LETTER TO WILLIAM J. CASEY FROM LIONEL H. OLMER

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP85M00364R001903730002-0
• Release Decision: RIPPUB
• Original Classification: C
• Document Page Count: 3
• Document Creation Date: December 21, 2016
• Document Release Date: May 7, 2008
• Sequence Number: 2
• Publication Date: July 16, 1983
• Content Type: LETTER

Approved For Release 2008/05/21: CIA-RDP85M00364R001903730002-0
EXECUTIVE SECRETARIAT
Routing Slip
ACTION INFO DATE
DCI X -L, IN11YAL
DDC1
D/ICS
Compt
SUSPENSE 1,
oat*
f,, ire,;)
Please prepare ack-nowl edgrient
for DCI's signature. !'dote wo
separate letters attache'--co;
-h i ch together with encl os::r ; ~.1er
sent directly to ,von,.)
Executive Secretary
:?d Elul y l r"'^
Oat.
Y0 encl e
Approved For Release 2008/05/21: CIA-RDP85M00364R001903730002-0
 Approved For Release 2008/05/21: CIA-RDP85M00364RO01903730002-0
Q~v \y '~pym
'9'rATES OF P~~P
CONFIDE
Honorable William J. Casey
Director
Central Intelligence Agency
Washington, D.C. 20505
UNITED STATES DEPARTMENT OF COMMERCE
The Under Secretary for International Trade
Washington, D.C. 20230
1TIAL
At the same time as the 14-member panel of Soviet experts from the private
sector was assessing CIA's work on the Soviet economy and producing its
report which I provided to you separately, I asked that a small component
of my staff undertake a separate analysis on the methodology used by CIA
to calculate Soviet GNP growth. My rationale was that consumers of
intelligence tend to ascribe greater significance to quantitative
measurements, compared to other forms of intelligence reporting, than is
warranted. As the panel observed, even though GNP indices provide only
limited insights particularly for closed societies like the USSR, policy-
makers focus on them perhaps excessively. Thus, continuing efforts must
be made to examine the validity of and improvement upon the methods which
are used in the intelligence process. (I know of your effort to do some-
thing similar regarding the methodology used to estimate Soviet defense
spending.)
The experiment we selected was to calculate the growth of the U.S. economy
and of the West Germany economy by applying the methodology CIA uses on the
Soviet economy, and then to compare the results with the official estimates
published by the U.S. and FRG governments. The U.S. and FRG economies were
chosen because both make available the most elaborate statistics. Production
data was assembled for a list of products as close as possible to those used
by the CIA for the USSR; the methods of estimation were also duplicated as
completely as possible.
The results of this experiment show a substantial downward bias in estimating
growth rates. These "shortfalls" are remarkably similar to the differences
between CIA's estimates for the USSR and the Soviet Union's official
estimates of its own economy. Thus, inferentially, this would indicate a
possible underestimation in the growth of Soviet industry at large by
2'2-27%; in the growth of Soviet machinery industries by 26-37%; and in the
growth of Soviet GNP by about 30%,
I hasten to point out that the panel of experts concluded that this experiment --
while interesting -- is not persuasive; and that the parallels drawn between
the U.S., FRG and Soviet economies are simplistic and misleading,
Classified by; Lionel H. Olmer
Declassify on: 7/16/89
CONFIDENTIAL
Approved For Release 2008/05/21: CIA-RDP85M00364RO01903730002-0
 Approved For Release 2008/05/21: CIA-RDP85M00364R001903730002-0
WUNFIUENIIAL
Especially because the panel included) who is recognized
throughout the world for the excellence of his work in the field of national
account measurements, and because he believes that the CIA's methodology
produces results within a tolerance that would not fundamentally alter our
assessment of the Soviet economy, the experiment may be of marginal utility.
Perhaps. The fact that estimating national accounts is an enormously
complicated process even for specialists with long experience, might be
sufficient reason, given the judgment of the panel of experts, for laying
the experiment aside and not troubling you with it. However, because the
issue is so important and because certain of the most striking results are
not readily explainable as mere probability, it does I think warrant your
attention and possibly additional analysis. If we can help in this regard,
I'd be prepared to make staff available.
Copies of this report have been provided to analysts within CIA's Office
of Soviet Affairs.
Sincerely,
Lionel H. Olmer
Enclosure
Classified by; Lionel H. Olmer
Declassify on: 7/18/89
CONFIDENTIAL
Approved For Release 2008/05/21: CIA-RDP85M00364R001903730002-0