ANALYSIS OF THE INTELLIGENCE COMMUNITY

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP79M00467A003100060004-8
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 17
• Document Creation Date: December 16, 2016
• Document Release Date: April 29, 2005
• Sequence Number: 4
• Publication Date: February 24, 1976
• Content Type: MF

Approved For Release 2005/05/23 : CIA-RDP79M00467A003100060004.8
h -//77?
'WASHINGTON
February 24, 1976
MEMORANDUM FOR JACK MARSH
PHIL BUCHEN
GEORGE BUSH/
BILL HYLAND
BOB ELLSWORTH
FROM: MIKE DUVAL
SUBJECT: ANALYSIS OF THE INTELLIGENCE
COMMUNITY
Attached is an article which will appear in the forthcoming
issue of Foreign Policy. While I disagree with some of its
conclusions, some of the analysis is well done and I think
it's worth reading.
13
l
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 J
:.,"I E_I-IGlNCL%pproved
SL171NG
TILE OPPORTUNITY
since in a style that reflects its origins. It
For Release 2005/05/23: 0Wf??1'7M004674k003a001M0QQ,4-4ociety
by Peter Szanton and Graham Allison
The revelations of the past two years pro-
vide a rare and important o_iportunity: to
rethink and restructure the U.S. ''intelli-
gence community." There has been no com-
parable opportunity since I947-the year
the CIA was established-and much needs to
be done. But seizing the moment is not
without danger: the community is diverse
and complex, shrouded by secrecy, and still
poorly understood. Recent investigations
have not produced a comprehensive and bal-
anced assessmer t, nor are they likely to. In
at least one respect they may have hindered
necessary reforms by creating the red herring
of a rogue elephant. The notion that the
community's covert operations ran wild. out
of the control of responsible elective leader-
ship. does not square with the evidence: the
fact is that elective leadership in the execu-
tive branch acquiesced or insisted on those
operations, and that the Congress displayed
a decided preference for ignorance.
The United States needs effective intelli-
gence. It also needs a government that abides
by the moral standards essential to a free
society. Assuring both is difficult, and sort-
ing out what the difficulties are, and what
they aren't, is an essential precondition.
The evidence now appearing about the
U.S. intelligence community is not complete,
but it reveals a great deal. It suggests that
the community-the CIA in particular-has
probably performed its assigned functions
more effectively than any other major for-
eign affairs bureaucracy in Washington. Yet
it shoe: s important failures within that
community and dangerous failures outside
it. It shows a community organized at the
height of the cold war. and operating ever
generally: er.ci., .1c, rich, technically master-
ful, fascinated with means. and forgetful of
ends. It shows data collected by sophisticat-
ed methods, and assessed by primitive ones.
It shows biased analyses used to support
policy, and sound analyses ignored. It shows
old questions answered repeatedly, and new
problems neglected. It shows agencies urged
to illegal acts by their political superiors. It
shows willful somnolence in the Congress.
Our analysis of that evidence leads to a
number of proposals. Chief among them are
to:
> disassemble CIA and assign its two princi-
pal functions to separate organizations:
> place the central analytic and estimating
responsibilities in a ncw agency, organized
and staffed to perform those functions solely
and divorced from all clandestine activity:
> create greater competition in the analysis
of intelligence by expanding the number and
improving the quality of analytic starts at-
tached to major intelligence consumers:
> cut back on both clandestine collection of
intelligence and covert action, and make
both subject to far clearer rules and tighter
control;
> assign central many emene of the intelli-
gence community, divided as it will continue
to be among many departments, to a presi-
dential assistant:
> face squarely the conflict between legiti-
mate needs for intelligence and covert action
on the one hand, and constitutional rights
and social values on the other. by establish-
ing processes of checks and balances for au-
thorizing clandestine activity and assuring
its accountability.
How are those conclusions derived.' What
form might such changes take? We start at
the beginning.
What Do We Want of Intelligence?
The principal purp,-~c of foreign intelli-
gence is to provide information and analyses
useful to decision-makers.. quite secondary
183. 184.
Multilar,guoge Typographers AA-65
REVISED REPRO
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 purpose is to m.1kAL roved For ~~ease~1005/05/23
interve: n in ev~c s abro, le stew y
illumination of policy choices through the
detailed und_rstanding of political. econom-
ic, and military developments abroad is es-
sential to a successful foreign policy. Covert
action is useful far less broadly, and is rarely
essential. (Indeed, a case can be made for
forswearing covert action entirely. We dis-
cuss that position below, and, on balance.
reject it.) Those goals must be pursued sub-
ject to two constraints: that the methods
used remain consistent with fundamental
rights and social values: and that the costs
of these activities be commensurate with
their benefits.
Achieving these purposes over the past de-
cade has been particularly hard, for reasons
that will persist. The shifting nature of the
foreign policy agenda-the growing impor-
tance of economic issues. for example-im-
poses requirements for novel kinds of infor-
mation and unfamiliar metiiods of analysis.
The emergence of new forms of threat-
organized terrorism, for example-creates
new needs for covert capabilities just as pub-
lic attention is being drawn to the risks that
clandestine activity can pose to constitu-
tional rights and basic values. The erosion
of the cold war consensus as to U.S. pur-
poses and the means legitimate to advance
them leaves vulnerable a community whose
objectives, size. and operating methods still
largely assume that consensus.
The Community .
How are we now organized to meet those
demands and observe those constraints? The
intelligence community is composed of a
number of separate agencies. diverse in their
histories, lines of command, modes of op-
eration, and forms of responsibility. It com-
prises more than 100,000 persons and ex-
pends some S6 billion annually, counting
about S2 billion devoted by the military
to "tactical" intelligence.
> Central Intelligence Agency. At the cen-
ter of the community is the diversified con-
Mullilr2nguage Typographers AA-66
jt:~60llA&467#~ '69666426ute in
ex-
ecutive office of the president. reporting to
the National Security Council (NSC), the
CIA was intended to provide the president
and his principal foreign policy officials with
independent, timely, and reliable analyses of
important national security issues. The chief
argument for creating an agency separate
from State. Defense. and the armed services
was that the president needed estimates and
analyses undistorted by the policy prefer-
ences and operational responsibilities that
colored the conclusions of both the Depart-
ment of State and the services. In addition,
the CIA was to perform "services of com-
mon concern"; to recommend methods for
intelligence coordination: and to perform
"such other functions and duties related to
intelligence affecting the national security
as the NSC may from time to time direct."
That final ambiguous clause has been taken
-and was probably intended-as author-
ity for the agency's covert actions.
To perform its assigned tasks. the CIA
now deploys roughly 15,000 persons and
expends some 5750 million annually. Very
roughly half that suns supports an exten-
sive network of clandestine agents and oper-
ations. Roughly a third is devoted to the
interpretation of- data, and the preparation
of analyses and estimates. Most of the re-
mainder supports a resourceful technical de-
sign and engineering arm. Each of the ca-
reer intelligence officers to have headed the
agency-Dulles. Helms, Colby-has previ-
ously directed clandestine operations.
In addition to serving as the nation's chief
substantive intelligence officer (his briefings
provide the customary opening of NSC meet-
ings) , the director of the CIA (the DCI) has
two distinct obligations. One is to command
the CIA. The other is to coordinate the ac-
tivities of the entire community. Predictably,
DCIs have had far greater success in the first
of those roles. \Vith respect to the CIA, their
authority is clear: controlling budget, pro-
motions, and assignments, they can direct
84-856-F.P. 22 (10; l 2 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467A003100060004-8
 the agency. %Virh respect to tt - ~~ u' r com- ) Dc-
municy hove c~;r. Alpproueld For F e~eases2005/05/23 : Cfh-f N, 60461f d39'0?d06000 8
Tense ntelli Agency. iDtAl reports to
Roughly 85 per cent of the coma unity lies the Joint Chiefs of Staff and the secretary
within the Dcfcnse Department. and the sec- of defense. It was established in 1961, prin-
retary of defense is a statutory member of the eipally to reduce interservice duplication and
body-the NSC-to which the Dr[ reports. disagreement by performing centrally much
As a result, the community has never been of the work then conducted separately by
effectively, centrally managed. the service intelligence branches. Its success
> Natioral Security ,Agency. In terms of in that mission has been limited. Each of
the number of people employed-reported- the service intelligence branches is now larger
ly more than 20,000 in the United States, than it was in 1961. The DtA collects lit-
with others manning some 2.000 overseas tle information, but publishes various in-
monitoring stations-the community's larg- telligence digests, performs analyses on a
est component is the National Security wide range of subjects, and represents the
Agency (NSA). Established by executive Joint Chiefs of Staff in community-wide
order in 1952 and lodged in the Defense analyses and estimates. Its performance is
Department, the NSA monitors and attempts constrained by two difficulties that it shares
to decode or analyze an enormous range of with the service arms. Intelligence assign-
foreign communications and other electronic ments have little promotion value in ser-
signals. It is also responsible for the security vice careers, and are generally- avoided by
of U.S. codes and communications. The promising officers. Intelligence assignments
NSA produces enormous masses of raw data. outside one's service. in particular. are viewed
> National Reconneiss::nceOr,ice. The larg- as dead ends.
est agency in terms of budget is Rhe Na- > Bureau of Int,'lliozrce and Research. The
tional Reconnaissance 01-ice (SRO), also State Department's Bureau of Intelligence
lodged in Defense. The NRO operates the and Research (INR) is the smallest b'v far
numerous "overhead" (principally satellite) of the foreign intelligence agencies, and the
reconnaissance programs for the communi- only one which engages in no collection
ty. working largely through the U.S. Air activities. But Foreign Service reports from
Force. Its products are medium-resolution posts abroad-not considered intelligence in
photographs of wide areas and high-resolu- the usual sense-arc probably the largest
Lion pictures of selected pclnts: these are and often the most important source of in-
useful to economic analysts and essential to formation on foreign political and econom-
those concerned with military and arms con- ic developments. I':R provides analyses to
trol issues. The NRO is subordinate to the State's principal officials and contributes
DCI and a deputy secretary of defense. to the national estimates made jointly by
Army, Nauy, and Air Force Intelligence. the intelligence community. Its budget of
Each of the armed services maintains its $8 million approximates one one-thousandth
own substantial intelligence organization. of the community resources. Like the ser-
Their combined staffs total some 50,000, vice intelligence arms, it is not viewed by
largely overseas. These are especially con- its department's professionals as a mainline
cerned with so-called "tactical intelligence," assignment.
the capabilities and disposition of their
Other Agencies. The intelligence units
counterpart forces in other countries. But of the FBI, the Treasury- Department, and
service staffs also participate in the produc- the Energy Research and Development Ad-
tion of national intelligence estimates and ministration also participate in the intelli-
functions, and maintain their own commu- gence community, contributing on matters
nications security arms. within their. jurisdictions.
187. 188.
Multilanguage Typographers AA-67 84-856-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 Community DirAfypto4@A Forl a ase 2005/05/23 : CIA- iIR79MO0467A00a1Qff06DOQ year,
congressional ,,?crsight, lodged in intelli-
Central direction of the community falls gence subcommittees of the Armed Services
to several committees chaired either by the
DCl or the assistant to the president for Na-
tional Security affairs.
> The U.S. Intelligence Board (turn) ,
chaired by the DCI and composed of all ma-
jor U.S. agencies with intelligence respon-
sibilities, works through various committees
to establish intelligence requirements and
priorities, to produce national intelligence
estimates, and to protect intelligence sources
and methods.
> The Intelligence Resources
Committee, chaired by the DCI
Advisory
and com-
posed of major intelligence agencies,
pro-
vides a forum for coordinating resource al-
location throughout the community. The
DCI's authority to shift resources within the
Defense Department budget, however, has
been quite limited.
> The Intelligence Committee of the NSC,
chaired by the assistant for National Secu-
rity Affairs. is intended to provide a forum
which major consumers of intelligence
can inform collectors and analysts
of their
interests and requirements. It has met in-
frequently and had little impact.
> The Forty Committee of the NSC. also
chaired by the assistant for National Secu-
rity Affairs, approves covert actions and
other high-ris'r, operations. It, too, has rare-
ly met and in recent years has largely pro-
vided pro forma approvals to recommen-
dations of its chairman.
Outside the community, the Office of
Management and Budget plays the key role
in reviewing agency budgets. On presiden-
tial request, it has also served as a source
of critiques of the community and of pro-
posals for restructuring. The president's
Foreign Intelligence Advisory Board is a
panel of distinguished citizens with a broad
oversight charter, but it meets infrequently
and has tended to focus on the targetting
of intelligence and on techniques of collec-
tion. It is served by a two-man staff.
flullilanguage Typographers ."+A-68
Committees, had
been similarly perfunctory.
The subcommittee chairmen have
to inquire deeply into activities
telligence community,
chosen not
of the in-
How Has the Community Performed?
No assessment of so diverse and shrouded
a community can be complete, or complete-
ly balanced. But enough is now known to
force a number of conclusions.
The capability of the community to col-
lect information by technical means is re-
markably good-in some respects almost
magical. That is of great significance. It
provides a hard basis for military-and to a
lesser extent for economic and diplomatic
-decisions of high importance. The SALT
agreements. for e%amnle, could not have
been concluded without independent verifi-
cation of the Soviet strategic posture. Sim-
ilarly, some analytic work specially in
the CIA-is of high quality. Yet the com-
munity's deficiencies are large. They fall in-
to three broad categories.
1. Inadequate analusis. Congressman
Pike's assertion that the intelligence agen-
cies would fail to warn of another Pearl
Harbor is almost certainly wrong, but short-
comings in the analyses and formal esti-
mates prepared by the community are real.
They include bias, irrelevance, and a judg-
mental rather than analytic orientation.
It is a law of bureaucratic behavior that
agencies with operating responsibility pro-
duce intelligence, analyses, and advice that
supports their own policies or programs.
The cause is not dishonesty-indeed, the
process may he barely conscious-but the
tendency is universal.
To counterbalance such tendencies, the
CIA was established as a neutral and inde-
pendent agency, having no important oper-
ational function, no responsibility for pol-
icy, and a direct link to the president. And
the CIA has, in fact, proved the most in-
84-856-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467A003100060004-8
 dependent and (AljSF1dvee# Fot` Rel sen2?05/05/23 : CbArfiWt7a9W04r371A9 19 QQQ4;
0 goy-
i
B
h
d
es.
agenc
ut t
e pro
ucts the commu-
nity as a whole have been biased by a num-
ber of factors. First, "national intelligence
estimates'' are a composite of judgments of
the CIA, the DIA. the INR. and the service
intelligence agencies. Compromises among
these perspectives often lead to estimates that
reflect an exaggerated, military-oriented view
"However accurate its infortuation
or prophetic its estimates, the in-
telii~crrce cottttnniiity roust cort-
fornt to acceptable standards of
contl uct."
of the threat. Second, agencies whose busi-
ness it is to collect information tend to be-
lieve that the capacity to collect more will
insure important results. As a result. some
estimates-as to our ability to monitor So-
viet adherence to arms control agreements,
for example-tend systematically to high
confidence in U.S. capabilities, provided
only that certain budget requirements are
met. Finally, in the 1960s, the CIA's own
clandestine functions evolved into large op-
erating programs. In Vietnam and Laos.
for example, the CIA provided support and
leadership for hill-tribe armies and man-
aged the attempt to identify and destroy
Vietcong leadership. Responsibility for any
such large-scale operation affects judgment
about its worth and effect. The DCI was
caught between his intelligence divisions'
quite pessimistic judgments and his operat-
ing arm's optimism not only about its own
programs but also about prospects for gen-
eral success in the war.
A second source of inadequate analysis
is simple irrelevance. Consumers of intelli-
gence, especially, at high levels, are often too
busy and sometimes too secretive to clearly
identify the issues on which analysis might
be most helpful. Compared to the perfor-
mance of other agencies, the CIA's record,
again, is relatively good. But like all other
erned largely ,,, the lav.'s of inertia. A de-
cade of O['EC, terrorism. devaluation, threats
to the ozone layer, the opening to China,
and wheat deals with the Soviets has dem-
onstrated that both threats and opportuni-
ties can take novel forms. Making intelli-
gence consistently useful requires clear and
early communication of revised priorities
from levels high enough to force response.
Even when appropriate questions are ad-
dressed and biases are absent, the quality of
analysis and estimates is often low. Though
sonic national intelligence estimates series are
of high quality, most deliver compromise
judgments in an ex cathedra fashion that
makes it next to impossible for policy-
makers to uncover the analytic basis for the
judgments offered, or to educate themselves
about the grounds for disagreement. Perfec-
tion is not the appropriate standard. Assess-
ments of complex situations will often he
mistaken: predictions must often he wrong.
But a better product would emerge from
agencies that trained their analysts with
care and rewarded them well: that main-
tained close links with sources of knowl-
edge outside the government: and that in-
vested significantly- in the refinement of an-
alytic techniques.
2. Unacceptable means. However accu-
rate its information or prrnhetic its esti-
mates, the intelligence community must
conform to acceptable standards of conduct.
It obviously has not done so. Over the
last 20 years, virtually all U.S. foreign in-
telligence agencies have been involved in
the surveillance of persons in the United
States having no relation to any foreign
power, or in monitoring the mail or tele-
phone and telegraph communications of
large numbers of U.S. citizens. The CIA
plotted-and perhaps effected-the deaths
of persons who had been adjudged guilty
of nothing, and with whose countries the
United States was not at war. The results
have been a blurring of the moral standards
that should distinguish the behavior of an
191. 192.
Multilanguage Typographers AA-69
I
84-856-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 oven societies I, and a hcmo:rnagc of con-
fidence in agencies previously held in gen-
eral respect. The ugliest truth is that these
were not the acts of a rogue bureaucracy.
They were not unauthorized. They were
ditected by the community's political supe-
tiors-including presidents.
Identifying villains and punishing them
may be necessary, but it will not be suffi-
cient. The roots of the problem remain.
They are basically three. First, neither law
hot tradition has established clear rules
of behavior; these will have to be specified.
Second, monitoring the observance of such
rules as exist has been left to interested par-
ties-parties doubly insulated, moreover, by
secrecy and the "deniability" of most covert
behavior. Presidents and cabinet members
cannot be permitted to press secret agencies
for difficult results without accepting 're-
Eponsibility for the measures such results
require. Finally, the only effective
such activities, the Congress. has
Until recently, to avoid implication
ing knowledge.
checa: on
managed,
by avoid-
3. Waste. The third problem is less crit-
ical. Limiting the costs of the intelligence
community and allocating its resources in
accordance with an integrated sense of na-
tional requirements have been objects of
pressure from the White House, and of per-
sonal concern to two DCIs over the past five
years. That pressure, and the partial suc-
cess of two institutional innovations-the
Intelligence Community Staff and the Intel-
ligence Resources Advisory Committee-
halve achieved a substantial cutback in in-
telligence personnel (chiefly from the NSA)
and a leveling of the budget. Moreover, in
an uncertain world, too large an intelligence
effort is preferable to one too small.
Still, important additional economies re-
main to be made. The community is com-
posed of many agencies funded through sep-
3tate budgets and pursuing overlap: ping as-
signments. It is tempted to large invest-
cessing tcchno,.)gics. And it gives little con-
sideration to the marginal value of addition-
al information. As a result. analytic work
(which is cheap) is underfinanced, while
collection, which is expensive, is performed
prodigally. It seems a reasonable guess that
of the $4 billion expended by the commu-
nityon "national" intelligence. at least $500
million could be saved without significant
loss.
The gains would not he merely mon-
etary. Large and activist intelligence ser-
vices inevitably press against the constraints
upon behavior which an open society must
impose. The ready availability of funds-
some unvouchercd-compounds the prob-
lem. Similarly, plentiful resources blur at-
tempts to focus on issues of high priority.
Tighter budget control would strengthen
both the observance of rules of behavior
and attention to high priority concerns.
What Is to Be Done?
The aim of organizational reform 'must
be to create institutions, processes, and in-
.eentives that together serve a number of par-
tially competing objectives:
> providing ''eyes and ears" of the high-
est capacity to Scquire timely and accurate
information about issues of inter_st:
> developing a "mind" capable of draw-
ing the most penetrating inferences about
the likelihood of future developments, and
so connected to the eyes and ears as to as-
sure timely access to all information:
> insuring independence and objectivity in
analyses and estimates-the mind must be
responsive to the questions of policy-makers,
but not to their preferred answers:
> assuring the mind's access to the pres-
ident and its application to policy:
> making the mind accessible and useful
to Congress:
> guaranteeing regard for constitutional
rights and social values: and
> providing relative efficiency.
193. 194.
Multilonguage Typographers AA-70 84-856-F.P. 22 (10112 Ben Bk) NIK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
ohon society, the setting of precedents that ments in exotic } r
increase risks toA. 0MVedlForcRalbalge 2005/05/23 : CIA-FJRot7c9 0046 ~b 1`11c, and d. ra.
[ c ectrcn and daproro-
 Many pr-noted reforms att.?-id to several
of these objccti,esAlaprpyAd,Forp~teta,S,1,2005/05/23
ers. Some proposals. for example. would
provide an analytic capability so indepen-
dent from the executive that its access to
the president or the relevance of its conclu-
sions to top policy-makers would suffer.
Here we propose, for discussion and debate,
a package of substantial reforms.
Splitting the CIA
The need to insure the neutrality and in-
dependence of advice to the president was
the main reason for creating the CIA. Un-
like the services and State, it was to be free
of operating responsibilities and a policy
role, and hence exempt from the commit-
ments and loyalties those responsibilities in
evitably create. The central analysts were
to share an organization with clandestine
operatives, but the advantages of keeping
those two functions together seemed sub-
stantial: collectors would keep the analysts
abreast of what was known: analysts would
keep collectors aware of what was needed.
Experience has proved otherwise. Dif-
ferences in style, temperament, lines of com-
mand, and requirements of secrecy have com-
partmentalized the two functions, limiting
communication between them. While the
advantages have proved slight, the disad-
vantages have been great.
The top position in the CIA has been
largely controlled by the clandestine side of
the agency, and clandestine work has in-
volved substantial programs. The result has
been a partial compromise of the agency's
freedom from institutional bias. one of the
very reasons for its existence. The surest
way to avoid such compromise in the fu-
ture is to split the CIA, placing the analysts
in a separate and autonomous organization.
That reason alone might not justify sur-
gery so radical. but it does not stand alone.
At least four other benefits would flow from
the separation of analytic and estimating
staffs from clandestine collection and covert
operations. First, in the study of many clues-
195.
Multilanguage Typographers AA-71
~7~1'0~~drblit~Mf~6'~f~b4'd in
t e eve opn;c; improved analytic meth-
ods, the most advanced work goes on not in
intelligence agencies, or anywhere in the
government, but in the academic commu-
nity and in "think tanks." During World
War 11 and the early years of the cold war.
the intelligence agencies drew heavily on
these resources. But as U.S. intelligence be-
came increasingly associated with disputed
policies and large-scale covert activities,
those connections became deeply strained.
They need to be restored, and a clear sepa-
ration of analysts from operators is prob-
ably an essential precondition. Second, such
a separation-placing the central analysts in
an organization focused wholly on produc-
ing studies, assessments. and estimates of the
highest quality-would also facilitate the re-
cruiting and training of superior analysts,
and promotion and reward practices better
adapted to retaining them. Third. splitting
the CIA, following revelations of clear mis-
conduct, would help insure future sensitiv-
ity in all U.S. clandestine services to the
importance of observing more demanding
standards of behavior. Finally, the disap-
pearance of the CIA would relieve us of a
name and an organization that would other-
wise remain a target at home and abroad.
A Foreign Assessment Agency
The analytic and estimative tasks of the
CIA would be assigned to a new entity or-
ganized and staffed solely for the purposes
of producing good analysis and providing
objective estimates. Its title might be the
Foreign Assessment Agency (FAA). While
this agency would be free to analyze any
issue of foreign intelligence its director
thought pertinent. its priorities would be set
at the NSC level under the direction of a
presidential assistant for Intelligence, as de-
scribed below. The agency would begin by
assuming most of the functions and person-
nel of the CIA's Directorate of Intelligence,
including key interpreters of photography
and other technically derived data, but it
84-856-F.P. 22 (10112 Ben Bk) NIK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 would be sh(-rn of 0 L
.A~fr ve For``# 646a'se1OO005/05/23 : CI P7bM004&Th O 400000041r uiring
any ccdlcetion its[ it wo -d seek good
h
working relations With university research
centers and think tanks. Indeed, it should
develop and sponsor one or more Rand-
like institutes specifically devoted to improv-
ing analytic and forecasting methods.
"lit spite Of the real risks i-tltcrent
ill tnaitttainiii covert caltahilitir .
the dan=ces of lackirt.t them in the
currcttt ti.orltl elivirounlcnt semi
larger."
The FAA would contain a Board of Na-
tional Estimates, staffed by senior analysts,
that would preside over the production of
estimates on subjects like U.S.-Soviet stra-
tegic balance, on which formal goverment-
wide consensus is agreed useful. As for sub-
jects on which formal agreement is not aD-
propriate or not possible. the board would
present its own assessments together with the
views of any differing agencies and full ex-
position of the nature of those differences.
The director of the FAA would rank. as
the community's senior "producer'' of anal-
yses and estimates, inheriting the DCI's cur-
rent role as intelligence adviser to the pres-
ident. He would have no responsibility for
clandestine collection or covert action, and
would not attempt to manage or direct the
community as a whole. His appoi lent
v c subic anon by the {` o
Dint Intelligcnce Committee of the Con-0: 1C
gress, proposed below.
The potential vulnerability of such an
agency is that in calling shots just as it
saw them it would make few.y friends: and
without substantial collection programs of
its own, or a supervisory responsibility for
the rest of the community, it might simply
be ignored. The main defense against this
would be the authoritative quality of its
analytic and estimative crk, backed by its
special capacities for -eting important
technical data. That -,se might be re-
Multilanguage Typographers AA-72
t
e NSC to core .,er FAA vice: s prior to Spec-
ified kinds of decisions. It Would be further
strengthened by the responsibility of the
FAA director to submit requested analyses
to the Joint Intelligence Committee of the
Congress.
Additional Intelligence Analysis Staffs
The FAA will not be able to meet all
day-to-day needs of senior officials for anal-
yses utilizing intelligence data. But there is
little reason why it should have to. Respon-
sibility for the production of analyses and
estimates, now highly centralized, can read-
ily be diffused. The collection of many
forms of information is complicated and
expensive, and therefore important not to
duplicate. But the analysis of information
shares none of those characteristics. It re-
quires relatively few people, little machin-
ery, and is useful to duplicate: nothing im-
proves the quality of analysis so power-
fully as the existence of competing sources
of analysis. Small analytic and estimating
staffs should therefore be assigned diicctly
to all key intelligence consumers who need
them. The additional costs of such staffs
would be trivial, but their benefits substan-
tial. They would give policy-level officials
an opportunity to pose the questions whose
answers would be most helpful to their own
work, and to receive these answers from
analysts whose performance they would be
able to reward or penalize. The associated
danger-that analysts so situated would
tend to produce the responses their bosses
found most congenial-would be largely
offset by the existence of competing staffs
elsewhere and by the retention of a well-in-
sulated central analytic agency.
A Special Services Agency
The case can be made that the United
States should abolish its clandestine agencies
and deny itself a capability for covert ac-
tion. This argument focuses on the risks
posed by failure or disclosure of covert ac-
84-856-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 bons and the dcc!A1ncovederFor6~`I taid1005/05/23
bilitict. It strcr,scs point such a cn-
eics have tended to infrintie constitutional
rights, and may main en .ige in actions,
like assassinations, that affront humane na-
tional values.
We find these arguments telling, but ul-
timately unpersuasive. In spite of the real
risks inherent in maintaining covert capa-
bilities, the dangers of lacking them in the
current world environment seem larger.
Consider the growing accessibility of the
materials and technology for manufactur-
ing crude nuclear weapons. It is frightening
enough if one considers only the possible
behavior of unstable states and irrational
rulers. But that is not the worst prespect:
A nuclear device entering the hands of the
Red Guerrilla Army or a national liberation
splinter group is not grossly improbable.
Supplementing the formal systems designed
to account for fissionable material should
surely become a high priority for the infor-
mation-gathering arms of U.S. intelligence.
And it would be a tragic peculiarity to for-
swear the capacity to act, if clearly neces-
sary, on what the intelligence disclosed.
So we judge the question "covert action:
yes or no?" badly posed. The appropriate
questions are: what ?rinds of covert action
may be justified, and under what circum-
stances: and how can the processes for their
approval assure restraint and accountabil-
ity?
We propose that needs for clandestine
collection of intelligence and for limited
covert action be met by a much slimmed
and substantially- restaffed agency incorpo-
rating elements and personnel of the CIA's
present Operations. Science. and Support
Directorates. A name like ''Special Services"
would emphasize the relatively narrow and
controlled nature of its mission. The proper
placement of such an agency is debatable;
in fact, any location has important defects.
But since it must be kept sensitive to the
foreign policy implications of its actions,
and also given some insulation from direct
Multilanguage Typographers AA-73
cD 1F~D 719MU046Td~ i03c1c0e0 0Qi4a ling it
an indcpendc;.. agency, reporting to the
president through the secretary of state, as
the Arms Control and Disarmament Aggcn-
cy does now. Its director should be held by
statute to a limited terra-perhaps six years.
It is essential that the process for approv-
ing and controlling covert actions and ether
high-risk clandestine activities be tightened.
The objective here is to make a secret pro-,
cess embody effective substitutes for the
forced consideration of various viewpoints
and the clear assignment of responsibility
that more public debate normally assures.
There must be clear accountability at all
points. The current process clearly fails in
this. It involves the submission of proposals
to the so-called Forty Committee, a sub-
committee of the NSC with narrow mem-
bership and-at least recently--extraordi-
narily informal procedures. Proposals have
typically been carried from department to
department, and committee members have
been polled by phone. Once approved, ac-
tions have rarely been evaluated or reviewed.
A number of changes are necessary. The
membership of the committee should be ex-
panded to include at least one person of
public stature with no other active associ-
ation with the administration. No covert
action or high-risk clandestine collection ac-
tivity should be author" ed except hr' writ-
ten presidential order given after consider-
ation of the action's risks and benefits by
all available committee members, and after
their signed recommendations have been re-
ceived by the president. Besides granting
initial approvals, the committee should be
required to regularly review the utility and
appropriateness of activities still being pur-
sued. It should be served in this work by
a staff drawn from outside the intelligence
agencies. And its actions should be subject
to the oversight of the Joint Committee of
the Congress, as discussed below.
To assure that collection and analysis are
targettcd on priority concerns, and that the
intelligence community's resources are ef-
84-356-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 fcetively dc;;lov cd,
the cornniur.it 's central nlana,ticr. Until
now, that resnonsihility has been under-
taken by the DC[. But a long series of stud-
ies has concluded that effective direction of
the community has never been achieved, for
two reasons. The first and forer-,lost, re-
ferred to earlier, is inadequate authority.
The great preponderance of the community
falls under the jurisdiction of the secretary
of defense, and the DCI is subordinate to the
NSC, of which the secretary of defense is a
statutory The second difficulty is
apparent bias. The DCI is viewed in the
community as necessarily partial to the in-
terests and perspectives of the organization
he heads: the CIA.
The long attempt to make the DCI re-
sponsible for community- vide direction has
not only largely failed, but its failure has
imposed an important cost. The DCI's devo-
tion to the development of a capacity for
analysis and estimates of the highest pos-
sible quality--the key responsibility of his
own agency-has inevitably been diluted.
Quite independent of our proposals for the
splitting of the CIA, therefore. we believe
that continuing to seek the direction of the
community from the DCI would be mis-
taken.
A Presidential Assistant for Intelligence
Who then should perform the task? We
believe that it can best be performed by a
special presidential assistant for intelligence,
for the fundamental reason that the pres-
ident is the only cf-racial to whom all agen-
cies in that disparate community report.
i The tasks of establishing consumer prior-
ities, assessing producer performance, and
developing budgets that allocate resources
across the community in accordance with
the national importance of the functions be-
ing performed can only be effectively carried
out by someone who speaks in the pres-
ident's name.
To help clarify consumer priorities, and
to specify the meaning of those priorities
Multilanguage Typographers AA-74
Approved: Foretceae 2005/05/23 f'opI Rppg7pq00467tAORAIIOOa0n60Ps9 i nc
should expand and activate the NSC's In-
telligence Committee, the little-used forum
designed to permit high-level consumers of
intelligence to regularly make producers
aware of current policy concerns and result-
ing intelligence priorities. Agriculture. Com-
merce, and Energy officials might appropri-
ately be added to its membership on at least
an ad hoc basis. The assistant should also
use the committee as a medium for review
of intelligence performance.
Beyond eliciting better specifications of
intelligence priorities and sharper assessments
of intelligence performance, the assistant
for intelligence should be responsible for
developing and defending a comprehensive
community-wide foreign intelligence bud-
get. The budget would be based on a clear
allocation of responsibilities among the pro-
ducing agencies, and would be subjec~ to
authorizing action by the Joint Committee
of the Congress on Intelligence discussed
below.
Setting Boundaries to Bch.auior
More demanding standards of behavior
must be set and enforced. That v.-11! require
dealing with the three main sources of past
failure: (1) Neither lank- nor tradition has
established clear rules of behavior: (2) mon-
itoring the obser: a::_e of such rules as exist
has been left to interested parties: and (3)
the only effective source of oversight. Con-
gress, has defaulted on its responsibilities.
What rules.' Centuries of reflection by
theologians, lawyers, and statesmen have
produced widely accepted rules of war. But
what are the rules of quasi-war' Is it proper
for the CIA to subsidize a foreign newspa-
per? To blackmail a foreign leader' To en-
courage assassins' Do the circumstances mat-
ter, and if so, hey' Is it permissible for
the NSA to intercept telephone or cable traf-
fic between foreigners and Americans' Be-
tween foreigners and American corpora-
tions? Is it avoidable' Amtorg, the Soviet
trade organization that performs espionage
84-856-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
 as "ell as eo n'Aop~ vedForcRelseaasL~'JF005/05/23 : CIS-RWW9M004f7d4,QQ 3$QQp6Q084c ,av of
) ork eorpor.stion. Does th. matter.'
None of the answers to these questions is
clear. but there is no way to hold men or
organizations resronsiblc for their behavior
unless rules of responsible behavior are made
known in advance. If we v: ish intelligence
agencies to accomplish difficult results in se-
cret and at the same time observe certain
standards, then we had better define the
standards, understand that we have issued
partially conflicting instructions. and pro-
vide some method for resolving the conflict.
What
adding other tn?rceptions and giving them
sonic weight. "1 he attorney general should be
made responsible. by statute. for determin-
ing whether actions with clear potential for
violating the rights of U.S. citizens can be
authorized. The statutes should specify civil
and criminal penalties applicable to viola-
tions of their terms. Finally, an indepen-
dent inspector general of intelligence should
be established. His appointment should be
subject to Senate confirmation, and his office
should be made responsible for reporting
any discovered violation of statute or exec-
utive order to both the president and the
Congress.
Active Congressional Or-,ersiaht
The proposed rules and procedures cover-
ing clandestine and covert activities would
begin to establish checks and balances against
abuse of power. But standing alone, they
could be short-circuited or ignored. In the
end, the only, effective check against exec-
utive abuse is the same for intelligence as
it is for other policy areas: an informed
Congress imposing effective oversight. We
have never had such oversight.
The span of intelligence responsibilities
has now broadened far beyond military
concerns, and the costs of con^ressional pas-
sivity are clear. It is time for genuine over-
sight, and that will require a congressional
body specifically concerned with intelligence,
capable of viewing the purposes of the com-
munity broadly and of assessing its perfor-
mance critically. Though it profits little for
outsiders to specify the nature of congres-
sional reform, such a body should almost
surely be a joint committee. on which the
leadership as well as the Foreign Relations,
Armed Services. Judiciary, and economic
policy committees of each house are repre-
sented. To keep the body representative, its
membership should rotate. It should exer-
cise jurisdiction over all intelligence agen-
cies and activities.
The committee should propose the statu-
Congress
fine and limit, by statute, the powers of any
government agency to intercept messages, to
engage in surveillance. or to commit any
other acts in possible violation of the civil
rights of U.S. citizens. For a wide range
of acts undertaken abroad and not directed
against U.S. citizens, explicit but probably
nonstatutory standards must be developed.
These mi ht hest be embodied in executive
__DLdzks. They cannot be written in the derail
appropriate to a tax code, but they might
reasonably be expected to distinguish among
types of behavior that involve no violation
of law in the country where undertaken
(such as the provision of financial support
to a friendly journal), cases that may he
technical violations of foreign statutes. and
actions clearly violating basic norms (such
as those intending to injure or
statutes and orders establishing
kill). The
these sub-
stantive rules must also set out clear-cut pro-
cedures for applying and enforcing them.
Enforcing the rules. The design of mech-
anisms for enforcing such rules should be
based on the principle that clandestine and
covert capabilities are too easily abused to
be controlled solely by the foreign policy
community. Ofiiitials of that community
must obviously be part of the decision pro-
cesses since the only legitimate purpose of
these activities is to advance U.S. foreign
policy objectives. But their perspectives are
not the only ones relevant. The addition of
a nongovernmental member to the Forty
203. 204.
Mullilanguage Typographers AA-75
84-856-F.P. 22 (10j12 Ben Bk) MK
Approved For Release 2005/05/23 : CIA-RDP79M00467A003100060004-8
ij
 Approved For.'l elease"ib05/05/23 : CIA-RDP79M00467A00310006000'4-8
tory charters of all major in,..iigence agen George A. Ca, r, Jr.:
tics, establish the rules accordin'. to which
they will operate, review the internal reg-
ulations of each agency, and monitor their
observance. It should confirm the appoint-
ments of directors of the Foreign Assess-
ment and Special Services agencies, and of
the inspector general of intelligence. It should
be empowered to receive (and staffed to re-
view) all requested estimates, analyses, or
information except policy advice to the pres-
ident, and should be regularly briefed on
all major issues of intelligence policy. It
might be empowered to require advance
notice of any particularly sensitive activity
it may designate. And it must bear the as-
sociated responsibility of proposing rules of
congressional procedure capable of protect-
ing the confidentiality of the information it
receives.
Finally, the joint committee should have
responsibility for reviewing and authorizing
the comprehensive community-v.wide intelli-
gence budget prepared by. the assistant for
intelligence. It will require staff fully cleared.
This is a formidable list of reforms.' But
an effective and controlled U.S. inrelli`,ence
capability y: ill require restructuring of at
least this scope and approxiniately this na-
ture. And the political conditions for such
reforms are now present-for the first time
since 1947, arid perhaps for the last time in
this century. It is important that the oppor-
tunity be seized.
'Yet it is for from complete. Mart irrpor:aot end
arguable que,tions remarr. LL' thee tie NS,\
and the ,\coonci Recnrr r.s n,e G,--:e h'ould be re-
moc'ed from the [ire-se 3,rartr-e-'t: uhc:hcr the DLk
should be cboGsl:c.! or sr,::rplu reduce:: in s:7e arid func-
tion; uhethcr the area:I Lnt't Forerun inte;l:am:e Ad-
visory Board can muJe more uuful ; and -: nether cl'e the current system of clussfymq infor,maucn, con-
structed almost en'ire/u on exeruizee orders, should not
be made both r,,,-,e l:m:re' ared r,-re enrorceahle bu
statute. We be(iece rho arsu'cr to e.:rh of these ques-
tions is probably "yes." -l boy are a questruns eh:ch
might usefully be explored by the joint committee.
Multilanguage Typographers AA-76
eter Szanton and Graham Allison's
provocative,
thoughtful article is a valu-
about the proper role,
scope. and structure
of American intelligence. Their discussion
of the intelligence community-though
wrong in some details-effectively, conveys
its complexity. And they do a real service in
laying to rest that taxonomist's delight,
the red herring of a rogue elephant."
Being a party at interest. I will not dwell
on their net assessment of the community's
performance. More than many, they make a
serious effort to be fair; but their assessment
(to use their words) is neither complete nor
completely balanced. The community's anal-
yses and formal estimates-few of which
Szanton or Allison has ever actually read-
arc far better than their assessment suggests.
Both need improvement; but shotgun alle-
gations of bias, irrelevance, poor intellectual
quality, and general inadequacy are rather
sweeping charges to make on hearsay or sec-
ondhand evidence.
Szanton and Allison commendably call
attention to the fact that this is an often un-
pleasant world in which mindless terrorism
and access to nuclear wwearonry are both
proliferating. They should have underlined
the fact that it is also a world in which meg-
aton warheads. could land on American
cities minutes after being launched, in which
openly declared wars are out of fashion, and
in which many nations energetically inter-
vene in the internal affairs of others (includ-
ing ours). They are also to be commended
for proposing that we squarely face the con-
flict posed by our open, democratic society's
legitimate needs for intelligence and covert
action. Unfortunately, in their discussion of
"unacceptable means" and "setting bound-
aries to behavior," they deflect their gaze
from some of this conflict's harsher realities
and their perspective thus goes awry.
Acquiring intelligence on the capabilities
84-856-F.P. 22 (10112 Ben Bk) MK
Approved For Release 2005105/23 : CIA-RDP79M00467AO03100060004-8
 Approved For Release 2005/05/23 : CIA-RDP79M00467AO03100060004-8
and intentions of potential ,ger,ssors-espe-
eially on their intentions-requires consid-
erably more than "technical violations of
foreign statutes." The secrets of dictator-
ships planning aggression are ruthlessly pro-
tected; and even in open societies-including
ours -cspion: ge is punishable by death.
American intelligence agencies have clear-
ly done sonic things that' Were stupid,
wrong, even criminal-thins deservedly
censured. In their zeal for "identifying vil-
lains and punishing them," however, Szan-
ton and Allison are careless about the dis-
tinction between allegation and proof, and
ignore the fact that carelessness about this
distinction on the part of many impassioned
critics has made its own contribution to the
"hemorrhage of confidence in agencies pre-
viously held in general res, ect." They are
also careless about historical accuracy. Amer-
ican intelligence agencies .have set no prece-
dents. Even if all the allegations about the
CIA were true--and they are net-its acts
would have been pallid compared to the
precedents set by (and elaborated by the
national, institutional, or ideological suc-
cessors of) net just Dzerzhinskv, Yagoda,
and Bcria, but also Saint Ignatius Lcyo!a,
Sir Francis