IC21: The Intelligence Community in the 21st Century
Staff Study
Permanent Select Committee on Intelligence
House of Representatives
One Hundred Fourth Congress
II. Intelligence Community Management
Executive Summary
One of the centerpieces of the Intelligence Community for the
21st Century (IC21) review is a hard look at Intelligence
Community (IC) management and the development of a proposed
community model that synthesizes the findings and recommendations
of the other staff studies. At the beginning of this undertaking,
a hypothesis was developed that the IC and its customers would
benefit, either through performance enhancement or cost reduction
or both, from a more corporate approach to intelligence. This
hypothesis was then "tested" in the following specific areas:
planning, programming and budgeting; collection management;
production management; personnel management; and research and
development. The goal was to identify what specifically would
improve management of these areas, and whether or not a more
corporate approach would be constructive. Then, if a more
corporate approach were dictated, to identify what changes in
organization, function, and authority would be required to achieve
it.
Perhaps not surprisingly, we discovered that the Intelligence
Community would benefit from a more corporate approach in each of
the major areas we addressed. In order to form a flexible "tool
kit" of capabilities for the future, the Director of Central
Intelligence (DCI) and his staff require additional authorities and
different management structures to create a unified, effective and
efficient community. Services of common concern should be
consolidated at the community level. Programming and budgeting and
personnel management must be more centrally managed. Collection
must be managed coherently across the disciplines, with
increasingly difficult resource trades made at the community level
in an informed, all-source process. Improved synergy during
collection operations, which will become more and more critical to
success in the 21st century, requires movement away from the
traditional stovepipe approach to collection. Research and
Development requires closer coordination with requirements, and a
contingency fund for "good ideas" should be established to allow
the community to take advantage of technological targets of
opportunity.
The community needs to become a corporate entity; personnel
reform that promotes lateral movement among agencies and a
community SES cadre is essential. The primacy of all-source
analysis needs to be reinforced, and strong links forged between
analysts and policy-makers and analysts and collectors. The
community should be, and to an extent already is, moving toward a
"virtual analytical environment" that requires a new set of skills
and management techniques. Increased centralization of management
functions must be balanced by a strengthened and independent
evaluative function.
Clandestine operations will continue to be both the riskiest
and potentially the highest-payoff intelligence operations,
becoming increasingly important in the 21st century due to the
likely nature of future targets. This aspect of the intelligence
community requires a more intensive level of management involvement
on the part of the DCI and should be housed in a separate
organization, with a direct reporting chain to the DCI.
The defense intelligence community also stands to benefit from
more coherent and centralized management. A Director of Military
Intelligence with enhanced control over defense intelligence
programs and operations would serve as both a senior military
advisor to the Secretary of Defense for intelligence, and a locus
for the close coordination required between the national and
tactical intelligence communities and budgets.
INTELLIGENCE COMMUNITY MANAGEMENT
I. Approach
One of the centerpieces of The Intelligence Community in the
21st Century (IC21) review is a hard look at Intelligence
Community (IC) management and the development of a proposed
community model that synthesizes the findings and recommendations
of the other staff studies. At the beginning of this undertaking,
a hypothesis was developed that the IC and its customers would
benefit, either through performance enhancement or cost reduction
or both, from a more corporate approach to intelligence. This
hypothesis was then "tested" in the following specific areas:
planning, programming and budgeting; collection management;
production management; personnel management; and research and
development. The goal was to identify what specifically would
improve management of these areas, and whether or not a more
corporate approach would be constructive. Then, if a more
corporate approach were dictated, to identify what changes in
organization, function, and authority would be required to achieve
it. Although they are presented first in this document, the role
and authorities of the Director of Central Intelligence (DCI) were
considered last, in the context of the needed changes in the
above-mentioned areas.
II. Introduction/Assumptions
It immediately became clear that it is impossible to measure
the effectiveness of something without a standard by which to
measure -- an understanding of the purpose and role of
intelligence, and its appropriate relationship to policy and
national strategy. With very little research it became apparent
that there has historically been disagreement on these topics, and
that the level of disagreement is greater today, in the post-Cold
War period, than it has been for some time. This makes it
necessary to examine these issues in at least a cursory way in
order to establish some assumptions without which the answers to
the questions posed by this study would be meaningless.
At the most basic level, there have been, and remain, two
diverging views of the appropriate role of intelligence in the
United States. One view maintains that intelligence provides
impartial and objective information to policy-makers; intelligence
is a truth-seeking profession and the policy community is a
customer who does not and should not influence the product. The
other, and less widely held, view is that intelligence is in fact
an instrument of policy and should be used to both shape and
further policy goals: the intelligence and policy communities must
act as partners. The question of whether intelligence informs
policy or serves it is truly a chicken-or-the-egg issue -- we
believe it must do both at different times. Tending too far in
either of these directions threatens lack of relevance on the one
hand, and politicization on the other. The challenge for the IC is
to maintain a balance of objectivity and involvement, a goal that
can only be met with the cooperation and understanding of the
policy community. This study assumes that the basic structure of
the United States government, including its policy apparatus, will
remain relatively stable at the departmental level, but that the
policy community may be influenced positively by recommended
changes in its formal relationship to the IC.
Another basic question that must be raised is that of the
evolving definition of national security. Although there may be a
consensus that intelligence exists primarily to identify potential
threats to the national security of the United States, the
definition of those threats, and perhaps the threats themselves,
change over time. We have seen an evolution from nation-based
threats and conflicts to trans-national threats and regional and
ethnic strife. New areas of intelligence emphasis, such as
proliferation and terrorism, clearly represent emergent threats to
our national security. Other, less clear-cut areas of endeavor,
such as economic and environmental intelligence, remain subjects of
debate concerning the closeness of their relationship with national
security, how much value intelligence actually adds to these areas,
and at what cost to other, higher priorities. Regardless, all of
these areas of endeavor represent a new level of complexity for the
IC, requiring an "interdisciplinary" approach to intelligence and
a different set of skills than that needed in the Cold War world.
Each Administration will be faced with defining threats to
national security, and the results will vary. In the absence of
definitive guidance, the IC will inevitably try to be all things to
all people. Therefore, it is a mistake to structure the community
to meet currently articulated or even projected future threats
except in the most general sense. In looking to the 21st century,
it is important to reach a consensus on the core missions and
capabilities of the IC, and to add to those missions only on a
pay-as-you-go basis. The new approach to mission-based budgeting,
which creates four primary mission areas (support to policy makers,
support to military operations, support to law enforcement, and
counterintelligence), and within those areas identifies core
capabilities, sustaining capabilities and supporting capabilities,
appears to be a move in the right direction. The community of the
future should be based on the capability and flexibility to perform
those basic functions -- a "tool kit," if you will, for the
challenges of the next millennium.
Within the IC, there are a series of checks and balances.
Starting at the top, the relationship between the DCI and the
Secretary of Defense (SECDEF) epitomizes an important tension in
the community: support to military operations (SMO) versus support
to national-level policy makers. Considering that military
operations are an instrument of policy, SMO is in fact another
facet of support to the policy-maker, but it is of a different and
potentially all-consuming sort. The Department of Defense (DoD) is
the largest customer of intelligence information, and that
justifies its significant voice in the process; the DCI, however,
must be able to protect the equities of the civilian policy-makers
and the longer-term interests of the nation (a more detailed
discussion of this tension is contained in both the Intelligence
Support to Military Operations and the Intelligence Community Surge
Capability staff studies). That much of the intelligence community
is a shared resource is at times problematic, but is in accord with
statutory direction to "eliminate waste and unnecessary duplication
within the intelligence community." It makes sense from a resource
perspective, as long as appropriate management safeguards exist to
ensure that no customer's needs are shortchanged in the process.
Another balance issue within the community is the role of the
program manager vis-a-vis the issue coordinator. The Needs Process
has established an increasing tension between the issue
coordinators, who are looking across programs to fund priority
activities that contribute to their individual areas of
responsibility adequately, and the program managers, who are faced
with satisfying the requirements of all of the issue managers and
must make internal trades to build a coherent and sustainable
program. This would be more of a contest if the issue coordinators
had any real leverage over the budget process, but currently they
do not. A similar case is the lesser, but still important, tension
between functional managers and program managers. Because the
program managers build the budget, and the issue coordinators and
functional managers can basically only advise and recommend, the
balance of power is skewed in favor of the program managers. In
any scheme of intelligence community management, there will be
competing requirements of this type. The challenge is to create a
programming and budgeting process that minimizes destructive
competition and can adjudicate competing requirements and
priorities in a balanced way.
Finally, the Congressional intelligence oversight function,
unique to this nation, represents one of the legislative checks on
the executive branch that is the hallmark of our system of
government. The two intelligence committees, in turn, provide a
check on each other in the performance of this function. Although
this makes for a complex and sometimes inefficient system, in the
long run it protects the interests of the American people. Within
the IC as within the government at large, some of these existing
balances may need to be recalibrated; overall, however, they serve
a useful purpose and should not be lightly set aside.
III. Summary of Findings:
Perhaps not surprisingly, we discovered that the IC would
benefit from a more corporate approach in each of the major areas
we addressed. In order to form a flexible "tool kit" of
capabilities for the future, the DCI and his staff require
additional authorities and different management structures to
create a unified, effective and efficient community. Services of
common concern should be consolidated at the community level.
Programming and budgeting and personnel management must be more
centrally managed. Requirements and collection must be managed
coherently across the disciplines, with increasingly difficult
resource trades made at the community level in an informed, all-source
process. Improved synergy during collection operations,
which will become more and more critical to success in the 21st
century, requires movement away from the traditional stovepipe
approach to collection. Research and Development (R&D) needs to be
more closely coordinated with requirements and a contingency fund
should be established to take advantage of technological targets of
opportunity.
The community needs to become a corporate entity; personnel
reform which promotes lateral movement among agencies and a
community SES cadre is essential. The primacy of all-source
analysis needs to be reinforced, and strong links forged between
analysts and policy-makers and analysts and collectors. The
community should be, and to an extent already is, moving toward a
"virtual analytical environment" that requires a new set of skills
and management techniques. Increased centralization of management
functions must be balanced by a strengthened and independent
evaluative function.
Clandestine operations will continue to be both the riskiest
and potentially the highest-payoff intelligence operations,
becoming increasingly important in the 21st century due to the
likely nature of future targets. This aspect of the IC requires a
more intensive level of management involvement on the part of the
DCI and should be housed in a separate organization, with a direct
reporting chain to the DCI.
The defense intelligence community also stands to benefit from
more coherent and centralized management. A Director of Military
Intelligence (DMI) with enhanced control over defense intelligence
programs and operations would serve as both a senior military
advisor to the SECDEF for intelligence, and as a locus for the
close coordination required between the national and tactical
intelligence communities and budgets.
IV. Roles, Relationships and Authorities
Role of the DCI
The role and authorities of the DCI are central to achieving
the goal of a more corporate IC. There are two broad areas at
issue: (1) the role of the DCI vis-a-vis the President; and (2)
the role of the DCI within the IC.
Several witnesses, including several past DCIs and Deputy
DCIs, noted that the degree to which the DCI visibly commands the
respect and confidence of the President is central to the DCI's
effectiveness. Realistically, however, there is no way to mandate
or to legislate a close working relationship between these two
officials. Two suggestions repeatedly surface regarding the status
of the DCI. The first is that he be made a cabinet-rank official.
The second is that he be given a fixed term of office. The study
group does not believe that either of these has sufficient merit or
would achieve the goal of a stronger DCI. The third is that he be
relieved of his responsibilities for the Central Intelligence
Agency (CIA) and elevated to a position over the entire IC.
Cabinet-rank for officials who are not members of the Cabinet
(i.e., the heads of departments) is merely an honorific. The
United States does not have Cabinet government; being designated a
member of the Cabinet does not in any real sense increase one's
authority. It certainly will not enhance or improve the DCI's
relationship with the President, which can only be based on an
existing level of trust and confidence. Indeed, mandating
Cabinet-rank for the DCI while doing anything less than creating a
true Intelligence Department -- which no one has contemplated -- only
calls more attention to the disparity between the DCI's
responsibilities and his authority, even with the enhancements
being proposed here.
The importance of the DCI's personal relationship with the
President is also the main argument against a fixed term.
Proponents of a fixed term argue that this would have several
benefits. Ten years is often suggested, as has been done with the
Director of the Federal Bureau of Investigation (FBI). First, and
perhaps foremost, a fixed term would provide for greater continuity
and stability than we now have. Until 1977, it was not customary
for the DCI to be replaced with a new administration. That is no
longer the case. Moreover, the DCI's position has since been
subjected to fairly frequent turn-overs over and above presidential
transitions. From 1973-1977 there were five DCIs; from 1991-1996
there have been four DCIs. However, a fixed term could create the
situation where a President would inherit a DCI with whom he could
not work. Although there would be greater continuity, the DCI's
effectiveness would diminish rapidly, a far greater loss. As
noted, an analogy is often drawn to the Director of the FBI. The
comparison is inapt. The DCI is the chief intelligence officer and
deals directly with the President. The Director of the FBI is not
the chief law enforcement officer; the Attorney General is and
serves at the President's pleasure. In sum, a fixed term would not
be an improvement.
The National Security Act states that the DCI is the head of
the IC and the President's principal intelligence adviser. Neither
of these designations for the DCI is the same as meaningful
control. If the IC is to achieve a greater degree of coherence and
corporate identity, then the role of the DCI has to be changed.
The glaring gap between his responsibilities and his authorities
has to be closed to the greatest extent possible. The DCI should
be viewed as a chief executive officer of the IC, with purview over
all of its major functions and a greater degree of control over
budgets, resources and major policy issues that are common to all
agencies. However, the testimony of former DCIs and other former
senior IC officials all concur that the DCI needs an agency
"of his own" -- i.e., the CIA -- if he is to have any real power
within the IC.
The National Security Council
The National Security Act also places the DCI under the
direction of the National Security Council (NSC). The NSC is
composed of four officials: the President, the Vice President, and
the Secretaries of State and Defense. The IC is a service
organization. It has no meaning without its relationship to policy
makers. Thus, the DCI must have regular contact with the NSC
members. However, it is not reasonable to expect that they can
give the DCI and, through him, the IC, the kind of regular
executive guidance that was envisioned by the National Security
Act. Indeed, in each successive Administration, there has been
some sort of sub-NSC group created to deal with intelligence,
reflecting the shortcomings of the NSC itself to carry out this
role.
Finally, many witnesses at hearings and staff panels and the
oversight experience of this Committee indicate that certain
intelligence activities -- clandestine operations and covert action
-- require special attention. These activities consume an
inordinate amount of the DCI's time, in terms of both management
and testimony before Congress. In the future, certain types of
offensive information warfare (IW) activities conducted in
peacetime or outside the context of a military operation may also
fall into this category. We do not question the utility of these
activities and believe that the United States must have recourse to
them. At the same time, executive control can and should be made
more direct. It is important for the DCI to maintain close control
over these activities.
The following recommendations are designed to resolve the
issues noted above. Beginning with the issue of executive
guidance, of the various sub-NSC bodies created to deal with
intelligence, the Committee on Foreign Intelligence (CFI) created
by President Ford in 1976 appeared to be among the more successful,
in terms of its stated role, its membership and its performance.
Interestingly, the Senate Select Committee on Intelligence proposed
re-establishing this group in legislation in 1992, as has the
Aspin-Brown Commission. We believe that the CFI, properly
constituted and empowered, can more usefully serve as a body to
provide the DCI and the IC with the necessary guidance and policy-maker
oversight. This is not meant to supplant the DCI's current
direct access to the NSC members; it is meant to give the DCI
access on a more regular basis to senior policy-makers who can give
direction to the IC and can listen to and relay IC concerns.
Two Deputy Directors of Central Intelligence
As noted, we do not find major flaws in the broader parameters
of the role of the DCI as currently described in legislation in
terms of his tenure or his responsibility for the CIA. The DCI
should continue to serve at the pleasure of the President and
continue to exercise control over the CIA and the Community
Management Staff (CMS), and have direct control over the
Clandestine Service. The DCI would, thus, continue to have
multiple major responsibilities. All DCIs have found this a broad
and sometimes difficult mandate. The ability to delegate is
important, although it has been done differently by virtually every
DCI. The current DCI, for example, relies on two executive
directors -- one for the CIA and one for the CMS. Their titles
belie their responsibilities. The positions responsible for these
two large parts of the DCI's portfolio should be enhanced and their
duties better defined. Given the importance of their positions,
Senate confirmation also appears necessary. Some permanence in the
DCI's supporting structure is needed and can be achieved without
losing necessary flexibility. It also allows for greater
institutional continuity, clearer definition of responsibilities
and improved congressional oversight.
In order to minimize superfluous bureaucratic layering, we
concluded that the current position of Deputy DCI (DDCI) should
specifically be given day-to-day responsibility for the CIA, whose
enhanced analytical responsibilities are discussed below. This
would reduce layering, would continue to give the DCI direct access
to his major bureaucratic and institutional base, and yet would
relieve the DCI of many lesser administrative concerns.
Paralleling this first DDCI, there should be a second DDCI for
Community Management, for much the same reasons, with purview over
the collection, acquisition and infrastructure elements of the IC.
There are also changes in the DCI's budget and personnel
authorities, noted below. As currently allowed by law, either the
DCI or one of his DDCIs -- but no more than one -- could be a
military officer. The DCI would select which of the DDCIs would
act as DCI in his absence.
As noted above, the importance of the DCI's relationship with
the President is such that few prerequisites for nominees should be
imposed. However, to the extent possible, these DDCI positions
should be considered as professional as well as political
appointments and should go to individuals with extensive national
security or intelligence background. This is especially important
if a DCI with less such background is chosen. The two DDCIs should
be confirmed by the Senate, just as is the current DDCI position.
The Central Intelligence Agency
The CIA, which would now be directed by the DDCI, was
envisioned by President Truman as a coordinator of disparate
intelligence being produced by other agencies. The CIA quickly
became a producer in its own right because of policy-maker demands,
the unwillingness of then-existent agencies to respond, and an
aggressive CIA leadership. Although this is different than
President Truman's vision, we do not believe that this development
should be reversed. Indeed, it would appear more profitable to
underscore the CIA's analytical role by confirming it as the
premier all-source (i.e., deriving its analysis from all
intelligence collection disciplines) analytical agency within the
IC.
We concur with the observation of former DCI Richard Helms
that the President needs his own analytical group and that if we
did not have the CIA today we would probably invent it.
Underscoring this role means more than words. The CIA should house
not only its analysts, but the second- and third-tier exploiters of
the various intelligence collection disciplines. By bringing them
closer together we can improve the efficiency of the all-source
analytical process and achieve a true synergy between collection
and analytical production.
The Clandestine Service
Given the political and administrative problems raised by
clandestine operations and covert action, their bureaucratic tie to
the DCI must be made more direct. At present as many as two or
three officials are between the DCI and the CIA's Directorate of
Operations (DO). Moreover, there is no compelling substantive
reason for the DO to be part of the same agency as the analytic
Directorate of Intelligence (DI). This is largely the product of
historical accident and the bureaucratic aggressiveness of DCI
Walter Bedell Smith, who expanded CIA activities into both
operations and analysis in the early 1950s, when other agencies
failed to meet policy-maker needs in these areas.
We believe that it would be better for the DO, renamed the
Clandestine Service, to be a distinct entity, under the direct
control of the DCI. This would rationalize the structure of the
CIA as the premier all-source analytical agency. The Clandestine
Service and the CIA can continue to be housed in the same building.
However, both the Clandestine Service and the CIA could also be
managed more effectively if they each had one major task. The
separation of the Clandestine Service should also reinforce the
fact that clandestine Human Intelligence (HUMINT) serves the entire
community and not just the CIA. The Clandestine Service would
conduct all clandestine HUMINT operations, even those undertaken by
military personnel, who would be integrated into the organization.
There should be a Director of the Clandestine Service,
reporting directly to the DCI. This individual should be an
intelligence professional. After much debate, we recommend that
this individual not be subject to confirmation by the Senate. The
sensitivity of this position is such that the DCI must be free to
choose the man or woman upon whom the utmost reliance can be
placed. Senate confirmation raises a number of other political
considerations that might best be avoided. This recommendation,
coupled with the role of the new DDCI/Community Management, should
also allow a closer integration of collection management and
operations, and should enhance oversight of clandestine operations.
The Director should have a deputy who is a two-star active duty
military officer (further details are contained in the Clandestine
Service staff study).
NFIP Defense Agencies
If the IC is going to achieve the goal of "corporateness," and
if the DCI is going to function as a true CEO, then he should have
a greater say in the selection of his "corporate team" -- the heads
of the other major intelligence components. Current law requires
that the SECDEF "consult" with the DCI in naming heads for National
Foreign Intelligence Program (NFIP) defense agencies. Although it
is unlikely that the SECDEF would nominate someone to whom the DCI
is strongly opposed, it is possible. Instead, the DCI's advice and
concurrence should be sought. In the unlikely event of
disagreement, the issue could be referred to the NSC Committee on
Foreign Intelligence or, ultimately, to the President. But the
importance of a truly corporate team requires a stronger DCI voice
in this process. The study group believes, however, that the role
of the NFIP defense agencies is so substantially different from
that of the other departmental elements of the NFIP that this
arrangement is not appropriate for the State, Energy or Justice
Departments. The defense agencies are primary collectors and
producers of intelligence without whom the DCI could not perform
his statutory functions, while the other departmental elements are
analytical efforts focused on tailoring intelligence products for
their departmental consumers. Therefore, we recommend no change in
the selection process for those activities.
Director of Military Intelligence
The Defense Department -- civilian policy makers and military
services at all levels -- is one of the largest components and
mostly important customers of the IC. Many of the larger
organizational issues noted for the IC at large are also found
within the defense-related part of the IC. Enhancing the DCI's
authority solves some, but not all, of the problems. It is
important that the defense intelligence establishment also have a
single, uniformed official who is both responsible for and
empowered to address these issues, or to advise the SECDEF about
them. We believe that this should be a three-star military
officer, carrying the title of Director of Military Intelligence
(DMI). The study group also believes that this individual should
be dual-hatted as the Director of the Defense Intelligence Agency
(DIA), the program manager of the Joint Military Intelligence
Program (JMIP), and program coordinator for the Tactical
Intelligence and Related Activities (TIARA). Although previous
proposals for a DMI have sought a four-star office, the study group
believes a four-star officer is neither appropriate nor likely to
be approved. For the senior military intelligence officer to be on
a par with the Chairman of the Joint Chiefs of Staff (CJCS) and the
Commander in Chief is not appropriate for a supporting function
such as intelligence, and could potentially promote an unhealthy
rivalry between the DMI and the DCI, particularly if the DCI were
to remain as currently constituted, i.e., not of cabinet rank. The
DMI would report to the DCI on IC-wide issues and activities.
The three-star DMI concept consolidates management of defense
intelligence across the NFIP (DIA), JMIP and TIARA and continues to
provide intelligence support to both OSD and CJCS, via the J-2, and
a unified J-2/DIA staff. The DMI would not control the DoD
agencies within the NFIP, but would be responsible, as currently,
for all defense analysis, production, and overt HUMINT operations.
As program manager for JMIP, the DMI would ensure a coherent
program that complemented national and tactical capabilities. As
program coordinator for TIARA, he would ensure that the services'
intelligence programs were interoperable and consistent with the
larger intelligence architecture. The DMI would need a
significantly enhanced staff element to handle program and budget
activities for the JMIP and TIARA formerly handled by the office of
the Assistant Secretary of Defense for Command, Control,
Communications and Intelligence (ASD(C3I)), and to be responsible
for defense intelligence architectures and coordination with the
community systems and architectures office.
Assistant Secretary of Defense for Intelligence
The position of ASD (C3I) is, in the study group's view, an
artificial construct. Although C3I for the Warrior and related
concepts have been constructive in encouraging the Services and DoD
to integrate intelligence and information handling techniques
better into Command, Control and Communications (C3) architectures,
integration of C3 and Intelligence as staff functions has simply
not happened, either in ASD(C3I) or in the Joint Chiefs of Staff
(JCS). One can also make an argument that in the Information Age,
intelligence needs to become increasingly linked to operations; C4I
for the Warrior may support this operational concept in theory, but
is of limited utility for staff planning purposes. To date, most,
if not all, Assistant Secretaries for C3I have placed primary
emphasis on the "C3" rather than the "I." Similar emphasis must be
placed on intelligence if doctrinal concepts such as Dominant
Battlefield Awareness are to be realized. One aspect of this
increased emphasis is a more corporate approach to intelligence as
embodied by a DMI. The other aspect is a stronger policy presence
in Defense. Consequently, the study group believes that defense
intelligence would be better served by having a separate Assistant
Secretary of Defense for Intelligence (ASD(I)), an option that the
SECDEF could exercise at any time. Regardless, the role of the
ASD(C3I) or ASD(I) should be policy, planning and oversight; the
programmatic and budgeting functions that have devolved to ASD(C3I)
should be handled by the DMI staff.
Infrastructure Management
Numerous studies and reviews of the community, including the
National Performance Review, have concluded that there are
efficiencies and potential cost-savings to be had by consolidating
infrastructure and "services of common concern." During the course
of this study, it became apparent that it makes sense to combine
under centralized management, although not necessarily in one
place, such community functions as personnel management, security,
certain types of training, communications, and automation./1/
Although many of the personnel performing these functions could
remain physically in place as support detachments, the study group
believes that an Infrastructure Support Office should be
established to manage these areas across the community. The growth
of the IC and proliferation of distinct agencies have led to
unwarranted duplication in what are, essentially, administrative
and logistical functions. This is not only duplicative and costly,
but also can harm the ability of the IC to operate as a corporate
whole.
Finally, these recommendations raise one final question about
oversight. There is, currently, a statutory Inspector General (IG)
for the CIA and for DoD. In order to ensure that major IC-wide
functions are available to necessary scrutiny, the current CIA IG
should serve as the IC IG, operating, when necessary, in
conjunction with the DoD IG for NFIP Defense agencies.
Recommendations:
1) Reestablish the Committee on Foreign Intelligence to provide
the DCI with necessary guidance and feedback. The Assistant
to the President for National Security should chair the CFI;
other members should be the Secretaries of State and Defense,
the Chairman of the JCS, and the Attorney General, or their
deputies.
2) Create two Deputy Directors of Central Intelligence: a DDCI
to manage the CIA, responsible for all IC production and
analysis; and a DDCI for Community Management, responsible for
requirements, collection and resource management. Both DDCIs
should have extensive national security experience; both
should be confirmed by the Senate. At no time should more
than one of the three (DCI and two DDCIs) be active duty
military. The DCI will designate one of the DDCIs to serve as
the acting DCI in his absence.
3) Designate the Director of DIA as the Director of Military
Intelligence (DMI). The DMI will be the program manager for
the JMIP and the program coordinator for TIARA.
4) Increase the DCI's role in the appointment of NFIP agency
directors by requiring the Secretary of Defense to obtain his
"advice and concurrence" for these appointments.
5) Urge the Secretary of Defense to consider creating an
Assistant Secretary of Defense for Intelligence.
6) Create a separate Clandestine Service directly accountable to
the DCI. The Director of the Clandestine Service should be
selected by the DCI from among intelligence professionals.
The Deputy Director should be a two-star military intelligence
officer.
7) Create an Infrastructure Support Office (ISO) which
consolidates services of common concern across the community,
to include at a minimum personnel, security, training,
communications and automation.
V. Collection and Requirements Management
One of the IC's main shortcomings is an inability to manage
collection optimally across disciplines or "INTs." This
shortcoming is reflected in two areas: in short-term collection
management against current intelligence problems, and, more
seriously, in longer-term resource reallocation between collection
disciplines based on an examination of intelligence needs, the most
appropriate mix of collection assets to fulfill those needs, and an
evaluation of how well those assets perform against their tasking.
Collection requirements and tasking are currently handled by
committees that make resource and tasking decisions in a single-source
context that does not promote an optimal all-source approach
to collection problems. In the global and resource environment
envisioned for the future, competition for collection assets,
already stiff, will only increase. Trans-national problems such as
proliferation require integrated, all-source solutions. With the
collapse of the Soviet Union, even as more information becomes
available from open sources, the remaining "hard" targets have
become tougher to crack, also necessitating a coordinated, multi-INT
approach. The tension between military requirements -- now
expanded to include humanitarian and peacekeeping missions -- and
longer-term national interests will become greater and the
mechanism for making decisions such as whether or not to move a
satellite from one region to another must become more robust. The
IC needs a management staff with the resources and authorities to
build and maintain a coordinated collection program, and keep it in
balance with the production and infrastructure elements of the
community.
What community management is currently provided comes from the
National Intelligence Collection Board, a companion organization to
the National Intelligence Producer's Board. Although this forum is
beginning to become more "energized" under its new chief, it is
not yet the body to compel the needed integration of the collection
process within the community. The fact that the Executive Director
(ExDir) for Community Affairs and the Associate Director of
Intelligence for Military Affairs are planning the establishment of
a Collection Operations Management Group indicates an awareness of
this problem. This organization, or something like it, needs to
exist at the community level, with representatives from the
programs and DoD/JCS, to provide an integrated forum for collection
decisions and to mediate conflicts between short-term military and
longer-term policy-maker support. This organization could either
supersede or be superimposed upon the current entities involved in
single-INT tasking: COMIREX, the Signals Intelligence (SIGINT)
Committee, the Measures and Signatures Intelligence (MASINT)
Committee, and the National HUMINT Requirements and Tasking Center.
For short-term collection against current intelligence
targets, there are two collection management centers within the
Community, one at the CIA and one at DIA. Although these centers
can be said to work reasonably well, the coordination mechanism
between them is not well-defined. Also, tasking collection or
requesting information within the current system is inefficient.
At some point in the requirements chain, a customer with a
requirement must submit a SIGINT or Imagery Intelligence (IMINT)
collection request, rather than a general request for information.
It is virtually impossible for a requestor to ascertain whether the
information he requires has already been collected and exists in a
database somewhere or must result in new collection tasking. The
IC needs a system that centrally manages information requests, and
a focal point for managing this process across the community.
Although some progress has been made towards this goal, it has
been done mostly on an "INT by INT" basis rather than as a
community-wide, all-source effort. However, the Intelligence
Systems Board (ISB) has proposed a Request For Information (RFI)
management system that would further this goal.
One cannot discuss collection without addressing
"stovepipes." To illustrate the long-standing nature of this
debate, the following is a quote from Community Management Task
Force Report commissioned by then-DCI Robert Gates and conducted by
Danny Childs and Rich Haver in 1991: "We have made one key
assumption -- that vertical collection management structures are
created. We should note, however, that there is a body of opinion
that strongly doubts the wisdom of creating such 'stovepipes.' One
concern is that powerful checks and balances will be needed to
compensate for the possible tendencies of such strong functional
managers to operate unilaterally and make decisions with an eye to
resource advantage. A second concern is the possibility that
community requirements will not be equitably addressed without the
aid of a strong independent body as a requirements authority."
Although the existence of stovepipes was an assumption for
that report, the study group believes that it is no longer wise or
even possible to accept stovepipes as a given. There are real
benefits to be achieved by creating a more unified management
structure for technical collection operations. MASINT, in
particular, which many view as the "INT of the future" because of
its potential application for some of the more difficult
intelligence problems such as proliferation, would benefit from an
approach that does not view it as a competitor to SIGINT and IMINT,
but rather as a complementary discipline making use of many of the
same sources of collection (see the MASINT : Measurement and
Signatures Intelligence staff study for more details). As noted
above, the key to future success against difficult collection
problems with shorter and shorter timelines is to achieve greater
synergy between the collection disciplines. Wherever this occurs,
the results are greater than the sum of the parts. Instead of
designing cumbersome systems "after the fact" to tip off collection
assets operating within a completely different conceptual and
operational framework, these operations need to be conceptually
integrated from the beginning and managed coherently. The target
environment itself is beginning to blur the lines between the
technical disciplines.
The truth is that, to a certain extent, stovepipes are
unavoidable; the issues are how far up they extend and whether or
not a mechanism exists to ensure interaction between them at the
operational level. Although the technical collection disciplines
share many elements (as several interviewees told us, "it's all
about bandwidth") and will undoubtedly become increasingly similar
in the future, there are nevertheless distinct skills and training
requirements associated with SIGINT, IMINT and MASINT -- and HUMINT
collection is significantly different from all the others.
Although the study group believes that all of the technical
disciplines would benefit from being managed in a coherent fashion,
the different endeavors are not, in the foreseeable future,
interchangeable, and it is important to maintain the levels of
expertise in each of these areas that have contributed to our
success to date. Therefore, if the technical collection
disciplines were combined into one agency, as we recommend, there
would in all likelihood be "mini-stovepipes" within it. This would
not necessarily be a bad thing as long as there was cross-leveling
activity both at the operator level and at the top, where it would
all "come together" under the control of one individual. Under a
consolidated collection concept, technical control of the various
collection disciplines would be vested in the director of the
collection agency and delegated to designated functional managers
for each discipline. The director of the collection agency would
thus assume the Director of the National Security Agency 's (NSA's)
responsibilities as SIGINT advisor to both the DCI and the SECDEF,
and perform similar functions for IMINT and MASINT.
Additionally, the best collection operations occur when
collectors and analysts work closely together, so it is important
to keep the "first-line" analysts or exploiters with the
collectors. These analysts provide immediate feedback to the
collectors, report on time-perishable information, and act as a
"bridge" to the all-source analytical community, with whom they
should be electronically linked. Although we acknowledge that the
dividing line between first-line exploiters and second- and third-tier analysts is not as clear-cut in the SIGINT arena as it is in
the imagery world, we nevertheless believe it is possible to
distinguish between these levels of analysis in a systematic way
(see the SIGINT: Signals Intelligence staff study for more
details). It is equally important to leave first-tier HUMINT
exploiters such as reports officers with the HUMINT collectors.
Although the technical collection disciplines could reasonably
and effectively be combined into one agency, it is the opinion of
the study group that HUMINT collection can and should remain apart,
with overt HUMINT collection continuing to be conducted by DIA and
the State Department, and all clandestine HUMINT collection
operations falling under the purview of the Clandestine Service
(see the Clandestine Service's staff study for more details on this
concept). HUMINT tasking and operations are different enough that
there is little to be gained by combining its management with that
of the technical collection disciplines, and, as mentioned earlier,
its risks are such that it warrants a more intensive level of
organizational oversight. There are, however, numerous instances
where HUMINT supports technical collection in extremely important
ways. To maintain effective cooperation in these areas, an
aggressive rotation policy is required to ensure that clandestine
operations personnel are employed in the collection areas supported
by their efforts, and that technical personnel are employed where
they can affect the tasking of HUMINT assets. It is also important
to note that clandestine HUMINT collection tasking and
requirements, along with all other collection operations, will be
managed by the CMS and reviewed by the National Intelligence
Evaluations Council (NIEC). (The NIEC is discussed in the
Intelligence Requirements Process staff study.
The study group also considered whether or not it was
advantageous to combine Open Source collection with the technical
collection disciplines. Although clearly areas of similarity
exist, we determined there was little to be gained from this
proposal. Since the primary focus of Open Source collection is the
management of huge amounts of information that are readily
available rather than the attempt to collect information from
denied areas or that the originator does not wish anyone to have,
it was decided to place responsibility for Open Source with the
analytical agencies, primarily the CIA.
Recommendations:
1) Create a community-level requirements and collection
management activity within the CMS responsible for directing
collection tasking to the appropriate organizations and
ensuring a coherent, multi-INT approach to collection
problems.
2) Create and centrally administer a community-wide system for
RFI management.
3) Create a Technical Collection Agency (TCA) that combines
SIGINT, IMINT and MASINT collection, processing and first-tier
exploitation and analysis. The TCA should be a Type 3 Combat
Support Agency, and its director should be either a senior
defense or intelligence civilian or a flag officer.
VI. Production Management
There are three primary, sanctioned producers of all-source
intelligence products in the IC: the CIA, DIA, the State
Department's Bureau of Intelligence and Research (the Department of
Energy's Intelligence Division is also an all-source producer of
tailored products for its departmental consumers). Although the
appropriateness of the State Department maintaining its own
analytical capability is rarely questioned, many have suggested
that the separate DIA and CIA efforts are not necessary. However,
in our view, reality dictates that the Defense community must have
its own analysis and reporting capability. If we were to do away
with DIA, it would be recreated in another form somewhere in DoD.
The study group also believes that the DIA/CIA balance is of value
to the community: they have largely deconflicted their analysis
and production, they have very different customer bases, and there
is inherent value to maintaining the ability within the overall
community to get a "second opinion." CIA correctly views one of
its roles as providing an independent assessment of the efficacy of
U.S. military operations. Although DIA has no formally constituted
charter to challenge CIA assessments, in those areas that most
threaten our national security, maintaining the ability to do
competitive analysis is prudent, as long as it is by design and not
a result of lack of management.
CIA and DIA, largely left to their own devices by the CMS
but questioned by Congress repeatedly over a period of years for
duplication of analysis and production, have made a great deal of
progress in coordinating and deconflicting their analytical efforts
and scheduled production. The fact that scheduled production
represents a smaller and smaller percentage of total intelligence
product in no way minimizes this achievement, but also shows that
this process is a moving target. The coordination of finished
products also does not address the issue of the community's other
analytical products, which are not (theoretically) all-source --
SIGINT and IMINT reports.
Elements of the community have been moving independently in a
positive direction in the analysis and reporting area -- this is
both the good news and the bad news. The good news is that the
community is using technology to work towards the types of products
that are most useful to the customer: multi-source, multi-media
products delivered electronically. The bad news is that this is
being done in a largely uncoordinated way, resulting in the births
of multiple, pseudo-all-source analysis centers using many of the
same sources of data and producing products that look a lot like
all-source products. What the community needs is a coordinated
approach to distributed and collaborative analysis, similar to the
concepts being developed at NSA (the Analyst Driven SIGINT System
being developed in conjunction with NIDL/Sarnoff Labs) and DIA (the
Joint Intelligence Virtual Architecture, or JIVA). The community
needs to create a "virtual analytical environment" that will
maximize the efficiency of an increasingly scarce and valuable
commodity -- the analyst. Although exploitation and first-level
analysis should remain with the individual collection disciplines,
many of the analysts currently doing SIGINT- and IMINT-centered
analysis should be moved, physically or, preferably,
electronically, to an all-source enclave (CIA or DIA) to provide
the understanding of the source data and collection process
required to produce high-quality all-source analysis and reporting,
with appropriate feedback to the collectors/exploiters. By
consolidating these efforts, we prevent the unnecessary replication
of analytic effort by ensuring that this second- and third-tier
analysis feeds directly into an all-source product, rather than
resulting in an intermediate product that contains information from
other sources but is not actually or officially all-source. This
maximizes the productivity of the analysts and provides the
customer with a faster and more comprehensive product.
The role of the CIA as the premier analysis and production
agency should be reinforced. The DDCI who manages the CIA should
also have primary responsibility for coordinating the community's
analytical efforts, to include determining when and for what
competitive analysis is justified. Most of the DCI's centers will
remain in the CIA except for those associated almost exclusively
with the current DO, which will become part of the Clandestine
Service (see the Intelligence Centers staff study for more
details). The CIA will also be the home of the National
Intelligence Officers (although one or two may reside elsewhere, at
DIA or State) and will be responsible for sponsoring the production
of National Intelligence Estimates when they are warranted. The
other role currently performed by the National Intelligence
Council, that of evaluation, should be assumed by a new
organization, the NIEC, which is independent of the CIA and is
chartered to evaluate both analysis/production and collection
against requirements. This evaluation activity needs to be linked
directly to both the community requirements management, collection
management and the program management activities (see the
Intelligence Requirements Process staff study for more details),
with the results of the evaluations going directly to the DCI, the
DDCI managing the CIA, the DDCI for Community Management and the
DMI.
Recommendations:
1) Move towards a "virtual analytical environment" within the IC
that electronically links collectors, exploiters, analysts,
and, where appropriate, customers.
2) Move second- and third- tier exploitation and analysis, either
physically or electronically, to the primary all-source
analytical agencies, CIA and DIA.
3) Create a National Intelligence Evaluation Council (NIEC) for
evaluating IC-wide collection and production, working closely
with the Community Management Staff. The Head of the NIEC
should be appointed by the DCI and report directly to him.
VII. Planning, Programming, and Budgeting
The vast majority of the NFIP budget is embedded in the DoD
budget. This was done partially for security reasons, in the case
of the CIA, but there are practical and historical reasons for this
as well. The DoD provides 86 percent of the personnel who conduct
intelligence activities, both military and civilian. Of the
statutory elements of the NFIP, only six do not belong to DoD: the
CMS, the CIA, and the other Departmental elements belonging to the
State Department, Justice Department (FBI), Energy Department and
Treasury Department. The "fungibility" of defense dollars -- i.e.,
the fact that every dollar saved in intelligence can be used to
fund other defense programs -- prompts concerns about the
motivation of DoD (and Congress) to adequately fund intelligence in
light of competing defense priorities. This raises the question as
to whether it might not be better for intelligence and the nation
to separate intelligence funding from defense funding, either
completely or partially.
Attempting to separate the intelligence budget from the
defense budget entirely would be extraordinarily difficult, and,
philosophically, it is difficult to argue that intelligence does
not belong in the defense account. In the view of the study group,
under no circumstances is it practical or advisable to separate the
joint and tactical intelligence programs from the rest of the force
structure that they support, so, at most, it would be part or all
of the NFIP that could be moved. However, we also believe that
moving intelligence activities out of DoD would result in increased
costs to the community that are now borne as services of common
concern by DoD. Although the programs would be immune to the
occasional across-the-board unallocated reductions applied to all
DoD programs, the costs of not being part of DoD would probably far
outweigh any savings in this regard. Another implication of this
change would be that the total amount of the intelligence budget
would, in all likelihood, have to be declassified. Although sound
arguments can be made for declassifying the top line of the budget,
and the SECDEF may make the decision to do this, the study group
remains of the opinion that this would inevitably lead to the
disclosure of more information about the IC than would be prudent.
If the goal of separating intelligence funding from the
defense budget is to "protect" the NFIP, within the Executive
Branch it is already, to all intents and purposes, protected. NFIP
dollars, once identified, are effectively fenced. Executive Order
12333 tasks the DCI to:
"(n) develop, with the advice of the program managers and
departments and agencies concerned, the consolidated National
Foreign Intelligence Program budget, and present it to the
President and Congress;
(o) Review and approve all requests for reprogramming National
Foreign Intelligence Program funds, in accordance with guidelines
established by the Office of Management and Budget;
(p) Monitor National Foreign Intelligence Program implementation,
and, as necessary, conduct program and performance audits and
evaluations." The National Security Act of 1947, as amended,
states that the SECDEF shall:
"(2) ensure appropriate implementation of the policies
and resource decisions of the Director of Central
Intelligence by elements of the Department of Defense
within the National Foreign Intelligence Program."
DoD internal guidance (Carlucci memorandum of April 17 1981) stated
the policy that NFIP "resources are 'fenced' and they are not to be
increased, decreased, or transferred at any point in the fiscal
cycle unless such action has been officially coordinated with the
DCI." This policy is deemed to continue and has never been
seriously challenged. Thus, the concept of the NFIP as a fenced
program is well-established and accepted in the Executive Branch.
The greatest risk to the NFIP comes from the Legislative Branch,
which is currently free to "trade" intelligence dollars for defense
dollars in the appropriations process.
One way to address this problem would be to create a separate
line in the President's budget for intelligence. A separate line
would lead to either an Intelligence and Defense Appropriations
Bill or a completely separate appropriations bill (and
appropriations subcommittee) for intelligence. However, separating
intelligence from the rest of DoD (and, by inference, the other
departments) into a separate appropriations bill, as was done with
Military Construction some time ago, could well make the
intelligence appropriations bill more vulnerable to political and
fiscal winds, without the "cover" of the larger DoD appropriation.
In all, the study group believes that it makes the most sense to
leave NFIP funding in the various departments' budgets, but
recommend a rules change within the House of Representatives that
establishes some kind of a firewall between intelligence and
defense funding in the appropriations process.
Assuming the intelligence budget is to remain in the defense
budget, the question of how many mini-intelligence budgets there
should be remains. There are currently three: the NFIP, the JMIP,
and TIARA. Theoretically, the TIARA programs are service-unique
and the JMIP programs support multiple services or the theater/JTF.
It is an article of faith in DoD that the military services have
the right to an organic intelligence capability as part of their
force structure to serve their unique needs. The study group does
not dispute this. This capability is logically composed of the
programs grouped into the TIARA aggregation. The JMIP was
established to provide more centralized control over intelligence
capabilities required for joint operations and that serve multiple
customers. These programs are at the intersection between national
and tactical intelligence and require a more intensive level of
management to ensure that the boundaries are "seamless." There
are, thus, logical reasons to retain both the JMIP and TIARA budget
categories; however, their composition is a different issue.
The JMIP and TIARA budgets differ mostly in how they are
constructed. Both are aggregates of MFP II programs, but while
TIARA is merely the compilation of those intelligence and
intelligence-related programs that the Services have elected to
fund, the JMIP is constructed as a formal program and the role of
the Deputy SECDEF as program executive protects the program from
being "raided" by the Services. In practice, both the JMIP and
TIARA are a hodgepodge of programs, the result of a series of
unrelated and/or compromise decisions rather than a coherent plan.
The composition of the NFIP, JMIP and TIARA was one of the nine key
issue areas being examined for presentation to the Expanded Defense
Resources Board (EDRB) for the fiscal year 1997 budget submission;
it is to be hoped that the results of that review will rationalize
the division of programs; regardless, the study group believes that
further guidance is required for DoD on the appropriate composition
of the JMIP and TIARA aggregation (see the Congressional Oversight
staff study for jurisdictional implications of these divisions).
In addition to the policy and jurisdictional issues concerning
the budget, there are serious problems with the mechanical process
as well. The Community has long suffered from a vacuum in planning
and guidance emanating from the DCI and his community-level staff.
Although DCI guidance to the various functional managers is
theoretically issued for each budget cycle, it is frequently either
not done, not received in time, and/or not specific enough to
affect the programming and budgeting of the various programs. In
addition, the requirements system for the community, although much
improved as a result of the evolution of the Needs Process, has
never been successfully linked to the resource allocation process.
Some of these issues are being addressed by the DCI and ExDir of
the CMS. The NFIP budget has not previously been built in tandem
with the DoD process; until fairly recently, there were not even
agreed upon budget categories so that expenditures could be tracked
across national and tactical programs. Assuming that most of the
intelligence budget will remain a part of the defense budget, it is
critical to apply similar processes to building the intelligence
program and budget. The current ExDir's new programming and
budgeting process is a positive step for several reasons. First,
it rests the DoD portion of the intelligence budget on a foundation
of program merit rather than relying on a good relationship between
the DCI and the SECDEF. Second, it forces the IC itself to do a
much more rigorous budget review than it has been able or tasked to
do in the past, and to integrate its review with the non-NFIP
defense intelligence programs, something that has never been done
in a systematic way. It also puts the IC on a better footing with
the Office of Management and Budget (OMB), which is beginning to
play a more active role in vetting IC budget submissions. Although
this may or may not continue, it will always be a possibility
depending on the inclination of each particular administration.
The disadvantage to this new process is perceived to be
"greater DoD control" over the IC budget. However, the DCI and his
staff control the development and review of issues and the
composition of the program that is presented to the Expanded
Defense Resources Board. Although all capabilities are included in
the EDRB review process, formal budget action for the non-DoD
programs is reserved for the DCI and review is done by the IC
Executive Committee (EXCOM). Along with the rest of the NFIP,
these activities are subject to OMB review. DoD has gained no new
powers or authorities through this new process, only more
visibility into some intelligence programs. As resources continue
to be constrained, having DoD "buy-in" to the intelligence budget
is not a bad thing. And, as has always been the case, in the final
analysis the DCI has recourse to the President if he views the
results of the process as unfair or inadequate.
A more subtle, but more important disadvantage to this process
is that it is still the "tail trying to wag the dog." Currently,
the program managers submit to the CMS a proposed budget based on
top-line guidance from the DCI that has been coordinated with the
SECDEF. The CMS does a largely surface review of the submissions
(often by personnel on temporary rotation from the agencies they
are reviewing) and may make some minor changes to accommodate DCI
priorities or some of the more vocal issue coordinators. When the
budget is finalized, it is sent to Congress as part of the
President's Budget. When the Congress authorizes and appropriates
the money, it is appropriated directly to the program managers.
The CMS has no control over -- indeed, no visibility into -- budget
execution. If the DCI is to manage the Community as a corporate
entity and ensure that resource trades are made to address
priorities, he and his staff need more authority in the
intelligence budgeting process.
Although IC funding should still be appropriated to the
various Departments, the CMS must have formal authority for
formulating the NFIP budget, including the ability to monitor
execution, withhold funds and reprogram funds within the NFIP.
Thus, the elements of the NFIP should provide budget inputs to the
CMS, but the CMS should build the budget in the functional
categories mentioned above and submit the Congressional Budget
Justification Books (CBJBs) to Congress. The authority to
reprogram should be limited to not more than five percent of the
losing agency's budget over a one-year period, subject to normal
OMB review. The ability to withhold funds as a result of execution
review should be accomplished by a formal arrangement between the
DCI and SECDEF, allowing the CMS to identify to the OSD comptroller
funds to be withheld. These recommendations require the CMS to be
significantly enlarged, and although rotational personnel should
continue to provide manpower and expertise to the staff, it must
have a robust cadre of core staff to perform these and other
functions recommended in this staff study.
The single most important change that needs to be made
concerns the organizing principle around which the budget is
constructed. Broadly speaking, the budget could be organized
around programs, missions, disciplines or functions.
Notwithstanding the existing budget structural categories, the
current budget is constructed around programs, even though each
program varies widely in mission and composition. Almost any other
solution would be an improvement; however the study group believes
that the most constructive way to build the budget is along
functional rather than programmatic or discipline lines, in the
broad categories of collection, processing and exploitation,
analysis and production, and infrastructure (to include R&D,
dissemination, etc). Building the budget this way would force the
types of trade-offs between like items that the IC has been largely
unable to achieve to date, and would eliminate the current hegemony
of the program managers in the budget process. It would also
present to Congress a more balanced picture of the budget and the
resource trades made to accommodate changing priorities. Building
the budget around disciplines hinders the cross-discipline trades
that need to occur, and building it around missions is difficult,
because so many capabilities serve multiple purposes. While
clearly any budget must start with missions and the required
capabilities to perform them, the budget would more constructively
be built around those capabilities rather than the missions
themselves.
Complicating the achievement of this goal is the community
method of budgeting and accounting itself. Although there are
standard budget accounting categories for the community, each
program defines these categories somewhat differently and has its
own unique budgeting and accounting system and infrastructure. In
addition, resource data are retrievable only under the established
budget categories, so there is no efficient way to do cross-mission
or cross-functional analyses -- for example, to determine how much
the community as a whole is spending on computer support. The
Committee has several times engaged the CMS in discussions about
how to do matrixed cost accounting so that resources could be
flexibly associated with more than one category, but designing and
implementing a system for the community that would meet those needs
while allowing the DoD agencies to maintain necessary compatibility
with DoD is not a trivial undertaking. If the CMS is given both
the responsibility and the authority for building the NFIP program
and conducting execution reviews, as it should be, a new
programming, budgeting and cost accounting methodology must
accompany these changes, which will standardize programming and
budgeting procedures across the IC.
Recommendations:
1) Retain but rationalize the NFIP, JMIP, TIARA budgets. Provide
guidance to DoD concerning the appropriate composition of
JMIP and TIARA.
2) Provide the CMS a program analysis and evaluation (PA&E) and
a limited comptroller capability which would allow them to
take responsibility for formulating and executing the NFIP
budget.
3) Provide the DCI limited authority to reprogram funds within
the NFIP, the amount not to exceed five percent of the losing
agency's budget for a one-year period (Section 14(d) of the
National Security Act).
4) Provide the CMS the ability to withhold funds through an
arrangement with the OSD comptroller.
5) Mandate that the budget be built along functional rather than
programmatic lines. Mandate and fund a new community
programming, budgeting and accounting system that can track
resources in multiple categories across the IC.
VIII. Personnel Management
The IC continues to face a major personnel crisis that it has,
thus far, not addressed in a coherent way. The mandated
downsizing, conducted as it has been on a voluntary basis, has left
holes in the workforce that cannot be filled because there is no
head room to hire new people. The demographic profiles of NSA and
DIA are a disaster waiting to happen in 5-10 years unless some way
is found to maintain a steady infusion of new blood into the
community. At the same time that the number of personnel is
declining, the cost of the remaining personnel is continually
increasing, meaning that there has been little if any real savings
associated with this painful process. As mentioned earlier, the
focus of our global interest is changing and requires a different
skill mix than the preponderance of political and military analysts
that were the bread-and-butter of the Cold War.
A related issue that cannot be ignored indefinitely is morale.
Without the creation of some head room, prospects for promotion are
grim. Without a reasonable demographic spread, meaningful career
development is virtually impossible. Again, resolving these
problems is dependent at least in part upon the ability to reduce
the current workforce faster and more selectively than the hitherto
voluntary, incentivized approaches. Further eroding morale is the
lack of clear standards in some agencies and the perception of
unfair advancement of certain segments of the population. A viable
performance appraisal system across the community is an important
step to improving this situation.
Much of the discussion about the problems in the IC, and
particularly the CIA, has revolved around the culture of the
community and how it needs to change. However, it is difficult to
change a culture by simply moving the same people around in an
agency. New blood and fresh perspectives are required, and they
can be attained in two basic ways: hiring new people, or
"borrowing" people from other agencies and sending your people to
those agencies so they come back with some new ideas. The IC
overall needs to develop a "corporate culture," and it needs to do
this primarily through personnel reform that promotes the concept
of a community of professionals rather than a loosely connected
group of agencies between which personnel movement is very
difficult, if not impossible. This was the whole idea behind the
personnel provisions of Goldwater-Nichols, which was designed
(largely successfully) to break down the walls between the insular
service personnel systems and promote a culture of "jointness."
There have been numerous studies done on personnel management
in the IC. As is pointed out in the report of the most recent
Intelligence Community Task Force on Personnel Reform, led by
Christopher Jehn, the same recommendations have been made again and
again, but never implemented. In the past, the community has been
unable to overcome the resistance of agencies or individuals to
address personnel policy issues at the community level. However,
we understand that the DCI and the Administration are drafting a
legislative proposal for inclusion in the fiscal year 1997
authorization bill that incorporates the recommendations of the
Jehn report. The study group is prepared to endorse all of these
recommendations, particularly the requirement for an effective
performance evaluation system and a coherently managed personnel
system that would promote rotations and lateral movement within the
community.
The Jehn report states that in the course of the task force's
review of current personnel systems in the IC, "four principal
problems emerged:
1) a largely dysfunctional system of performance appraisal and
management;
2) a lack of systematic career planning and professional
development across the IC;
3) the variety and complexity of the various systems; and
4) inadequate promotion of a sense of community among the agencies,
including a lack of tools and incentives for managers to promote
diversity and make full use of the intellectual and cultural
diversity in the IC's workforce."
The task force's recommendations to counteract these problems
were:
1) create an effective performance management system, encouraging
the adoption of common performance criteria and standards across
agencies;
2) employ broadbanding for compensation and position management to
give more flexibility to local managers and immediate supervisors;
3) adopt a system of systematic initial appointment and separation
management;
4) standardize recruiting practices, much of career training and
elements of the performance management system across agencies, to
include a career development program that includes joint training,
rotational assignments, and dual tracks for substantive experts and
managers.
It is important to emphasize that a performance management
system would not be identical for each agency or skill area.
However, community-wide standards for performance appraisals,
compatible pay banding systems, centrally-managed personnel
security and a career development program are essential elements
for reducing duplication and facilitating lateral movement within
the community, thus promoting jointness and improving morale. At
a minimum, the SES system should be standardized at the community
level, and a rotational assignment should be a prerequisite for
achieving SES rank except in rare circumstances. Dual tracks
should be available for those personnel who do not aspire to high
levels of management but would rather remain in specialized areas
such as clandestine operations or cryptomathematics. In addition,
we believe the DCI should be able to detail personnel within the
community as required to meet short-term surge requirements (see
Intelligence Community Surge Capability staff study). However,
this authority should be limited to no more than 180 days without
the concurrence of the parent agency.
The issue of how to reduce further the numbers of personnel is
a complicated one and no single solution will effect the required
change. Many of the recommendations in the Jehn report would, over
time, improve the community's ability to identify and terminate
poor performers, particularly if the DCI's termination authority
were expanded to the entire community. The problem is how to
address the critical time period of the next 2-5 years before these
recommendations, if implemented, could begin to have an effect.
The agencies of the IC already have certain expanded
authorities beyond those accorded to other government agencies.
They have termination authorities (although only the CIA has a
truly unambiguous termination authority), but they have no special
RIF authorities or exemptions from the rules governing RIFs of
civil service personnel. The termination authorities are not
currently used for fear of lawsuits, a not unreasonable fear in the
absence of a performance appraisal system that could produce a
documentary record and justification for action. Limited
legislative authorities, such as the two percent waiver and
directed retirements of annuity-eligible personnel, could provide
some relief but could be extremely difficult to get through
Congress because of jurisdiction, fiscal and legal challenges.
These programs need to be approached as pilot projects with the
full cooperation of OMB in order to have some chance of being
instituted, and even then cannot be guaranteed. However, it is the
belief of the study group that the importance of this issue makes
these efforts worth making and we recommend legislation for the
Fiscal Year 1997 Intelligence Authorization Act establishing pilot
programs for the two percent waiver and directed retirement of
annuity-eligible personnel. Proposals for one-time dispensations
to either reduce personnel or temporarily exceed mandated
downsizing goals in order to allow hiring of essential new
personnel were rejected because, although they may be effective in
the short term, they do not provide the DCI with tools to prevent
a recurrence of the current situation and to enable to IC to
continually restructure its workforce in response to changing
priorities and targets.
Recommendations:
1) Implement recommendations of the Intelligence Community Task
Force on Personnel Reform.
2) Standardize SES system across the community and make a
rotational assignment a prerequisite for SES rank.
3) Authorize pilot programs to further reduce numbers of
intelligence personnel, to include the waiver of the two percent
retirement penalty and directed retirement of retirement-eligible
personnel.
4) Provide the DCI enhanced control over NFIP personnel, to
include the ability to detail as required for up to 180 days.
IX. Research, Development and Acquisition
Numerous interviews, panels and hearings confirmed the need
for better management of increasingly scarce R&D dollars. Reports
by an independent review panel on NSA's Advanced Research and
Development Program, the results of the Exploitation Technology
Working Group's review of R&D efforts in the imagery processing and
exploitation field, and a wealth of anecdotal information support
the contention that advanced R&D efforts are not adequately focused
on the highest priority technical problems facing the IC. The
individual discipline staff studies identify the critical areas
requiring attention. Currently, although there is an individual on
the CMS charged with looking at Advanced Technologies, R&D efforts
remain fragmented under the control of individual program
managers. The community coordinator has no budgetary authority
and, thus, a limited effect on the various programs of the
community.
The various R&D efforts in the community require closer
coordination with the requirements management element to ensure
that R&D dollars are focused on the problems that are the most
critical, not the most topical or the easiest. It is the study
group's belief that the community also needs an R&D fund, similar
to the Military Exploitation of Reconnaissance and Intelligence
Technology (MERIT) program run by the NRO, to fund promising R&D
projects. Under this concept, a fund would be established and
elements of the IC could submit proposals on an annual basis for
low-cost, potentially high pay-off technology demonstrations or
experiments. These would be evaluated by a formally constituted
review board and the available funds allocated to the projects
based on merit. The MERIT program has been an extremely effective,
albeit limited, response to the conundrum within DoD that it is
harder to get $2 million now for a good idea than to get a $20
million project into the planning cycle for two years down the
road.
Another issue that must be addressed by the IC is the
cumbersome acquisition process and the need to find a way to keep
pace with commercial technology developments, particularly in the
automation area. Each agency has automation plans and
recapitalization plans of varying degrees of effectiveness. The
result is that the community has a bewildering mixture of
automation support hardware and software, almost none of it
compatible and little of it state of the art. An important
function of the ISO, mentioned earlier, would be to establish
standards and information architectures for the entire community,
building on the role played by the Intelligence Systems Board
today. The community also needs a centralized fund for the life-cycle
replacement and upgrade of community automation equipment,and a
contracting vehicle that does not require the full-blown DoD
procurement process to be followed.
Consistent with the move towards corporateness and
consolidation where practical and efficient, the study group
believes that many R&D and acquisition activities should be
consolidated for greater efficiency and coherence. Portions of the
NRO would form the core of a new agency, but its scope would be
broadened to include development of all reconnaissance systems,
including airborne systems, and the sensor development and
acquisition activities currently undertaken by the Directorate of
Science and Technology (DS&T) within the CIA. This agency would be
called the Technology Development Office (TDO) and would be funded
via the NFIP and the JMIP (for programs currently within the
Defense Airborne Reconnaissance Office (DARO)). The inclusion of
the DARO in this concept would facilitate the development of a
truly unified air/space reconnaissance architecture, an elusive
goal thus far. The TDO would have Section 8 acquisition
authorities for NFIP monies to ensure that the NRO's and CIA's
traditional ability to conduct streamlined acquisition is not lost,
and would serve as the acquisition executive with milestone
approval authority for the DARO programs. As with most of our IC21
proposals, this would not necessarily require the physical
relocation of these elements, but would rely upon a unified
management approach to the overall reconnaissance architecture and
sensor R&D arena.
Other areas of R&D, such as those conducted at NSA in the
signal processing area and specialized R&D in support of
clandestine HUMINT operations, would remain associated with the
agencies they specifically support, but come under greater
management review in the process of building the budget
functionally. The imagery and MASINT processing R&D currently done
at the NRO and DS&T would migrate to the TCA.
Recommendations:
1) Create a Technology Development Office that combines R&D and
procurement functions for reconnaissance and sensor
technologies, to include elements of the NRO, DARO, CIA, and
NSA. Maintain Section 8 authorities for NFIP funds; serve as
acquisition executive for DARO programs.
2) Establish a MERIT-like contingency fund for the IC to exploit
technological targets of opportunity.
3) Establish a fund and a funding mechanism for rapid and
continuous update of information systems and automation
technologies.
4) Empower the Infrastructure Support Office (ISO) to establish
standards and develop architectures for the IC. Make the ISO
responsible for the life-cycle management of community ADP
systems.
------------------------------
FOOTNOTES
/1/The INFOSEC function, that is currently a non-NFIP MFP
III program, could also be managed by this consolidated activity
in better cooperation with communications and ADP; it could
remain at physically at NSA or the TCA, as later discussed, to
continue to enjoy the synergy between the "makers and the
breakers" of codes, but would respond to community direction.
Funding could be split between JMIP and TIARA, and management
coordinated with the DMI staff and DMI.