LETTER TO WILLIAM CASEY (SANITIZED)

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP86M00886R000400110001-8
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 7
• Document Creation Date: December 21, 2016
• Document Release Date: August 28, 2008
• Sequence Number: 1
• Publication Date: December 17, 1984
• Content Type: LETTER

EXECUTIVE SECRETARIAT
ROUTING SLIP
ACTION
INFO
DATE
INITIAL
1
DCI
X
2
DDCI
X
3
EXDIR
X
4
D/ICS
5
DDI
X
6
DDA
7
DDO
8
DDS&T
9
Chm/NIC
10
GC
11
IG
12
Compt
13
D/Pans
14
D/OLL
15
D/PAO
16
SA/IA
17.
AO/DCI
18
C/IPD/OIS
19
20
21
22
 The Honorable
William Casey
Director
Central Intelligence Agency
Washington, D.C. 20505
Your MEAP held its fall meeting on November 30 and December 1. A
number of points came out of this meeting, likely to be of interest to you
and your senior staff. In particular, we followed your request and spent
half the meeting at DIA. This produced a very illuminating session
concerning their ways of doing business; we have several suggestions on
ways to improve CIA/DIA interaction.
We have been increasingly concerned about the apparent inability
of CIA and DIA to reconcile the differences between their respective
estimates of Soviet military expenditure. Our meeting with DIA leads us to
believe that there are difficult problems of narrowing the differences in
the major area of DIA's military economic activity, Soviet ruble outlays
and the burden of defense, which we discuss below. However, in the area of
dollar estimates of procurement, there are considerably brighter prospects,
if the principals can demonstrate the required good will.
Dollar Value of Procurement. First, there are simple differences
of methodo ogy between the two organizations. Both use the same base price
to value the same piece of Soviet hardware, but they handle learning curves
differently. CIA takes the marginal cost for the units of a weapon system
bought in a given year, while DIA uses the lifetime average costs for the
weapons system. If there were no other differences than this one, DIA's
estimates of procurement growth would always be higher than CIA's.
The two agencies also calculate learning curves differently. DIA
treats a weapon system (e.g. the Mig 27) as a unit, whereas CIA calculates
learning on individual major sub-systems, such as each engine used in more
than one aircraft model. These differences revolve around important con-
ceptual issues, but we believe they can be resolved. There is no reason in
principle why the two agencies should not be able to agree on a common
methodology. They should be tasked to do so.
STAT
 A second area of differences is in production quantities. Often
both agencies agree on the order of battle for a weapon system, but not on
the production quantities or the start dates for future systems. Here a
solution is made more difficult by the administrative separation of the
responsibility for OB and production estimates in DIA. At the very least,
analysts from both organizations should continue to meet regularly to
discuss differences and narrow the range of disagreement as well as they
can. More formally, it would be desirable to reactivate the joint CIA/DIA
Military Costing Review Board, or create another working group in its
place, to deal with these problems on an ongoing basis. In any case, it is
highly desirable that in the event of CIA/DIA disagreement, the policy
consumers know where the agencies disagree and to what effect. That can be
accomplished by appropriate appendixes in the reports of both
organizations.
Ruble Military Expenditure. The basic methodologies of the two
organizations are totally different, so there can be no direct reconcilia-
tion of the estimates. DIA's estimates are based on an assumption that is
said to be supported by several other approaches, each of which is
individually subject to considerable estimating error. Our initial
impression is that there are important questions about the reliability of
the DIA approaches, but we will have specific recommendations to make to
DIA over the coming months as we familiarize ourselves with the details of
their methodology (we have been promised complete documentation of the
estimates).
The only way to compare directly the CIA estimates -- valued in
1970 prices and estimated from the bottom-up -- with those of DIA --
derived from Soviet aggregated data in current prices -- is by a set of
indexes translating 1970 prices into those of a recent year, say 1982. as
you know, SOVA is now attempting to do just that through a variety of
approaches. From the viewpoint of improving CIA/DIA communication and
understanding of their respective estimating approaches, it will be impor-
tant at an appropriate stage to subject the CIA study of price changes to
peer review in DIA.
CIA Policy
There are two areas of CIA policy where we think changes could
serve the interests of all parties concerned. The first relates to pre-
paring and publishing, at suitable levels of classification, estimates of
the dollar cost of Soviet military activity.
Put simply, the current policy is to do the analytical work
necessary for estimating the dollar values, but not to prepare or publish
the comprehensive document. Instead, CIA stands ready to respond to re-
quests for information from qualified consumers, e.g., DoD staff or con-
gressional committees. The stated reason for this policy is to prevent
misuse of CIA work.
 While we can sympathize with the objective, the policy is self-
defeating. On the practical level, if the information is to be made
available on request anyway, CIA would be better advised to keep the
initiative and determine the form of presentation, not merely react to
questions. CIA publication permits introduction of suitable caveats; it
also makes possible alternative estimates and a much richer presentation.
Of course the results presented in a CIA publication can still be misused,
but not as easily as a response to an outside request. In addition, we do
have some suggestions on how the published reports can minimize some of the
more flagrant abuses, which are appended at the end of this letter.
But our objections to this policy go deeper. They have to do with
the credibility and consistency problems implicit in a stop-and-go policy
of changing the estimates published from year to year in an erratic
fashion. The CIA has published a standard set of estimates, consisting of
various dollar and ruble series, as well as reports on manpower and other
resource categories, as the community-wide basis for a wide variety of
intelligence products. Ruble Soviet outlays, the "burden" rates, dollar
and ruble comparisons of the U.S. and the U.S.S.R., manpower and resource
comparisons -- these are components of an integral product -- a compre-
hensive, economic description of Soviet military activity. Each component
individually makes its own contribution to the all-round analysis.
Together, fitted into a particular conceptual framework, they provide the
most thoroughgoing picture of the Soviet military effort that can be
produced in the West. To delete one or another of these components is to
damage the integrity and utility of the estimate package as a whole.
The closest analogy would be the Bureau of Labor Statistics not
publishing some unemployment series because of misuse by a congressional
committee, an event that the Department of Labor would go to great pains to
avoid. The CIA estimates have attained a high degree of credibility and
utility, both in the government and, when appropriate, outside. We
strongly urge that the Agency return to its tried and proven program of
research and publication in military-economics.
The second policy which we would like to discuss concerns open
publication. You are aware of our concerns in this area, and our belief
that is possible to fashion a policy that continues to maintain an embargo
on publishing policy analyses, but does contribute non-policy, unclassified
research to the general body of knowledge. We have suggested to Bob Gates
a number of reports, published from 1982 to the present, which could have
been presented in unclassified form with varying but still reasonable sani-
tization.
We have discussed both these topics with Bob and were encouraged
by his response. He has also asked us to review the progress of the in-
creased efforts that you have invested into understanding and explaining
Soviet defense industries. We will work on this question over the next six
months, and report after the spring MEAP meeting.
 Specific Topics
1. When you met with the Panel in the spring you asked us about
new members. We have since added two -H n expert STAT
on Eastern Europe, and 0 who combines a background SIAI
in political science with good technical and industry back-
ground. F ___]will be a help in many areas, including the STAT
area of Soviet defense industries.
2. We hope that kill help in one area which needs further STAT
emphasis, i.e. the coordination of Soviet-Eastern Europe
economic analysis. It is clear that this has become a very
important area and, because of the SOVA/EUR partition, one
that requires greater management emphasis to achieve a level
of coordination and joint work.
3. Another area where further integration is required is in the
melding of policy and political analysis with economic anal-
ysis. In the major area of Soviet decision making in a time
of economic stringency, it is essential that the mode of
thinking of the economic analysts be more deeply affected by
political considerations and an awareness of policy alterna-
tives. We will work directly with SOYA in this area to give
concrete examples and recommendations.
4. We were pleased to see continued, encouraging evidence of
SOVA's attempts to use economic analysis and resource con-
siderations, to inform the projections of the force analysts.
5. R&D estimates are another area where important progress
appears to be continuing, particularly in shifting away from
the old, inconsistent methodology towards one more consistent
with other estimates. We were also pleased to have learned
that SOYA will be sponsoring a conference of outside experts
to help in charting a course in this difficult area.
6. One of the most fruitful ways to improve projections is
retrospective analyses -- systematically reviewing previous
years' projections to understand where and why mistakes were
made. This is particularly important when major changes in
trends are expected. Currently, there is a strong feeling
that Soviet procurement is starting to increase again. Pro-
jections for each of the last several years have shown the
same ramp effect, but no such change has been seen. SOYA is
already studying the flattening of procurement from 1976 to
1982 to explain why the Soviets did what they did; we would
like to see the study broadened to see why we were slow to
recognize the change, and why we have forecast but not seen
an upturn for each of the last three years.
 7. Another topic that deserves increased emphasis is that of
indicators of change in military-economic activity. When the
data are ambiguous and the intelligence community decides on
balance that one of several plausible projections is the most
likely, the analysts should not stop there. Rather they
should continue to home in on differences that should be
observed if the accepted projection proved wrong.
For instance the community now feels, based on highly ambigu-
ous data, that Soviet military procurement is starting once
again to rise. Some effort should be expended now to sharpen
the projection by identifying indicators which should be
examined in the next year or so to see if this growth pro-
jection appears to be coming true.
These are the major points of our review. We are working directly
with SOVA and with DIA on particular points of more localized impact on
each respective organization. Please do not hesitate to call on us for
further information, support, or examination of any of these points.
 Approved For Release 2008/08/28: CIA-RDP86M00886R000400110001-8
2. Toward the same end, present dollar cost comparisons
(U.S.-U.S.S.R.) only in the form of ratios, not absolute doT air values.
Show ratios of cumulative sums along with ratios for particular years.
3. Better still, always juxtapose ruble with dollar cost
comparisons -- i.e., ratios of Soviet-to-American (or American-to-Soviet)
costs of activity based on ruble as well as dollar valuation of both sides.
This, of course, meets the problem of misuse no. 3. But it also accom-
plishes two other purposes: it makes it much more difficult to talk about'
Soviet "spending" in a comparative framework, and the user who wishes to
interpret comparative value estimates as indicators of capability (misuse
no. 1) will be forced to confront the difference in magnitude between
ruble-based and dollar-based ratios.
4. Where possible, avoid using rates of growth in the analysis of
dollar costs of Soviet activity, presented without the counterpart ruble
expenditure. Wherever possible and sensible, cite the rates of growth of
both dollar costs and ruble expenditure and explain the reasons for the
differences.
5. Present alternative or variant calculations along with the
basic series. Such variants not only enrich the analysis but also help
prevent misinterpretation. For example, analysis of the dollar costs of
Soviet personnel at present-day prices might be complemented with a variant
using U.S. pay and allowance rates in a year when the draft was still in
effect, and compare that series with (a) one for the United States using
the same weights and (b) ruble-weighted valuations of U.S. and Soviet
costs. Such a set of calculations would not only deal with the often-heard
charge that CIA exaggerates Soviet manpower costs by using high U.S. pay
rates, but it would also call attention to the resource cost basis of the
calculations; one cannot easily discuss the relative "capability" of two
forces when its measure has at least four different answers.