STRATEGIC INTELLIGENCE: PROBLEMS AND REMEDIES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP86B00985R000200130016-0
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 36
• Document Creation Date: December 16, 2016
• Document Release Date: December 17, 2004
• Sequence Number: 16
• Content Type: PAPER

Approved For Release 2Q05101/16 CIA-RDP8.6B00985R000200130016-0
Strategic Intelligence:
Problems and Remedies
intelligence is one of the most critical elements of statecraft directed
toward the outside world; it permeates the making of all foreign: policy that
is not regarded as.routine and unproblematic; and although it is also
relevant and important to estimating international opportunities for coopera-
tion, intelligence has been cultivated most attentively in matters of threat
perception because gross misestimates'in this area risk the very survival
of states. While intelligence is concerned` with other than military threats
(e.g. economic), the plan of this volume demands concentration on military
threats. Both overestimates and underestimates of external threats can
produce calamitous consequences. Underestimates can produce ''disaster and,
even if they do not, can make war more probable and costly. Overestimates
can lead to excessive military build-up that is economically wasteful and,
by causing anxiety, additional military preparations and animosity abroad,
can make international conflict and the outbreak of hostilities more probable.
In the following, we will (1) conceptualize the objectives of inter-
national threat perception; (2) indicate the historical record of intelligence
activities; (3) present a particular case of strategic surprise in order to
introduce an analysis of the fundamental problems besetting acts of threat
perception; and (4) discuss possible improvements of statecraft in this
area of concern.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 CIA-RDP86B00985R000200130016-0
Strategic intelligence attempts to estimate acute and potential threats.-~
Acute or actual threats are those that are believed to be probably imminent.
Potent3.al threats are attributed to states that have or are developing the
capability to proceed to actual threats and attack, and whose peaceful inten-
tions in the future cannot be taken for granted. Although potential threats
are regarded as more hypothetical than actual threats, they are not necessarily
a matter of the longer run and therefore safely of lesser concern. As many
historical examples remind"us, a potential threat can turn acute suddenly
and unexpectedly. If Weimar Germany was at worst a potential threat to-its
neighbors in 1930, Hitler's accession to power and rapid military expansion
under his rule made this threat acute in a very few years. The split between
the Soviet Union and Communist China occurred very quickly and abruptly
raised serious security problems in Peking.
The objective of intelligence is not only to estimate whether or not a
threat exists, but also to assess its precise nature, e.g. the likely places
and modes. as well as the timing of attack.? S i n c e many attacks in the
past succeeded quickly because the victor achieved strategic surprise, the
prevention of surprise by means of intelligence estimates that give timely
and proper warning is a central objective of threat perception. The correct
estimate of an actual threat facilitates the design of proper responses,
whether by alerting and mobilizing. forces of defense and attempting to
Essentially the same analysis applies to states intent on the aggressive
use of force.
2/Avi Shlaim, "Failures in National Intelligence Estimates: The Case of the
Yom Kippur War," World Politics XX3JIII, 1976, p. 348.
Approved For Release 2005/01/10 : CIA-REiP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
reinforce deterrence, or by encouraging a course of accommodation. The correct
estimate of potential threats is a basis for providing adequate forces for deter-
rence and defense in'the hope that the potential threat will not become acute while,
on the other hand, avoiding provocative responses that are apt--by way of self-
fulfilling prophecy--to increase the threat. In either case, threat perception
is oriented toward the capabilities and intentions of actual or potential opponents,
0
Both can change over time, and this makes conjecture about future developments a
part of the estimative process. Indeed, because it takes a great deal of time to
bolster forces for deterrence and defense by developing and producing new weapons
systems, .conjectures about the capabilities of actual or -potential opponents,
including arms, military budgets and even national,-economic output, range far into
the future. Defense budgets depend on such estimates.
Even good intelligence. estimates--clear, timely. and valid'"-cannot, of course,.
insure good policy; and when estimates are not good in any or all of these terms,
it is sometimes difficult to distinguish between failures of threat perception
and of policy responserr But without proper intelligence estimates, states--unless
they are lucky--can at best hope to improvise and muddle through in the face of
surprise, which is risky and costly, and which only basically very strong and re-
silient countries can afford. Reasonably accurate threat perception is clearly a
precondition of any effective posture for survival in the larger sense and the
longer-run.
/For- example, the huge losses suffered by the Soviet army in 1941 resulted from
both the German achievement of strategic surprise and Stalin's insistence on a
linear defense of the Soviet boundary, permitting no strategic withdrawal, even
though the Germans had previously executed dashing Blitz Krieg in Poland and France.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
It is precisely because international threat perception plays so critical a
role in international affairs that it would be interesting to assess its quality
as a matter of statecraft.' Unfortunately, this is difficult to do. Relatively
few recent estimates, for the most part American, are in the public domain; and
the historical record of threat perception when estimates were made by leaders
4
and governments without elaborate intelligence services has not so far been examined
systematically..
Moreover, any useful appraisal of the ex-post facto success of past estimates
depends on one's criteria for judgment and a suitable classification of cases.'.
In the area of strategic intelligence, these problems can be illumined by com-,
paring estimates of the capabilities of states with estimates of their intentions
in matters touching on peace and war. Capability estimates are usually about
-continuous things raising questions of more or less (e.g. GNP, military budget,
numbers of divisions), and can be more or less right or sarong. On the other hand,
estimates regarding critical intentions of governments concern also questions
of either/or and then are either right or wrong (although the attribution of
degrees of probability mitigates this dichotomous character). Moreover, capa-
bilities that move on a continuum are more conjecturable (that is, more predict-
able with some degree of confidence) than are estimates of critical intentions
and decisions, because many components of capability are observable and countable
and not only do not usually change much from year to year, but are incapable of
abrupt and substantial change,4/ whereas-the intentions of governments and their
expression in observable behavior can change very abruptly and with great consequence.
4/Certain capability changes that can be sudden and substantial as a result of
mobilization and new deployment express sudden change in intentions.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
Of course-, capability estimates that overestimate or underestimate foreign
capabilities considerably and persistently can have great cumulative consequence.
The fact that critical government intentions can change abruptly does not mean
that they change with great frequency. The intelligence service of country A
may estimate year after year that a military attack on its ally X by state Y is
highly improbable, prove correct in nine years and incorrect in the tenth. But
the consequence of the one failure may be enormous.
Crude success-failure ratios tell us little about the quality of intelligence
estimates. While the obstacles to a statistical evaluation of strategic intelli-
gence seem to me insuperable,5' there is an alternative approach to assessing the
quality of statecraft in this area. This is to start with very consequential
past events and to examine the record of threat perception preceding them case
by case. Such study suggests that the quality of intelligence has been extremely
and surprisingly poor.'/ Although this is not known as a general phenomenon,
misperceptions have apparently played a significant, and often crucial, role: in
the precipitation of wars and during war as well, regardless of period of time or
part of the world one turns to. However, this approach naturally tells us a
great deal only about underestimates of external threats. It is more difficult
to use historical search in the pursuit of overestimates. Nothing dramatic is
5'For an attempt that examined 289 international crises involving the United
States from 1946 to 1975 and attributes an element of surprise to about half of
them, see Leo Hazlewood, John H. Hayes and James R. Brownell, Jr., "Planning
for Problems in Crisis Management," International Studies uarterly, vol. .XT,
March 1977, op. 75-106.
"For an extensive discussion of historical cases, see Klaus Knorr, "Threat Per-
ception," in Klaus Knorr (ed.), Historical Dimensions of National Security Problems,
Lawrence, Ka., University Press of Kansas, 1976, pp. 78-119.
Approved For Release 2005/01/10 : CIA-RbP86B00985R000200130016-0
 Approved For Release 2005/01/10 CIA-RDP86B00985R000200130016-0
apt to happen if a state overestimates a threat and increases its military forces
as a result. For example, there is some evidence-that the Soviet leaders over-
estimated threats from the United States and a reconstituted Germany during the
late 1940's and early 1950's, but the consequences of this misperception, if it
was real, are hard to trace among all the factors that actuated Soviet behavior.
Even if, following excessive military preparations in reaction to an overestimate,
a notable event, perhaps even war, does ensue, it is historically difficult to
attribute it to the failure of the strategic estimate. The subsequent event may
then seem to confirm the. original estimate even if, by inducing excessive military
responses, it contributed to its genesis. The historical approach is even less
productive when it comes to identifying cases of correct strategic estimates.
If such perceptions lead town improved posture of deterrence and no actual threat
materialized, we do not usually know whether or to what extent this absence of
crisis or war was determined by the adequate response to.a good estimate or by
other factors.. It seems to this author nevertheless reasonable to deduce from
historical experience that the record of international threat perception in the
vital'strategic area is disconcertingly poor.
In the following section, we will refer to some cases which contribute-to
the strong impression that the record of international threat perception is far
from good, and. which will serve us in analyzing the-difficulties encountered in
the estimating process. Some of these difficulties are inherent in the nature
of the job while others of an aggravating character result from situational condi-
tions that may or may not be present in particular instances. The analysis will
concentrate first on estimates of intentions and then, more briefly, on estimates
of capabilities.
I
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For lease 200.5101/10.: CIA-RDP86B00985R000200130016-0
Prior to the.Y'om Kippur War of 1973, Israeli intelligence--which-enjoyed
an enviable reputation on account of past efficiency-had-given a guarantee .to
the Government that there would be'ample advance warning of any Arab attack, and
the Government had accepted this guarantee. 2/ On this-basis, Israel maintained
only thin forces along the Suez Canal and-on the Golan Heigats, planning to
mobilize its highly trained reserves only in the event an attack seemed imminent.
Yet, on Oct. 6, the Egyptian and Syrian forces struck and achieved strategic sur-
prise, invading territory controlled by Israel. Once the Israel Defense Forces
were mobilized, they counter-attacked vigorously. But while they succeeded in
throwing the enemy back, they suffered very sizeable casualties in the process.
American intelligence, which had kept a close watch on the situation, also had
failed to predict an Arab attack.
Although this intelligence failure deeply shocked Israel, it must be empha-
sized that this kind of error?is.common in the annals of intelligence. Some -
recent examples may be cited to support this important point. Strategic surprise
was suffered by the Soviet Union when Germany attacked in 1941, by.the United
States in the same year when the Japanese, attacked Pearl Harbor, by this country
in 1950 when North Korea invaded South Korea and subsequently when China inter-
vened in the Korean War, by India in 1962 when Chinese forces crossed the bound-
ary between the two states, and by Syria and Egypt when the Israelis attacked in
.1967. British leaders had underrated Hitler's threat before World War IT.. Hitler
7P
-- For a detailed analysis of this case of surprise see Shlaim, op. 380. Other similarly instructive and well-researched case studies are?Roberta48
Wohlstetter's book on Pearl Harbor and Barton Whaley's on the German attack on
the Soviet Union (see the Appendix for the bibliographic references).
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 elease 2005/01/10: CIA-RDP86B00985R000200130016-0
App roved For
himself was surprised by the British execution of 1941 of their commitment to go
to war in the event Germany attacked Poland. Following World War II, the United
States was surprised by the outbreak of several severe crises precipitated by the
Soviet Union (over Berlin in 1948, 1958-59, 1961, and over the Cuban missiles in
1962). It was also surprised by the landing of Turkish troops on Cyprus in
1974, and by the Cuban intervention in the Angolan civil war. in 1975. While
we know little about recent Soviet intelligence errors, it is plausible that
the Soviet government was surprised by the strong American reaction to the
emplacement of nuclear missiles on Cuba in 1972. Gross underestimates of the
opponent's strength have also occurred often. Thus, the Soviet Union was sir
prised by the staunch Finnish resistance in the Winter War of 1939-40, the
United States erred repeatedly in underestimating t'he strength of North Vietnam
when it had intervened in support of the Saigon regime, and the Indians grossly
underrated Chinese military strength in the Himalayas before 1962. Although
European governments were alert to the possibility of war in 1914, they had been
advised by experts that, as a result of advanced international economic inter
dependence, a war between industrial nations could not be sustained for more
than a few weeks or months, and therefore acted with extremely unrealistic expecta-
tions about the requirements and destructiveness of war.
The surprise experienced by Israel in 1973 was evidently not exceptional.
How can we account for it and many other cases? Which factors make threat
perception apparently so difficult a task? Unfortunately, the available empir-
ical and analytical literature is quite small, and much of it of very recent
origin, published only after 1973.-21 There is even-now only the beginning of
8/See the Appendix for a selective listing of the literature.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
a useful theory of intelligence that can enlighten us about the inherent impedi-
ments to the making of good estimates. Fortunately, some of the few studies that
have been made are very helpful and permit us to identify the essential problems.
If an estimate is made at all, it is, in principle, either lack of informa-
tion or its misinterpretation that accounts for faulty threat perception. Intelli-
gence officers often do attribute failure to lack of information. A lag
between foreign events and the supply of relevant information can be a
serious handicap even in this age of rapid communications. It is also
trivially true that foreign governments do not supply all the information that
intelligence officers like to have. Indeed, they often try to maintain secrecy
and to issue disinformation. But, then, threat perception is a matter of "estimates"
that would not be needed if all the pertinent information were unambiguously on
hand. One estimates when one does not know. For this purpose one needs only
enough'.information that, if correctly interpreted, permits a good estimate to be
made.
Misperception and surprise do not usually result from lack of relevant informa-
tion. It was all there to be used in the Israeli case and, in fact, a subordinate
Israeli intelligence officer put it together correctly and predicted an impend-
ing Arab attack on Oct. I and again on Oct. 3. But his superiors rejected his
estimate and, *as late as Oct. 5, held that the chance of an Arab attack was of
"low probability" or"even lower than low."
.In every other case I have studied, it is easy to see in retrospect that the
relevant information for making a correct estimate was available. But as Roberta
Wohlstetterg-' put it, these "facts"--which she calls "signals"--are embedded in a
great deal of "noise," that is, irrelevant information. Moreover, as Barton Whaley.
9
--Roberta Wohlstetter, Pearl Harbor: Warning and Decision (Stanford, Stanford
University PreWr;v )For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 App ved Forlease 2005/01/10 : CIA-RDP86B00985R000200130016-0
has emphasized,- there often is also "disinformation," that is, misleading infor-
mation introduced by the opponent for purposes of deception. The Egyptians
resorted to this in 1973, for example by spreading rumors about the unreddiness
of their forces. Hitler used deception in 1941 in order to confuse the Soviets.
The problem of separating the correct information from the rest--which is so easy
in retrospect, with the benefit of hindsight--is inherent in the fundamental ambi-
guity of the total information that is being received. "Facts" do not speak for
themselves. All that can be derived from ambiguous information are inferences,
and no one inference can ever be compelling because any ambiguous behavior can
be explained by more than one motivational pattern. The Israeli and American
intelligence services were, of course, aware in 1973. of the massive deployment
of Egyptian and Syrian troops. But both decided that the deployments indicated
no more than elaborate military. maneuvers. That inference did fit the "facts."
The Arab countries had staged such maneuvers before and touched off an Israeli
reaction that, in retrospect, was deemed unnessary because no attack ensued.
As Hitler amaaaed German troops on the Soviet boundary in 1.941, five different
hypotheses were entertained by knowledgeable European officials, and Stalin's
conclusion that the Germans would not attack fitted the information as well as
any other inference.!-'/ After all, a military threat can be a bluff and military
'movements that look menacing can be preparatory to making demands rather than
starting hostilities.
Providing correct estimates in such situations is complicated further by
the possibility that the "opponent" is undecided on further steps while making
'Barton Whaley, Codeword Barbarossa (Cambridge, Mass., MIT Press, 1973), p. 244
I~VIbid., p. 223 -
Approved For Release 2005/0.1/10 : CIA-RDP86B00985R000200130016-0
 Approved For4 elease 2005/01/10 : CIA-RDP86B00985R000200130016-0
threatening moves. He may also be acting on the basis of multiple options and
proceed to attack only on certain contingencies, such as that the other side
fails to take proper counter-measures. Thus, when the Japanese fleet was steam-
ing toward Hawaii in 1941., its admiral was under contingency orders to attack only
if the U.S. fleet was at anchor in Pearl Harbor and if he could achieve surprise. 121
The fact that the opponent's intentions are unsettled or contingent must be taken
into account by intelligence officers but does not preclude an estimate that
issues appropriate warning. The intentions of foreign actors are necessarily
estimated in terms of probabilities. Yet the ambiguity of observed behavior
constitutes a profound problem inherent in threat perception.
In drawing inferences from available information that is usually fragmentary
and ambiguous, intelligence officers use certain assumptions about the behavioral
pattern of the potential enemy. These assumptions or preconceptions guide them
in distinguishing between signals and. noise and in arriving at a conclusion.
Intelligence bureaucracies formulate these assumptions, usually with care, on the
basis of his past behavior; and individual leaders are similarly guided by pre-
conceptions. or images-about what the potential opponent "is like." 31 Stalin's
error in 1941 resulted from the observation that, in every past case, Hitler's
.Z1Thomas G. Belden, "Indications, Warning, and Crisis Operations," International
Studies Quarterly, vol. XXI, March 1977, pp. 185-186.
131A classic example of the sophisticated development of underlying assumptions--
indeed two alternative hypotheses regarding the German threat to the United
Kingdom are carefully delineated--is Sir Eyre Crowe's "Memorandum on the Present
State of British Relations with France and Germany" of Jan. 1, 1907. This essay
in threat perception is very much worth rereading both for its analytical peg--
spicacity and the presence of some preconceptions that come easily to the repre-
sentative of a premier power and are questionable in retrospect. There is also
the larger but surely unanswerable question of whether this influential docu-
ment, by affecting British policy, contributed to the outbreak of World War I.
The main parts of the memorandum have been reprinted in Kenneth Bourne, The
Foreign Policy of Victorian England, 1830-1902 (Oxford, Clarendon Press, 1970),
pp. 481-493.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005101/10 : CIA-RDP86B00985R000200130016-0
aggression was preceded by a period of intense crisis during which he made inor-
dinate demands. This had not happened in 1941 and it was therefore assumed that
no real threat of war existed. Stalin also assumed that the warnings about an
imminent German attack he received from the governments of-Britain and the United
States were faked because he believed that these governments were trying to bring
about war between Germany and the Soviet Union. The chief assumption of the
Israelis in 1973 was that Arab leaders, because they knew that their forces
were inferior to those of Israel, would not go to war, unless they expected
Israel to launch a strike; and since Israel was maintaining only very small.
forces on their perimeter, the Arab governments were given no cause for fearing
a surprise attack by Israel. A subsidiary Israeli assumption, based on past
experience, was that the Arab countries were incapable of planning a joint
attack without these plans leaking to Israel intelligence. In 1941, American
leaders did not believe that the Japanese would commence war against the United
States because they assumed that their Japanese counterparts must have known
that this country had a militarry potential vastly superior to that of Japan.
On the other hand, Japanese leaders decided to attack the United States because
they assumed that a war in the Pacific with the United States was sooner or later
inevitable and preferred to fight at a time and under circumstances of their own
choosing. In 1962, the CIA disbelieved, despite much evidenceto the contrary,
that the Soviets were installing nuclear missiles in Cuba because past Soviet
military behavior had been doctrinally disinclined toward embarking on adventurist
courses of action and because it was assumed that Soviet leaders would perceive
any missile deployment threatening targets in the United States as an adventurist
move. United States intelligence was surprised by the Turkish military inter-
vention in the civil strife on Cyprus because the Turks had threatened to do so,
and assembled forces in nearby ports, on the occasion of previous crises over
Approved For Release 2005/01/10 : CIA=RDP86B00985R000200130016-0
 Approved For lease 2005/01/10 : CIA-RDP86B00985R000200130016-0,
Cyprus, but had never executed the threat. It was therefore assumed that the
Turks were bluffing once again.
It is inevitable that conceptions or assumptions structure international
threat perception because all human perception can approach reality only selec-
tively. They ;are indispensable to defining situations. It must also be under-
stood that the assumptions that guide expectations are usually not stupid. They
are often carefully reasoned. What is clearly wrong, however, in view of the
many times that they have led perception astray, is the degree of trust they
commonly. command. The simple fact is that human actors can produce unexpected
behavior for any number of reasons. Hitler wanted to achieve strategic-surprise
by means of deception and hence broke the behavioral pattern he had preferred
in the past. The Japanese leaders thought that they had no choice because they
.had assumed that war with the United States was inevitable and that they had a
better chance to win. now than in the future. The Arab leaders resorted to war
mainly in order to set the diplomatic world in motion, and especially to induce
the great powers to exert themselves, toward breaking the impasse vis--vis
Israel. For that purpose they did not need to win the war they started; a reason-
ably good military showing would suffice. In Turkey, competitive domestic politics
it .1974 made.ii difficult for the government not to act with great determination in
solving the Cyprus problem. In short, actors can and not seldom chooose to do
the unexpected because their objectives have changed as a result of new domestic
pressure, changes in relative capabilities, the individual influence of person-
alities, failures of intelligence on their part, deviations from rationality, and
other variable conditions including the possibility that, in the life of govern-
ments as elsewhere, things happen by accident as well as design. It is even
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
riskier to base prediction not on the careful study of a particular actor's past
behavior but on general "lessons of history" that tie the behavior of kinds of
actors to kinds of stimuli, the lesson, for example, that appeasement encourages
aggression. There are no easy and reliable lessons to be learned from history.14/
Although the future, within some framework of particulars, can be estimated,
it cannot be known. To estimate is to guess in order to reduce uncertainty
dictated by lack of knowledge. The assumptions and preconceptions about reality
that structure the guesswork can be more or less rigorously deduced from past
behavior. But--as the historical record discloses and for the reasons we mentioned
even the most sophisticated assumptions can lead threat perception astray.
depend on any one preconception or set of assumptions is to court surprise. This
risk is magnified by the tendency that the selection of an assumption about the
real world is an act of cognitive closure that easily leads the perceiver to be
close-minded and to ignore or explain away discrepant information. It must there-
fore be accepted that although good estimates can reduce uncertainty about the
future, even the best cannot be depended on to remove it.
The inherently difficult problems of international threat perception, we
have so far discussed are not the only obstacles to the formulation of realistic
1:i/The temptation is to conclude from some memorable past experience that: X
will follow if A or A and B happen, e.g. that aggression will be encouraged
by appeasement. But close inspection is apt to reveal that the earlier causal
pattern was complicated by the operation of other variable conditions and,
therefore, that the accepted lesson is a simplification likely to cause errone-
ous expectations when projected into the future. It is the very complexity of
unfolding events that, after all, accounts for the conflicting explanations
of historians and the endless rewriting of history. For an interesting study,
of the problem, see Ernest R. May,. "Lessons" of the Past, The Use and Misuse
of History in Ameridan Foreign Policy (New York, Oxford University Press, 1973).
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 t5
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
estimates. There is an entire further dimension of factors that tend to cause
mistakes. It is apparent from many historical instances that the perceiver,
far from being unbiased, often approaches his task under the influence, usually
unwittingly, of predispositions that affect his choice of assumptions and his
receptivity of incoming information, and thus are apt to distort his estimates.
Several kinds of intervening predispositions have been observed to operate.
First, emotions can condition the act of threat percepeion. For instance,
It is easy to overestimate potential threats from an actor who is hated. Racial
animosity expressed in such -slogans as "The Yellow Peril" were a factor in
American threat perceptions visa-vis Japan after the Russo-Japanese War. Com-
placency and basic anxiety are attitudes that can impinge on threat perception.
Complacency was a factor in Israel following her brilliant victory. :over the
Arabs in 1967. It tends to encourage underestimates of external threats. Anxiety
tends to do the opposite.
Second, misperceptions can be generated by strong ideological commitments be-
cause rigid general beliefs about the nature of the outside world are likely to
govern the selection of guiding assumptions. These are then based less, or not at
all, on objective empirical analysis than automatically deduced from prior
beliefs. An adherent to Leninist ideology will readily assume that capitalist
countries constitute the mortal enemy of communist societies. When the Cold
War posture had.become rigid in the United States, it was not hard for Americans
to overestimate the threat emanating from monolithic world communism even after
Yugloslavia and Albania had separated themselves from Moscow tutelage and control
and the Soviet-Chinese split had deepened.'
'Third, bureaucratic behavior, including bureaucratic politics, in intelli-
gence agencieh:Vprff@ r? I ase ZUOb/b fflb~ Fi"% 0(r 985R000200130016-0storting
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
predisposition. In addition to the factor to be taken up shortly, it tends
to be functional in such organizations to protect assumptions once they
have been accepted. Powerful incentives induce individual members to conform
and not to challenge such guiding preconceptions. Bureaucratic inertia thus
perpetuates expectations about foreign actors that fitted the real world earlier
on but have ceased to do so because of relevant changes in the environment.
Fourth, there is wishfulness which historical study suggests to be the
predisposition most frequently at work. Psychologists tell us that to predict
what one wants to happen is an insistent human tendency. On the -one hand,
people who wish to cut defense expenditure because they want to reduce taxes
.or increase welfare outlays come easily to the belief that external threats
are low even when these estimates are not based on any qualifications for. en-
gaging in threat perception. On the other hand, people who profit one way or
another from rising defense spending, including the military, tend to over-
estimate foreign threats. I am not speaking here of the deliberate misrepre-
sentation of threats but of sub-conscious leanings. The British appeasers of
Germany in the 1930's inclined toward.underestimating the Hitler menace in part
because they found the prospect of a major war abominable. They wanted peace.
In disbelieving all intelligence information foretelling a German attack, and
there was plenty of it, Stalin may have been influenced by his recognition
that the Soviet Union needed time for strengthening its armed forces, greatly
weakened by the preceding purge of the officer corps, before taking on the
German army. Of particular importance is that leaders and foreign-policy offi-
cials tend to like intelligence estimates that permit them to pursue favorite
policies. Once a particular foreign policy is found desirable on other grounds,
it becomes painful to accept evidence of foreign threats in conflict with that
policy.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0 _
 "Approved ForwRelease 2005/01/10 : CIA-RDP86B00985R000200130016-0
It is then more rewarding to ignore unwelcome signals or to interpret them in
ways that allow them to be assimilated to governing assumptions. This predilec-
tion is so ubiquitous according to historical experience that it has led to
the belief that intelligence production must be separated from the formulation
`)f po-U_cy.
Two reasons account for the frequency with which these intervening disposi-
tions operate, and for the remarkable staying power they display. One results
from the purely intellectual problems of threat perception, in particular the
fact that relevant information is usually ambiguous and hence capable to multi-
pie explanations. No-one inference can be proved in advance to be correct.
The implied choice of interpretation gives intervening predispositions great
ease of entry. The other reason is precisely that the actor is unaware of
their intervention.
Foreign threats can obviously also be misrepresented deliberately. His=
torical cases are not rare in which a ruler or ruling group tried to bolster
.waning domestic authority by diverting public attention to external threats and
profitting from the sense of solidarity that is often triggered in the face of
a foreign menace. More generally, individuals and groups that expect to gain
from national response to a threatening environment may assert the presence of
such threats not-as a result of sub-conscious urges but.as a. deliberate act of
exaggeration.. For the same sort of reason, threat perception can be manipu-
lated also in the opposite direction. In this area, however, we are in the realm
of supposition because firm evidence of deliberate misrepresentation is naturally
hard to find. But few will deny the plausibility of the hypothesis.
In addition to these structural problems of international threat perception,
there can also be situational conditions that may aggravate the difficulty of the
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R0O0200130016-0
operation.- As is generally observed in the literature, this aggravation tends to
occur at times of profound and fast-moving international crises when time constraints
and stress can reduce analytical effort and distort perceptions. Yet it needs
to be asked whether crises cannot also.produce positive effects by facilitating
the correction of preceding perceptions. To the extent that a serious international
crisis comes as a surprise or moves in surprising ways, it discredits intelligence
estimates previously, made and casts doubt on the conceptual assumptions on which
these estimates were premised. At the same time, the conduct of other governments
in the crisis supplies new information about their possible intentions. As a
result, crises offer an opportunity for learning and this is an advantage even
if the learning may have to be done very quickly. Indeed, the need for speed often
means that top leaders rather than intelligence bureaucracies will do the learning
in the first place. Whether the harmful or helpful effects of crisis pervail is
evidently an empirical question.
Estimates of the capabilities of states are an integral part of international
threat perception. Whenever the presently peaceful conduct of countries cannot
.be regarded as permanent, the estimate of potential military threats rests on the
estimate of capabilities. When a state is regarded as presenting an actual threat,
capability estimates attempt to assess the precise dimensions of the threat,.
One sometimes hears it said that--because it is difficult to estimate the
intentions of foreign governments, and risky to rely-on such estimates--prudence
is served best by an estimate of capabilities and the assumption that foreign actors
are apt to do the worst to us they are capable of doing. Cri.ticg of this prescrip-
tion point out that worst-case assumptions automatically overestimate external
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For,?YRelease 2005/01/10 : CIA-RDP86B00985R000200130016-0
threats and that action based on them provokes foreign insecurity, arms races
and confrontations, and thus ends up by making the world more threatening and
dangerous,.
It may well be that military services (but not necessarily military intelli-
gence) routinely design contingency war plans on the basis of worst-case assump-
tions; and, in the light of our preceding analysis, such practice is not neces-
.aari'ly unsound. However, this analysis and the historical case studies on which
it rests also suggest that the prescription is not widely followed in national
intelligence efforts. We are unable to say whether underestimates of threats
have been more or less common than overestimates. But the large number of
recorded cases of strategic surprise indicates that worst-case assumptions were
often not considered seriously enough. In some instances, e.g. Stalin's`in 1941,
best-case assumptions seem to have prevailed. To consider worst outcomes, or for
that matter best ones, is not of course the same as fashioning policy solely on
estimates based on either extreme assumption.
To think that foreign capabilities are estimated easily.is an illusion, and the
claim that capability estimates as substantially more reliable than estimates of
foreign intentions is highly debatable. To be sure, there are items in capability
analysis than can be directly observed, and counted or measured within narrow
limits, such as changes in defense budgets, number of armored divisions, fighter
aircraft, and certain weapons characteristics. However, while these sorts of
things are important, they represent only a fraction of overall capabilities.
Capability estimates are ultimately needed in order to decide what, if anything,
can and should be done in order to deter external threats when the need for
deterrence cannot be ruled out, and to defend if deterrence was needed but failed.
In the final analysis, therefore, capability estimates refer to the test of war
which is the only true test of capabilities.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 CIA-RDP86B00985R000200130016-0
As the his of warfare demonstrates abundantly, the outcome of war depends
not only on quantitative things known beforehand but, on a host of qualitative
factors, such as troop training and morale, military leadership, strategy and
tactics, military intelligence and communications, the performance of arms under
wartime conditions, the behavior of allies, the ability of belligerent governments
and publics to absorb casualties, ardso forth. And these qualitative things are
difficult to estimate. Moreover, all the elements that determine the military
capability of a potential opponent in wartime are significant only relative to
the capabilities of one's own side. The implied requirement to estimate one's
own capabilities, including the many qualitative components, opens another dimen-
sion on which realistic estimates are hard to achieve. Is there not a strong
inclination, reflecting fairly obvious reasons, to overestimate the capabilities
of one's own side, and especially the qualitative aspects for which evidence is
ambiguous? L5'/
We conclude, then, that--taking the entire range of ingredients into account--
the estimate of military and related capabilities is necessarily based on infor-
mation that, as a whole, is fragmentary, obsolescent and, above all, ambiguous.
The properties of this information are not very different from those that are
.relevant to the estimate of foreign intentions, and. the problems of threat per-
ception discussed in the preceding analysis apply also to capability estimates.
Again, how the information is. interpreted depends crucially on assumptions and
preconceptions, as is made clear by the following four examples of mistaken
capability estimates.
5/Military planners not rarely overestimate certain components of foreign forces
(e.g. numbers of men, numbers and types of arms) in order to appeal for more funds
from their governments. They are less likely to say that their troops suffer
from bad training and morale, and that their leadership, strategy and tactics are
inept.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130d16-0
The concept of the famous French Maginot Line of fortification (including
the missing link to the sea in the northwest) originated in strategic studies
undertaken in the early 1920's. The concept--on which French planning for deter-
ring and defeating German aggression came to rest increasingly--assumed that mili-
tary technology continued to favor the defense over the offense, as it had during
World Jar I. This assumption was not revised when improved tanks and aircraft,
and tactical uses that capitalized on these improvements, made it doubtful in
the 1930's. German capabilities were grossly underrated as a result. In 1940,
the Germans achieved surprise and, though.using fewer troops and tanks than the
French, quickly crushed resistance by making an end run around the Line.16/ Even
though some French officers (including Charles de (;ualle) had themselves been
thinking about the Blitzkrieg tactics developed by the German army, and even
though these tactics had been used
against Poland in 1939, the official French
strategy could be defended on the grounds that the superiority of Blitzkrieg
against staunch defenses was doubtful or improbable. New information was am-
biguous enough in its implications to protect old assumptions that led to defec-
tive estimates of foreign capability.
During World War II, the Americans and the British adopted strategies of air
bombardment that were based on serious overestimates of German vulnerability.17/
The American bombing offensive, which emphasized precision-bombing of key capital
structures in the war economy, was prompted'by the mistaken assumption that the
16/Paul Bracken, Unintended Consequences of Strategic Gaming, Hudson Institute
Paper, May 2, 1977.
1.7/
Harold L. Wilensky, Organizational Intelligence (New York, Basic Books, 1967,
pp. 24-3' .
J
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 proved For Release 2005/01/10 : CIA-RDP86B00985F,000200130016-0
o ca
es
-
t
h
stretc
G
p
.i
--.
y .
-
am
-___
erman s productive capacity
should be overtaxed-after several years of destructive warfare was no doubt a
reasonable hypothesis. But it turned out that the German economy actually had
plenty of slack. The British air offensive againt German cities was guided by
mistaken expectations about the brittleness of German civilian morale. This
M
underestimate of German capability, in turn, was derived from assumptions, later
shown to be unrealistic, about what German bombers would do to British civilian.
morale. And this overestimate was derived by British planners from dubiously
selected information relating civilian casualties to the weight of bombs dropped
on London by German zeppelins in World War I. These estimates, incidentally,
affected the behavior of the British government toward Germany during the late
1930's. 18/
The Soviet economy and Soviet defense spending have been important subjects
of American estimates. In 1976, the CIA suddenly revised its previous estimates
for 1970-1975 that apparently had underestimated the Soviet defense budget by
half and the growth of that budget by about two-thirds. The underestimates had
resulted from assumptions about the content and structure of the Soviet defense
budget that were suddenly recognized as incorrect.l9/ It was not new information
but a conceptual challenge thatcaused a long-employed assumption to be reviewed.
As Albert Wohlstetter has demonstrated in a number of studies, American intelligence
consistently overestimted Soviet deployment of inter-continental bombers and ICBM's
from the 1950's to 1961, only to underestimate Soviet ICBM deployment after 1961
18/ Bracken, op. cit., pp. 28ff.
19I
William T.Lee, Understanding the Soviet Military Threat, National Strategy
Information Center, Agenda Paper No. 6, New York, 1977, pp. '7-22..
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
with equal consistence. The faulty estimates in both periods ultimately rested
on the expectation that Soviet strategic thinking was like American strategic
thought. Lo/ Assumptions that reflect such mirror-imaging are a frequent source
of misperception.
Subconscious predispositions intervene in the choice and use of guiding
assumptions as readily in estimates of capabilities as they do in the perception
of foreign intentions. To refer to the examples of capability estimates just
presented, the change in American estimates of Soviet strategic nuclear forces
suggests the familiar learning process that overreacts to past error. It has
also been intimated that the later underestimates resulted in part from an
institutional desire to live down a reputation for producing overestimates.-
The data from which British overestimates of air bombardment were derived would
The
not have remained unchallenged for as long as they did if they had not been kept
highly classified. ?
Wishfulness, which is probably the chief culprit in distorting international
threat perception, was evidently at work in producing and protecting the estimates
of the German threat that justified the Maginot idea of deterrence and defense.
In the future, the French wanted to avoid the frightful losses of manpower suffered
in World War II. They hoped that, unable to break through the defenses of its
opponent, Germany would be beaten through economic strangulation.2 " This hope
to win without a long grinding war on land also inspired American and British
planners whose estimates exaggerated Germany's vulnerability to attack from the
201 Ibid., pp. 24-32. See also the papers by Wohlstetter listed in the Appendix.
?l/
Lee, o ci prHvOFor Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
22/ Bracken. on. cit.. V. 7.
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
V
We are now ready to turn to the question of how statecraft dealing with
international threat perception should be organized in order to cope with the
problems we have described. It is not surprising that the literature on reme-
dies is as yet only in its infancy. The severity of the problems had not been
realized until quite recently. We have been at pains to demonstrate that bad
intelligence performance--one is almost inclined to say unlucky intelligence
performance--does not usually result from stupidity or lack of effort. Indeed,
it seems facile to blame intelligence services for having done a poor job of
prediction in particular cases. They probably did the best they could in want
of a trustworthy crystal ball. Nor do the intrinsic difficulties encourage
hopes of radical improvement. What can be expected at best is auoderate im-
provement of average performance. Relevant statecraft, therefore, concerns
not only the production of intelligence but also the question of what can be
prudently expected from it, the matter of intelligence consumption and, ultimately,
adaptation to the fact that the intelligence product can never rise securely above
suspicion.
Two conclusions that are sometimes drawn from the recognition that the
-obstacles to perfect threat perception are insuperable, should be firmly rejected.
First, to eliminate or seriously downgrade formal intelligence efforts could make
matters only worse, for no policies can be designed without some sort of assump-
tions about the future. To do away with professional intelligence officers would
be replacing them by amateurs who are up against the-same problems with less
awareness and aptitude. Second, it would be wrong to conclude that, predicting
the intentions of foreign actors being hopeless, one' better focus-entirely on esti-
mating their capabilities and. assume that their intentions will be the worst
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
imaginable. To imagine the worst is by no means easy and surprises could not
be avoided if one tried. As demonstrated above,.. to estimate capabilities is
not all that easier than estimating intentions. And to act on worst-case assump-
tions about foreign intentions might well increase rather than reduce national
insecurity.
Measures to insure. the improvement of intelligence production can be
institutional and doctrinal. Institutional remedies are essentially not a matter
of organii?at$onal structure but of practices which, if adopted, will entail
certain administrative requirements. The key problem in threat perception
is clearly the quality of the assumptions that are brought to the information
and guide the selective perceptions of intelligence officers. Several remedial
practices seem worth considering. (1) It is obviously important that these
assumptions be made explicit and that they be continuously reviewed in the light
of new'information. The danger is that, if these things are done, they will be
done routinely and without keen alertness to the likely obsolescence of all pre-
conceptions.
(2) As a further check on the fitness of preconceptions, it would seem useful
that an explicit attempt be made to identify and evaluate various motivational
patterns that could explain the observed behavior of foreign actors,-and various
assumptions that, when applied to information about foreign. capabilities, would
produce different estimative conclusions. Because no estimate can do more than
-reduce uncertainty and every estimate can be wrong and, if believed in, cause
subsequent surprise, it would seem sound practice to.present every probabilistic
estimate on matters of great consequence within the framework of accompanying
worst case and best case interpretation of the same information.- Again, the
danger is that the two limiting interpretations would be made unappealing and
turned into caricatures.
Approved For Release 2005/01/10 CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
(3) To achieve something of the same purpose, it has also been suggested
that intelligence services practice multiple advocacy.by appointing a devil's
advocate within their organizations. But it is doubtful that such a person
or group can be given enough autonomy of operation,and bureaucratic incentives
and influence, to do the job effectively..?'
(4) Another approach to this problem--and one already practiced--is not to
0
set up one bureaucracy that enjoys a monopoly of intelligence production. A
degree ofompetition is some safeguard against the hardening of assumptions.
One danger of this particular solution. is that the desire of separate bureau-
cracies for what economists call product differentiation--a normal organizational
proclivity--encourages dissenting estimates for thi wrong reasons. Dissent for the
wrong reason can also result when different.-intelligence groups have different
institutional customers with competing vested interests. A third danger is
that the different groups seek to limit competition by negotiating informally
about the degree and nature of estimative disagreement. Remedies (2), (3) and
(4) mean that the consumer will not receive a single clear-cut estimate. In
principle, this result is to,be welcomed because the consumer should understand
the limited reliability of all estimates, and should not be allowed to escape
facing up to uncertainty. Indeed, he should recognize that prediction is not
only the most risky but also the least important function of intelligence.24
233/According to newspaper reports, the CIA experimented recently wit., appoint-
ing an outside team for estimating Soviet intentions as a check cr: its own
"in-house" estimate. The outside team was rumored to have been one of "hard-
liners" and in any case came to the conclusion that, the Soviet threat to the
United States was greater than the one presented by the regular CIA team.
The value of this practice would seem to depend on who picks the teams, on
which criteria_of_selection. For a description of this experiment, see Lee,
op. cit., Appendix B.
24/Wilensky, op. cit., p. 64.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
There is the danger, however, that,the supply of alternative estimates will let
the consumer follow his predispositions and simply pick the one he likes best
because it permits him to do what he wants to do anyhow.
(5) To evaluate intelligence predictions after the event would seem to be
critical to organizational learning. But to make sure that postmortems are
done thoroughly and without prejudice, they would have to be undertaken by an
independent group, and making estimating staffs react to theoreviews in detail
would be necessary in order to assure that feedback generates learning. The re-
view process'should include estimates that turn out to have been correct because
they may have done so for the wrong reasons, and therefore have just been lucky.
(6) Intelligence services become sometimes reluctant to alert governments
to disruptive events they think might occur because the same warnings have. been
expressed repeatedly before when nothing of the kind ensued. The "cry-wolf"
syndrome is obviously dangerous when the events in question are of grave conse-
quence, and intelligence officials should not be deterred from reissuing such
warnings with a considered indication of probabilities. If the warning problem
is properly understood, intelligence consumers in the government will accept the
necessity of repeated estimates that, expressed in probability terms, warn of
events that do not later take place (possibly because of a state of alertness
produced by the warning).
(7) Because the behavior of other countries commonly depends on the policies
pursued by one's own government, it seems advisable that intelligence estimates
take such policies into account. This may not be easy to arrange and even be
frowned upon when the policies are still secret and.the intelligence service is
institutionally separated from policy-making.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 CIA-RDP86B00985R00020013001.6-0
(8) The development of new techniques and equipment for processing and
analyzing information (especially computers), though worthwhile for some special
tasks, are not nearly as helpful as might be assumed. They often supply a
flood of information-that is more than an individual or small group can digest,
and hence leads to greater specialization. The resulting compartmentalization
impedes the reintegration of informational inputs. Given the difficulties of
using ambiguous information in matters of threat perception, computers are useless.
for the larger analytical tasks. How do we program them for the simulated
enactment of*real-life sequences? How do we program computers for processing
ambiguous information? To hope for the discovery of indicators that remove
ambiguity seems extremely far-fetched at this stage of knowledge. No doubt,
computers are and will be useful in accomplishing clear-cut subsidiary functions.
But the central problems in international threat perception are not susceptible
to technological solutions.
None of the remedial practices we have listed are without problems of their
own. It is possible that some will be counterproductive in the real bureaucratic
world. At this point, therefore, the question is one of experimenting with various
remedial procedures in order to discover their cost-effectiveness.
Yet, in any case, the introduction of such practices that promise improve-
ment is unlikely to yield good results, and might not even take place, in the
absence of proper professional doctrines among intelligence producers and con-
sumers alike. Professional self-indoctrination among producers would encompass
a sense of responsibilities that is realistic in terms of all the grave diffi-
culties that beset international threat perception. The preceding discussion
of practices indicates the specific issues which doctrine should, engage.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
However, there is, in my opinion, one pre-condition that must be met squarely
before sound doctrine can develop. This is a full understanding of the histor-
ical record of statecraft concerned with international threat perception--a
record that now exists only in fragments and demands more analysis as well as
completion. Only historical knowledge and-analytical penetration can give
intelligence officers a realistic conception of the daunting difficulties they
must professionally.face, the pitfalls to be avoided as much as possible and
the approaches to be cultivated.
If kribwledgd about the problems of intelligence production has only begun
to develop, even more needs to be done about understanding the problems of intelli-
gence consumption and of bridging the gap between output and consumption by means
of proper guidance. The following suggestions are therefore necessarily sketchy.
For reasons we have spelled out, trouble is inevitable if the main produc-
tion of intelligence is closely integrated with the design of foreign policy.
The temptation to tailor intelligence estimates to the desires of the policy-
maker is then exceedingly hard to resist. Separation, however, also creates
problems. One is that intelligence production may not be properly directed
toward the issues that are important to policy making. The intelligence people
therefore need guidance to make their work sensitive to consumption needs.
This does not mean that intelligence should be dominated by policy. All that is
needed is sufficient coordination. Even if this separation of functions is
adopted, the policy maker and especially top leaders should not rely exclusively
on the output of the main intelligence organ because, as we have shown, even
the most carefully made estimates can turn out to be seriously misleading.
Consumers, therefore, should also have access to smaller intelligence units
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 in the governme PSQY?dt i+'d:~fllsaneol 'N&0tKklAdfat6g88lB0G985RCCD1612*X(+ 16Qgnopoly.
In the United States, for instance, intelligence groups in the Department of
States and Defense Department, and -- closer to the White House -- attached to
the Security Council, are available in addition to the CIA. Such pluralism also
entails drawbacks. If conflicting estimates are produced, consumers have a
choice and may well pick the one that fits their own predispositions. Yet the
disadvantage of relying on one single source seems to be even greater. In
addition, top consumers should-insist on estimates (concerning questions of great
consequence) in which the one favored by the producer is bracketed by competing
interpretations of the available information that are rejected. If it is objected
that this would overburden the consumer, the answer surely is that to accept any
lesser burden--a short. and single that is supposedly reliable estimate--is to
have surrendered to illusion.
Such practices will not be instituted unless intelligence consumers are
properly indoctrinated in the uses of estimates. Consumers also require considerable
education in the problems that are inherent in international threat perception..
Once they have internalized this knowledge, they will be less likely to substitute
their own estimates for intelligence products that frustrate their expectations,
will tolerate the fact that no one estimate is sure to be correct and will apprec-
iate and yet not abuse the presentation of diverging products. In the light of
past performance and of analysis, there can be no excusable expectation of perfect
intelligence and no justification for expecting the resolution of uncertainty
about the future.
The acceptance.of unavoidable uncertainty greatly complicates the conduct
of foreign policy, but need not paralyze it. If we understand that all forecasts
have a substantial chance of proving faulty, we have accepted the real possibility
of surprise and we are free to make provision for minimizing the ill consequences
Approved For Release 2005/01/10 : CIA-RDP86500985R000200130016-0
 Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
of surprise. Such provision may be costly in terms of resources, as insurance
always is, and even in terms of downgrading or abandoning po:Licy options that
look too risky. But to make such provision would seem to be prudent and conducive
to*national interest in an insecure world.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 Approved For Release 20 p X IA-RDP86B00985R00020013O016-0
Selective List of Literature
(with Selective Annotations)
Belden, Thomas G., "Indications, Warnings, and Crisis Operations," International
Studies Quarterly, XXI, March 1977, pp. 181-198.
--A thoughtful analysis of the problem of indicators to be used in the warning
process.
Commission on the Organization of the Government for the Conduct of Foreign
Polite, Appendices, Vol. VII, Washington, D.C., Government Printing
Office, June 1975.
--Several papers of interest on many problems of foreign intelligence,
but ma.inly?on organization matters.
George, Alexander L..and Richard-Smoke, Deterrence in.American Foreign Policy-
Theory and Practice, New York; Columbia University Press, 1974..
--The case studies on deterrence failure are also case. studies on inade-?
quate threat perception. The theoretical part researches especially'
attitudinal barriers to correct threat perception.
Rilsman, Roger, Strategic Intelligence and National Decisions; Glencoe, Ill.,
Free Press, 1956.
--A formulation of intelligence doctrines based on interviews with intelli-
.,gerice operators.
Jervis, Robert, Perception and Misperception in International Politics,
Princeton, Princeton University Press, 1976.
--The most systematic analysis of problems of international perception.
Kent, Sherman, Strategic Intelligence for American World Polite, Princeton, N.J.,
Princeton University Press, 1949.
-One of the earliest American introductions to'the subject, written by
an academic historian who subsequently became?a high-ranking intelligence
officer.
Knorr, Klaus, "Failures in National Intelligence Estimates: The Case of the
Cuban Missiles," World Politics, XVI, 1964, pp. 455-467.
--The influence of preconceptions and assumptions on intelligence production
is examined in this case study.
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 4pprovec For RJease 2005/1/ 0 : IA-RDP> 6Q00985R0020013001 0
Knorr, Klaus,. oreign me igence and the Social Sciences, esearc onograph
No. 17, Center of International Studies, Princeton University, June 1964.
-- In addition to discussing the relevance of social science knowledge to
intelligence work, this paper emphasized the need for a descriptive
and normative theory of intelligence.
------, "Threat Perception". in Klaus Knorr (ed.), Historical Dimensions of
National Security Problems, Lawrence, Ka., Kansas University Press, 1976,
pp. 78-119.
--A historical assessment of the record of threat perception and an analy-
sis of its problems, including predispositional factors.'
Lee, William T., Understanding the Soviet Military Threat (National Strategy
InforQation.Center, Agenda Paper No. 6, New York, 1977).
--An examination of faulty CIA estimates on the Soviet defense budget.
Leites, Nathan C., The Operational Code of the Politburo, New York, McGraw
Hill, 1956.
--A fascinating attempt at identifying systematically the doctrines that
govern foreign policy-making in a major country.
Morgenstern, Oskar, Klaus Knorr and Klaus Heiss, Long-Term Projections of
Power: Political, Economic, and Military Forecasting, Cambridge, Mass.,
Ballinger, 1973.
--A critique of several techniques of forecasting.
Shlaim, Avi, "Failures in National Intelligence Estimates: The Case of the
Yom Kippur War," World Politics, XXVIII, 1976, pp. 348-380.
--An excellent case study placed within a conceptual framework and illus-
trating several key hypotheses. Also good on remedies.
Whaley, Barton, Codeword Barbarossa, Cambridge, Mass., MIT Press, 1973.
--A detailed case study of Stalin's failure to foresee the German attack
in 1941. It adds to the study of Roberta Wohlstetter by emphasizing
the factor of deception.
Wilensky, Harold L., Organizational Intelligence (New York, Basic Books, 1967).
--Fine analysis by a sociologist of the intelligence function in all organiza-
tions.
Wohlstetter, RRbbe to
providd
Press,
University
D
~r -
ecision Stanford Stanford
B_6688rc000200130046-0
 H
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
Highlights from Klaus Knorr, "Strategic Intelligence:
Problems and Remedies"
". . . the record of international threat perception in
the vital strategic area is disconcertingly poor." (6)
2. "Misperception and surprise do not usually result from
lack of information." (9) In the Israeli, October 1973, case
an intelligence officer made the correct estimate, predicting
attack on Oct. 1 and again on Oct. 3. But his superiors rejected
his judgment. "In every other case I have studied, it is easy to
see in retrospect that the relevant information for making a
correct estimate was available." The problem is to separate the
correct information from the rest in a situation that is essentially
ambiguous and may be further confused by deliberate "disinformation."
"All that can be derived from ambiguous information are inferences,
and no one inference can ever be compelling because any ambiguous
behavior can be explained by more than one motivational pattern." (10)
Correct estimates are further complicated by the fact that the other
side is itself undecided or "acting on the basis of multiple options"
- as was the case in the attack on Pearl Harbor (the Japanese
admiral was to attack only if the U.S. fl et was at anchor and he was
able to achieve surprise). (10-11)
3. "In drawing inferences from available information that is
usually fragmentary and ambiguous, Intelligence officers use certain
assumptions about the behavioral pattern of the potential enemy."
(12) These assumptions are not stupid usually; what is wrong about
them is "the degree of trust they commonly command. . . human actors
can produce unexpected behavior for any number of reasons." (13) Thus
the patterns may be changed deliberately, for purposes of deception,
but they are also likely to be changed for other reasons, such as
"new domestic pressure, changes in relative capabilities, the individual
influence of personalities, failures of intelligence on their part,
deviations from rationality, and other variable conditions including
the possibility that, in the life of governments as elsewhere, things
happen by accident as well as design." (13) Reliance on "the lessons
of history" (e.g. that "appeasement encourages aggression") can be
even riskier. (14). In sums "although good estimates can reduce
uncertainty about the future, even the best cannot be depended on to
remove it." (14)
4. "There is an entire further dimension of factors that tend
to cause mistakes. . . . the perceiver, far from being unbiased,
often approaches his task under the influence, usually unwittingly,
of predispositions that affect his choice of assumptions and his
receptivity o incoming information. . ." (15) The sources of these
predispositions may be emotions, ideological commitments, bureaucratic
politics, or simple wishful thinking ("the predisposition most
frequently at work"). (16-17)
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0
 5. "APprbhcc#t~ortf l4asfeomign/l?zi-Pbi OSBO0 i6ROO%m.N @-0easily is ai
illusion, and the claim that capability estimates are substantially
more reliable than estimates of foreign intentions is highly debatable
(19) Further, as to estimate enemy capabilities requires an estimate
of one's own,, another dimension is opened on which "realistic
estimates are hard to achieve." (20)
6. "We have been at pains to demonstrate that bad intelligence
performance -- one is almost inclined to say unlucky intelligence
pwerformance -- does not usually result from stupidity or lack of
effort.. . Nor do the intrinsic difficulties encourage hopes of
radical improvement. What can be expected at best is a moderate
improvement of average performance. Relevant statecraft, therefore,
concerns not only the production of intelligence but also the question
of what can be prudently expected from it, the matter of intelligence
con.umption and, ultimately, adaptation to the fact that the
intelligence product can never rise securely above suspicion." (24)
7. ". . . to eliminate or seriously downgrade formal intell-
igence efforts could make matters only worse, for no policies can be
designed without some sort of assumptions about the future. To
do away with professional intelligence officers would be replacing
them by amatenrzn who are up against the same problems with
less awareness and aptitude(;"' (24)
8. Institutional remedies. "[These] are essentially not
a matter of organizational structure but of practices which, if"
adopted, will entail certain administrative requirements." (1) Ass-
umptions should be made ex -ic-it and "continuously reviewed in tie
ight of new information." (2) In every case "an attempt should be
made to identify and evaluate various motivational patterns. and assumptions that. . . would produce i ferent estimative
conclusions." Meanwhile avoiding the fallacy of pre~entzng t]i.e
snl a[' a-h-fe...est.im- to between unappealing extremes must be
achieved. (25) (3) Employment of "a cLeevil's advocate" within the
intelligence institution is a doubtful. recourse. (Y4) It is useful
to up "one bureaucrat mono f% V
intelligence production." 6 If the consumer receives competing
estimates, is
result is to be welcomed because the consumer
should understand the limited reliability of all estimates, and
should not be allowed to escape facing up to uncertainty. Indeed,
he should recognize that prediction is not only the most risky but
also the ast im ortan _ ce." (26) There is
a danger, though, that multiple estimates may encourage the consumer
to "follow his predispositions and simply pick the one he ,`lce~ sec
(5)Independent postmortems are a useful exercise. (6) Intelligence-
officials shouldnt e deterred (by the "cry-wolf syndrome")
from reissuing warnings with a considered indication of probabilities.
(7) " . . it seems advisable that intelligence estimates take . . .
ie_s..into account (even though it may not be easy and may be
frowned upon while the policies are still secret). (27) (8) "The
development of new techniques and equipment for processing and
analyzing information (especially computers), though worthwhile for
some special tasks, are not nearly as helpful as might be assumed."
Approved For Release 2005/01?/10 : CIA-RDP86B00985R000200130016-0
 9. The institutional rememdies are unlikely to be effective
Jithout"prcA VoOM4 1e%Sg120( x1@'1'(3.nCAA-RDP& 9.85RQ01Dg@6k' 90 r@ducers
and Consumers alike." "There is, in my opinion, one pre-condition
that must be met squarely before sound doctrine can develop. This isl
a full understanding of the hi t"r~ re.co-r-d-_of fie- -af_ t concerned
with international threat erce do -- a record that now exists only
i ragmen's and eman s more analysis as well as completion." (28,29) G
10. ". . . the policy maker and especially top leaders should
not rely exclusively on the output of the main intelligence organ
because, as we have shown, even the most carefully made estimates
can turn out to be seriously misleading. Consumers, therefore, should
also have access .gsmaller intelligence units in the government
structure as a check on the dangers of intelligence mono oly." (30)
"In addition, top consumers s Would insist on estimates concerning
questions of great consequence) in which the one favored by the
producer is bracketed by competing interpretations of the available
information that are rejected. . . a`cce`pt-any lesser burden-
a short and single that is supposedly reliable estimate -- is to
have surrendered t o illusion."(30)
11. "In the light of past performance and of analysis, there can
be no excusable expectation of perfect intelligence and no justificati
for expecting the resolution of uncertainty about the future." (3)
12. "The acceptance of unavoidable uncertainty greatly
complicates the conduct of foreign policy, but need not paralyze it.
If we understand that all forecasts have a substantial chance of
proving faulty, we have accepted the real possibility of surprise
and are free to make provision for minimizing the ill consequences
of surprise." (31)
COMMENT: Klaus seems to me to have hit all the high spots without
exhausting the subject. His argument is, of course, fully illustrated
from historical cases. It would be fascinating to me and I suspect
to the DCI as well, to hear him debate with the subject'. 25X1
Intelligence and Policy Decisions."
25X1
Approved For Release 2005/01/10 : CIA-RDP86B00985R000200130016-0