WARSAW PACT THEATER FORCES

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP80M01082A000800030002-3
• Release Decision: RIPPUB
• Original Classification: S
• Document Page Count: 31
• Document Creation Date: December 15, 2016
• Document Release Date: March 26, 2004
• Sequence Number: 2
• Publication Date: November 18, 1974
• Content Type: MF

mNNI Vv=u I VI IIGIGQJG LVVY/VJ/J I clr V1VFV 1VVL/1VVVVVVVJ VL-J
THE DIRECTOR OF CENTRAL INTELLIGENCE
WASHINGTON, D. C. 20505
18 November 1974
MEMORANDUM FOR: Lt. General Samuel V. Wilson, USA
Deputy to the DCI for the
Intelligence Community
SUBJECT Warsaw Pact Theater Forces
1. Ever since the DCI gave me my present responsi-
bilities, I have been concerned about the complex problems
inherent in the Community's discharge of its continuing
responsibility for maintaining a valid, current assessment
of the size, strength and capabilities of the Warsaw Pact's
theater forces -- an intelligence conundrum about which
25X1 Vol] knnUr ('nnCjrlor0}.J.1 ____ a-1..,.,. T T
I took advantage of his dedication and good nature to ask'
if he would be willing to take a thorough, thoughtful look
at this range of problems once he was unchained from an
in-box and the pressures of deadlines or flaps. No one
has been more immersed in these questions than he, or
done more to shed light on them. I asked him -- at his
own pace and in his own way -- to take a hard look at this
conundrum and its associated issues, to draw on the totality
of his background and experience, and then to set down
his views, conclusions, and suggestions candidly and in
some detail.
2. Attached for your personal information and
consideration are two copies of the results of 0 work, 25X1
an essay entitled "Beans and Bean Counters." It is a
document I commend to you as well worth reading. In it,
I makes a number of suggestions about changes that
might be made in various aspects of the Community's
address to this problem which, in his opinion, would make
CL BY
Approved For Release 2004/03/ c1A= 2DP80MO1082A000800030002-3
25X1
 'FJh~L'rtiLl~------ _ --- - ------------ -
that address more effective and efficient. Given the near
real time systems just over the horizon, the necessity
for putting ourselves in the best possible posture before
we are inundated in a flood of new data is self evident.
3. Q suggestions have the merit of being well
thought out and coherent. I would not necessarily advo-
cate their adoption in their entirety but believe they
can be extremely useful as a stimulus (or even provocation)
to the ideas of others. By working together under this
stimulus, the Community may be able to come up with an
even better package.
4. I would like to get in touch with you after I
return from Europe to get your reactions to Otreat-
ment of this problem and his recommendations. At this
stage, this is a very informal exercise and should be kept
as such. I am not broadcasting this study throughout the
Community and believe it is too early to do so. I have
sent it on a personal basis in a similar note to
IHank Knoche.
25X1
25X1
25X1
~1George A. Carver, Jr.
Deputy `or National Intelligence Officers
Approved For Release 2004/ ,: G1A-RDP80M01082A000800030002-3
'W~~j 'm-4p.
 Approvad For Release 2004/03/31 : CIA-RDP80MO1082A00Q800030002-3
STAT
STAT
STAT
MEMORANDUM FOR:
W46. A44
A& Gin 1p
(DATE)
AMC.
t7LL(
a C"
F^RM NO.
AUG 54
=_11 Ix L)
REPLACES FORM 10.101
WHICH MAY BE USED.
FORM GN 54 I0I WHICH RELACES FORM
BE US 10-01
I AU
ACTION
STAT
Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
 Next 1 Page(s) In Document Exempt
Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
 TOP SECRET
14 November 1974
BEANS AND BEAN COUNTERS
Foreword
Keeping the books on Communist theater forces, a task usually
referred to as order of battle intelligence but better described
as bean counting, has been a long time subject of controversy.
It is a task that nobody and everybody wants. No one doing the
work hasn't wished he could push it off on someone else because
it requires an increasing input of resources at a time when other
demands have grown also. At the same time, control of production
of the basic data on theater forces, data on which most analysis
must rest, is something everyone wants. It's felt to be too im-
portant to be trusted to the other fellow.
gies of other producers.
selves and finally end up questioning the capabilities and methodolo-
There are both political and substantive reasons for the.
dichotomy. The numerology of the theater forces is important in
balance of power arguments, force capability studies, interservice
rivalry, and budget competition. Because no one is free of bias,
conscious or unconscious, no consumer completely trusts data pro-
duced by someone else or presentations based on them. But trust
aside, no series of numbers ever seems to be quite what everyone
can use in analysis. The salami is never sliced quite right. The
result is that most analytical shops try to produce series to
meet their own needs, and in doing. so get into the numbers them-
The duplication of effort which results involves more than
just additional sets of analysts. Collectors are confronted with
competing and conflicting requirements -- as are the processors.
Indeed, the duplication has extended into processing as well as
analysis. But most troubled are the consumers whose anguished
cries for someone to get the numbers right are seldom satisfied.
This paper examines the order of battle function, who wants
data, how they are produced, and proposals for changes in the
systems that produce them.
TOP SECRET
 TOP SECRET
1. The demands for data on the Warsaw Pact theater forces
have increased steadily for the past 10 years.
-- forces capability studies have gone far beyond the
old ground division, airplane and manpower simpli-
cities.
-- Operations analysis and gaming requires greater de-
tail on all parts of the theater forces.
-- Comparative East-West force tabulations and studies
need variations from traditional data breakdowns.
-- MBFR focuses attention on standing forces, especially
manpower.
top level policymakers.
2. Consumers have come to include, in addition to military
planners, a large array of politico-military groups including
-- The Secretary of Defense and the OSD interest in
forces data reflected heightened awareness of the
relative importance of tactical nuclear and con-
ventional forces for the defense of Europe in the
context of strategic nuclear standoff.
-- The Assistant for National Security Affairs and
the Secretary of State were brought into the data
picture by the necessity to evaluate relative US
and Allied contributions to NATO because of:
- Congressional pressure to reduce US-forces;
- the need for NATO force improvements, and
- movement toward mutual East-West force re-
ductions.
3. CIA/ORR-OSR came to duplicate in part the data production
of DIA and to furnish data directly to consumers.
TOP SECRET
 Approved For Release 2004/03/31 : Cl - 82A000800030002-3 25
TOP SECRET
-- DIA remained wedded to the data and data formats
favored by military planners, particularly tactical-
logistic offices.
-- CIA, seeking to accommodate the wider range of
consumers, questioned not only DIA data but use of
evidence from which data were produced and methodolo-
gies used to make estimates.
-- US data vs. NATO 161 data became an increasingly
important issue in US policy debates with NATO
countries, aggravating internal US differences.
-- MBFR forced previously unattained cooperation and
coordination between CIA and DIA while generating
a new set of NATO data outside the MC-161 context.
-- Basic differences between CIA and DIA in outlook,
methods, response and attitudes toward innovation
survive and will outlast MBFR.
4. Suggestions to relieve the burden of data production in-
clude decentralization of part of the work to field and centrali-
zation of the work in Washington in one agency.
25
TOP SECRET
Approved For Release 2004/03/31 : CIA
25
 Approved For Release 2004/03/31 :
TOP SECREC
M PqRFISAMMAJ
r-r-tur'OVIVIV 2A000800030002-3
- Estimates are focusing on current policy problems,
and both data and analysis must be supportive in
both scope and depth.
-- Complete centralization would deprive either the DCI
or the DOD of direct data support -- an unthinkable
situation.
- It would also remove valuable competition in
analysis.
- It would deprive consumers of alternate views
and enlarged perspective.
5. Even though radical structural change is not an answer,
new approaches to the production of forces data, or bean counting,
are required if tight resources are to be used efficiently and
if the products of new collection systems are not to be wasted.
-- The lack of unified, integrated concepts for
collecting, processing, analyzing and estimating
is wasting resources, causing the wrong kind of
duplication, and opening dangerous gaps.
-- The NIO function and the KIQ operation offer the
potential for getting the needed supra-agency
guidance and direction if allowed to carry through
into affecting agency operations.
TOP SECRET
Approved For Release 2004/03/31 : CI4
25
25
25
 Approved For Release 2004/03/31 : CIA-RDP80M01082A000800030002-3
TOP SECRET
7. Processing's interpretive function should be broadened
so that each of the major processors of technically-collected
information produces a data base for community use derived from
information yielded by his source.
-- This recommendation runs counter to past efforts
to restrict the processors to disseminating raw
information.
8. Analysis is strengthened by competition and contention
and provides the stuff of better estimates when encouraged into
independent innovation and exploration.
-- Duplication, per se, has become a target for
those aiming to increase efficiency.
-- In analysis, however, competition offers consumers
balance and perspective unavailable in products
from a single source.
-- The development of a strong common view of the
physical dimensions of a theater force as well as
its capabilities requires independent analysis and
judgment.
-- Complementary analytical teams have developed com-
plementary data storage and retrieval systems which
are providing the foundation for a "national" data
base.
9. Estimating is embracing a wide variety of "national"
products because of NSSMs and MBFR.
25
25
25
Approved For Reliaojp20 1E' ~-RDP80M01482A000800030002-3
 Approved For Release 00 WR : CIA A000800030002-3
TET
-- The estimating system which now taps analysts
as well as NIOs is providing timely support for
MBFR negotiations.
-- The variety of products and the current ad hoc
dissemination of them, however, needs monitoring
by a guiding authority to see that a "national"
position is identifiable, and is the result of
a fair and balanced coordinating system, and that
dissemination is proper.
TOP SECRET
Approved For Release 2004/03/31 : CIAI-RDP80MO108
25
25
 Approved For Release TOPO SEe//E A000800030002-3
T 7
25
A. What are the Beans?
There has been a considerable evolution in the quantity and
types of data required of the bean counters. Twenty years ago
interest in data on theater ground and air forces on the part of
higher levels of government did not extend much beyond numbers
of ground divisions, numbers of airplanes, and gross numbers of
men. General requirements for data most often could be satisfied
by recourse to tables in the National Intelligence. Survey, and
the accompanying text was taken as the word on other military
resources and on capabilities. The ground and air sections were
written by the intelligence components of the army and air force,
neither of whom felt any great pressure from critics. The result
was ritualized, stagnated methodologies, the details of which
were known to few outside order of battle shops. Changes occurred
infrequently. New evidence was allowed a considerable period
of aging before acceptance. Such arguments over evidence as there
were occurred principally between ACSI, Army, and USAREUR, but
neither felt required to coordinate final judgments with the other.
The initiation of the 11-4 series of National Intelligence
Estimates in the mid-1950s began the evolution toward more compre-
hensive military estimating and began a continuing agony for the
military intelligence components. Not only did members of the
Board of National Estimates ask difficult questions about how
specific estimates were arrived at, but CIA began delving into
military detail through its work on military expenditures. And
as luck would have it, the Khrushchev era began in the USSR, resulting
in a rapid series of changes in the Soviet military posture. The
military estimating apparatus did not adapt easily to the new
situation. The general reaction was to fall back into defensive
positions rather than to advance to meet the challenges through
reexamination of methodologies and production of new forms of
intelligence. Pressures to improve estimates and from new develop-
ments were evaded to the maximum extent possible despite the
evident inadequacies of the existing set of data on the Warsaw Pact
theater forces.
TOP SECRET 25
Approved For Release 2004/03/31 : CIA A000800030002-3
 Approved For Release 2004/03/31: Cl - muroulV17 2A000800030002-3 25
TOP SECRET
A good example of how the system reacted was the refusal to
accept evidence that the Soviets were reducing military manpower
during the years 1954-1959. The Soviets announced cuts of over
two million men during this period, and good evidence was availa-
ble to show that considerable cuts were in fact being made. US
estimates of Soviet military manpower actually speculated that
there might be increases at one point during this period. The
main problem was that US estimates had failed to follow the in-
creases that the Soviets had made during the Korean war, so that
if the announced cuts or something approximating them were accepted,
the US estimates would have been obviously too low. So the
military estimators kept the level of 4.2 million men carried since
1948. The mounting evidence that reductions were being made be-
came increasingly embarassing so that by the mid-1960s most
military shops were ready to accept Khrushchev's figure
of
3.6
million for January 1, 1960. Even then, however, each
of
the
US services insisted the cuts had been made not in its
counterpart but in one of the other services.
Soviet
In 1963, Defense Secretary McNamara's impatience with unrespon-
sive estimating resulted in setting up a combined CIA-DIA task
force to reexamine the evidence on the number of Soviet divisions
and other field units and on the production and inventories of
ground force equipment. The study exposed the inflexibilities
of the estimating process. A good share of the problem lay with
the "experts". An expert was responsible for data on some small
substantive area and he alone made the estimate for that area.
Aggregations were the summation of a number of detailed estimates
by these experts. Challenging an aggregate was difficult because
that was the way the pieces added up even though the summation
might appear curious or even absurd. In fact, the experts were
often people newly assigned to intelligence or were old hands
wedded to some long standing methodology or particular: type of
evidence. For example, analysis of Soviet field post numbers
played a key role in OB work long after it was clear that the
analysis was only marginally useful in formulating an OB. Another
example was the curious reluctance or outright resistance to the
acceptance of satellite photography as grade A evidence. Although
early photography lacked detail, its usefulness seemed obvious to
all but some of the OB analysts. Nonetheless, because of managerial
decisions first by ACSI (Army) and then by DIA its predecessor in
the OB business, there was no process for getting overhead photo-
graphy into OB shops.
25
TOP SECRET
Approved For Release 2004/03/31 : CI 2A000800030002-3
 Approved For Release 2004/03/31 : Cl - 82A000800030002-3
TOP SECRET 1"B"Oeme 11
The introduction of systems analysis into military analysis
also helped stimulate the era of sophisticated intelligence
assessment of the Pact theater forces. The systems analysts
brought into OSD proved to be no respectors of experts. In
addition to promoting new methodologies, they dug into the
evidence and analysis behind "published data if the data didn't
look right. While the systems analysts also had some preconceptions
of what the analysis should show, their probings stimulated better
developed estimates as well as fostering long overdue consistency
between estimates. The new approaches demonstrated the value of
arms-length, broad scale analysis along with the dangers of not
being able to see the army for the tanks on the part of narrowly
focused experts. The successors to McNamara's systems analysts
have continued the valuable role of critics as they have moved
into various gaming techniques. The games require lots of data
and whether the games themselves have been able to prove anything,
the: process spotlights soft spots remaining in theater force
estimating, while keeping analysis probing far beyond the ground
division-aircraft superficies.
Other developments also have fostered breadth and depth of
analysis on theater forces. Two longer term factors are the
advances in weapon development and the front line role of the
East Europeans. The more recent, and most demanding, is the work
required to support MBFR analysis.
Weapons development since the early 1960s have given increased
importance to non-divisional elements because much of the advance-
ment has been in weapons not carried in divisions. Surface-to-
surface and surface-to-air missiles are the best examples, but
antitank missiles and rapid-firing mobile artillery pieces using
improved munitions are also of increasing importance. The latest
arrival is the antitank helicopters. The new weaponry has not
only prompted interest in combat support units but also in logistic
units because of the high dependence of"the new weapons systems
on fast resupply and maintenance. The overall increase in mechaniza-
tion and mobility has had the same effect for general transport and
POL handling units.
The East Europeans have had an independent front line role
since 1964 when a reorganization of Warsaw Pact front assignments
gave the Poles responsibility for the northern front and the Czechs
Approved For Release'2 00/ff373'I-`TA
 Approved For Release 00S4 (3,1 - 2A000800030002-3
the southern front. At the same time, the East Germans appeared
in exercises arrayed in wholly--East German field armies in the
central front area. Since that time the East Europeans have
acquired an array of weaponry similar to the Soviets giving them
an increasing potential for engaging in mobile ground warfare
and in aerial combat.
It has been MBFR, however, that has forced a quantum jump
in the depth of analysis and a new emphasis on the peacetime
Warsaw Pact forces. Earlier analysis was concentrated on the
forces which were expected to comprise the wartime threat with
the peacetime force of interest primarily in its role as a
mobilization base. Thus, until MBFR, detailed data on the standing
force as such were of only passing interest and were easy to
neglect. Curiously, the Soviet-',s"-*forces in Eastern Europe were
not far behind the peacetime East European forces in the race
for being the most neglected. Soviet forces in Eastern Europe
were ascribed an estimated wartime strength with comparatively
little specific attention given them. Although it might have
been expected that USAREUR and USAFE could be relied upon for
details, that proved to be misplaced confidence. Neither US nor
Allied military intelligence in Europe could make useful contribu-
tions to MBFR analysis because of their preoccupation with threat-
oriented, and thus worst case analysis.
MBFR caused a further upheaval. Manpower estimating has
always been secondary to counts of weapons and units, and that
effort placed on assessing manpower was concentrated on estimating
wartime strengths as a part of the wartime threat analysis. CIA
has always had a very difficult time getting decent estimates of
peacetime manpower for costing purposes and knew well the dangers
of relying too heavily on any of the available detail. The fact
is that few of the peacetime manning factors available were backed
up by any decent evidence. When MBFR analysis was getting under-
way in 1970-1971, intelligence continually cautioned NSC staffers,
ACDA, and ISA about developing reduction options which involved
manpower unless manpower was linked to units. A focus on manpower
per se was inevitable, however, because men are the one common
denominator all can agree upon. The result has been a continuing
trauma for both intelligence analysts and MBFR officials. The
latter have refused to believe that a lack of information rather
than neglect and ineptitude continue to be responsible for the
situation. Earlier cautions about wide margins of error have been
forgotten. A crash effort to shape up the manpower estimates has
25
TOP SECRET 25
Approved For Release 2004/03/31 : CIA DP80M010 2A000800030002-3
 X1
Approved For Release 2004/03/31 : C
TOP SECRETI
25
absorbed a very large share of the theater force analysis time
in both CIA and DIA for the past two years 25
Pressure will
continue, however, as negotiations progress toward the specifics
of reductions.
Conclusion or cancellation of MBFR negotiations will not,
however, relieve the pressure for better and more comprehensive
data on the theater forces. Efforts to develop useful comparative
measures of theater force capabilities, largely unsuccessful so
far, will go on and will represent a continuing requirement for
better data. Other consumers, increasingly aware of the complexities
of modern theater forces will not be satisfied with data sets
restricted to a few types of front line units or items of equip-
ment.
B. Who Wants to Know?
There has been a constant increase in the demand for intelligence
on the Warsaw Pact theater forces in the past 15 years. The change
began with McNamara's discovery of how much of his budget went for
conventional forces and with the growing realization':that strategic
nuclear standoff didn't necessarily preclude war in Europe. Prior
to that time, intelligence on theater forces was of interest to
few outside of a small circle of military planners. And there was
little in the requests for intelligence from the military-planners
to foster innovations in analysis. Indeed, changes were not
accepted readily. All is different now with the analysts under
constant pressure to develop new ways of measuring and comparing
theater force capabilities. And the pressure comes from the
highest civilian as well as military levels.
Although McNamara failed in his initial efforts to get the
other members of NATO to acknowledge that strong theater forces
were a worthwhile investment, he finally won the day when NATO
was persuaded to adopt the strategy of flexible response in 1967.
Actually the Europeans had long doubted that a conventional con-
flict with the USSR in Europe would trigger an automatic US stra-
tegic nuclear strike on the USSR. To admit such doubts publicly,
however, would have reopened questions about unilateral strengthen-
ing of European conventional forces. By agreeing to flexible
response, the Europeans got a commitment from the US to increase
both its conventional and tactical nuclear forces in Europe in re-
turn for greater European contributions. The new situation created
a greatly increased demand for intelligence on the conventional
and tactical nuclear forces of the Warsaw Pact.
Approved For Rel Tease 0 d cAEGl14
25
 Approved For ReleasseP20 4/0
R1 ~ ELj -RDP80M010 2A000800030002-3 25
NATO intelligence on the Warsaw Pact is codified in the
MC-161 document series intended as guidance for planning by
SACEUR. Despite the ostensible purpose of the document, it
has evolved to be essentially political in nature and has assumed
a character which best suits the needs of the European allies
for a source of inputs to their internal political military
budget processes. While there is good reason to record the
presence and nature of the threat posed by the Warsaw Pact for
these purposes, as happens in such exercises, an inflationary
factor works over time. To make matters more complicated,
each member desires to see the threat against himself described
in the largest possible terms. Thus, through the 1950s and
into the 1960s the series documented the Europeans' contention
that the overwhelming capabilities of the Soviet theater forces
made efforts on the part of NATO to build a theater counterforce
hopeless and a waste of resources. Although the threat described
by the 161 series in the 1950s was not out of line, there was
difficulty in acknowledging changes in the Soviet force posture
and strategy which took place in the late 1950s and early 1960s.
The US reanalysis of Soviet capabilities which started in
1963 had little impact on the 161 series. The basic reason for this
was the political nature of the document, but the US did not make
a sustained effort to introduce judgments reflecting new evidence
and reanalysis. MC-161 is not used by the US in its national
planning and the content has had little direct impact here.
Essentially, it was more politic to let the Europeans have their
threat while the US used its own judgments in making its
estimates.
When it became important to the US to bring NATO's view of
the threat more into line with the current evidence, there was
resistance. The Europeans were determined to have little to do
with an exercise they saw as being essentially self-serving to
the US. Missionary work done in Europe by land
others was counterproductive, with the US accused of cooking the
books to support its campaign to get the Europeans to spend more
on conventional forces. There has been progress since that time,
but it has been painful and distressingly slow.
The basic problem remains the political nature of MC-161, and
the US hesitance to make an issue of change. Some of those who
have led the US delegations have worked to introduce more reality
into MC-161 but not only have they not had strong back-up from
25
25
Approved For Releasee OO T(f f3~ECIA[RDP80M010812A000800030002-3
 Approved For Rel
so~0~f ( 1 J1 - 2A000800030002-3
To F 1 i
Washington, they have not had good teams in the working sessions.
US representatives in the working groups have all been from DIA
because the drafting of a military committee document is almost
entirely a JCS show. Even those in sympathy with the changes
made in US estimates have been a poor match in maneuvering against
the Europeans because of political naivete and a lack of sub-
stantive background. Regardless of the path the US chooses to
follow in the future, it should take steps to select working
members who understand the relative importance of contentious
issues, have depth in substance, and can handle debates with the
Europeans.
Having to approve as accurate a set of data as possible for
MBFR may be bringing the Europeans around on the MC-161 issue,
at least on the numbers. Data introduced for MBFR are being
debated on their merits in Brussels and relayed on to Vienna.
For the most part these data are as the US would have them,
,,,reflecting all-source evidence available to the US. With
approval having been given to the MBFR numbers, it might be
hard for any national representatives to support different data
for MC-161.
A related problem which won't be taken up here but underlies
some of the differences between the US and the continental NATO
allies in both MC-161 and in MBFR is the sharing of evidence.
As analyses of Warsaw Pact forces become more critical to national
decision making within NATO, the pressure on the US to prove its
case by sharing the output and/or evidence derived from its
technical collection will increase. The recent move to make a
partial exposure to the FRG will not help much and may make
matters worse. The Germans will soon get impatient with their
meager fare and the other Europeans will be jealous and petulant.
Some other action will be required within the NATO framework.
When action is decided on, it should be controlled and operated
from the national level. If it should be kept within strictly
military channels problems similar to those encountered in
MC-161 over the years will surely arise. In other words, the US
must have a national representation which can accommodate both
the military and political ramifications of such an operation.
25
25
TOP SECRET
Approved For Release 2004/03/31 : CIA
25
 Approved For ReTOe?2 O9ECRE I
25
25
MBFR options analyses and negotiations top the priority list
in requirements for extensive detail on the ground and tactical
air forces. In this case the counts of beans are needed for
themselves rather than as inputs to further analyses. The ina-
bility of intelligence to supply requested counts of men, units
and equipment within narrow ranges of uncertainty has made MBFR
planners and negotiators severe critics of the capabilities of
the intelligence community. Although some of the criticism has
been the result of (possibly) willful ignorance of intelligence
collection, processing and analysis, it has stimulated intel-
ligence producers to the best estimates they have ever made.
There is a danger, however, that under the MBFR goad, analysts
can be driven to refine estimates well beyond the level of detail
warranted by the available evidence. The pressure to drive for
detail also means that each new piece of evidence requires a
re-tweaking of the estimate. This results in further cries from
the MBFR consumer that intelligence doesn't know what it is
doing and is unable to get even a simple count right. As it
happens, most of what the MBFR people want hasn't yet had a direct
effect on negotiations so that the margins of error which so
distress them have been unimportant. There is a payoff for intel-
ligence, however, because when the details of an agreement are
negotiated, intelligence will have been driven into a position
where it probably can support the requirements. It could not
have done so from the data base in hand before MBFR negotiations
started.
As noted at the beginning of this section the demands for
data on the Warsaw Pact have increased remarkably. A common
characteristic of much of the new demand both in the US and
abroad is that the data are for use in comparisons with U.S or NATO
TOP SECRET
Approved For Release 2004/03/31 : Cl
25
 Approved For Relpfle090WH1 -
forces, from simple tabular comparisons to complex, dynamic
gaming. Because of a lack of commonality between force compo-
nents, deriving useable definitions for purposes of comparison
has put a new burden on analysts. Their knowledge not only of
the numbers and composition of force elements but of how they
are used'has increased in importance. And because comparison
is the object, new knowledge about US and NATO forces must be
acquired.
C. Who Counts the Beans Now?
The bulk of bean counting for publication as such -- orders
of battle, tables of organization and equipment, equipment inven-
tories, and logistic guides and tables -- is done by DIA. The
workload is formidable, resulting from a set of requirements
which seemingly are subject only to addition, not subtraction.
DIA's consumers for these products are primarily military and
include military components all over the world. Some of the
latter also publish local orders of battle, usually based on
the DIA publications, but sometimes they introduce analysis of
their own in the form of supporting detail.
Other bean counts are issued by DIA in answer to specific
requirements such as contributions to NIEs, MBFR analysis,
OSD/SA (ODDPA&E) studies and as parts of studies of some aspect
of Warsaw Pact capabilities. CIA/OSR also issues bean counts for
these purposes and it is here that conflict often arises. Other
organizations like NSA or NPIC disseminating counts are recog-
nized to be single source organizations and while differences of
opinion and argument result, the clash is different than between
the producers of all-source finished intelligence.
Initial substantive differences between CIA and DIA probably
are unavoidable. DIA's contributions to meet specific national
requirements, not surprisingly, are usually kept consistent with
their most recent published listings. OSR's contributions on:the
other hand usually have been made for the immediate purpose, and
OSR does not pay much heed to previous listings if new evidence
is in hand or a new methodology has resulted in new figures.
Institutionally, DIA must observe continuing commitments to military
planners while OSR is free to make those changes it feels are
justified to give the consumer the benefit of the latest informa-
tion and analysis and for his specific need and purpose.
TOP SECRET
Approved For Release 2004/03/31 : CI 2A000800030002-3
25
25
 Approved For ReIq 20 1ETl -RDP80M01 p82A000800030002-3 25
Substantive differences growing out of varying interpretations
of evidence are, for the most part, reconcilable. Dissemination
of information by collectors is now universal and complete so
that only minor irritations of timing of receipt remain. Thus
both agencies are working from the same raw material and when
compelled by circumstances to coordinate output, most substantive
differences can be reconciled or can be subsumed even though
different forms of interpretation have been used. The considerable
time required to do_..this, however, could be reduced if common data
bases were at hand per the recommendations made below.
Other differences in counting relate to definitional varia-
tions or to the results of application of different analytic
methodologies. These are not so easily eliminated and constitute'
the basis for some long standing conflicts. The fact that DIA is
under military control and uses mostly military personnel and OSR
is under civilian control with mostly civilian analysts only
partly explains the situation. As noted above, institutional
responsibilities are quite different. There are also, however,
significant differences in size, in access to support such as ref-
erence and computer services, and in management philosophy. As
now organized, DIA could not do what OSR does but neither could
OSR do DIA's job. Thus, more than 10 years of coexistence.
DIA has made substantial progress in the past few years. It
has been able to acquire better officers motivated toward intel-
ligence assignments. It turns out better basic data and has
shown more flekibility in its approaches to analysis. Compared
to OSR, however, DIA's response time remains slow, and its ability
to work out approaches to new problems is limited. DIA remains
best organized and substantively prepared to meet military-type
requirements for finished intelligence. Requests, particularly
from civilian policy makers, which depart from the customary,
stylized needs of military planners are much harder for the DIA
system to meet satisfactorily. Frank statements to this effect
by consumers of DIA's products continue to this day.
There are some good reasons for DIA's continuing problems.
It has too many masters with OSD, the joint chiefs, the services
and the major field commanders all able to crowd its docket with
projects. It is hobbled by a severe briefing load which absorbs
the time of many of its best people. The burden of publishing
OBs, TOES, guides, manuals and handbooks continues. And perhaps,
most importantly, its personnel system is still saddled with many
officers in important and supervisory positions doing their first
TOP SECRET
Approved For Release 2004/03/31 : UA-KUVt$UMU 82A000800030002-3
25
 Approved For D- e1MI) 1W ErjC
082A000800030002-3 25
tours in intelligence. Given the situation it is not surprising
that DIA responses so frequently have an "off the shelf" feel to
them. If DIA were to be the sole producer of basic data for
the Warsaw Pact theater forces, Washington consumers would have
to become accustomed to more scissors-and-paste estimating.
Although OSR-DIA relationships had both smooth and rough
patches over the years, relationships stayed basically distant
and antagonistic until MBFR negotiations got under way in Vienna.
A considerable mess occurred during the first negotiating session
with OSR, DIA and the DOD MBFR Task Force all supplying the
delegation with data, much of it conflicting. The agonies of this
experience force formation of an ad hoc coordinating arrangement
which has resulted in unparalleled cooperation between OSR and
,DIA in data work. The agreed data now in hand were generated
in a fraction of the time which would have been required through
the usual coordinating routines. It is doubtful, however, that
the arrangement would survive if MBFR pressure were relieved.
The bureaucratic, professional and personal differences are great.
A recommendation on how to institutionalize cooperation and
coordination on data work under the aegis of the NIO is made below.
TOP SECREpproved For Release 2004/03/31 : Cl CI
82A000800030002-3
25
 Approved For Re" 2803 IA-RDP80M0
D. How are the Beans Counted?
Bean counting is, with rare exception, an arduous, unending
process of fitting together many bits and pieces of all-source
information. The keys to good counts are continuity and com-
prehensiveness in filing systems and the analysts' understanding
of the behavior patterns of the forces with which he is dealing
as well as an understanding of the strengths and weaknesses of
the sources he has at his disposal. Most analysts find the work
boring as a steady diet and unrewarding because their small
triumphs seldom evoke high level interest. Changes are as apt
to bring criticism as praise because changes upset consumers'
calculations. There is a general reaction that the analyst has
discovered something he should have already known and accounted for.
Consumers rarely appreciate that despite the fact that military
beans are finite and quantifiable a large measure of judgment is
required in evaluating and appraising the inflow of information.
A lot of incomplete or misleading information comes across every
.analyst's desk.
Because of the nature of the function there are all too many
plodding, unimaginative analysts who are happy at the work aided
and abetted by good analysts who are all too frequently rotated
through the bean counting positions. The latter have seldom been
around long enough to acquire the necessary understanding of the
data and forces to do a satisfactory job. Fortunately computer
storage and retrieval systems can now go a long way toward
supplying a good analyst with the history and force behavior pat-
terns he needs, as well as making source comparison easier and
more thorough. This offers management the alternative of giving
good analysts bean counting responsibilities in addition to other
more stimulating tasks.
TOP SECRET
Approved For Release 2004/03/31: C
2A000800030002-3
25
25
25
 25X1 Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
Next 1 Page(s) In Document Exempt
Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
 Approved For Release 2004/03/31 : CI 082A000800030002-3
TOP SECRET 25
E. Who Should Count the Beans?
The thoroughly military nature of the work, the duplication
of effort, and the increasing drain on resources represented by
bean counting have given rise to proposals either to decentralize
the work or to rationalize efforts at the center. It seems
unarguable that the present system which is the result of a series
of disconnected responses to developments over time cannot be
improved upon. Flaws and inefficiencies are easy to point out,
and disgruntled consumers are not hard to find. Before tinkering,
however, architects of a new system should not only remember
that the old system works, contains useful checks and balances,
and has learned a good deal of late about how to serve high level
consumers. There are also some physical and bureaucratic realities
which have to be taken into account.
If there were decentralization, it would be a short time
before duplication would reoccur with Washington agencies finding
that they could not do the analysis required of them without the
ability to directly control the production and formulation of the
basic data they need. They would feel the same forces that origi-
nally put ORR-OSR into bean counting. Competence and bias aside,
it is a truism that no numerical series in hand ever quite fits
the next task. Realignment and reinterpretation are always needed,
but they cannot be done without complete knowledge of the evidence
behind the basic figures and the definitions used to aggregate
them.
There are other strong reasons for not decentralizing. The
success which analytical shops have in getting the kind of col-
lection, processing and support they need is in direct proportion
to their proximity to the collectors, processors, and reference
and computer people. These elements are of course all located in
Washington and fast, sympathetic response from them depends on
continuing personal contact. It is most critical in the case
of the collection and processing of satellite photography where
everything is geared to mission schedules. Quick contact will be
even more critical with the new systems. Any of the analysts who
were involved in the crash effort for MBFR to shape up the estimates
on the Pact forces in the forward area will testify to the respon-
siveness of both the collectors and processors when the problem
was explained to them and to the cooperation given in working out
and monitoring the effectiveness of the effort. The monitoring
of progress was the key to the all-important follow-through to
cover hard-to-get areas.
Approved For RelTal h3R1ET14
25
 Approved For Release 2004/03/31 : Cl fk-RDPROMO 82A000800030002-3
TOP SECRET
Separation from the other elements in the chain would strike
most analysts as unthinkable. Forces analysis, including bean
counting, is a vertical process involving continuing interplay
with collectors, processors, and consumers. Disconnection of
any one of these would reduce-responsiveness, encourage errors
of interpretation and lengthen the time required to do a project.
It would also preclude the use of links between machine systems,
a step back into the dark ages of ledgers, file cards and calcula-
tors.
There are two excellent examples of what happens when proces-
sing and analysis for forces in specified geographical areas
is separated from that done for other areas. The problems en-
countered would be likely to occur if the separation were made
on other bases than geography because geography as such was not
a problem -- in either example.
25
Two 1 r Z ojects involving CIA, DIA 25
were set up to share the establishment
of new common data bases on ground forces -- one on the USSR
and the other on China. All parties entered the project with
the highest resolve to achieve consistency in their products,
shared confidence in their judgments and common rates of progress.
The projects had the backing of senior officials of each parti-
cipating organization. Preliminary meetings were held to ensure
consistency in criteria, standards of evaluation, and commonality
in reporting. The military districts of the USSR and the mili-
tary regions of China were divided up between the participants
wherever possible to take advantage of any unique capability
or special interest. The studies were to be all-source but would
be based on satellite photography furnished by the US which under-
took to ensure that the collection requirements of all were looked
after. The project on China had the further advantage of experience
gained in establishing the project on the USSR. Annual con-
ferences were held to review performances. In sum, it would be
hard to visualize projects undertaken under more favorable circum-
stances.
Although much useful work was accomplished, both projects
failed to achieve their principal objectives: acceptable inputs
to common data bases. The bulk of the work has been redone or
remains to be redone. Despite the measures taken to guard against
diverse criteria, reporting, etc., each participant in fact sailed
a different course which in the end made his reports unus.eable by
Approved For Reif 2
II .E:7El
25
 X1
Approved For Release 2004/03/31 : C A-RQP80M0j 082A000800030002-3
TOP SECRET
25
the others without reworking. The absence of a single controlling
authority and of continued contact, and the presence of different
biases and analytical interests caused reports to be written which
not only were not interchangeable but were not trusted by those
who did not write them. Few schedules were met, often because
local priorities got in the way. Sadly, differences between CIA
and DIA reports caused n ly as much difficulty 25
The lesson is that the benefits of
division of effort cannot make up for the loss of control and
common perspective in work so complex and so dependent on human
performance.
The most important factors driving the bean counting process
today all militate toward greater centralization, not dispersion.
A shift of the work, or any significant part of it to organizations
outside the Washington area, would result in less relevant collection,
less responsive processing, a loss of control over analytical
methods, and a diminished capability to accommodate the policy-
making consumers.
Complete centralization, unifying analytical responsibilities
in one organization, not only would deprive either the DCI or the
DOD of immediate support on theater force questions -- an unthink-
able situation -- but also would involve losses in two substantive
respects, losses so significant that they alone provide reason
enough for not centralizing. One loss would be in competition.
The benefits derived from having competing organizations are hard
to overstate. Competition between the agencies brings an outside
challenge to each -- a challenge from another group which also has
access to the information but has a different routine of analysis
and a different outlook. Neither can afford to be asleep or to
be indifferent.as long as such a.challenge exists. As in the
production of economic goods and services, the routine prime mover
in product improvement, willingness to serve and devotion to duty
in intelligence is competition, both personal and organizational.
Competition, even if it is no more than bouncing ideas off one
another in meetings in conferences, stimulates action. Obviously,
competition can get out of hand and it has in the past in dealings
between ORR-OSR and DIA. It has led to thievery, deceit and
double-dealing. Both organizations have matured, however, and the
confrontation of problems larger than both of them such as MBFR
has encouraged cooperation which probably would have been unat-
tainable in earlier times.
Approved For Rely 2 ORE' l
82A000800030002-3 25
 Approved For Release 2004/03/31 : CIA-RDP80M01082A000800030002-3
TOP SECRET
The other loss which would come from centralization would
be the absence of the checks and balances provided by the
participation of organizations with offsetting biases. Several
references to military bias have been made earlier. OSR has
its bias too, principally impatience with certain military realities
which OSR considers illogical. Whether one set of'biases is
equal in weight to the weight of the other is less important than
the fact that a counterweight exists. The availability to the
consumer of more than one view provides him with perspective
otherwise unobtainable. Redundancy in analysis is worth the
price.
25
25
Approved For Rele fb0$E,f RF jRDP80M01182A000800030002-3
 Approved For Release 2004/03/31 : CIA-RDP80M01082A000800030002-3
TOP SECRET
The foregoing discussion has focused primarily on
problems related to the analytical function in the produc-
tion of forces data. Confusion, duplication, and par-
tisanship are most apparent in analysis, and it is the
target for most suggestions on how to improve production
of data. As was pointed out earlier, however, the pro-
duction of data is a vertical system closely linking
collection, processing, analysis, and estimating (Chart
1). Those intimately acquainted with the system know
that there are problems throughout and that some of the
blame aimed at analysis had better be put on the system
itself or another part of the system. Recommended ac-
tions, therefore, cover both the system as a whole and
each of the major functions.
A. The System
As industrial manager familiar with the operations
of a vertically integrated industry would find incompre-
ensible the fact that there has been no overall direct-
ing and coordinating authority for the major substantive
intelligence activities. He would view the system of
independent operators in the production of forces data
as an invitation to waste and inefficiency at best and
as risk for bankruptcy at worst. And he would be right.
This history of the production of forces data by
the intelligence community is analogous to that of a
family cottage industry that grew into a multimillion
dollar concern but kept the old intuitive management
concepts. It would make an excellent case study for a
business management course. The collectors have worked
against requirements not formally intercoordinated be-
tween collection systems and in many cases are uncon-
scious of the intelligence problems driving the analysts
and estimators. The processors are for the most part
captives of developments in the processing systems as
such and remain independent or semi-independent, jealous
of their mystiques and of each other. The analytical
Approved For Rele OO0$ECjZEg.`R-PDP80M010> 2A000800030002-3
25
25
 25X1 Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
 Approved For Release 2004/03/31 : CIA- 2A000800030002-3
TOP SECRET
shops are on the one hand too hidebound, and on the
other over eager to chase interesting, but possible ir-
relevant, avenues of analysis. The estimators respon-
sible for NIEs haven't been sure how to satisfy consumers
while analytical shops have ground out a stream of
bucket-shop estimates for NSSMs and MBFR.
The NIO function and the KIQ operation offer the
potential for getting the badly needed supra-agency
guidance and direction. The systems needs substantive
orientation toward key national intelligence issues
with a corresponding reduction in the ability of the
various fiefdoms to chart their own courses whether
dictated by their views of consumer needs or by. what
they see as overriding operational considerations. If
the NIO can't take on the job of overseeing the entire
system, he might form a standing council empowered to
investigate and recommend.
The system obviously does not lend itself to finely
tuned orchestration, but on the other hand, each of the
parts could be better brought into line so they were
at least following the same objectives and their re-
spective roles were better defined relative to each
other. This is the place to start rather than with a
piecemeal centralization or specialization as has been
suggested for the analysis function.
25
There are several ways to increase efficiency in
the system as it is now comprised. Some components are
wasting effort because of misorientation or because some-
one else can do the job still better even though the
current operation may be doing a splendid job. Else-
where, duplication or overlapping is failing to yield
the benefits of competition because of non-complementary
analysis. Desirable duplication in one area or part of
the system has been extended into other parts which
should not be duplicative. Finally, a component which
can do several jobs well may be sub-optimizing by trying
to do all of them rather than shifting some of the load
to underemployed components.
Approved For Relea`F 04 R -
DP80M0108 A000800030002-3
25
 25X1 Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3
Next 17 Page(s) In Document Exempt
Approved For Release 2004/03/31 : CIA-RDP80MO1082A000800030002-3