POSTMORTEMS

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP98S00099R000400800007-1
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 6
• Document Creation Date: December 27, 2016
• Document Release Date: May 23, 2012
• Sequence Number: 7
• Publication Date: August 16, 1985
• Content Type: MEMO

Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
THE DIRECTOR OF CENTRAL INTELLIGENCE
NIC-04141-85
16 August 1985
MEMORANDUM FOR: Deputy Director of Central Intelligence
FROM: Senior Review Panel
SUBJECT: Postmortems
1. Your memorandum of 8 August highlights the impact of
abrupt policy shifts on the estimative process. The point is
well taken, and deserves careful consideration.
2. Our direct response is that we entirely agree that
postmortems should be sensitive to the intersections of policy
and intelligence and that in the future the Panel will seek more
intensively to integrate the chronology of policy formulation and
execution into its rear-view intelligence assessments. This
reply, while responsive to the issue, does not in our view do
sufficient justice to the complexity of the problem and the
likelihood of its recurrence.
3. At the outset, it seems to us important to recognize
that the system is working generally quite well. None of our
surveys has uncovered a significant intelligence failure in the
1980s--a record which contrasts significantly with that of the
1970s. We think this owes chiefly to two circumstances:
a. An improved dialogue on policymaker needs and
interests between the NIOs (and their DDO chief
counterparts), on the one hand, and State Department
colleagues, usually at the Assistant Secretary or DAS level
on the other. Such exchanges, when forthcoming and
straightforward, provide our people with a good understanding
Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
 Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
not only of the policy, but also of the process by which it
has been evolved. Similarly, regular meetings between Agency
personnel and other members of the policymaking apparatus at
the NSC, DoD, and other senior levels have broadened our
understanding of policy, its formulation, and the centers and
weight of opposing views. Three critical follow-on
assumptions are that the NIOs: (a) are sensitive to signs of
policy obsolescence; (b) do not become overly dependent on
single institutional arrangements or policy advocates; and
(c) share their knowledge on a timely basis with the drafters
of Estimates.
b. The inclusion of the DCI in the Cabinet and among
the inner circle of the President's advisers. Even though he
may be constrained from sharing the full extent of his
knowledge, the DCI's intimate familiarity with the process,
the methods, the personalities and the conclusions allows him
to guide the intelligence process with a sure hand, requiring
that pertinent Estimates examine all reasonable courses of
action and the likely responses to them. And perceptive
NIOs, in turn, guided by shared knowledge of the process of
policy formulation and execution have been far better
equipped to manage estimative production, develop relevant
intelligence advice, explore meaningful alternatives, and
assess their likelihood and their impact on US interests.
4. But policy under the best of circumstances is an
amorphous commodity, seldom clearly articulated in concise terms,
often the product of strongly-held opposing views, and frequently
subject to interpretation by its executors. Worse, in times of
urgency, policy inevitably tends to become the province of the
very few, reducing the paper trail to a bare minimum. Secrecy in
the policy process begets confusion, imprecision and false
starts. Few people know or understand what is intended and why,
and they may be sworn to secrecy. All of which recognizes that
there can be no absolute guarantee that abrupt and spur-of-the-
moment policy shifts in fast-breaking situations may not on
occasions confound the unsuspecting analyst.
5. Such cases may be the exceptions, but they are also
likely to be the ones which involve major US interests, attract
at some point significant Congressional and public attention, and
become, however unfairly, benchmarks for miscellaneous appraisals
of Intelligence Community performance. One can think of a half
dozen candidates at the present time.
6. We have no panaceas for such circumstances--other than a
continued patient effort to increase the sensitivity of all
2
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
 Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
concerned to the problem of dramatic policy shifts and the
special requirements they levy for our estimative work before the
fact and our postmortems afterwards.
7. For the analyst, there must be recognition that there
will be fast-moving, g -interest times in which his
involvement--or his information--will be sporadic, spotty, or
nonexistent on some of the key elements. He has an obvious need
to maintain close touch with, and seek guidance from, the NIOs
concerned and with the leadership of the Agency. He has an
initial requirement to produce Concept Papers and TORs which pose
the real intelligence issues, describe an intended approach, and
define a range of alternatives. The requirement is not met by a
routine listing of a standardized table of contents outline. He
has additionally two benchmarks for the possibility of policy
shifts of an abrupt and dramatic nature:
a. When US influence and interests in a given situation
are major, the locus of crisis management is clearly in the
White House, and special Presidential agents replace more
conventional channels for dealing with overseas situations.
b. Closely related, but perhaps more as an information
inhibitor: When the Secretary of State himself undertakes a
direct negotiating or mediating role (Schultz to the Middle
East, Haig to a Falklands shuttle, etc.)
No two crises are identical, but either of the above is a
reasonably clear signal, and one likely to have been preceded by
a considerable run-up period in which there will have been a
number of indications that, in your words, "the policy winds are
blowing in different directions." It is time for the analyst to
shift his gears, recognize the hazards of a status quo bias,
overconfidence in the continuity of previous trends, or frozen
analytical frameworks. The sooner this occurs the better, for
when the force of the gale strikes, current information will
almost certainly be sharply circumscribed or cut. Persistence in
estimative judgments based mainly on past experience becomes in
such circumstances quite hazardous.
8. The best remedies now at hand would appear to be
avoidance of single-outcome projections, a greater willingness to
try alternative scenarios, and more recourse to the sort of
speculative "What if?" papers we have found useful and thought-
provoking whenever they have been attempted. (Parenthetically,
and not only for urgent cases, we hope the DDI and the VC/NIC
will encourage greater use of such speculative pieces.) The aim
is not to second guess the policymaker but to confront him in
3
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
 Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
time with intelligence advice before he acts--whether he thinks
he needs it or not.
9. For the post-mortician, we think he should continue to
see his task as the production of essentially in-house products,
directed at "lessons to be learned." We believe there are a few
specific additions to his process which should go a considerable
way towards meeting your concerns:
a. A new effort to reconstruct the policy process with
precise chronological benchmarks. To do so will require
interviewing available senior levels of the policymaking
process, as well as the NIOs and drafters of the Estimates
involved, in order to answer, as precisely as possible, one
vital question: "Who knew what and when?" Without
unambiguous answers, the attempt to correlate the estimate
process with policy formulation and execution becomes
exceedingly problematic.
b. A clearer record of the extent to which antecedents
to policy shifts have been made known in advance,
subsequently, or not at all, to the analysts responsible for
the preparation of intelligence advice.
c. An increased emphasis on what drafters have done to
cope with, or compensate for, uncertainty or exclusion from
advance knowledge of significant policy shifts. This would
mean a closer examination of the presence or absence of
analyses of critical variables, the use of alternative
scenarios, and the extent of attempts to develop "What if"
implications for US policy.
10. The foregoing observations make no pretense of being
able to provide all the answers, but we believe that at the
least, they will add an important and revealing dimension to
future postmortems. In these comments, we have to a considerable
extent worked the problem backwards--a sort of reverse
engineering on postmortems. What stands out at the end most
clearly is that the Community--at all levels--should position
itself by close and amiably relentless contact to persuade
policymakers that seeking the Community's advice before decisions
are taken is a more rewarding practice than reading its
retroactive appraisals. This means, among other things,
validating the quality and the dependability of Agency and
Community assessments, demonstrating a capacity to protect and
limit access to information, and manifesting a quick-response
capability. (Such considerations account for a number of our
crotchets: short, readable papers, accelerated preparation, and
4
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
 Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
a campaign to restore the SNIE art form to its original
purpose: rapid response to specific critical issues.) However
done, there needs to be developed a more telling conviction that
a policymaking process which excludes its intelligence arm at
crucial points is a high-risk, disaster-prone business.
11 We would welcome your reaction to these thoughts.
2rX1
25X1
c
c:
DCI
DDI
5
SECRET
Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
 Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1
SECRET
Executive Regi,~try
8 August 1985
MEMORANDUM FOR: Ambassador William Leonhart, Senior Review Panel
FROM: Deputy Director of Central Intelligence
SUBJECT: Postmortems
1. In looking at some of the postmortems to see how intelligence
alerted policymakers to unfolding events, I read with interest your
observations on Iran and the fall of the Shah as well as Nicaragua
centering around the Sandinistas taking over Somoza. Each subject was
treated from an intelligence standpoint alone. In the case of the Shah,
we did note in some publications the unraveling nature of events
happening in Iran. The assessment was that the Shah could hang in there,
this undoubtedly prompted by the history of 1953 events. In Nicaragua,
we assessed Somoza as having a 50/50 chance of hanging in there and one
could judge that our intelligence missed the boat.
2. What is missing in the postmortems, however, is the role that US
policy shifts played in altering the course, of events dramatically far
beyond any prognosis for intelligence to predict.
3. In Iran, the judgment was when things really got tough, the Shah
would call out the military much as he did in 1953 and squash Khomeini
and the dissident movement. In Iran, the analysts could not predict that
the US Government would end up sending General Huyser on a mission to
quell the military from any involvement. In Nicaragua, no one predicted
that the US policy would shift from supporting Somoza to the point where
we cut off all aid and assistance and took effort diplomatically to
undermine the Somoza rule.
4. I guess my bottom line is that I feel that postmortems must take
into account the shifts of policy which cannot be predicted by the
intelligence analyst and somehow point out that the failures were really
not intelligence but the results of the policy winds blowing in different
directions.
e-
5. I would like your comments on how we might tackle this illusive
factor in future postmortems.
STAT
John N-. McMahon
cc: DDI
STAT
Declassified in Part - Sanitized Copy Approved for Release 2012/05/23: CIA-RDP98S00099R000400800007-1