AbbreviationsManagerial Support for Spending Estimates ErodesWider Use of Estimates Generates Distrust
Through the early 1980s the demand for CIA's estimates of Soviet military spending continued at a high level. The dollar comparisons of the cost of Soviet and U.S. programs were especially in demand; they proved useful in supporting the U.S. defense buildup during the Carter and early Reagan presidencies. William Kaufman, a longtime Pentagon consultant, recalled that during the Nixon and Ford administrations he "turned very strongly" to the CIA's estimates "as a basis for indicating why the U.S. defense budget ought to be increased."
To this end the comparisons were inserted into the annual DoD posture statements. They also appeared in Soviet Military Power, the DoD series documenting various facets of the Soviet military threat.
Indeed, the increasing use of the CIA estimates in the debate over the U.S. defense budget aroused suspicion about their integrity in Congress and academia. James Locher, a staff member of the Senate Armed Services Committee, told the Selin panel "that there are a number of people who believe this analysis is intentionally 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." Depending on what the United States is doing with its defense budget, either side of the "political spectrum—(will be)—unhappy with the analysis." Writing for the Stockholm International Peace Research Institute (SIPRI), Carl Jacobsen charged that the level of CIA's dollar estimates of Soviet military program costs coupled with high rates of growth of these costs "were seen to legitimize and compel the Reagan Administration's historically unprecedented defense budget increases." Holzman asserted that "over the past 15 years, the CIA's estimates have been riddled with errors and misrepresentations, all making Soviet defense expenditures appear larger than they actually are." On the other hand, Rosefielde wrote in 1982, "my impression is that the political pressures on the CIA to keep their estimates low are very strong."Agency's New Senior Managers Skeptical
When Casey became DCI in 1981 he was not on record regarding CIA's estimates of Soviet defense spending. Early on, however, he indicated his doubts about the dollar estimates by asking Huffstutler how much could be saved by eliminating them. Very likely, Casey's views were influenced greatly by those of Gates, who served him first as his special assistant, later as his DDI, and finally as DDCI. In 1984 when Gates was preparing to brief Weinberger on the latest dollar comparisons, which showed that the difference between U.S. and Soviet defense spending had narrowed considerably, he told Casey he planned to urge DoD to rely on physical comparisons rather than dollar comparisons of the defense programs. According to meeting records, "the DCI agreed that this would be a better approach than using a lot of artificial cost figures." Nonetheless, Casey kept his distance from decisions over the spending estimates and CIA's economic papers in general. Gates insists, for example, that Casey never interfered with the publication of any economic paper. This claim is disputed by John Reynolds, a former manager of CIA's military-economic analysis, who reviewed this book in draft and noted that he personally drafted memorandums for Casey's signature refusing requests from the Senate Select Committee on Intelligence (SSCI) for public release of CIA's dollar cost comparisons, even though they had been made available in earlier years. Both Gates and Reynolds could be right in the sense that Casey at the outset clearly intended to stop the dissemination of unclassified CIA reports. Gates proved to be the driving force in turning agency management against the defense spending estimates.
When he became DDI in 1981 he already had "some fundamental misgivings about some of the CIA's work on costing of the Soviet defense effort" based on the fundamental difference in the U.S. and Soviet economies and his feeling that sufficient data did not exist to make accurate comparisons. He distrusted the dollar comparisons as only an "analytical construct" and pressed to substitute physical comparisons (numbers of missiles, aircraft, and tanks) for the dollar comparisons. He also thought the ruble defense spending estimates did not take into account the many ways the Soviet economy was skewed toward military purposes. "No one around town," he said, believes CIA and DIA when they say defense is only 13-16 percent of GNP. Instead Gates said that, when asked, he would answer that the Soviet economy was about 15-20 percent purely military and 20 percent purely civilian, and the remainder could not be broken down into civilian or military components. Gates's intuitive views were generally shared by those who followed him as DDI. In fact, his appointment as DDI signaled a marked change in the background and attitude of the people who occupied that position. Whereas DDIs of the 1970s like Proctor and Clarke came from intelligence backgrounds of military and military-economic analysis, understood the estimates thoroughly, and felt comfortable explaining what they meant and didn't mean, DDIs of the 1980s didn't understand the estimates and were not at ease discussing or defending them.
Because of his discomfort with the estimates, Gates decided to commission Selin, then chair of MEAP, to establish a working group of outside experts, including some MEAP members, to review the utility and reliability of the estimates. As we have discussed, the resulting working group made several recommendations for improvements in the estimates, but it fundamentally endorsed them. The group's findings were reported to Weinberger and, according to Casey, received wide distribution within the administration. Still, the intuitive discomfort with the spending estimates and the apparent lack of confidence in subordinate experts persisted.Estimates No Longer Helpful
The distrust of the estimates stemmed in large part from the lingering embarrassment over the 1976 major revision of the ruble estimate and the post-1982 controversy surrounding the announcement of the slowdown in the real growth of Soviet defense outlays after 1975. Gates delayed several months before approving the SOVA paper dealing with the slowdown
, "awaiting the outcome of the Selin panel report and SOVA's work with DIA" (comparing CIA and DIA estimates). In his forwarding memorandum to DCI Casey and DDCI McMahon, Gates said, "this is the paper over which I have had the greatest doubts," but he now found it persuasive because of SOVA's work on the Soviet economy and the endorsement of SOVA's economic and military-economic analysis by the Selin panel and the PFIAB panel, which he called "the two most far-reaching, comprehensive external evaluations of our analysis in many years."
When the paper's findings became known, there was general consternation. The Senate Armed Services Committee, for example, canceled a major hearing because the idea of slower growth in Soviet defense spending generated enough controversy to make "our hearings less useful for our purposes." Just as the concept of a plateau was being digested, DIA poured fuel on the fire by reporting in mid-1984 that Soviet procurement (valued in constant dollars) had increased by 5-10 percent in 1983 while CIA's preliminary estimate was 2-3 percent. In testimony before the JEC, DIA was questioned intensely about this difference. CIA's representatives were grilled on the same discrepancy and testified that they disagreed with DIA on the estimates of physical production in 1983. They said the DIA estimate of spending growth was based on a smaller number of Soviet procurement items than CIA's estimates and that the DIA methodology overweighted higher growth items. After a comparison of their respective estimates, DIA scaled back its estimate, but the 1983 upturn in procurement in the CIA and DIA estimates turned out to be caused by lead costs of future systems that were never procured at anticipated levels.
Meanwhile, the CIA/DIA differences were awkward for the intelligence community. In part they led Gates to ask whether the paper announcing the slowdown in Soviet defense spending should be delayed. When the agency's classified dollar-comparison paper was published in March, the findings were leaked to the press—findings that were sharply at variance with DIA's estimates. To defuse headlines heralding a CIA/DIA split, Ikle arranged a March 1983 press conference involving DoD, DIA, and CIA to set the record straight. DIA and CIA reported their positions while minimizing their differences and cautioned that even with unchanged levels of spending, the USSR would buy enormous quantities of military equipment.
In 1986 when CIA and DIA testified jointly before the JEC, the two agencies still disagreed on trends in 1983-84 procurement. Even so, the degree of convergence in views and the fact that the testimony was joint led some to charge that CIA and DIA had fudged their differences. In this vein, Holzman argued that CIA "may have moved, however reluctantly, to harmonize its published estimates of ME/GNP with those of the DIA." Nothing so sinister occurred. The agency's estimate as presented to the JEC continued to show slow growth in real Soviet defense outlays until an acceleration forecast for 1985-88. The estimate of the burden increased because of the changeover to 1982 prices. Beginning in 1985, however, CIA/ DIA consultations over defense-economic matters were held on a regular basis with a revival of the Joint Military Costing Review Board, which met quarterly to review estimates and other work in progress.
From some consumers' standpoint, the finding that the gap, measured in dollars, between higher Soviet and lower U.S. defense spending was narrowing also meant that the CIA estimates were becoming a liability. Gates, already dubious about the utility of the dollar estimates, briefed Weinberger in July 1984. In a draft memorandum summarizing the sub-stance of the session with Weinberger, Gates referred to "future use of expenditure analysis by the Department of Defense" and noted "the use of the dollar comparison figures of Soviet defense expenditures by DoD to defend the defense budget was becoming less useful." He urged DoD to use physical comparisons of forces and procurement and noted that as the U.S. buildup continued, the U.S.-Soviet gap in spending would continue to narrow "and will not show the kind of gaps that Defense has used in the past to underscore U.S. defense needs."
Weinberger and Casey agreed with Gates's recommendation to stop publication of the dollar comparisons
. But this proved difficult, and soon problems materialized. SSCI asked for the defense spending comparisons, and a decision had to be made on permitting publication of an overall comparison of U.S. and Soviet GNP (including defense) for the JEC. SOVA began meeting with JEC staff about the upcoming DDI's testimony in November 1984. Pressed hard by Senator Proxmire, Gates agreed to give the comparisons to the JEC. SOVA soon told Gates, however, that the JEC request was only one of many pending, including requests from the Department of Commerce, Supreme Allied Command Europe (SACEUR), the Office of Management and Budget (OMB), the Deputy for Defense Research and Engineering (OSD/DDR&E), the Director of Net Assessment in the Office of the Secretary of Defense (OSD), Strategic Air Command (SAC), and SSCI. A policy on release was needed, especially because some customers suspected that the decision to stop publication was based at least partly on what the trends showed.
111 Requests from DoD continued to pour in. DIA sent a memorandum to the DDI asking for more detailed dollar comparisons. While understanding his "discomfort" with the dollar estimates, DIA said "unfortunately, no acceptable substitute has been found for the measure of the size, direction, and significance of the overall Soviet military effort."
SOVA Director Douglas MacEachin recommended in January 1985 that publication of the dollar comparison be resumed, arguing that the past several months had shown that customers would create their own estimates and that CIA needed to again be in position to control the quality of the comparisons and state the proper caveats. Gates reluctantly agreed. As he wrote to Deputy National Security Adviser Robert MacFarlane, "Recent developments have made it necessary to reverse (my earlier) decision." "I wanted to apprise you of these developments because of the earlier interest shown by you and John Poindexter." But Gates promised changes in the dollar-comparison paper,
including a beginning section comparing accumulated physical stocks of military hardware, inserting in the key judgments section of the paper warnings of the limitations and proper uses of the comparisons, and relying as much as possible on percentage rather than absolute comparisons. At the DoD, Weinberger was briefed by Andrew Marshall on the preliminary results of the latest CIA dollar comparison. He told Weinberger that the most important trend showed U.S. procurement overtaking Soviet procurement in 1983 and that DoD should be prepared to respond. On the positive side, Marshall said it was unlikely that the USSR would return to earlier rates of growth of defense spending for any sustained period, giving the United States an opportunity to challenge the Soviets in a way they would be hard pressed to match.
During the late 1970s and early 1980s demand increased for an expanded definition of national security spending, on the grounds that the traditional definitions used for calculating U.S. and Soviet defense outlays did not capture the entire national security-related burden. Gates asked SOVA to research the question so that the dollar-comparison paper could include a "broader view of the Soviet burden."
 This effort, described in the next chapter, increased the Soviet burden relative to that of the United States—measured in their respective currencies. The ratio of the cost of Soviet national security programs to comparable U.S. programs, measured in dollars, was marginally higher than the ratio reflecting the more traditional definition of defense.
The minor relief that comparisons based on an expanded definition of 93 defense provided to defenders of the U.S. defense budget was offset by the results of an expansion in another direction—a comparison of the dollar cost of NATO and WP defense programs which is discussed further in chapter 5. When DoD asked the agency for permission to give the finding to the JEC on an unclassified basis, DDI Richard Kerr demurred, saying he was reluctant to issue the paper in an unclassified form, and "in a form that will almost certainly be used against us."
 SOVA replied that the agency should avoid being accused of selectivity (having already released the results of the U.S.-Soviet comparison to the JEC). In any case, "our dollar estimates have long been both in demand and controversial." In reviewing the classified version of the paper, Gates, now DDCI, again voiced his concerns. Although he had few specific problems, he said it contained the same old caveats
, which would be ignored: "One more time, our approach is an analytical construct, not a reflection of the real world." He predicted that the findings showing that the cumulative dollar costs of NATO defense programs in 1976-86 greatly exceeded the comparable costs for WP countries would be "precisely the part that will be leaked by the opponents of defense spending and used." Clearly, high-level opposition to the dollar comparison was building again.Intracommunity Disputes over Economic Constraints on Soviet Military Programs
Through most of the 1970s and 1980s an underlying tension existed between the analysis of the Soviet economy and intelligence community expectations about the development of the USSR's armed forces. A study prepared for the JEC in 1977 summarized the agency's views on the outlook for the economy: reduced growth with no easy options to prevent the slowdown. Still, the implications for defense were hardly drastic, according to the paper: "The slowdown in economic growth could trigger intense debate in Moscow over the future levels and pattern of military expenditures. Military programs enjoy great momentum and powerful political and bureaucratic support. We expect defense spending to continue to increase in the next few years at something like the recent annual rates of 4 to 5 percent because of programs in train. As the economy slows, however, ways to reduce the growth of defense expenditures could become increasingly pressing for some elements of the Soviet leadership." The message that the Soviet economy was in trouble, although not new, began to percolate in wider circles, causing some to wonder whether the economic difficulties could lead to greater Soviet accommodation in arms control and restraint in military programs generally.
A few years later an interoffice task force led by OSR tackled the question of what these military programs would look like in the 1980s. More than forty individual research projects contributed to a major report published in early 1981. In the process OER argued that the paper's projections of defense programs for the 1980s did not take seriously enough the possibility that economic constraints would force slower growth or even a downturn in Soviet defense spending. The approved paper concluded that programs in progress implied continued growth in defense spending in real terms through the 1980s, although "political strains resulting from growing economic problems could lead them (the Soviets) to moderate the growth of spending, particularly late in the decade." Alternative scenarios considered two versions of stepped-up military programs: one featuring strategic forces and other general purpose forces, and one in which spending was reduced. The paper warned, however, that there was no evidence that the USSR planned to cut military spending. All three alternative scenarios to the baseline projection were considered unlikely.
Undoubtedly, if the late-1970s CIA estimates of Soviet military outlays had shown the slowdown in spending beginning in 1975 that later estimates did, the analysis in The Development of Soviet Military Power would have proceeded somewhat differently, giving greater weight to the downside of the force projections. The 1983 SOVA paper announcing the slowdown, for example, thought it more likely that the Soviets would continue the trend of little or no growth in procurement "in the hope of strengthening the economy for a long-term military competition." But the idea of economic constraints on Soviet military programs did not win many adherents in other parts of the intelligence community, partly because many thought CIA's analysis of the Soviet economy was too bleak. The PFIAB review of the agency's work on the economy was sparked by suspicions of this kind. There was a desire not to underrate the Soviet Union's capacity to sustain the military competition with the West. Thus Rowen, NIC chair in the early Casey years, turned back draft JEC testimony prepared in SOVA and ordered that it begin with a listing of basic Soviet economic strengths. Gates, as DDI, rejected a draft economic paper, saying, "The continuing litany of 2 percent growth for the 1980s has less and less credibility with me."
What, he asked, is the potential for policy changes to improve the economy?
In the NIE arena, the proposition that economic constraints would impede the implementation of ongoing military programs had even less weight. We have already discussed the one-sided debate over the strategic forces estimate (NIE II-3/8-85). The wide-ranging survey of Soviet strategic force developments delivered to two senate defense subcommittees by Gates and Lawrence Gershwin, NIO for Strategic Programs, in June 1985 illustrates the official view in the early Gorbachev years. They told the senators that the projected Soviet forces would "result in a growth of total Soviet strategic force expenditures of between 5 and 7 percent a year over the next five years," and 7-10 percent a year if Moscow decided to deploy ABM defenses on a wide scale. Soviet economic problems, while severe, would not compel the USSR to give up important strategic programs or make significant concessions in arms control, although "the stark economic realities might cause some deployment programs to be stretched out." (The disparity between the flat spending on strategic programs of the recent past and the vigorous growth they projected for the future was not addressed.) Rather than using arms control to relieve economic pressures, the Soviet leadership would orchestrate its arms control policy to advance its "strategy of achieving strategic advantage" by limiting U.S. force modernization and blocking the Strategic Defense Initiative (SDI).
When Gorbachev announced his plans for industrial modernization, SOVA pointed out the conflict between his program and Soviet military objectives, especially in the machine-building and metal-working (MBMW) branch of industry, which was the source of most defense hardware and the machinery and equipment required for investment programs. Indeed, as we noted in chapter 5, Gorbachev evidently planned to step up the growth of defense spending in the 1986-90 plan period. Also, the high-technology branches of MBMW (electronics, computers, communications equipment) were being counted on to upgrade the stock of fixed capital in the economy and close the technological gap between western and Soviet weaponry. A March 1986 SOVA paper described this civil-military competition. It said the near-term effects on defense programs would probably not be large but that a crunch could develop in the late 1980s or early 1990s because Gorbachev's chances of success were dim: "Unless Gorbachev's efforts to modernize industry pay off in greater numbers of more advanced, high-quality equipment and in substantially increased productivity, the battle between civilian and defense interests will become more severe."
The Washington policy community's reaction to this message was mixed. The paper was briefed to Secretary of State George Schultz and Weinberger in the spring of 1986. Schultz was taken with the idea of an impending crunch and what it portended for Soviet interest in arms control. Weinberger and his aides, on the other hand, thought the Soviet industrial modernization program, after a few years, would put the USSR in a better position to carry out its force modernization plans. Even when it became clear that Gorbachev's economic program was foundering, there was generally great reluctance—whether within the intelligence community or in policy circles—to accept the idea that economic constraints would impinge seriously on Soviet defense programs.
In this vein, Gates, then acting DCI, asked in 1987 that a memorandum on the state of the economy focusing on Soviet strengths be sent to the secretary of defense. The lead sentence of the resulting paper declared, "The Soviet economy, despite some ugly warts that make its operation costly and inefficient, is vested with great crude strength from the enormous resource base, and remains a viable system, capable of producing large quantities of goods and services annually, especially for industrial and military applications." The memorandum reported that even with no growth in GNP the Soviet Union could still spend the equivalent of $275 billion on its military. But zero growth was unlikely, and—although a relaxation of international tensions would help the modernization program "by freeing some resources that would otherwise be spent on defense"— Moscow had "the resources and the will" to increase defense outlays if the international climate did not improve. Industrial modernization and gains in living standards would be held down, "but it would not cause economic collapse or social upheaval."The End Game
As the Soviet Union disintegrated, the support and rationale for CIA estimates of the cost of Soviet military programs evaporated. First to go were the dollar comparisons. In August 1990 a SOVA memorandum considered the pros and cons of continuing the dollar estimates and recommended their continuance while putting consumers on notice that they would be eliminated in one to two years. By then SOVA had dissolved its Defense Comparisons Branch and put the analysts to work on other accounts. Shortly thereafter, SOVA provided talking points to DDI John Helgerson to justify dropping the dollar estimates. Marshall in mid-1991 urged that the dollar database be maintained because it "has been a major foundation for my office's work on a military investment balance report that top defense officials have found very useful in characterizing broad trends in our security situation." In his reply to Marshall, Helgerson reaffirmed his decision to discontinue the dollar comparisons. Letters to this effect were sent to the principal customers of the estimates on April 1, 1992.
Meanwhile the effort devoted to estimating the ruble cost of Soviet defense programs was scaled back to free analysts to work on problems connected with the unprecedented changes in the USSR. The comprehensive SCAM database was mothballed in favor of a spreadsheet model requiring much less data and operating at a higher level of aggregation. The costing contracts were also dropped, and the agency's commitment to defense costing was reduced from forty-five analysts, managers, and on-site contractors in the early 1980s to fourteen in early 1992. CIA management wanted the ruble work to continue, but at a much lower level of effort. Ironically more than a decade of refining the Soviet defense spending estimates was marked by increasing official doubts and disavowal. Still, the post-1975 research developed an improved set of estimates of the cost of Soviet defense programs. We present these in the next chapter.
(Firth, Noel E., Noren, James H, Soviet Defense Spending -A History of CIA Estimates, 1950-1990-
, Texas A&M University Press, 1998, pp.87-97)