Analyzing Soviet Defense Programs, 1951 1990 .pdf
Nombre del archivo original: Analyzing Soviet Defense Programs, 1951-1990.PDF
Este documento en formato PDF 1.4 fue generado por Xerox WorkCentre 5665 / Adobe Acrobat 10.1.13 Paper Capture Plug-in, y fue enviado en caja-pdf.es el 13/06/2015 a las 06:54, desde la dirección IP 173.178.x.x.
La página de descarga de documentos ha sido vista 951 veces.
Tamaño del archivo: 1.4 MB (15 páginas).
Privacidad: archivo público
Descargar el documento PDF
Vista previa del documento
http://www. csi .cia/studics/vol42no3/art7 .him I
Determining the Cost and the Burden
Analyzing Soviet Defense Programs, 1951-1990
From the early days of the Cold War, the economic potential of the Soviet Union was a prime
intelligence question. In 1951, Max Millikan, the first director of the Office of Research and Reports
(ORR) in CIA, identified five objectives for foreign economic intelligence in support of US national
• Estimating the magnitude of potential military threats to the United States and its allies by
calculating the level of military operations a potential enemy's economy could sustain .
• Estimating the character and location of future military threats by examining how the potential
enemy is allocating its resources.
• Assisting in judging the intentions of a potential enemy by monitoring changes in its economic
• Helping policymakers to reduce possible military threats by telling them how economic measures
could impair a potential enemy's economic capabilities.
• Assisting in the estimation of probable development of the relative strength of the East and West
blocs in the near term, assuming hostilities are avoided._!_
When Millikan formulated these objectives, the Korean war was on, and the possibility of conflict with
the USSR was real. At the same time, knowledge of the Soviet economy was sketchy. As Millikan put it,
his office in the first half of 1951 "was engaged in taking an inventory of its ignorance concerning the
economy of the Soviet Bloc."1 Soviet economic statistics were of little help. Publication of detailed
statistics had ceased for all practical purposes after World War II. Data on national aggregates such as
national income or industrial production greatly exaggerated real growth, while information on defense
outlays was limited to a single line of uncertain coverage in the state budget.
Within a few years, dozens of analysts were focused on putting together a picture of the Soviet economy
industry by industry. The military economics branch in ORR tried to estimate Soviet defense spending.
Another branch began reconstructing Soviet national accounts and compiling independent estimates of
real growth in the components of GNP--industrial and agricultural production, consumption, and
Such was the beginning of CIA's building-block measures of Soviet national output and defense
expenditures. Essentially, this approach meant compiling estimates from the ground up by filling in the
empty boxes in accounting frameworks encompassing Soviet defense-related economic activity and total
economic activity. The work on Soviet GNP was able to borrow heavily from the pioneering work
carried out at Columbia, the RAND Corporation, and Harvard by a group of specialists on the Soviet
Mcoa•Y"'"'I6 .:>uV!et Defense Programs, 1951 1990
http://www .csi. cia/stud ies/vol4 2no3/art 7 .h tm l
economy. The effort to cost Soviet defense programs, on the other hand, had to start essentially from
scratch by assembling information on all the physical measures of defense activity and multiplying them
by the estimated costs of these activities.
These costs were then aggregated by military service; resource category (procurement, construction,
personnel, operations and maintenance, and R&D); and military mission (strategic offense and defense,
general purpose forces, and command and general support). Thus, the Agency came to grips with
Millikan's second objective for foreign economic intelligence--information on how a potential enemy
was allocating its resources. Soviet GNP and defense programs were valued in rubles to assess Soviet
decisions on the allocation of resources and to measure real growth in defense and in dollars to permit
comparisons with the GNP and defense programs of the United States and other countries.
Building an Evidentiary Base
The evidentiary base for the defense and GNP estimates at first varied greatly but improved considerably
over time. After the mid-1950s, a great deal of information relevant to the GNP calculations became
available. By the mid-1960s, the methodology and the data were available to estimate GNP by sector of
origin and end use in 1955 ruble prices and in current prices and to compare US and Soviet GNP in both
ruble and dollar prices. From then until the breakup of the Soviet Union, the estimates were almost
continuously reviewed and improved, but these were variations on a theme and not radical new
departures. Throughout their history, the estimates of real growth in GNP rested on the premise that
while Soviet value statistics--that is, the ruble value of industrial output were badly flawed, statistics on
physical output--tons of steel, square meters of cotton fabric--were sufficiently reliable to support the
The evidence for the defense estimates was quite different. National intelligence collection provided the
information on Soviet defense production and procurement and order of battle, and this information
became increasingly accurate after the early 1960s with each enhancement in US overhead
reconnaissance capabilities. But operating rates, maintenance practices, and the like had to be estimated
by a variety of methods, and the acquisition of several Soviet maintenance manuals helped. Identifying
the resources committed to military R&D, however, was a problem until almost the end of CIA's costing
effort. For most of the period, military R&D had to be inferred from a small sample of Soviet statistics
on the financing of science. In the 1980s, an effort to build an R&D estimate from the bottom up by
collecting intelligence on more than a thousand Soviet R&D establishments finally succeeded--and
reduced the previous estimate of its growth of military R&D markedly.
Finding prices to attach to the estimates of military activity was another problem. Prices were generally
available for military pay and allowances and other personnel costs. Defector reports supplied the
information on military pay in rubles, and US dollar pay rates for equivalent ranks or occupational
specialties were obtained from the Pentagon. Ruble and dollar prices for much of operations and
maintenance (O&M) and construction could be found in Soviet publications and in CIA research on the
relative ruble and dollar costs of the different kinds of construction work and some of O&M. During the
1970s and early 1980s, a large number of contracts with US defense firms provided estimates of the
dollar cost of manufacturing most of the more expensive items in the Soviet procurement and spare parts
accounts. Intelligence on ruble prices was scarce, however. Not until the 1980s had enough ruble prices
been collected to firm up the ruble estimates of Soviet military procurement and the maintenance part of
The question of what to include in the definition of defense outlays was an issue from the beginning. For
comparisons with US programs, the Soviet activities were selected to match as well as possible US
spending on Department of Defense (DoD) national security programs, defense-related nuclear activities
of the Coast Guard, and the present value costs of retirement, disability, and survivors' benefits for
active-duty military personnel. When estimating the cost of Soviet military programs, a broader
definition was used, one that included civil defense troops, railroad troops, construction troops, MVD
troops, and civilian space outlays.
2 of !5
I !i!2/02 7 39 AM
'."""J'"'"i!> •JV••et Defense Programs, 1951-1990 (U)
Figure 1. Soviet GNP and defense outlays, 1950-90 (1982 rubles). (U)
What the Estimates Showed
It is instructive to review what CIA's defense spending and GNP estimates, as they matured in the late
1980s, suggest about the relation between defense and the economy in the postwar USSR. The first thing
to look at are the estimates of Soviet defense spending and GNP in 1982 rubles (Figure I). If prices of an
earlier year are used, the rates of growth of both defense and GNP would be marginally lower and the
ratio of defense to GNP a bit higher, but the overall picture would be much the same. In constant 1982
prices, the share of defense in GNP declines from 24 percent in 1951 to 14 percent in 1959. It then varies
only within the range of 14 to 16 percent between 1960 and 1990. This calculation provoked a good deal
of controversy over the years as it was taken to represent the burden of defense on the Soviet economy.
One can isolate different stages in the growth of Soviet defense spending and relate them to
developments in the economy generally. The building-block estimates set out in considerable detail the
changes in the composition of GNP or defense spending. This detail permitted analysis of changes in the
USSR's economic policy--the third of Max Millikan's objectives for US foreign economic intelligence.
The average annual rates of growth shown for defense in Figure 1 certainly do not reflect the monotonic
rapid growth that is perhaps the prevailing public perception. Meanwhile, the continuous decline in the
rate of growth of GNP suggests the underlying forces that led Soviet leaders to constrain defense
spending and finally caused Gorbachev to introduce his ill-fated reforms.
The immediate explanation of the changes in the growth of defense spending are complicated, but a few
of the prime movers can be identified. In the 1950s, Khrushchev's demobilization was the major factor.
In the 1960s, defense spending grew rapidly, propelled by growth in procurement of missiles, ships and
submarines, space vehicles, and aircraft as well as the post-Khrushchev buildup in general purpose
forces. Thereafter, the growth of defense subsided as procurement increased less rapidly in 1970-74 and
then leveled off in 1975-84. Outlays for O&M and R&D continued to rise, although at a slower pace
then formerly. After Gorbachev came to power, there was a three-year acceleration in defense spending
marked by a spurt in outlays for aircraft and missiles, and then a sharp decline in defense, spurred by a
downturn in outlays for aircraft, land arms, and space-related equipment.
Figure 2. Dollar cost of US and Soviet military programs, 1951-64 (1972 dollars}. (U)
At minimum, the trends described in Soviet defense spending suggest some questions for historical
research on Soviet military policy during the Cold War. For example:
• Was the slower growth in real defense spending after 1975--and the plateau in procurement--the
result of conscious policy decisions? If so, what was the cause?
• What role did the US defense buildup beginning in the late 1970s play in the spurt in Soviet
defense spending in 1985-87?
Interest in CIA's estimates of Soviet defense outlays and GNP within the US Government and in
Congress, however, centered more on the dollar comparisons than the ruble measures. Unfortunately, the
records necessary to put the defense comparisons on a single-dollar price base are no longer available.
Figure 2 shows the comparison of the dollar cost of Soviet and US defense programs from 1951-64
expressed in 1972 dollar prices, and Figure 3 covers the period 1965-89 in 1988 dollars. According to
these comparisons, the cost ofUS defense programs exceeded the dollar equivalent of Soviet programs
by roughly one-fifth in the earlier period, while the dollar equivalent of Soviet military spending in
1965-89 was slightly greater than US spending.
In the 1980s, some variants of these comparisons were constructed--one to take in a broader definition of
national security outlays and another to construct a NATO- Warsaw Pact comparison. The comparison
expanded to include the so-called cost of empire, civil defense, and the like pushed up Soviet spending
relative to US outlays by a couple of percentage points in the 1980s, but the NATO-Warsaw Pact
:''"''.Y "-""~'> "vv u.:t Defense Programs, 1951-1990 (U)
http://www. csL cia/studics/vol42no3/art7. htm I
comparisons showed the dollar-equivalent cost ofNA TO defense programs 15 percent larger than
Warsaw Pact programs over the period 1976-86.
Figure 3. Dollar cost of US and Soviet military programs, 1965-89 (1988 dollars). (U)
Comparisons of Soviet and US GNP first presented by CIA in the 1950s had a receptive audience, given
the concern then that the Soviet Union was on its way to overtaking the United States in the economic
realm. When DCI Allen Dulles testified before Congress in 1959, he projected Soviet GNP growth
through 1965 at 6 percent per year and industrial growth at 8 to 9 percent per year. These rates, he said,
would raise Soviet GNP to about 55 percent of the US level by 1970 and Soviet industrial production to
perhaps 60 percent of the US level. Dulles concluded by saying:
The Communists are not about to inherit the world economically. But while we debunk the
distortions of their propaganda, we should frankly face up to the very sobering implications ofthe
Soviets' economic program and the striking progress they made over the last decade.l
The most recent CIA estimates of Soviet and US GNP show Soviet GNP rising from a little more than a
third of US GNP in 1950 to a little less than half in 1965 to 60 percent in 1975 before falling to less than
half in 1990. These calculations, which represent the geometric average of comparisons made
alternatively in ruble and dollar prices, have been criticized as being overly generous to the Soviets, but
subsequent research by Russian statisticians suggests that they are not badly out of line. The point is
that, by the mid-1970s, it was clear from the Agency's publications that the fears of the USSR's
becoming the foremost world economic power were unfounded. The dollar comparisons of Soviet and
US GNP--and later of NATO and Warsaw Pact GNP--proved to be invaluable in forecasting the relative
strengths of the East and West blocs--Millikan's fifth objective for the US Government's foreign
economic intelligence ..:!
How the CIA Estimates Were Used
Analysis of Soviet economic prospects provided the framework for analysis of the Soviet economy both
in the US Government and in the academic community after the original partners--Harvard and the
RAND Corporation--left the field. In the early 1960s, the Agency's estimates disclosed a sharp slump in
Soviet economic growth to a skeptical world. In December 1963, DCI John McCone told President
Johnson and the NSC about the Agency's most recent assessment of the Soviet economy: economic
growth had slowed drastically, and the USSR's grain and gold reserves were lower than previously
estimated, accounting for Soviet attempts to find long-term credits in the West.
Gradually, the CIA estimates of slumping Soviet growth and then stagnation gained general acceptance.
Between the early 1960s and the early 1980s, the few critics ofthe GNP estimates thought they
understated Soviet achievements. Only recently have revisionist claims been heard that the CIA
estimates overstated the rate of increase in Soviet GNP.
The Defense Spending Estimates
The focus of interest in CIA's estimates of Soviet defense spending shifted over time. During the 1950s,
when the USSR's economy was on a roll and the Soviet armed forces were in the midst of a partial
demobilization, Max Millikan's first objective of foreign economic intelligence had been met. The
estimates of the cost of defense programs and the size of the Soviet economy demonstrated that the
economy could continue to support the then existing level of defense effort and more.
Then, in the Kennedy administration, the Pentagon pressed hard for finer breakdowns of Soviet defense
spending as an aid in defense planning. This was the era of cost effectiveness calculation in the DoD.
Fred Kaplan, in his book The Wizards ofArmageddon, describes the use of CIA's dollar estimates in
Pentagon deliberations over a damage-limiting strategy to employ against Soviet strategic forces ..2 The
calculations indicated that any combination of civil defense, an ABM system, and anti bomber defense
4 of I
'.\ltalyzmg ::.oviet Defense Programs, 1951·1990 (U)
http://www .csi .cia/stud ieslvol42no3/art7 .htm I
system was a losing proposition. As Kaplan put it, the studies showed:
For each extra dollar that the Soviets added to the attack forces, the US would have to spend $3 to
protect 70 percent of its industry, $2 to save 60 percent ... and the same $1 to defend a mere 40
Until the end of the Cold War, the DoD continued to be an eager customer for the estimates of the dollar
equivalent of Soviet military programs. In 1977, Secretary of Defense Harold Brown, in a memorandum
to the DCI, declared that the reports and analysis being produced on military economics were "the basis
of the comparative economic analysis employed by Defense." He added, "The dollar estimates provide
the best, single aggregated measure of US and Soviet defense efforts."_§. In 1991, when CIA announced
its intention to discontinue the dollar estimates, Andrew Marshall, Director of Net Assessment in the
Office of the Secretary of Defense, wrote Deputy Director for Intelligence (DOl) John Helgerson "to
urge that CIA continue its work on dollar cost estimates of Soviet defense programs.".l Marshall
supported his request by stating:
Dollar cost data has [sic} been the major foundation for my office's work on a military investment
balance report that top defense officials have found very usefol in characterizing broad trends in
our security situation. For all its limitations as an indicator of the military balance, dollar costing
allows dissimilar systems and activities to be aggregated so that broad trends in national military
capabilities can be depicted Moreover, some activities not ordinarily compared in "bean counts"
can be measured in dollars, for example hardening and sheltering programs, or command,
control, and communications activities. Even intangible factors such as training and readiness
can be valued at the dollar cost of the activities that foster them.
CIA analysts used the ruble estimates to help determine whether certain large Soviet military programs
would be undertaken, and, if so, what the likely schedule would be. For example, an air defense analyst
decided cost constraints would prevent the USSR from deploying more than 600 AMB launchers instead
of the 6,000 sites then proposed by many in the Intelligence Community (IC).J!. CIA projections of
certain large aircraft programs were also cost constrained.
As CIA's Office of Strategic Research (OSR) proceeded with its research on Soviet military spending, it
found that, in the process of filling all the boxes in its accounting framework, it was uncovering
information--as Marshall noted--on Soviet defense activities that had been neglected in intelligence
analysis. It learned about operating rates and maintenance practices, which shed considerable light on
Soviet military readiness. By contracting with US defense firms to provide dollar costs of Soviet
weapons, OSR also learned a good bit about Soviet manufacturing technology and capabilities.
Perhaps the most visible contribution of the defense spending estimates, however, was to the debates
over US defense budgets. The US-Soviet comparisons figured heavily in the presentations that
successive presidential administrations made to Congress and the country. While Secretary of Defense,
James Schlesinger at the beginning of his 1975 annual report said that, "The Soviets now devote more
resources than the United States in most of the significant categories of defense" and showed a chart
comparing US and Soviet defense spending.-2 He emphasized even more strongly the adverse trends
embodied in the spending estimates by telling the armed services committees, "If they continue to grow
at 5 percent or 7 percent per annum, and we continue to shrink, then it is plain that sooner or later the
divergence will become so great that we will be in very substantial trouble." JQ
Under Democratic administrations, the pitch was much the same. William Kaufman, a consultant to
DoD through most ofthe 1970s, said that, beginning in 1973 and 1974, he "turned very strongly to the
CIA estimates as a basis for indicating why the US defense budget ought to be increased. "Jl
Thus, the US-Soviet defense comparisons were featured in DoD Congressional testimony, Defense
posture statements, and in the DoD publication Soviet lvfilitary Power. The comparisons, however, lost
their popularity when the dollar-equivalent cost of Soviet defense programs began to fall below the cost
of US programs in the 1980s. Robert Gates, as DOl and later as Deputy Director of CIA, had long been
7 6 0
ovv.et Defense Programs, 1951-1990 (U)
http://www .csi .cialst udies!vo142no3/art7. h tm I
skeptical of the dollar estimates and tried to abolish the!ll in ~he early 1980s~ pressure from_ cons~~ners in
the Pentagon and Congress, however, forced a resumptiOn ot the dollar costmg. \Yhen Sovtet m1htary
spending moved downward in the late 1980s, consumer demand for the dollar estimates weakened, and
Agency managers decided the time was right to drop them.
Connection Between Defense Spending and the Economy
With the slowdown in Soviet economic growth, Max Millikan's first objective for foreign economic
intelligence again came to the forefront--the calculation of how much military spending a potential
enemy's economy could sustain. Generally, the Agency's economic and military analyses of the Soviet
Union were imperfectly integrated. Even when they were, and when they highlighted economic
pressures on Soviet military programs, there was great skepticism in the IC.
In 1964, I 976, and the early 1980s, the IC did focus on the question of economic constraints on Soviet
defense spending. A Special National Intelligence Estimate (SNIE) was commissioned in 1964
following the turnaround in the analysis of Soviet economic prospects. After reviewing the deteriorating
economic situation, it said there had been a "continuing struggle between those who believe that the
national interest requires greater attention to an invigorated agriculture and those who oppose any
weakening of the priority accorded the development of heavy industry and a preeminent defense
establishment. ".11 Khrushchev's plans for boosting the production of chemicals and consumer goods and
an announced 4-percent reduction in its official defense budget convinced the US Intelligence Board that
the USSR "would make every effort" to hold down outlays on defense and space. The estimate declared
that, while defense outlays might decline, it was more likely that they would grow at a slower pace.fl In
any case, Khruschev was ousted within the year, and the decade of the 1960s produced the highest
average annual growth in real Soviet defense expenditures in the postwar period.
In 1965, another NIE on Soviet defense policy argued that, even if defense outlays increased by 20
percent in the new five-year plan period ( 1966-70), the Soviet economy could shoulder the burden and at
the same time gradually modernize Soviet industry and raise the population's standard of living ...!:! It
added, however, that defense programs would be "increasingly subjected" to reviews of their cost and
When Soviet economic growth slowed still more in the 1970s and CIA published an extremely
pessimistic appraisal of Soviet economic prospects, the view that military programs were not threatened
remained entrenched in CIA publications. Thus, a memorandum to holders ofNIE 11-4-78 provided a
grim description of the USSR's economic situation, including a "growing possibility of social
instability," and reported that "if military spending continues to grow at 4 percent per year, the increment
in national output going to defense could rise from l/5 to as much as 3/4," severely limiting the
leadership's scope for maneuver. Still, the "projection of Soviet military spending most consistent with
available evidence suggests that pressures in favor of continuing the existing arms buildup are likely to
offset any inclination toward change that might arise from the leadership's growing economic concerns."
If necessary, "appeals to a more extreme patriotism" and "repressive measures" would be used to
increase production and maintain domestic control in face of mounting military budgets.E
How tenuous the connection was between defense burden measures and CIA's projections of Soviet
defense programs was apparent in the Agency's reaction to the abrupt increase in the estimate of Soviet
defense outlays in 1976. Asked by Senator Charles Percy of the Joint Economic Committee of the US
Congress how defense was impinging on the civilian economy and what effect growing consumer
demands would have on an economy still spending so much on defense, the Agency replied:
The new CIA estimates ofSoviet defense spending--50 to 55 billion rubles in 1975--have altered
significantly our perceptions of the economic costs of the Soviet defense effort. Analysis of the
complex issues of economic burden and resource allocations is still in its preliminary stage.
However, if is clear that the Soviets are far more willing than we thought to forego growth in the
civilian sector (and consumer satisfaction} in favor of expanding military capabilities.
6 of l
Defense Programs, 1951-1990 (U)
The response went on to say that "there are no indications that the leadership has seriously considered
diverting resources from military use in response to consumer demands." It asserted that the Soviet
population had ''traditionally" recognized that programs to boost industrial production and defend the
country justified "a slow grovv1h in living standards."~
Developments in the Soviet Union in the 1980s and the process of the IC's coming to terms with the
deceleration in the estimated growth of Soviet defense spending restored the defense burden as a factor
in projecting Soviet defense programs. In 1983-84, the discussion of the origins of the slowdown
unearthed an array of possible causes, but they centered first on adaptations to current difficulties in the
economy while playing down the possibility that a strategic decision had been made to limit the growth
in the defense budget.
The debate over Soviet defense policy heated up when Gorbachev became General Secretary in 1985.
CIA analysts thought that his ambitious investment program in support ofhis goal of modernizing the
USSR's production base would "leave little or no room for increasing defense spending procurement
above its recent rates."Jl In fact, according to CIA's estimates, Soviet defense spending accelerated in
the first year of the 1986-90 plan, and Gorbachev later said that this growth had been planned.
Gorbachev probably believed his focus on labor discipline and a shakeup of the party and managerial
ranks would result in productivity gains that would permit substantial growth in both defense
procurement and the production of equipment for investment.
CIA's analytic line at the time was that the capacity to produce the strategic weapons coming on line was
already in place, suggesting "that Gorbachev can handle the defense/growth tradeoff for the time being."
The real test for Gorbachev would be when demands to expand and renovate defense industries are
renewed. "If the modernization program is not going well, Gorbachev will have to deal with military
leaders asking for more defense-related investment and advocates of devoting even more resources to
modernization--here the plan begins to unravel and choices have to be made."J! The reforms in the
economy introduced in mid-1987 proved to be disruptive rather than corrective. According to Marshall
Sergey Akhromeyev, a decision to plan a cutback in defense was taken in mid- I 988.12 At this point at
least, the connection among developments in the economy, the defense burden, and decisions on defense
spending was clear enough.
So, in every case, the IC concluded that Soviet economic difficulties would impinge only marginally, if
at all, on Soviet defense plans. What mattered was the technological competition. How would lagging
technology limit improvements in Soviet military capabilities? In general, the IC reasoned that in the
Soviet Union defense had such an overwhelming priority that economic problems would not constrain
growth in the allocation of resources to military programs. Only when Gorbachev's perestrokyka was
foundering did the idea of economic constraints on the defense budget gain a foothold in the national
estimates arena, and even then the majority opinion rejected the notion that the USSR would unilaterally
reduce its defense spending as it did in 1989.
Erosion in the Credibility of the Estimates
Why the concept of economic constraints on the USSR's defense budget was not taken more seriously,
especially after the mid- I 970s, is still unclear. Before 1976, one main reason surely was the woefully
incorrect CIA estimate of the level of Soviet defense spending. A joint analysis of the Soviet defense
burden in 1975 by the Office of Economic Research and OSR reported that the Soviet defense effort
took less than 8 percent of GNP and only 10 to I 5 percent of industrial production. Moreover, even
though defense expenditures had been growing by 3 percent per year, GNP had been increasing faster.
The paper's key judgments were:
The Soviet leaders have not acted as though costs have been a major factor in their military
decisions. Defense programs have been well funded, even during periods of lagging economic
growth, and the followthrough on new programs has been strong
00 7 6 0
Analyzmg Soviet Defen$e Programs, l95l-l990
Defense costs are unlikely to constrain the Soviets unduly in the }itture. Political priorities favor
continued emphasis on defense, and expected economic growth should permit continued increases
in defense spending. 20
The sharp upward revision in the ruble defense spending estimates in 1976 changed the estimate of the
defense burden radically. Instead of6 to 8 percent of GNP, defense spending was now 12 to 14 percent,
more than twice the share of defense in US GNP. Moreover, analysts were forced to revise their
description of Soviet defense industry as impressively efficient, and able to turn out hardware at
relatively low cost. By the time the paper was being written, new information on Soviet prices for
military hardware indicated that the ruble prices being used in the CIA estimates were much too low.
Then a Soviet emigre provided a figure for procurement that was much higher than the current CIA
estimate. A standdown in the spending estimates was ordered while the new intelligence was digested.
When new estimates were produced in the fall of 1976-77, procurement in rubles was higher by
two-thirds in the 1971-75 period, and total defense spending had increased by 55 percent.
The abruptness of the change in the defense spending estimates compiled in rubles badly eroded their
credibility among policymakers and in Congress. One Senate Armed Services Committee staff member
told an outside panel reviewing CIA's defense spending estimates that "there are a number of people
who believe this analysis is intentionaJly misleading, that it somehow comes out to exactly what it is that
the administration would like to argue on Capitol Hill. I mean they totally reject it."l.! President Nixon
complained that American presidents were being supplied by CIA with figures on Russian defense
spending that were "only half of what the Agency later decided they had been." He added, "Thanks to
this intelligence blunder, we will find ourselves looking down the nuclear barrel in the 1980s."
Policymakers were also confused by the Agency's delayed call on a slowdown in the growth of real
defense spending after 1975. At the annual convention of the American Newspaper Publishers'
Association in 1979, President Carter insisted that at that time neither the United States nor the USSR
had the upper hand in the strategic competition, but he then said, "What causes us concern is not the
current balance but the momentum of the Soviet strategic buildup. During the past decade, the Soviets
have steadily increased their real defense spending, year by year, while our own defense spending over
that decade has had a net decrease." In January 1979, Carter had defended his decision to push higher
US defense budgets by saying, "In the last I 0, 15 years, the Soviets every year have had above and
beyond inflationary costs a 4- or 5-percent increase in allocation of funds for defense purposes." 24
The Agency's Soviet defense spending estimates continued to be used in support of rising US defense
budgets in the first Reagan administration. Then, in February 1983, CIA published a dollar comparison
paper that said, "This year's review shows a period of almost no growth in the dollar costs for Soviet
procurement from 1976 to 1981." It said that new evidence on major procurement items that "became
available in 1982" led to a changed assessment. The result was a "procurement plateau" that had lasted
longer than previous pauses in procurement and one that "may signal a secular change in the pace or
composition of procurement."~ Moreover, largely because of the lower estimates for procurement, the
new Agency estimate showed substantially slower growth in total defense spending after 1976.
Before it was published, the paper was briefed to the National Intelligence Council (NIC) and to
Secretary of Defense Casper Weinberger. DCI Casey brought the news to President Reagan, saying in
effect, "We have a problem,"~ at a time when the Secretary of Defense was testifYing before Congress
on the need for increased spending for defense. The draft encountered strong resistance in the NIC, while
the Secretary of Defense reportedly "went nuts." 27 So the controversy continued through most of 1983,
although the focus shifted to the reasons for and likely duration of the slowdown rather than its
existence--at least through 1982.1!!
Meanwhile, critics of administration foreign policy seized on the estimates as evidence that the policy
was wrong. Raymond Garthoffwrote that talk of a "relentless Soviet buildup "was mistaken" and that
the "spending gap" was no more solid than other gaps (bombers, missiles, ABM civil defense) that had
7 , ... 80
u,,et Defense Programs, 1951-1990 (U)
been headline news. 29 In a later analysis of Soviet foreign policy, Garthoff emphasized that President
Reagan's spending gap was "deflated by its own intelligence estimates."
At the same time, there were alternative estimates of Soviet defense spending. The Stockholm
International Peace Research Institute presented a series that replicated the Soviet official defense budget
for almost every year from 1955 to 1988. London's International Institute for Strategic Studies took the
official defense budget from 1955 to 1974 as a base and added some rubles for military RDT &E and
military security forces.
On the other hand, some other estimates had Soviet defense spending increasing much more rapidly and
to much higher levels than CIA's estimates.ll. These more threatening estimates were seized on by those
advocating larger defense budgets because they buttressed the position of those who believed Soviet
defense spending had been rising rapidly ever since the early 1960s and supposedly demonstrated that
mounting economic difficulties had not influenced defense policy.B
The organizations that identified themselves with the alternative estimates--most prominently the
Committee on the Present Danger--were led by influential people in the foreign policy field--for
example, Paul Nitze, Eugene Rostow, and David Packard. According to Fred Kaplan's count, 32
members of the Committee on the Present Danger joined the new Reagan administration in 1981
(including the President himself).ll Their presence could not help but further weaken the authority of
CIA's defense spending estimates in policymaking circles.
How Good Were the Estimates?
Now the estimates of Soviet defense spending and GNP are history and potentially useful only for
historical research--depending on how reliable they are. In this context, the revisions in GNP and
defense spending estimates that Soviet or Russian statistical agencies have made over the last decade,
the judgments of external review panels and Western experts, and some alternative estimates of Soviet
and Russian leaders and scholars are of interest.
Official and semiofficial revisions. Little has been done by Russian or other Confederation of
Independent States (CIS) statistical agencies to remove the distortions in Soviet-era statistics on GNP or
production generally. The statistical agencies of the CIS states have other fish to fry. 34 The problem is
not with the official data on GNP in current prices, incomplete as they are. They agree quite well with
the Agency's estimates. But Western and Russian observers reject the price indexes that have been used
to convert GNP in current prices to GNP in constant prices. International agencies have been working
with CIS statistical authorities to improve the price indexes in the post-Soviet era. There seems to be
little hope, however, that these authorities will go back in time and redo the indexes of the Soviet period.
The official statistics on defense spending are even worse. Not until 1989 did the Soviets abandon the
transparently false figures that had been reported as allocations to defense in successive state budgets.
The new figures given for 1988 and 1989 are more than three times higher than the old budget entries,
but still much less than the CIA's estimates. Part of the difference is because some security-related
outlays are reported elsewhere in the state budget (KGB border guards, militarized internal security
troops, and the like), but the revised defense budget has not been generally accepted in the West or by
many Russian officials and academics.E
Moreover, this dispute concerns Soviet defense spending in current prices. Neither the Soviets nor the
have reported a series representing the growth of defense spending in constant prices. If they
had, It would have been suspect for the same reason as the official estimates of growth of GNP, national
income, or industrial production in constant prices--the lack of reliable price indexes to convert defense
outlays in current prices to a constant-price basis.
External reviews and assessments. The CIA estimates of both Soviet GNP and Soviet defense have been
repeatedly examined and assessed by outside experts. The methodology underlying the GNP estimates
c>vVIet Defense Programs, 1951·1990 (U)
was set out in a series of unclassified publications and reviewed at three conferences of academic
experts.12 In addition, four independent external review panels examined and assessed CIA's research
on Soviet GNP.l!
In broad terms, the review panels (except for the GAO report) found little to quarrel with the estimates
of GNP in rubles but thought the estimates of Soviet GNP in dollars were probably overstated due to the
failure to account fully for the qualitative superiority of US goods and services as compared with the
Soviet counterparts (Figure 4 ). As two of the panels reported, the GNP measures have stood the test of
the time and the marketplace. Two major groups investigating the economic history of the Soviet Union
are using the Agency's GNP estimates.1.2
From the early 1970s, the defense spending estimates came under the scrutiny of a panel of outside
experts established by DCI Richard Helms. The DCI's Military-Economic Advisory Panel, meeting
twice a year, reviewed the methodology underlying the estimates and recommended improvements. In
1982, DDI Robert Gates commissioned a special panel of outside experts (the Working Group on Soviet
Military-Economic Analysis) to revisit the question of the quality and the utility ofthe defense spending
estimates. Two reports were given to the DDI in July 1983.40 The Working Group said, "Overall, and in
spite of the range of deficiencies that we will delineate ... the CIA does an excellent job of estimating
Soviet military expenditures.".±! The panel established at the request of the House Permanent Select
Subcommittee on Intelligence (HPSCI) came to a quite different conclusion. It thought the available
information "insufficient to support any reliable estimate of Soviet defense spending by any known
method of estimation. 42
Soviet and Russian statements. During the past several years, a number of prominent Soviets have given
figures for Soviet defense spending--usually given as a share of Marxist-style national income or GNP.
Former Soviet leaders President Gorbachev, Foreign Minister Shevardnadze, Politburo member
Ligachev, and Minister of Defense Yazov put Soviet defense spending at or near the lower range of CIA
estimates of Soviet military spending in current rubles (Figure 5). Their pronouncements in tum imply
defense spending substantially higher than the new official defense budget announced in 1989. Some
other--mainly nonofficial--Soviets believed the defense burden to be much greater than that suggested by
Gorbachev, Shevardnadze, and Ligachev--as high as 200 billion rubles.
PFIAB Panel Report*
• "CIA's quantitative estimates of past Soviet output levels and rates of growth are the best
available." (p. 3)
--"Have stood the test of the marketplace." (p. 3)
--"Non~ of the alternative measures have been proven to be superior to those of the CIA." (p. 4)
• The estimates have greater margin of error than those for Western economics, but in Panel's
judgment, "Error is not likely to be so large that its elimination would change one's assessment of
the strengths and weakness of the Soviet economy."
*President's Foreign Intelligence Advisory Board Panel, "An Evaluation of CIA Work on Soviet
Economic Capabilities, Problems, and Prospects," 7 July 1983.
Rowen Panel Report*
• SWG concludes that there are reasons to believe CIA has overestimated Soviet economic growth.
• "In any case, several members of the group doubt that the cumulative distortion between reported
10 of l
http://www .csi.cia!studies/vo142no3/art 7 ./Jtm l
::.uviet Defense Programs, 1951 ~ 1990 (U)
Soviet performance and reality could be more than 10 to 15 percentage points, in the aggregate
over the past decade." (pp. 17 -18)
*Rowen Panel, Report of the Soviet Working Group, May 1985.
• "The main findings (re GNP growth) have stood the test of time welL" (p. 37)
• "Our most significant critical conclusion is that certain CIA products during the 1970s and 80s
may indeed have tended to overstate the size of the Soviet economy relative to the United States ..
. ." (p. 47)
*Daniel M. Berkowitz, JosephS. Berliner, Paul R. Gregory, Susan J. Linz, and James R. Millar.
"An Evaluation of the CIA's Analysis of Soviet Economic Performance, 1970-90." Comparative
Economic Studies, 35, 2:33-38, summer 1993.
• CIA used "generally reliable methodologies" in measuring GNP but had to employ a
"considerable degree of judgment."
• --"As a result, these methods are unlikely to produce accurate estimates of Soviet GNP in rubles."
• "CIA has probably overestimated the relative size of Soviet GNP as compared with US GNP,
although any failure to capture a significant portion of the second economy in its estimates could
offset the amount of overestimation."
• *United States GAO, Soviet Economy: Assessment of How Well the CIA Has Estimated the Size
ofthe Economy, GAO/NSIAD-91-274, September 1991.
Figure 5. Soviet defense expenditures selected Soviet statements and CIA's estimate. (U)
Three observations seem warranted regarding the conflicting Soviet declarations concerning the size of
the USSR's defense budget. First, there are enough statements by top Soviet leaders to indicate that they
believed the new official defense budget understated the real amount by one-third to two-fifths. Second,
the various references to the defense burden are sufficiently vague and different to indicate that even the
Politburo was relying on incomplete and unverified evidence. Third, if the leadership was uncertain
about the level of Soviet defense spending after having ordered the Ministry of Defense to develop an
honest total, little credence should be given to the back-of-the-envelope calculations of academics and
journalists who had no access to the accounts of the Ministry of Defense, the security organs, or the
Whether efforts to reconstruct Soviet defense spending will ever succeed is questionable. The
impediments include the elusive subsidies granted to defense production, the possible off-budget
financing of defense programs, and the difficulty of estimating the cost of services provided free or at
large discounts to the military by local governments and nondefense ministries._fl_
What can be said of CIA's building-block estimates is that they reflected the best intelligence available
on the physical elements of Soviet military programs and forces. It was these quantities that determined
the trends in CIA's spending estimates. When 1970 ruble prices were substituted for 1955 ruble prices,
the level of defense spending in CIA estimates shifted upward, but the rate of growth was little affected.
c'""'YLII'JS Jvviet Defense l'rograms, 1951-1990 (U)
Similarly, when 1982 ruble prices were substituted for 1970 prices, the level of spending jumped by
more than two-thirds, but the rate of growth of defense outlays in 1956-75 increased from only 3.5 to 3.6
percent per year. Even a much more radical shift in relative price weights (from 1982 ruble prices to
1989 dollar prices) makes little difference in the estimated rate of growth: defense outlays in 1966-88
rise by 2.8 percent per year in rubles and by 2.0 percent per year in dollars.
The controversy over the levels and rates of growth of Soviet defense spending probably will never be
settled to everyone's satisfaction. The statistical authorities of the CIS states show no inclination or
capacity to revisit this problem. The building-block approach, however, avoided most of the dead ends
that alternative approaches based on Soviet aggregative statistics encountered. The building-block
estimates also are consistent with revelations by Soviet leaders in a position to have informed judgments
about the size ofthe USSR's defense budget and the volume of production of military hardware. If these
estimates are to be challenged--especially with respect to trend--the underlying estimates of force levels,
operating practices, and production must be overturned.
The post-1985 discussion of Soviet military budgets, however, does suggest that one of the virtues
claimed for the ruble estimates--that they were what the Soviet leaders were looking at in making their
decisions--may lack substance. It seems clear that the leadership thought there was little inflation so that
the distinction between spending in current as opposed to constant or "comparable" prices was of small
importance. But the weight of the evidence seems to point to the leadership's not knowing what the
all-inclusive defense budgets were. Thus, the principal usefulness of the Western estimates probably was
to give the West a sense of the trends in real Soviet defense outlays and the role they played in the
growth of national output and in the decades-long economic slowdown in the Soviet Union.
~ Statement by Dulles to the Subcommittee on Economic Statistics of the Joint Economic Committee of
the Congress of the United States, 13 November 1959, in Washington, DC, pp. 20-25.
~Fred Kaplan, The Wizards of Armageddon, New York: Touchstone, 1983; pp. 321-327.
~Ibid., p. I 0.
20 May 1977.
9 Interview with Rae Huffstutler, former chief of CIA's Office of Strategic Research and of the Office of
Soviet Analysis, 3 June 1994. In the event, only 60 were deployed.
10 Report of Schlesinger to the Congress on the FY 1976 and Transition Budgets FY 1977
Authorization Request and FY 1976-1980 Defense Programs, 5 February 1975,
l of l
• t'tt<HYL~'~'E; .:>vv 1ct
Defense Programs, 1951-1990 (U)
!! Interview with Schlesinger, 9 March 1995.
g Testimony before the Selin Panel of the Working Group on Soviet Military Economics, 24 March
13 SNIE 11-5-64, Soviet Economic Problems and Outlook, 8 January 1964, pp. 5-6.
~ NIE ll-4-65, Main Trends in Soviet Military Policy, 14 April 1965, p. l 0.
~ NIE ll-4-78, Soviet Goals and Expectations in the Global Power Arena, 7 July 1981, pp. 17-18.
18 Douglas MacEachin, "Gorbachev's Economic Strategy and Soviet Defense Spending," 13 December
20 "Topical Interview with Marshal Sergey Fedorovich Akhromeyev," Moscow television service in
Russian, 1700 GMT, 9 October 1989.
Richard Nixon, The Real World, New York: Random House, 1980; p. 26. The "nuclear barrel"
reference was a gross non sequitur. To the extent that US defense budgets were influenced by the
Agency's estimates, it was the dollar estimates that were in play, and these changed little when the ruble
estimates were revised.
24 Public Papers of the Presidents, Jimmy Carter, 1979,
p. 695. Washington, DC: US Government
Printing Office, 1980.
Ibid., p. 176.
~1 Interview with Rae Huffstutler, 3 June 1984. Huffstutler was then Director of the Office of Soviet
Analysis in CIA.
Interview with Robert Gates, 12 May 1994.
00 7 6 0
':'""'YL"l'!!. ,~ov1et Defense Programs, 1951-\990 (U)
30 Raymond L. Garthoff, Detente and Confrontation. Washington, DC: The Brookings Institution, 1985;
~. --, The Great Transition, Washington, DC: The Brookings Institution, 1994; p. 41,
See, e.g., William T. Lee, CIA Estimates of Soviet Military Expenditures: Errors and Waste.
Washington, DC: The American Enterprise Institute, 1995; pp. 119-153 and Steven Rosefield, False
Science. New Brunswick and London: Transaction Books, 1982; pp. 178-196.
In this connection, the National Strategy Information Center published one of Lee's monographs in
1977 as well as Rosenfield's 1982 book.
Fred Kaplan, The Wizards of Armageddon. New York: Touchstone, 1984; p, 386.
35 The Chairman ofthe Russian State Statistical Committee suggested in 1992 that this was a task best
left to academies (P. Guzhvin, "Doveriya k Statistike Trust in Statistics," Vestnik Statistiki, 9: 16
36 See James H. Noren, "The Controversy Over Western Measures of Soviet Defense Expenditures,"
Post-Soviet Affairs, 11, 3:328-296, 1995, for a discussion of the circumstances under which the revised
defense budget was compiled and the doubts that knowledgeable Soviets/Russians had about its
An Evaluation of CIA Work on Soviet Economic Capabilities, Problems, and Prospects. A report
commissioned by the President's Foreign Intelligence Advisory Board, 7 July 1983. Report of the Soviet
Working Group, a report of a panel chaired by Henry Rowen at the request of Robert Gates, Deputy
Director of CIA, May 1985; United States General Accounting Office, Soviet Economy -Assessment of
How Well the CIA Has Estimated the Size of the Economy, report to the Honorable Daniel P.
Moynihan, US Senate, GAO/NSIAD-91-274, September 1991; and the HPSCI report (cited in reference
40 Thus, the International Comparison of Productivity (I COP) Project in the Netherlands uses the CIA
estimates. ("In my view the RAND-CIA material in the growth of Soviet output by industry of origin for
the period 1928-90 is as good as we are likely to geL," Angus Maddison, "Economic Performances of
the Former USSR, a Quantitative Evaluation").
!4 of l
Defense Programs, 1951-1990 (U)
http://www .cs i.cia!studies/vol42no3/art7.htm I
42 Working Group on Soviet Military Economic Analysis, Report of the Working Group on Soviet
Military Economic Analysis, p.8. The report was reproduced in Berkowitz, Daniel M. Joseph S.
Berliner, Paul R. Gregory, Susan J. Linz, and James R. Millar, "An Evaluation of the CIA's Analysis of
Soviet Economic Performance, 1970-90, Comparative Economic Studies, 35, 2:33-58, summer 1993.
43 The subsidies issue is particularly troubling. As noted earlier, Soviet civilian and military leaders
argued that a true defense budget could not be compiled until a price reform had been completed.
According to Vadim Kuznetzov of the Foreign Ministry's USA desk, it would take several years (the
early 1990s) to introduce a new system of accounting in which defense industries have to pay their own
way rather than live on state subsidies (Washington Times, 1 August 1998). He explained, "Under the
old system, a truck could cost many times less for the military than for a collective farm."
'served in the Directorate of Intelligence.