenditure data have therefore been adjusted to reflect this price change and the results of this adjustment are characterized in this report as being in8 rubles.

Finally, all of the ruble data derived by the procedures described above are converted to dollar estimates by applying ruble-dollar ratios developed by the Office of Strategic Research specificallyctivity. (See the discussion beginning on)

Table X

Announced Soviet "Science" Expenditures and Related Indicators

I960 1 2 3 * 5 6 7 8 9

Total announced (billion current


Wage bill forscience services a


4 7 2 6 4 0 4 4 8


Totaln civiliar. economy

Research institutes, their branches and

Calculated from published Soviet <iahd on "average wages and employment in science

and ectenceb. Interpolated*

Reliability of the Ruble Estimates

There is no precise method for confirming the accuracy of the estimates derived by the methodology described above. There are, however, some aggregate measures available for checking the reasonableness of the estimates ofwith respect to trends.

The rate of growth in announced Soviet scientific and technical manpower over the pastears hasthe same general trend as announced total "science" expenditures. age bill for science and scienceon che basis of published manpower and salaryparallels the growth in total "science" expenditures over this period. Similarly, there has been substantial growth in the number of scientific research institutes in the USSRaken together, these indicators lend credibility to estimates of rapid growthpending based on data announced by the USSR

To check the reliability of the announcedon Soviet "science" expenditures, several data series representing different aspects ofwere investigated. Statistical correlation analysis was conducted between the announced total "science" expenditures and each of fourseries derived from published Soviet statisticsnumber of scientists, the number of engineers, the number of scientific research institutes,l series calculated on the basis of published Soviet scientific employment and salary information (see

These statistical tests show that thereigh degree of internal consistency between each of the indicators and the reported total "science"lending strong support to the conclusion that the announced totalalid indication of the trend in the Soviet scientific effort.

In addition to supporting the validity of the trend in total expenditures, the calculated wage bill also servesough check on the level of reported

total "science" expenditures. An article in the0 issue of Pinansyofficial organ of the Soviet Ministry ofwith theaspects of scientific research noted that "wages make up approximatelyercent" of the expenses of

carrying out scientific research for the nationhole." The fact that the calculated wage bill amounts to aboutercent of total "science" expenditures ineriod supports the conclusion that the published Soviet totals for "science" expendituresdo include allis,

In spite of these checks, however, thestill is limited by its dependence on the Soviets' intention and capability to report accurately. published expenditure data have frequentlydiscontinuities and elements of confusion that have taken considerable time and effort to unravel, there has been no indication that the numbers were falsified. The Soviet practice over the years has been to withhold data or to confuse with ambiguous or even changing definitions, rather than to falsify.*

The uncertainty of these estimates is greater for recent years. Whereas for the earlier years USofnd space piograms provides someinformation on the trend and levels of effort, the time lag betweenctivities and later phases such as testing makes such monitoring observatiors incomplete since The es-.inatc0 is particularly tenuous for this reason, and also0 budget Information has not yet become available. The estimates for the last two years should be regarded as preliminary and subject to revision.

Another form of indirect and incomplete evidence of growth trends in high priority areas ofs found in the book Economic Problems of the Effectiveness

The assumption that the. Soviets nadv ery the magnitude and trend of their budgetfor "classified" activities in8 handbook nas given supportdition of the same handbook uhich was identical in coverage to8 edition aith the single exception that it omitted the table with the unexplained residual.


of Science, published1 by an organ of the Central Committee of the Communist Party. This source contains informationxpenditures during they four major branches of Sovietmachinebuilding, electrotechnical, radioelectronic, andto be heavily involved in military and space programs* as well asn fields of top priority Soviet interest. Although there is some ambiguity in exactly what thexpenditures of these industries cover, the growth in the reported total for the four industries corresponds well with tho growth in estimated militarynd space expenditures. This does not directly support anylevel ofnd space expenditures. It does, however, support the fact that growth rates in some areas of Sovietpriority civil fields as well as in military and spacebe substantially above tho average growth for sciencehole. Further analysis of these data is required, however, before their full meaning can be understood.

The separate expenditure estimate for space relies heavily on the monitoring capabilities of US The intelligence coverage of those aspects of the Soviet space program most responsible for determiningtime phasing, test facilities, launch vehicles, other hardware,is excellent. This coverage makes it possible to use the direct costing technique for estimating Soviet space expenditures.

Because the estimate of militarys calculated by subtracting space program costs (with the exception of those for military space RDT&E) from the combined total for militarynd space, it too is dependent, at leaston the quality of US intelligence monitoring capabilities. It is also affected by uncertainties of the dollar costing procedures and of the estimated

' I ::

on avionicsj guidance Systems, radars, and ;'V of Defense. They would not be

."i'1 sits temp, and

for missiles and aircraft as veiln other weapons, military vehicles, and naval vessels.

- 17


conversion ratio by which the space program costs are converted to rubles.

In sum, the Office of Strategic Research believes that the estimates of Soviet spendingnd spacein rubleaaccurate representation of the Soviet effort in terms of both trends and absolute levels of spending, although less confidence can bo attached to the breakdown among, military RDT&E, and space.

Ruble-Dollar Conversions

To provide an appreciation of the magnitude of therogram the ruble estimates areto dollar values, the validity of which is critically dependent upon the accuracy of the ruble-dollar ratio used. Conceptually the ratio should quantify precisely the outlays required in the US and in the USSR toroduct. Unfortunately, no one has yetatisfactory means ofevenoutput ofoctor.

utput can be viewed in several ways, and some of the "products" are more quantifiable than others, although clearly none can be measured precisely. The product of theoretical research is usually considered to be new knowledge, while the end result of applied research and development is commonly viewed as the design and construction of prototype models of equipment.. Finally, at tho most general level, the technology embodied in operational equipment represents the useful output of therocess.

Looking at the product of theffort in this last, moat general sense makes it clear that conversion 'at the official rate of exchange ofublesollars would gronsly understate the dollar outlays required to reproduce theroduct in the US. Accordingly, the official rate of exchange is useful only in that it helps to define the upper limitalue for an appropriate ruble-dollar ratio foralue thai.

is clearly too high. (The higher the ruble-dollar ratio, the lower the dollar outlay implied.)

The lower limiteasonable ruble-dollar ratio has been established in another way. If the resourceis, manpower, capitalandthe US andfforts wore identical in their composition and quality, and if these resources were organized and managed with equal efficiency in the two countries, then one unit of physicalas one scientific worker and all of the equipment, facilities, material, and management that go withresult in an equal level of output in the two countries. Measurementroduct becomes unnecessary because inputs, which can be measured and priced, could be used as aproxy for outputs and would provide theratio for valuing oneroduct (total cost of inputs) in the other country's currency.

Direct pricing of Soviet andnput unitsuble-dollar ratio of aboutublesollars. The application of this ratio to the ruble estimates, however, would overstate therequired in the US to reproduce theroduct, because thenput unit is clearly less productive than the counterpart us input unit. This judgment is based on Soviet acknowledgments of lower productivity, and on the observations of US specialists who have visited and examined theof Soviet laboratories and other scientific establishments. The input ratio ofublesrs, therefore, dciino.s o lower to reasonable value for the ruble-dollar ratio forector.

A6 article by a leading Soviet scientistanager, Peter Kapitsa, helps to quantify the productivity differential and thereby narrows the range of reasonable ruble-dollar ratios. Kapitsa acknowledged the difficulty of measuring scientific productivity, but quoted approvingly the resultsan unidentified US study and noted that;

ne must conclude that we produce half the scientific output produced by the

Americans with nearly the same number of scientists. It must bo Inferred,that the productivity of our scientists is lower than that ofin the US. [Komsomolskaya Pravda,

Taken at face value, this statement suggests that the computed input cost ratio ofublesollars should be adjusted upward to nearlyublesollars. This would, however,an overly harsh judgment on The context of Kapitsa'sgreater efforts by the Soviot scientific community--suggests that he would have tended to exaggerate the differential. Bven Kapitsa did not carry hisso far as to state flatly that the productivity of Soviet science was half that of the US.

The lower productivity attributed to thenput unit is not based on the judgment that Soviet scientists and engineers are less competent overall than their US counterparts. Rather, they are not able to produce as much output because they have less capital equipment to workexample, tho extensive computer facilities that support the US scientist and engineer are simply not available to the Soviets. Moreover, by all reports--both from tho Soviets themselves and Western observers--the Soviet scientist worksess effective managerial environ-meht, particularly with respect to the adoption and implementation of new technology. From all reports this applies to both military and civil riclds, but is particularly true inrograms.

These considerations have led the Office ofResearch to select differential ruble-dollar ratios ofublesollars for theector andublesollars for the militarynd space* sectors. Compared with the input ratio ofublesollars, these ratios imply that the US is aboutercent more efficient than the USSR in military and space research and developmentand aboutercent more efficient in.

Viewed from the Soviet side, the militarynd space effort is more efficient than thereflection of the clear priority the Soviets give to defense and space in the allocation ofmaterials, and managerial talent.

As noted above, there is no way to measure the productsctivity with precision. Therefore, there is no way to test precisely the validity of the productivity adjustments or the resulting ruble-dollar ratios. For this reason, the estimated dollar values ofctivity in this report should be viewed as approximations rather than precise measures.

Trends in Spending

Ruble Estimates

The Soviets maintained an active military RDT&Ethroughout the Sixties, although the strong upsurge in space spending in the middle of the decade is believed

-JXf JiT accoraPanfedemporary, slight reduction ot expenditures for military.

(See chart below and Tableext page.) 0 the

ostinated spending level for militaryincluding


Ifeiiu* INK fcjbatJ


Table 2

Estimatednd Space

8 Rubles

0 1 2 3 4 5 6

Military RDTGEwhich:

Military space


" Civilspace


Total RAD and space 4 8 2 6 8 3 6 4

3". because of rounding, cort-ponento may not add to totals Covers all military systems RDTSE including stilitaru space sustens and all RDTtS* Outlaws for members of active nilitaru forces aa/iinned

ties ce lieII ae for civilian personnel enployed in such activities are included.

or militaryillion rubles was about twiceillion rubles estimated to have been spent The annual rate of growth in militarypending00 in8 rublesercent.

In terms of total spendingnd space during the Sixties, militarys estimated to have declined somewhat in relative importance. ilitaryxpenditures accounted for an estimatedercent ofnd space spending. At that timeccounted for aboutercent of the total and the fledgling civil space program onlyercent. The militaryhare declined steadilyow of aboutercentithtercent and the space share (excluding military space RDT&E)eak ofercent. 7 the share of militaryas increased and0 accounted for aboutercent.

07 militaryxpenditures amounted to aboutoercent of total Sovietexpenditures. pending on militaryas been increasing and accounted for aboutercent of the total Soviet defense bill in

The growth in spending for militaryompared with that for weapons procurement during the SixtiesSoviet willingness to forgo additional current hardware procurement for the development of improved future weapons systems. Compared with expenditures for hardware procurement, militarypending was about half as much0 but by0 had grown to B0 percent as much.

Dollar Valuations

Dollar valuations ofrt designed toan appreciation of the level ofinwould be required to reproduce the Soviet effort in the US. The dollar estimatesasis for approximate comparisons with US spending for equivalent activities.

Table 3

Dollar Valuation of Estimatednd Space Expenditures,



. 16


Military space , r.:




space operations



nd space 9

a.Because of rounding, ccponanta may not add tc totala aho-Jn.

o. Covera all military ayateme RDTIS including military apace euatena and all nuclear wnirau FDTIS. Outlava for member* of active militaruaiaied to BDTtS activitiaa a* veil aa for civilian, peraonnel employed in auah activitiee art included.

It is estimated that the Soviets spent theofillion for military00 {sea The estimated annual rate of expenditures rose from the equivalent of almost S5 billion0 to the equivalent of4 billion It declined slightlyut began increasing after that and0 reached the equivalent of8 billion.

And it is important to note that the relative levels of militaryfforts of the US andas measured in dollarboeliable guide to the current or future effectiveness of US and Soviet military technology. The Soviet space program, for example, is the largest in the world in terms of expenditures, but it has not approached the US level of technical sophistication and capabilities. In military RDT&E, the Soviets are buying today in many areas what the USumber of years ago. In short, expenditures do not equate to capabilities, and nowhere is this caution more likely to have morethanctivities.

It is clear, nevertheless, that the Soviets are increasing the resourcos devoted to military RDT&E. The level of resource allocation is obviously related to the product of therocess, so that it isto understand the general magnitude of the Soviet commitment. The Soviet effort is cast in dollar terms noteans of judging the relative positions of the US and the USSR in advanced military technology, but for the purpose ofeaningful basis for understanding tho magnitude of militaryctivity in the US and the Soviet Union- .r. understanding that would not be possible if US and Soviet expenditures wore not expressedommon currency. omprehensive assessment of the state of Soviet military technologyis the US, however, cannot be made solely on the basis of expenditure comparisons.


Original document.

Comment about this article or add new information about this topic: