THE PFIAB CONCERN ABOUT STRATEGIC ESTIMATES: CONVERSATIONS WITH MR. ROBERT GALVIN

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP91M00696R000700010001-8
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 9
• Document Creation Date: December 15, 2016
• Document Release Date: March 1, 2004
• Sequence Number: 1
• Publication Date: June 30, 1976
• Content Type: MFR

SECRET/ 25X1
Approved For Release 2004/05/13 : - 696R00W00010001-8
SUBJECT: The PFIAB Concern About Strategic Estimates:
Conversations with Mr. Robert Galvin
1. Mr. Robert W. Galvin (Chairman of the Board of the
Motorola Corporation) is a member of the President's Foreign
Intelligence Advisory Board and the Chairman of its NIE
Evaluation Committee, which was established by the PFIAB's former
chairman (Admiral Anderson) during the dialogue between the
Board and Mr. Colby which took place during the summer and
fall of 1975. The Committee is a concrete manifestation of the
PFIAB's overall concern about the estimative process and, in
particular, strategic estimates. Its other two members are
Dr. Edward Teller and Dr. John S. Foster, Jr.
2. As part of its activities, the Committee produced,
for the PFIAB, a 1 April 1976 study entitled "A Review of the
National Intelligence Estimates on Soviet Forces for Intercontinental
Conflict (NIE 11-3/8 series) and of the Intelligence Estimating
Process." Mr. Galvin briefed its contents to the DCI, Mr. Stoertz,
and myself on 28 April and sent the DCI and me
copies of it on 29 April. To minimize the risk of a paper war, the
study was not forwarded to the President and we in the Intelligence
Community decided not to make formal written comments on it or its
recommendations.
3. On Wednesday, 12 May 1976, I flew to Chicago to talk to
Mr. Galvin in the latter's office about the study and its recommenda-
tions. Our session, which lasted for over two hours, was cordial
throughout, with both of us agreeing that the PFIAB and the con-
cerned members of the Intelligence Community had a common goal --
better national estimates, particularly in the strategic force
arena -- and that our chances of attaining that goal would be
enhanced by a maximum of informal personal dialogue accompanied by
a minimum exchange of formal written memoranda.
Approved For Release 2004/05/13 : CI4
25X1
SECRET/
 SECT/ 25X1
Approved For ease 2004/05/13 : 6lA-RE)P9 0690fip00700010001-8
paragraph 25 of the Committee's study) that no group
could be convened competent to address the complex
issues involved in strategic estimates.
-- While acknowledging that the Community's track
record in producing such estimates over the past decade
was spotty, I underlined the fact that in no instance
have the Soviets ever deployed a new weapons system
not flagged by the Intelligence Community well before
that system's IOC date, in other words the Community
had certainly kept its consumers free from this type
of surprise.
5. After discussing the Committee's actual study, I turned
to the general question of intelligence estimates, their function
and some of the problems inherent in their production, making the
following points (with which Mr. Galvin generally concurred):
-- It was not repeat not -- the function of
intelligence to support (or oppose) any particular set
of policy programs, departmental budgets, etc. Those
who produce intelligence have a responsibility to be
objective, and not to indulge any form of advocacy.
-- In the production of intelligence, particularly
estimates, there is one inherent problem: estimates
are inevitably produced for multiple audiences, with
different interests, backgrounds, perspectives and
concerns. It is fairly easy to produce an intelligence
product which satisfies none of these audiences, and
impossible to produce a single product which will satisfy
them all simultaneously. Almost by definition, a level of
detail and mode of treatment which one senior consumer finds
right for his needs, will be regarded as less than
satisfactory by many other consumers -- usually including
the staff of the consumer who is himself satisfied.
-- One of the difficulties inherent in producing
intelligence is the fact that -- particularly on important
questions -- the evidence (no matter how good) is never
complete. Consequently, intelligence assessments always
embody some measure of judgment and extrapolation from
evidence which is fragmentary or ambiguous. As a result,
analysis inevitably yields judgments over which honorable
-3-
Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8
SECP.ET/ 25X1
 SECRET
Approved For ease 2004/05/13 :
069gigi900700010001-8
men can.reasonably differ. In the strategic field,
there is a related problem of equal importance.
Judgments in this area are often driven or shaped by
highly technical considerations and highly technical
arguments, the relative merits of which are almost
impossible for non-technical laymen to assess.
-- These aspects of the intelligence process
heighten the importance of the interaction between
technically versed analysts and policy-level users or
consumers of intelligence. It is essential that these
consumers not be or become the captives of any particular
school of technicians and equally essential that the
final judgments developed by technically versed analysts
be critiqued by other analysts with equal technical
expertise, with the consumer being put on clear notice
whenever there is-a significant difference of equally
expert technical opinion.
6. Some of the above factors and considerations, at least
to me, raised problems with what I perceived as part of the
philosophical approach underlying some of the PFIAB's recommenda-
tions. I had great difficulty with the thesis that intelligence
(on strategic questions) should simply lay out all the. alternative
courses of action which the Soviets could pursue (i.e., all
hypotheses not positively refuted by available evidence) and then
make no effort to assess the relative probabilities of the
Soviets pursuing some of these courses of action in lieu of
others. This approach, to my mind, left the ultimate users of
intelligence -- especially the President and his most senior
advisors -- at the mercy of technical shamans. I felt it essential
not to suppress divergence of opinion but equally essential to
force the technicians to compete with each other and to give the
ultimate consumers some feel for the way those versed in the
technical details of intelligence problems assessed the relative
probabilities of alternative judgments.
7. With respect to the Board's concrete recommendations,
I made the following points:
-- The Board's principal recommendation --
the creation of a competitive analysis group -- is
handled and discussed separately. In sum, Mr. Galvin
and I worked out a procedure for developing an
arrangement that would be mutually satisfactory but
Approved For Release 2004/05/13: CIA-RDP91M00696R000700010001-8
25X1
SECRET/ 25X1
 SECRET /
Approved For RJase 2004/05/13 : CIA-KUVUI 6100700010001-8
which would not involve anyone outside the Board
and the DCI, i.e., we agreed that there would be no
necessity for NSDMs signed by the President directing
the DCI to do what the DCI was already doing.
-- On the other recommendations contained in
paragraph 40 of the Board's study, I made the
following observations:
"a. The subject of Soviet intentions,
objectives and tactics in the broadest sense
is deserving of more comprehensive treatment than
it now receives. in NIE 11-3/8. Perhaps a separate
NIE on this central topic should be commissioned."
COMMENT: I agreed and noted that such a
separate NIE was already in train.
"b. Consideration should be given to establishing
a small (no more than six), part-time group of 'elder
statesmen' who, under the DCI's aegis, would review
and comment on selected NIEs or on other crucial
intelligence products -- prior to publication and
after being given full access to all of the evidence
used by the analysts in formulating their appraisals."
COMMENT: Picking up from my earlier conversa-
tions with Mr. Galvin on this topic, I explained
again the concept of the Intelligence Advisory
Panel to the DCI, which Mr. Galvin said would more
than meet the Board's concerns in this area.
"c. A thorough study should be made to determine
whether the intelligence Community has an affirmative
obligation to declassify and provide information to
the public. As a related matter, whether the
intelligence community should be required, upon
the publication of each annual strategic force
estimate, to specify in the document which of the key
judgments it is willing to be held publicly accountable
for five years hence, should also be considered."
COMMENT: I agreed with the first sentence of this
recommendation but took strong exception to the second,
noting that I had little sympathy for the concept of
-5-
Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8
25X1
SEC ',T 25X1
 ECRET
Approved For Rase 200g /05/13 : AIA P01294 69600700010001-8
publicly pillorying officers for the exercise of
discretionary judgment some years after the
judgment had been exercised, particularly since
in many cases intelligence judgments could stimulate
policy actions which in turn would make those
judgments "incorrect." Where this occurred, however,
intelligence officers were not wrong but were doing
their job in the best way possible.
"d. The question of the time period to be
covered by the NIE 11-3/8 series should be reconsidered
in light of the consumer's desire for timely informa-
tion and in view of the limitations of the intelligence
community with regard to accurate, long-term predictions."
COMMENT: I agreed that the time frame covered by
strategic estimates should be continually reviewed with
the principal consumers and users of its estimates
but noted that, in the final analysis, the needs and
preferences of these consumers would have to be controlling.
"e. The Board should consider encouraging policy-
makers and decisionmakers to schedule oral intelligence
briefings on topics of interest as principal means of
receiving intelligence. The purpose would be to
develop a direct relationship with the knowledgeable
intelligence officer, and cultivate a better understanding
between the user and the producer."
COMMENT: I agreed that we needed to give continuous
attention to more innovative and flexible ways of
presenting intelligence and that certainly a lot more
could be done with oral briefings.
"f. The intelligence community should reassess
the function of the NIE, the variety of readership
that must be served, and the kinds of topics that are
most important to each. For example, in lieu of a
single NIE on Soviet offensive and defensive forces for
intercontinental conflict, it may be preferable to place
greater analytic emphasis on addressing narrower topics
in varying degrees of detail, depending upon the
principal audience of interest."
COMMENT: I agreed with the first sentence, but
reserved on the second -- noting that a series of
Approved For Release 2004/05/3: CIA-RDP91M00696R000700010001-8
25X1
SECRET/ 25X1
 SECRET
Approved For R*kase 2004/05/13 : CIA 0700010001-8
excessively narrow and detailed estimates would meet
consumer resistance since consumers were always
pressing for comprehensive studies which synthesized
the Community's views on broad areas of consumer
interest.
"g.. Awareness of the efforts of this Committee
served as a stimulus for a number of activities by
the intelligence community with regard to observations
in the Board's letter of 8 August 1975. The full
Board should consider establishing an WE Evaluation
Committee' as a permanent body of the PFIAB and, to
aid in the maintenance of 'fresh ideas,' the membership
should be rotated periodically."
COMMENT: I strongly endorsed the approach reflected
in this recommendation, noting that continual interaction
and consultation between the Board and the Intelligence
Community would be beneficial to both and would help
both achieve their common goals.
-- As indicated above, I persuaded at least Mr. Galvin that
no NSDM was necessary and hence did not specifically address
the text of the draft NSDM appended to his Committee's
study. I did note, however, that that draft NSDM talked about
two things which were quite distinct: competitive analysis
in producing an intelligence estimate and a fresh approach
to net assessments. The former was an Intelligence Community
problem which we were in fact addressing. The latter was an
important issue, but not one which the Intelligence Community
could or properly should address.
Geo e A. Carver, Jr.
cc: Mr. Howard Stoertz
NIO for Strategic Programs
Approved For Release 2004/051.3 : CIA-RDP91M00696R000700010001-8
25X1
25X1
SECRET/I 25X1
 25X1 Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8
Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8
 SECRET
Approved For Rele,?a 2004/05/13: CIA- DP91 M00696R
P04W00010001-8
4. I opened the 12 May session with some general comments
on the NIE Evaluation Committee's study, paying particular
attention to the "Users' Survey" the Committee had conducted,
the results of which were summarized in the study's paragraphs
16-29. I agreed with Mr. Galvin that the fact that people whom
the Committee had consulted had certain opinions was in itself
a matter of intrinsic significance, quite apart from whether
or not one felt that those opinions were warranted or had any
solid basis in evidence. I did note, however, that the summary
of the survey presented in the Committee's study was a little
hard to read since the time period to which various criticisms
were keyed was not clear -- i.e., were they criticisms of all
strategic estimates, or were some of these criticisms directed
specifically at estimates of a certain period, such as the
late 50s, mid-60s, etc. -- and it was hard to evaluate some of
the criticisms without knowing who was making them, since various
interested parties had special concerns or perspectives which would
inevitably shape their reactions to the estimates in question.
While stressing that I did not want to engage in a sterile debate,
I did orally take exception to some of the more outrageous
assertions in the Committee's study. For example:
-- I flatly denied that estimates were ever
prepared by the Intelligence Community "to keep the
lid on defense spending by minimizing the threat"
or "to help rationalize an administration's foreign
and domestic policy."
-- I agreed that various people making inputs to
estimates might have their views shaped by departmental
constraints or considerations, but also flatly denied
that, so far as the DCI or the Intelligence Community
as a whole was concerned, there were any judgments
or topics which were "off limits." The analysts who
produced these estimates were fallible mortals, but
they did their best to call the shots as they thought
evidence dictated and any attempt to constrain their
judgments would have produced open revolt.
-- I agreed that, perhaps inevitably, a disproportionate
share of the total intelligence workload fell on the
shoulders of a relatively small group of able analysts,
but rejected as self-defeating a notion (reflected in
-2-
Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8
25X1
BECRET/I 25X1
 25X1 Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8
Next 21 Page(s) In Document Exempt
Approved For Release 2004/05/13 : CIA-RDP91M00696R000700010001-8