INTL613Wk2

September 11 and the Adaptation Failure of U.S. Intelligence Agencies

Author(s): Amy B. Zegart

Source: International Security , Spring, 2005, Vol. 29, No. 4 (Spring, 2005), pp. 78-111

Published by: The MIT Press

Stable URL: http://www.jstor.com/stable/4137498

REFERENCES Linked references are available on JSTOR for this article: http://www.jstor.com/stable/4137498?seq=1&cid=pdf- reference#references_tab_contents You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to International Security

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

September II and the Adaptation Failure of

U.S. Intelligence Agencies

Amy B. Zegart

In January 2000, al- Qaida operatives gathered secretly in Malaysia for a planning meeting. The Central Intelligence Agency (CIA) was watching. Among the participants was Khalid al-Mihdhar, one of the hijackers who would later help to crash Ameri- can Airlines flight 77 into the Pentagon. By the time the meeting disbanded, the CIA had taken a photograph of al-Mihdhar, learned his full name, obtained his passport number, and uncovered one other critical piece of information: al- Mihdhar held a multiple-entry visa to the United States.' It was twenty months before the September 11, 2001, terrorist attacks on the World Trade Center and the Pentagon. George Tenet, the director of central intelligence (DCI), later admitted that the CIA should have placed al-Mihdhar on the State Department's watch list denying him entry into the United States.2 It did not until August 23, 2001, just nineteen days before the terrorist attacks and months after al-Mihdhar had entered the country, obtained a California motor vehicle photo identification card (using his real name), and started taking flying lessons.

The case of Khalid al-Mihdhar provides a chilling example of the subtle yet powerful effects of organization-that is, the routines, structures, and cultures that critically influence what government agencies do and how well they do it. Why did the CIA take so long to put this suspected al-Qaida operative on the State Department's watch list, especially given Director Tenet's earlier declara- tion that the United States was "at war" with al-Qaida, and when U.S. intelli-

Amy Zegart is Assistant Professor of Public Policy at the School of Public Affairs at the University of Cali- fornia, Los Angeles.

The author would like to thank Bruce Berkowitz, Richard Betts, Lynn Eden, Stephen Krasner, Eric Patashnik, Charles Perrow, Scott Sagan, and the journal's anonymous reviewers for their insights and help. She is also indebted to the fifty U.S. government officials who agreed to be interviewed for this project on the condition they remain anonymous.

1. Eleanor Hill, "The Intelligence Community's Knowledge of the September 11 Hijackers Prior to September 11, 2001," testimony before the Senate Select Committee on Intelligence [SSCI] and House Permanent Select Committee on Intelligence [HPSCI], 107th Cong., 2nd sess., September 20, 2002, p. 6. 2. George Tenet, testimony before the SSCI and HPSCI, 107th Cong., 2nd sess., October 17, 2002.

International Security, Vol. 29, No. 4 (Spring 2005), pp. 78-111 @ 2005 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology.

78

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies ? 79

gence reporting throughout the spring and summer of 2001 revealed a dramatic spike in "chatter" about an upcoming terrorist attack?3 The simplest answer is that keeping track of the whereabouts of foreign terrorists had never been standard practice or a high priority. For more than forty years, the Cold War had dominated both the thinking and operation of the CIA and the thir- teen other agencies of the U.S. intelligence community.4 When the Soviet Union fell in 1991 and the principal threat to U.S. national security changed, U.S. intelligence agencies were slow to change with it. Before September 11, none of these agencies had formal training programs or well-honed proce- dures to assist their intelligence officers in identifying dangerous terrorists and warning other U.S. government agencies about them before they reached the United States.s As one CIA employee told congressional investigators a year after the September 11 attacks, he believed it was "not incumbent" even on the CIA's Osama bin Laden unit to place individuals such as al-Mihdhar on the State Department's watch list.6

No organization is failure-proof, and no one will ever know whether the World Trade Center and Pentagon attacks could have been prevented. Evi- dence suggests, however, that the U.S. intelligence community showed a stun- ning inability to adapt to the rise of terrorism after the Cold War ended.

This article attributes the adaptation failure of U.S. intelligence agencies to three factors: the nature of bureaucratic organizations, which makes internal

3. Tenet issued the memo on December 4, 1998.Excerpts can be found in Eleanor Hill, "Joint Inquiry Staff Statement, Part I," SSCI and HPSCI, 107th Cong., 2nd sess., September 18, 2002, p. 12. For threat reporting, see The 9/11 Commission Report: Final Report of the National Commission on Ter- rorist Attacks upon the United States (New York: W.W. Norton, 2004), pp. 254-263. 4. In addition to the CIA, the intelligence community consists of the Federal Bureau of Investiga- tion, the Defense Intelligence Agency, the National Geospatial-Intelligence Agency (formerly the National Imagery and Mapping Agency), the National Reconnaissance Office, the National Secu- rity Agency, and intelligence units in the Air Force, the Army, the Coast Guard, the Navy, the Ma- rine Corps, the State Department, the Energy Department, and the Treasury Department. In 2003 the creation of the Department of Homeland Security added a fifteenth agency to the community. 5. Hill, "The Intelligence Community's Knowledge of the September 11 Hijackers Prior to Septem- ber 11, 2001," pp. 7-8. The Joint Inquiry unearthed a December 1999 memo providing general gui- dance about placing terrorists on the State Department's watch list, but the investigation found no routine or formal programs in place to ensure that adding names to the watch list received regular attention. Nor were there any consistent guidelines about thresholds for placing suspected terror- ists on the watch list. Tenet testimony before SSCI and HPSCI, October 17, 2002; Joint Inquiry into Intelligence Community Activities before and after the Terrorist Attacks of September 11, 2001, report of the U.S. Select Committee on Intelligence and U.S. House Permanent Select Committee on Intelli- gence, S. Report No. 107-351, H. Report No. 107-792, 107th Cong., 2d sess., December 2002 [here- after Joint Inquiry Final Report], p. 147. 6. Hill, "The Intelligence Community's Knowledge of the September 11 Hijackers Prior to Septem- ber 11, 2001," p. 8.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 180

reform exceedingly difficult; the self-interest of presidents, legislators, and government bureaucrats, which works against executive branch reform; and the fragmented structure of the federal government, which erects high barriers to legislative reform.

The first section of the article considers whether the U.S. intelligence com- munity adapted as well as could be expected during the 1990s, given the chal- lenges and constraints that it faced. The second section examines the literature on organizational change and develops a framework for understanding why organizations fail to adapt. The third section, a case study of the CIA, describes how and why the agency adapted poorly to the growing terrorist threat be- tween 1991 and 2001. The fourth section offers three conclusions from the pre- ceding analysis. First, major reform of the U.S. intelligence community is difficult even after catastrophic failure. Second, reform is likely to continue lagging behind external environmental demands. And third, dramatic im- provements in U.S. intelligence capabilities require changing organizational routines and cultures as well as structures.

Determining Adaptation Failure: Change versus Adaptation

The failure of U.S. intelligence agencies to meet the terrorist threat following the end of the Cold War was not immediately apparent. Some foreign policy leaders and intelligence officials argue that the dangers of the post-Cold War world were too opaque, too numerous, and too fluid for U.S. intelligence agen- cies to assess the terrorist threat more effectively than they did. According to this view, the danger posed by al-Qaida and other terrorist organizations may appear obvious in hindsight, but this was not the case before the September 11 attacks. As President Bill Clinton's national security adviser, Samuel Berger, put it, "History is written through a rearview mirror but it unfolds through a foggy windshield."'

Others point to evidence that U.S. intelligence agencies did, in fact, recog- nize the gravity of the terrorist threat, allocating resources and launching new programs to combat it years before September 11. According to former DCI Robert Gates, the U.S. intelligence community began to shift resources away from Soviet-related missions soon after the fall of the Soviet Union. Whereas in

7. Samuel L. Berger, testimony before the National Commission on Terrorist Acts upon the United States, eighth public hearing, "Counterterrorism Policy," March 24, 2004, http://www .9-11 commission.gov/hearings/hearing8/berger_statement.pdf.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 81

1980, 58 percent of all intelligence resources was devoted to Soviet-related is- sues, by 1993, the figure had dropped to just 13 percent.8 Although specific budgets are classified, resources appear to have been redirected to combat ter- rorism. Despite declining intelligence budgets during the 1990s,9 direct spend- ing on counterterrorism roughly quintupled."'

In addition, the CIA, the Federal Bureau of Investigation (FBI), and other intelligence agencies launched a number of new counterterrorism initiatives. These included the creation, in January 1996, of a special multiagency intelli- gence unit to track the activities of bin Laden and his network"; dramatic in- creases in the number of FBI offices overseas, with a focus on countries critical

to fighting terrorism'2; and a concerted effort to forge closer relationships with foreign intelligence services, which resulted in the disruption of terrorist cells in roughly twenty countries after 1997."3 As Director Tenet concluded in Feb- ruary 2002, "This community has worked diligently over the last five years, and the American people need to understand that with the resources and au- thorities and priorities, the men and women of the FBI and the CIA performed heroically." Tenet, in fact, strongly objected to the idea that the September 11 terrorist attacks signified an intelligence failure, adding, "When people use the word 'failure'--'failure' means no focus, no attention, no discipline-and those were not present in what either we or the FBI did here and around the world."'4 Tenet was not the only senior intelligence official to resist accusa- tions that the September 11 attacks represented failure." When asked how well the intelligence community had adapted to meet the terrorist threat, for exam- ple, another senior intelligence official answered, "I think before September 11th, I would have said exceptionally well.... [now] I think we've done very, very well."'6 As these statements and examples indicate, U.S. intelligence

8. Robert Gates, quoted in John H. Hedley, "The Intelligence Community: Is it Broken? How to Fix It?" Studies in Intelligence, Vol. 39, No. 5 (1996), p. 14. 9. National foreign intelligence program budgets declined every year between 1990 and 1996 and, with the exception of one large 1999 supplemental, remained basically flat between 1996 and 2000. The 9/11 Commission Report, p. 93. 10. Joint Inquiry Final Report, pp. 254, 257. 11. Ibid., p. 4. 12. The 9/11 Commission Report, p. 76. 13. This information was provided to the Joint Inquiry by former National Security Adviser Sam- uel Berger. See ibid., p. 12. 14. George Tenet, "Worldwide Threat--Converging Dangers in a Post 9/11 World," testimony be- fore the SSCI, 107th Cong., 1st sess., February 6, 2002. 15. See in particular Thomas Powers, "The Trouble with the CIA," New York Review of Books, Janu- ary 17, 2002, http://www.nybooks.com/articles/15109. 16. Interview, Washington, D.C., April 17, 2002.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 182

agencies did make some internal changes in response to the end of the Cold War and the rise of the terrorist challenge.

Change, however, is not the same as adaptation. As sociologists have long pointed out, organizations are always changing.'7 The key issue is whether those changes matter, or more precisely, whether the rate of change within an organization keeps pace (or lags behind) the rate of change in its external envi- ronment.'8 Manifestation of this concept is more easily observed in the private sector, where responding to shifting market forces, consumer tastes, and com- petitive pressures can mean life or death for a firm. The concept may be less obvious, but no less important, for evaluating public sector organizations. The question is not: Are you doing anything differently today? But: Are you doing enough differently today to meet the challenges you face? Adaptation must be judged relative to external demands.

In the case of U.S. intelligence agencies, determining adaptation failure re- quires answering three questions: (1) did intelligence officials and policy- makers recognize the gravity of the threat posed by al-Qaida before September 11, and if so, when? (2) did they understand the connection between the terror- ist threat and the imperative for organizational change? and (3) to what extent did they achieve the organizational changes they believed were necessary? Timing is a crucial component. September 11 post mortems may highlight or- ganizational deficiencies, but to make a strong case for adaptation failure, it is necessary to show that before September 11, intelligence officials and policymakers identified key organizational problems but did not succeed in fixing them.

The answers to the three questions above appear to be yes, yes, and only to a small degree. Many U.S. intelligence officials and policymakers recognized the threat, but were unable to achieve the intelligence reforms they believed were vital several years before the World Trade Center and Pentagon attacks.

RECOGNIZING THE THREAT

In 1994 the DCI began delivering unclassified annual threat assessments to Congress. Analysis of these reports reveals that terrorism was identified as a significant threat to the United States every year from 1994 to 2001.19 By 1998

17. James G. March, "Footnotes to Organizational Change," Administrative Science Quarterly, Vol. 26, No. 4 (December 1981), p. 563. 18. Michael T. Hannan and John Freeman, "Structural Inertia and Organizational Change," Ameri- can Sociological Review, Vol. 49, No. 2 (April 1984), p. 151. 19. R. James Woolsey, "World Trouble Spots," testimony before the SSCI, 103d Cong., 2d sess., Jan-

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 183

terrorism ranked in the top tier of threats, alongside other transnational dan- gers such as the proliferation of weapons of mass destruction. The U.S. intelli- gence community became aware of bin Laden in the early 1990s, soon after he had founded al-Qaida, and was aggressively collecting intelligence on him by 1996.2() A number of terrorist attacks and plots from 1991 to 2001 associated with Islamist groups also raised the profile of terrorism within the intelligence community and indicated that targets included the U.S. homeland.21

In 1998 concern for and warnings about a possible al-Qaida attack reached a heightened level inside the community. In February bin Laden issued a public fatwa encouraging attacks on Americans anywhere in the world.22 In May he discussed "bringing the war home to America" in a public press conference.23 And in August his terrorist network succeeded in carrying out two sophisti- cated, simultaneous, and devastating truck bomb attacks against the U.S. em- bassies in Nairobi, Kenya, and Dar es Salaam, Tanzania, killing 224 people and injuring 5,000 more.24 Over the course of the year, U.S. intelligence agencies also received a number of reports indicating possible al-Qaida terrorist plots inside the United States. These events led CIA Director Tenet in December

1998 to issue a memo declaring war against Osama bin Laden. He wrote, "We must now enter a new phase in our effort against Bin Ladin. . . . We are at war ... I want no resources or people spared in this effort, either inside CIA or the Community."25 The House and Senate Intelligence Committees' Joint In-

uary 25, 1994; R. James Woolsey, "World Threat Assessment Brief," testimony before the SSCI, 104th Cong., 1st sess., January 10, 1995; John Deutch, "Worldwide Threats to U.S. National Secu- rity," testimony before the SSCI, 104th Cong., 2d sess., February 22, 1996; George Tenet, "World- wide Threats to National Security," testimony before the Senate Armed Services Committee, 105th Cong., 1st sess., February 6, 1997; George Tenet, "Worldwide Threats to National Security," testi- mony before the SSCI, 105th Cong., 2d sess., January 28, 1998; George Tenet, "Worldwide Threats to U.S. National Security," testimony before the Senate Armed Services Committee, 106th Cong., 1st sess., February 2, 1999; George Tenet, "Annual Assessment of Security Threats against the United States," testimony before the SSCI, 106th Cong., 2d sess., February 2, 2000; and George Tenet, "Worldwide Threats to National Security," testimony before the SSCI, 107th Cong., 1st sess., February 7, 2001. 20. Hill, "Joint Inquiry Staff Statement, Part I." 21. Among these were the 1993 World Trade Center attack, the 1993 New York City landmarks plot; the 1995 Bojinka plot, in which plans were uncovered in the Philippines to assassinate Pope John Paul II, bomb the U.S. and Israeli embassies in Manila, blow up several American aircraft flying Asian routes, and crash an airplane into CIA headquarters; and the 1999 disruption of the millennium attack against Los Angeles International Airport. 22. For an English translation of the fatwa, see http://www.fas.org/irp/world/para/docs/ 980223-fatwa.htm.

23. Hill, "Joint Inquiry Staff Statement, Part I," p. 9. 24. The 9/11 Commission Report, p. 70. 25. Quotation from memo included in Hill, "Joint Inquiry Staff Statement, Part I," p. 12.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 184

quiry into the September 11 attacks, a ten-month investigation that examined nearly 500,000 pages of documents and conducted 300 interviews, concluded that "Bin Ladin's declaration of war in February 1998 and intelligence reports indicating possible terrorist plots inside the United States did not go unnoticed by the Intelligence Community, which, in turn, advised senior officials in the U.S. Government of the serious nature of the threat."26

Tenet maintained his heightened level of concern in public addresses over the next three years. In 1999 he testified in open session before the Senate Armed Services Committee: "Looking out over the next year . . . there is not the slightest doubt that Usama Bin Ladin, his worldwide allies, and his sympa- thizers are planning further attacks against us .... I must tell you we are con- cerned that one or more of Bin Ladin's attacks could occur at any time."27 In 2000 he told the Senate Select Intelligence Committee, again in open session, "Everything we have learned recently confirms our conviction that [bin Laden] wants to strike further blows against the United States.... We still be- lieve he could strike without additional warning."28 In his 2001 public threat assessment, the DCI bluntly warned, "The threat from terrorism is real, imme- diate, and evolving."29As one senior U.S. intelligence official lamented after September 11, "You know, we've been saying it forever, [bin Laden] wants to bring the fight here." "

Public statements and actions by policymakers during the 1990s suggest that they shared the intelligence community's assessment of the growing terrorist threat long before September 11. In 1993, after Islamic terrorists detonated a bomb in the World Trade Center parking garage, killing 6 and wounding more than 1,000, Attorney General Janet Reno declared that terrorism had become a major threat to U.S. national security interests.3' Beginning in 1994, President Clinton mentioned terrorism in every one of his State of the Union addresses. In June 1997 the FBI's chief of International Terrorism Operations warned about the threat of an Islamic attack on U.S. soil in a public speech.32 That same

26. Hill, "Joint Inquiry Staff Statement, Part I," p. 12. 27. Tenet, "Worldwide Threats to U.S. National Security," testimony before the Senate Armed Ser- vices Committee, 106th Cong., 1st sess., February 2, 1999. 28. Tenet, "Annual Assessment of Security Threats against the United States." 29. Tenet, "Worldwide Threats to National Security," testimony before the Senate Select Commit- tee on Intelligence, 107th Cong., 1st sess., February 7, 2001. 30. Interview, Washington, D.C., April 17, 2002. 31. R. James Woolsey, remarks to the American Bar Association, Washington, D.C., April 29, 1994. 32. John P. O'Neill, excerpts from a speech to the National Strategy Forum, Chicago, Illinois, June 11, 1997, http:/ /www.nationalstrategy.com/speakers/oneill.html.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 185

year, two different strategic assessments, the Pentagon's Quadrennial Defense Review and a report by the National Defense Panel, also included strong warnings about terrorist threats to the U.S. homeland.33 In 1998 President Clinton delivered a major address at the opening session of the United Nations General Assembly in which he issued a forceful call to combat terrorism. Re- ferring to terrorism as "a clear and present danger," the president said that the issue ranked "at the top of the American agenda and should be at the top of the world's agenda."34 Less than a year later, U.S. Secretary of Defense William Cohen wrote an op-ed in the Washington Post in which he predicted a terrorist attack on U.S. soil.35 In sum, the U.S. intelligence community's assessments of a growing terrorist threat did not go unnoticed. Senior policymakers across the national security establishment appear to have agreed with them.

UNDERSTANDING THE IMPERATIVE FOR ORGANIZATIONAL CHANGE

Evidence also suggests that U.S. policymakers understood the need for organi- zational changes to meet the terrorist threat. Between the fall of the Soviet Union and the September 11 attacks, six bipartisan blue-ribbon commissions,36 three major unclassified governmental initiatives,37 and three think tank task

33. National Defense Panel, Transforming Defense: National Security in the 21st Century, December 1, 1997, http://www.fas.org/man/docs/ndp; and William S. Cohen, U.S. Department of Defense, Report of the Quadrennial Defense Review, May 1997, http://www.defenselink.mil/pubs/qdr. 34. President William Clinton, "Remarks to the Opening Session of the 53rd United Nations Gen- eral Assembly, United Nations," New York, New York, September 21, 1998, http://www.state .gov/ www/ global/ terrorism/ 980921 pres_terror.html. 35. William S. Cohen, "Preparing for a Grave New World," Washington Post, July 26, 1999, p. A19. 36. Commission on the Roles and Capabilities of the United States Intelligence Community [Aspin-Brown commission], Preparing for the 21st Century: An Appraisal of U.S. Intelligence (Wash- ington, D.C.: U.S. Government Printing Office [U.S. GPO], 1996); Commission to Assess the Orga- nization of the Federal Government to Combat the Proliferation of Weapons of Mass Destruction [Deutch commission], Combating Proliferation of Weapons of Mass Destruction (Washington, D.C.: U.S. GPO, 1999); Advisory Panel to Assess Domestic Response Capabilities for Terrorism In- volving Weapons of Mass Destruction [1999 Gilmore commission], First Annual Report to the Presi- dent and the Congress: Assessing the Threat (Washington, D.C.: U.S. GPO, 1999); Advisory Panel to Assess Domestic Response Capabilities for Terrorism Involving Weapons of Mass Destruction [2000 Gilmore commission], Second Annual Report to the President and the Congress: Toward a National Strategy for Combating Terrorism (Washington, D.C.: U.S. GPO, 2000); National Commission on Ter- rorism [Bremer commission], Countering the Changing Threat of International Terrorism (Washington, D.C.: U.S. GPO, 2000); U.S. Commission on National Security/21st Century [Hart-Rudman com- mission], Road Map for National Security: Imperative for Change (Washington, D.C.: U.S. GPO, 2001); and Commission on the Advancement of Federal Law Enforcement [Webster commission], Law Enforcement in a New Century and a Changing World: Improving the Administration of Federal Law En- forcement (Washington, D.C.: U.S. GPO, 2000). 37. National Performance Review, "The Intelligence Community: Recommendations and Ac- tions," in From Red Tape to Results: Creating a Government That Works Better and Costs Less (Washing-

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 186

forces38 examined the U.S. intelligence community, the FBI, and U.S. counter- terrorism efforts. The commissions were chaired by well-respected leaders such as former Defense Secretary Harold Brown, Ambassador Paul Bremer, former Senators Gary Hart and Warren Rudman, and William Webster, former director of both the CIA and FBI. The three reports issued by governmental ini- tiatives were President Clinton's 1993 and 1995 interagency National Perfor- mance Reviews; a 1996 House Intelligence Committee staff study that was the most comprehensive review of the intelligence community since the Church committee investigations into CIA abuses of the 1970s; and the FBI's 1998 Draft Strategic Plan. The think tank reports were produced by the Council on For- eign Relations, the Twentieth Century Fund, and the National Institute for Public Policy. All twelve were detailed, substantive examinations.

Although the reports addressed a variety of intelligence issues and prob- lems, the common theme was the need for major change. The Council on For- eign Relations report noted in 1996 that "the intelligence community has been adjusting to the changed demands of the post-Cold War world for several years ... [but] additional reform is necessary."39 The report listed nearly forty recommendations that ranged from significant structural reforms to changes in personnel recruiting, training, and assignments. The 1996 House Intelligence Committee staff study found that the intelligence community suffered from a lack of "corporateness," or integration between individual agencies. The re- port noted, "Only intelligence, of all major government functions, is carried out by a very disparate number of agencies and organizations that are either independent of one another or housed in separate departments by officials whose main concerns are policy, not intelligence."4) In particular, the report criticized what it saw as "the glaring gap" between the DCI's responsibilities

ton, D.C.: U.S. GPO, September 1993); National Performance Review, National Performance Review Phase II Initiatives: An Intelligence Community Report (Washington, D.C.: U.S. GPO, 1995); House Permanent Select Committee on Intelligence, IC21: The Intelligence Community in the 21st Century (Washington, D.C.: U.S. GPO, 1996); and Federal Bureau of Investigation, Draft FBI Strategic Plan: 1998-2003, Keeping Tomorrow Safe (Washington, D.C.: U.S. GPO, 1998). 38. Council on Foreign Relations Independent Task Force, Making Intelligence Smarter: The Future of U.S. Intelligence (New York: Council on Foreign Relations, 1996); Twentieth Century Fund Task Force on the Future of U.S. Intelligence, In from the Cold: The Report of the Twentieth Century Fund Task Force on the Future of U.S. Intelligence (New York: Twentieth Century Fund Press, 1996); and National Institute for Public Policy, Modernizing Intelligence: Structure and Change for the 21st Cen- tury (Fairfax, Va.: National Institute for Public Policy, 2002). 39. Council on Foreign Relations Independent Task Force, Making Intelligence Smarter, p. 1. 40. House Permanent Select Committee on Intelligence, IC21, "Overview and Summary," p. 5.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 187

and his authorities,41 the "fundamental and urgent" need to improve the intel- ligence requirements process that sets agency priorities,42 and the "internecine competition" between the various intelligence collection disciplines such as signals intelligence and human intelligence.43 It issued eighty-two recom- mended reforms. Four years later, the Bremer commission warned that "inter- national terrorism poses an increasingly dangerous and difficult threat to America." It urged the government to take immediate "steps to reinvigorate the collection of intelligence about terrorists' plans."44 The commission's key recommendations included clarifying the FBI's authorities to investigate suspected terrorist groups; rescinding CIA guidelines that hindered the re- cruitment of terrorist informants; giving higher funding priority to counterter- rorism efforts in the CIA, FBI, and National Security Agency; and establishing a new cadre of reports officers to distill and disseminate terrorism-related information quickly once it is collected. Indeed, the commission noted with concern that "U.S. intelligence and law enforcement communities lack the ability to prioritize, translate, and understand in a timely fashion all of the in- formation to which they have access."45 In total, the reports issued 340 recom- mendations to improve U.S. intelligence capabilities."4 The imperative for organizational change was clear.

FAILING TO CHANGE

To what extent were the studies' recommendations achieved? To be sure, gaug- ing adaptation failure by examining the adoption of study recommendations has its limitations. Commissions may be created for the sole purpose of deflecting blame or delaying action rather than generating change, though this is far less often the case in national security affairs than most believe.47 Even earnest efforts at reform often take a variety of forms, with some focusing on "the art of the possible" and others proposing more ideal and unlikely solu-

41. Ibid., p. 7. 42. Ibid., p. 26. 43. Ibid., p. 28. 44. National Commission on Terrorism, Countering the Changing Threat of International Terrorism, p. iv. 45. Ibid., p. 13. 46. The Gilmore commission recommendations are from its 1999 and 2000 reports. The panel is- sued two other reports after September 11 that were not included. 47. Amy Zegart, "Blue Ribbons, Black Boxes: Toward a Better Understanding of Presidential Com- missions," Presidential Studies Quarterly, Vol. 35, No. 2 (June 2004), pp. 366-393.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 188

tions. Examining the totality of study recommendations and their success, however, has the advantage of providing a rough, macro view of adaptation that does not rely on hindsight or impose ex post personal judgments of which reforms were better ideas than others. Asking what major consensus recom- mendations were made and whether these were implemented provides a use- ful, systematic first cut at the problem that goes beyond anecdotal evidence of failure.

The data indicate a widespread inability of U.S. intelligence agencies to adapt to the terrorist threat before the September 11 attacks. Of 340 recommen- dations for changes in the intelligence community, only 35 were successfully implemented, and 268-or 79 percent of the total-resulted in no action at all. Closer examination reveals surprising agreement on four major problems: the intelligence community's lack of coherence or "corporateness"; insufficient hu- man intelligence; personnel systems that failed to align intelligence needs with personnel skills or encourage information sharing; and weaknesses in setting intelligence priorities. The twelve studies provided a menu of options to address these commonly identified problems. Yet policymakers left the intelli- gence community essentially intact. As the Hart-Rudman commission con- cluded in January 2001, "The dramatic changes in the world since the end of the Cold War ... have not been accompanied by any major institutional changes in the Executive branch of the U.S. government."48 The commission presciently predicted that these institutional deficiencies left the United States exceptionally vulnerable to a catastrophic terrorist attack.49

Explaining Failed Adaptation: Crossing the Theoretical Divide

Existing academic work does not offer a ready-made explanation for the fail- ure of U.S. intelligence agencies to adapt to a changed external environment. On the one hand, organization theorists study organizational pathologies but focus almost exclusively on firms. On the other hand, political scientists exam- ine national security affairs and bureaucracy but rarely treat intelligence agen- cies as dependent variables.5" Taken together, however, these literatures

48. U.S. Commission on National Security/21st Century, Road Map for National Security, p. x. 49. U.S. Commission on National Security/21st Century, New World Coming: American Security in the 21st Century: Major Themes and Implications, September 15, 1999, p. 138. 50. Notable exceptions are Scott D. Sagan, The Limits of Safety: Organizations, Accidents, and Nuclear Weapons (Princeton, N.J.: Princeton University Press 1993); and Lynn Eden, Whole World on Fire:

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 89

provide the building blocks to construct a model of intelligence agency adap- tation failure.

THE BOTTOM LINE OF ORGANIZATION THEORY

Organization theory has long drawn a multidisciplinary crowd united by a common interest in understanding what organizations do, how they do it, and how well. Questions of organizational inertia,"' evolution,52 and learning53 have generated rich and dynamic research programs over the past thirty years. Although organization theory has much to offer, particularly about why or-

ganizations resist change even when it is clearly beneficial, the literature can- not be applied easily to the political realm. This is understandable. The field emerged with firms in mind and has remained focused on the private sector ever since. As Richard Cyert and James March noted in the introduction to their 1963 classic, A Behavioral Theory of the Firm, "We had an agenda .... We thought that research on economics and research on organizations should have something to say to each other."54 The focus on business organizations contin- ues. In his 1999 book, Organizations Evolving, Howard Aldrich criticizes the se- lection bias of organization theory. Notably, though, he is troubled by the inordinate attention paid to large publicly traded corporations instead of pri- vately held businesses. Aldrich's idea of broadening organization theory is to include different kinds of businesses, not political organizations.5s Because of

Organizations, Knowledge, and Nuclear Weapons Devastation (Ithaca, N.Y.: Cornell University Press, 2003).

51. Glenn R. Carroll, "Organizational Ecology," Annual Review of Sociology, Vol. 10 (1984), pp. 71- 93; Michael T. Hannan and John Henry Freeman, "The Population Ecology of Organizations," American Journal of Sociology, Vol. 82, No. 5 (March 1977), pp. 929-964; Michael T. Hannan and John Henry Freeman, "Structural Inertia and Organizational Change," American Sociological Review, Vol. 49, No. 2 (April 1984), pp. 149-164; and Michael T. Hannan and John Henry Freeman, Organiza- tional Ecology (Cambridge, Mass.: Harvard University Press, 1989). 52. Joel A.C. Baum and Jitendra V. Singh, Evolutionary Dynamics of Organizations (Oxford: Oxford University Press, 1994); and Howard Aldrich, Organizations Evolving (London: Sage, 1999). 53. Barbara Levitt and James G. March, "Organizational Learning," Annual Review of Sociology, Vol. 14 (1988), pp. 319-340; Richard M. Cyert and James G. March, A Behavioral Theory of the Firm (London: Blackwell, 1963); Mary Ann Glynn, Theresa K. Lant, and Frances J. Milliken, "Mapping Learning Processes in Organizations: A Multi-Level Framework Linking Learning and Orga- nizing," in James R. Meindl, Joseph F. Porac, and Chuck Stubbart, eds., Advances in Managerial Cognition and Organizational Information Processing, Vol. 5 (Greenwich, Conn.: JAI Press, 1994), pp. 43-83; Anne S. Miner and Stephen J. Mezias, "Ugly Duckling No More: Pasts and Futures of Organizational Learning Research," Organization Science, Vol. 7, No. 1 (January-February 1996), pp. 88-99; and James G. March, "Exploration and Exploitation in Organizational Learning," Orga- nization Science, Vol. 2, No. 1 (February 1991), pp. 71-87. 54. Cyert and March, A Behavioral Theory of the Firm, p. xi. 55. Aldrich's point should not be minimized. He notes that large, publicly traded corporations

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 90

the focus on firms, organization theory has developed without much attention to politics and power, forces that are crucial for understanding the develop- ment of government agencies."6

Consider, for example, the literature that appears most centrally related to the question of agency adaptation failure: population ecology. According to population ecologists, most organizations face strong pressures that make them resistant to change. Market forces, the competitive environment, technol- ogy, and consumer preferences all shift over time, but organizations are rarely good at shifting with them. Instead, population ecologists argue that innova- tions in organizational strategies and structures occur at the population level, through the birth of new organizations and the death of others. In other words, individual organizations do not adapt; populations do, with newer, fitter firms constantly replacing older, outdated ones through a process of natural selection."7

This approach is helpful in explaining the private sector, where firms rou- tinely come and go. It is much less helpful, however, in understanding the public sector, where there is substantially less organizational churn. As many scholars have observed, government agencies are notoriously hard to elimi- nate because there are always interest groups and elected officials who have vested interests at stake.5" In his pilot study comparing U.S. government agen- cies between 1923 and 1973, for example, Herbert Kaufman found that 85 per- cent of those in the 1923 sample were still in existence fifty years later.5'9

constitute less than 0.05 percent of all U.S. organizations using a corporate form. Aldrich, Organi- zations Evolving, p. 8. 56. For more on the difference between the private and public sectors, see James Q. Wilson, Bu- reaucracy: What Government Agencies Do and Why They Do It (New York: Basic Books, 2000); Terry M. Moe, "The Politics of Structural Choice: Toward a Theory of Public Bureaucracy," in Oliver E. Williamson, ed., Organization Theory: From Chester Barnard to the Present and Beyond (Oxford: Ox- ford University Press, 1990), pp. 116-153; and Terry M. Moe, "The New Economics of Organiza- tion," American Journal of Political Science, Vol. 28, No. 4 (November 1984), pp. 739-777. 57. For an overview of this approach, see Hannan and Freeman, "Structural Inertia and Organiza- tion Change"; and Hannan and Freeman, "The Population of Ecology Organizations." 58. Anthony Downs, Inside Bureaucracy (Boston: Little, Brown, 1967); Arthur Stinchcombe, "Social Structures and Organizations," in James G. March, ed., Handbook of Organizations (Chicago: Rand McNally, 1965), pp. 142-193; Herbert Kaufman, Are Government Organizations Immortal? (Washing- ton, D.C.: Brookings, 1976); and Theodore J. Lowi, The End of Liberalism: The Second Republic of the United States (New York: W.W. Norton, 1979). 59. Kaufman, Are Government Agencies Immortal? p. 34. The durability of agencies has more re- cently come into question. See David E. Lewis, Presidents and the Politics of Agency Design: Political Insulation in the United States Government Bureaucracy, 1946-1997 (Stanford, Calif.: Stanford Univer- sity Press, 2003); and Daniel P. Carpenter, The Forging of Bureaucratic Autonomy: Reputations, Net- works, and Policy Innovation in Executive Agencies, 1862-1928 (Princeton, N.J.: Princeton University Press, 2000).

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 91

By contrast, Aldrich found that in 1957, 398,000 businesses were created in the United States; about the same number were transferred to new owners;

and almost as many failed.6" In addition, population ecology takes organizational inertia as a starting as-

sumption rather than a dependent variable.61 This is a serious weakness. If population ecologists are correct, and innovations are generated through the replacement of old organizations by new ones, then the imperative to under- stand what it is that keeps any single agency from adapting to environmental demands is all the greater in the public sector: the government is likely to be riddled with poorly performing agencies that persist because they are unchal- lenged by the threat of new entrants.

Finally, most research in population ecology consists of large-n studies of entire organizational populations, with little attention given to the role of indi- vidual choice, decisionmaking processes, or politics. The result is a strangely antiseptic theory of natural selection that seems to ignore the role of real peo- ple exercising real power. As Charles Perrow notes, "It is almost as if God does the negative and positive selecting."'2

THE TALLY FOR POLITICAL SCIENCE

The political science literature suffers from a different problem. Rational choice approaches focus squarely on politics and power. Still largely unexamined, however, are the routines and cultures that develop within government bu- reaucracies and their resistance to change. Major work on U.S. government agencies in the 1980s and 1990s was produced by Congress scholars who were interested primarily in agencies as objects of congressional control, not as sub- jects of study in their own right. Using ideas from transaction cost economics, Mathew McCubbins, Barry Weingast, and other congressional dominance scholars argue that Congress controls the federal bureaucracy, and in surpris- ingly efficient ways.63 Indeed, they contend that evidence usually thought to suggest poor oversight, such as sparsely attended congressional committee

60. Howard E. Aldrich, Organizations and Environments (Englewood Cliffs, N.J.: Prentice Hall, 1979).

61. Aldrich, Organizations Evolving, p. 43. 62. Charles Perrow, Complex Organizations: A Critical Essay, 3d ed. (New York: McGraw-Hill, 1986), p. 213. 63. Mathew D. McCubbins, "The Legislative Design of Regulatory Structure," American Journal of Political Science, Vol. 29, No. 4 (November 1985), pp. 721-748; and Barry Weingast and Mark Moran, "Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Fed- eral Trade Commission," Journal of Political Economy, Vol. 91, No. 5 (October 1983), pp. 775-800.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 192

hearings, actually reveals oversight hard at work. Government bureaucrats know they must follow their congressional mandate or face the consequences. Congress therefore does not have to expend much effort monitoring them.

For the purposes of this analysis, the most serious limitation of the congres- sional dominance literature is its assumption that agencies can and do adapt. To say that Congress controls the bureaucracy is to say that agencies are, by and large, responsive to the preferences and demands of legislators. Much of the literature examines how legislators ensure that these preferences and de- mands are heeded-by building control mechanisms into the very design of government agencies, or by using (or threatening to use) existing controls such as withholding appropriations, or both.64 The mechanisms may vary, but the logic is the same: government bureaucrats usually heed legislators' demands because they know, as McCubbins puts it, that "Congress holds the power of life or death in the most elemental terms."'5 The logic also suggests that when legislative preferences change significantly, bureaucracies should change, too. There is some evidence that they do.66

Notably, even critics of the congressional dominance literature agree that government agencies can adapt. Terry Moe, for example, argues that Congress is not alone in determining agency behavior; many factors, including presi- dents, interest groups, and courts, influence what agencies do." Others argue

64. David Epstein and Sharyn O'Halloran make the distinction between ex ante and ongoing oversight mechanisms. Epstein and O'Halloran, "Administrative Procedures, Information, and Agency Discretion," American Journal of Political Science, Vol. 38, No. 3 (August 1994), pp. 697-722. For ex ante controls, see Murray J. Horn, The Political Economy of Public Administration: Institutional Choice in the Public Sector (New York: Cambridge University Press, 1995); Mathew D. McCubbins and Thomas Schwartz, "Congressional Oversight Overlooked: Police Patrol versus Fire Alarm," American Journal of Political Science, Vol. 28, No. 1 (February 1984), pp. 165-179; and Mathew D. McCubbins, Roger Noll, and Barry Weingast, "Administrative Procedures as Instruments of Politi- cal Control," Journal of Law, Economics, and Organization, Vol. 3, No. 2 (Autumn 1987), pp. 243-277; For ongoing controls, see Randall Calvert, Mark Moran, and Barry Weingast, "Congressional Influence over Policy Making: The Case of the FTC," in Mathew D. McCubbins and Terry Sullivan, eds., Congress: Structure and Policy (Cambridge: Cambridge University Press, 1987); and Randall Calvert, Mathew D. McCubbins, and Barry Weingast, "A Theory of Political Control and Agency Discretion," American Journal of Political Science, Vol. 33, No. 3 (August 1989), pp. 588-611. For how legislators choose between options, see Kathleen Bawn, "Political Control versus Expertise: Con- gressional Choices about Administrative Procedures," American Political Science Review, Vol. 89, No. 1 (March 1995), pp. 62-73; and David Epstein and Sharyn O'Halloran, Delegating Powers (New York: Cambridge University Press, 1999). 65. Mathew D. McCubbins, "The Legislative Design of Regulatory Structure," American Journal of Political Science, Vol. 29, No. 4 (November 1985), p. 728. 66. Calvert, Moran, and Weingast, "Congressional Influence over Policy Making." 67. Terry M. Moe, "An Assessment of the Positive Theory of 'Congressional Dominance,"' Legisla- tive Studies Quarterly, Vol. 12, No. 4 (November 1987), pp. 475-520.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 93

that congressional dominance does not capture the degree of bureaucratic shirking, when agencies opt to pursue their own organizational interests in- stead of the interests of their congressional overseers. As David Epstein and Sharyn O'Halloran conclude, the literature on congressional-bureaucratic rela- tions has reached something of an impasse in recent years: "The general con- sensus has been reached that legislators have more effective means of control than was previously realized, but bureaucrats still retain significant amounts of discretion in setting policy."68

All of this work suggests that agencies are out there, on the move, doing things. Usually agencies respond to congressional wishes; sometimes they pursue the interests of presidents and others in the political system; and some- times they shirk to serve their own interests. Nowhere, however, is there a sense that government agencies may be unable to change. According to this literature, the challenge is to keep them from running amok. But the greater danger may be that agencies are stuck running in place.

A MODEL OF AGENCY ADAPTATION FAILURE

Although neither organization theory nor rational choice theory offers a ready- made explanation of adaptation, each does provide the foundations for a gen- eral model of agency adaptation failure. Organization theory offers insights about intraorganizational impediments to reform, while rational choice theory helps to explain external impediments to reform. Rather than viewing government agencies from without (i.e., as more lim-

ited public sector versions of firms or as objects of congressional control), I be- gin by assuming the bureaucracy's perspective. Heads of government agencies that are confronting a changing environment must answer two questions. First, how can they get the reforms they need in their agencies to keep pace with environmental demands? And second, what forces will hinder them from

succeeding?69 There are three major sources of bureaucratic reform: (1) internal reforms

made by the agency itself, whether in memos, speeches, revised guidelines, or sanctions of undesired behavior; (2) executive branch action, for example, executive orders, presidential directives, or efforts by executive branch officials

68. Epstein and O'Halloran, Delegating Powers, p. 29. 69. I do not mean to suggest that all agency heads are so well intentioned and interested in maxi- mizing organizational performance. Instead, the device is a heuristic used to tease out the ways in which agencies can be reformed and the forces that are likely to keep reforms from succeeding.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 94

outside the agency in question, such as the National Security Council; and (3) statutory reforms that require the involvement of both Congress and the executive branch.

Impediments to reform may emerge from both inside and outside the agency. Some reforms may fail because they challenge deeply held organiza- tional values and threaten to alter established routines. Others may trigger opposition from competing government agencies that stand to gain or lose de- pending on the outcome. Proposed statutory changes that require the consent of multiple congressional majorities and the president bring institutional forces more centrally into play. Thus, developing a better understanding of agency adaptation failure requires combining the enduring realities operating within organizations with those operating outside them. More specifically, these are (1) the nature of organizations; (2) the rational self-interest of political officials; and (3) the fragmented structure of the U.S. federal government. Tak- en together, these three realities raise exceptionally high obstacles to agency change.

This approach, though not theoretically elegant, is analytically useful. The persistence of organizational habits, the unwillingness to veer from estab- lished routines, and the role of culture say little about why legislators have so rarely proposed statutory changes to the intelligence community, and why they have been opposed by colleagues in the House and Senate Armed Services Committees when they did. Conversely, examining only rational self- interest does not explain why, before the September 11 attacks, FBI agents pre- ferred to keep case files in shoeboxes under their desks rather than entering them into computer databases; nor does it explain why, for a year and half, in- telligence officials neglected to tell the State Department to include on its watch list a suspected al-Qaida operative with a multiple-entry U.S. visa in his passport.

THE NATURE OF ORGANIZATIONS. The first route to agency reform is through the adoption of internal changes. Yet much of the work in organization theory argues that organizations do not change easily by themselves.70 Examples abound. The U.S. Army kept a horse cavalry until World War II. Until the mid- 1990s, U.S. Customs forms asked ships entering American ports to list the number of cannons on board, and federal law required the U.S. Agriculture Department to keep field offices within a day's horseback ride to everyplace in

70. For an alternative view, see James G. March and Johan P. Olsen, Ambiguity and Choice in Orga- nizations (Bergen, Norway: Universitetsforlaget, 1976).

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 195

the United States. Even private sector firms, which have considerably more leeway over personnel decisions, more access to capital, and fewer manage- ment constraints than government agencies, do not fare well when changing circumstances require adjustment. Consider the most basic case: whether firms adapt enough to survive. Of the 5.5 million businesses tracked by the U.S. Cen- sus Bureau in 1990, only 3.8 million were still in existence four years later, a failure rate of 31 percent.71 In New York City, more than 60 percent of all res- taurants surveyed in the Zagat guide between 1979 and 1999 folded.72 Between 2000 and 2003 more than 400 public companies, including Enron, Global Crossing, and Kmart, declared bankruptcy. As Aldrich points out, these and other findings about organizational adaptation failure are most likely under- stated because they tend to focus only on surviving firms and exclude all the organizations that never made it past the start-up phase.73

Government agencies are even less able than private sector businesses to make internal adjustments. First, they have more constraints, facing more conflicting missions with less managerial discretion and fewer resources than private sector firms do.74 Any manager working for Coca Cola knows his mis- sion is to sell soda-the more, the better. By contrast, for years U.S. Immigra- tion and Naturalization Service officials were charged with helping some foreigners enter the United States while keeping others out. The most embat- tled chief executive officer answers only to a small group of influential inves- tors, all of whom share the same goal: maximizing shareholder value of the firm. By contrast, even the new director of national intelligence (DNI), a posi- tion created in December 2004 to manage the U.S. intelligence community, serves a variety of intelligence consumers, from the president and his senior foreign policy advisers to the secretary of defense to 535 members of Con- gress.75 Unlike a firm's shareholders, these policymakers have preferences and interests that often conflict. Caught between Democrats and Republicans, mili- tary hawks and civil libertarians, interventionists and isolationists, policy ad- vocates and opponents, the new DNI, like DCIs before him, must navigate carefully, knowing that not everyone will be satisfied with his decisions.

71. Aldrich, Organizations Evolving, p. 262. 72. Proprietary data provided by Zagat to author, February 2, 2001. 73. Aldrich, Organizations Evolving, p. 257. See also Joseph W. Duncan and Douglas P. Handler, "The Misunderstood Role of Small Business," Business Economics, Vol. 29, No. 3 (July 1994), pp. 1- 6.

74. Wilson, Bureaucracy. 75. The Intelligence Reform and Terrorism Prevention Act of 2004, Public Law 108-458, 108th Cong., 2d sess., December 17, 2004.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 196

Second, government agencies were not built to be adaptable. As James Q. Wilson and Joel Aberbach and Bert Rockman note, agencies are not designed to be especially nimble or innovative.76 Instead, they are designed to be reli- able and fair. Above all, reliability means ensuring that tasks are completed consistently and predictably, which is best guaranteed by having one organiza- tion repeatedly performing the same functions. The U.S. Postal Service has built the capacity to surge at Christmas time. Each state has a department of motor vehicles to issue driver's licenses. Assigning these tasks to a different agency every year would almost certainly result in more delays, lost mail, and perhaps even greater driver frustration. Government agencies are also de- signed to ensure fairness, by creating and adhering to rules and procedures about what benefits or services are provided to which citizens, regardless of their personal wealth, power, or political connections. The lines for govern- ment services may be long, but in principle everyone must stand in them.

Ironically, these attributes of bureaucracies create strong impediments to reform.77 Standard operating procedures guarantee that the Social Security Administration mails benefit checks on time each month, and they help to en- sure that every military pilot has the same rules of engagement in wartime. But they can also reduce the willingness and ability of agencies to deviate from established practices, even when such deviation may be more beneficial-a phenomenon that Barbara Levitt and James March refer to as a "competency trap."78 The very characteristics that give an organization reliability and fair- ness reduce the probability of change.7'

Finally, time is almost never on the side of government agencies that must adapt. All organizations become more resistant to change as routines, norms, and relationships become firmly established.8" In addition, as Herbert Simon notes, the larger the number of organizations in a given system and the greater their level of interconnectedness, the harder it is for that entire system to

76. Wilson, Bureaucracy; and Joel Aberbach and Bert Rockman, In the Web of Politics: Three Decades of the U.S. Federal Executive (Washington, D.C.: Brookings, 2000). 77. Hannan and Freeman, "Structural Inertia and Organizational Change." 78. Barbara Levitt and James G. March, "Organizational Learning," Annual Review of Sociology, Vol. 14 (1988), pp. 322-323. 79. Hannan and Freeman, "Structural Inertia and Organizational Change," pp. 149-164. 80. Major works arguing that administrative agencies become more durable as they get older in- clude Downs, Inside Bureaucracy; Kaufman, Are Government Organizations Immortal?; and Lowi, The End of Liberalism. A great deal of work suggests that government agencies are like firms and other organizations in this regard. See Stinchcombe, "Social Structures and Organizations."

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 97

adjust. Improvements must occur throughout the system at the same time to produce results.81

Government agencies that do not adapt on their own may be subjected to change from the outside, either through executive branch action or through legislation. In such cases, the rational self-interest of political actors and the fragmented structure of the federal government work to block success.

RATIONAL SELF-INTEREST OF PRESIDENTS, LEGISLATORS, AND NATIONAL SECU-

RITY BUREAUCRATS. Government officials are constrained by the incentives and capabilities that come with their positions. Although individuals have their own ideas, skills, and policy preferences, institutional incentives and capabili- ties exert a powerful influence, making some courses of action easier and less costly than others. These incentives and capabilities explain why, before the September 11 attacks, no president championed intelligence reform, why legislators largely avoided or blocked it, and why national security agency bureaucrats opposed it.

Presidents have strong incentives to improve organizational effectiveness. To make their mark on history, they must make the bureaucracy work well for them. Perhaps even more important, presidents are also driven to enhance organizational effectiveness by the electorate, which expects far more of them than they can possibly deliver. Held responsible for everything from inflation to Iraqi democratization, presidents have good reason to ensure that govern- ment agencies adapt to changing demands as much and as fast as possible.82

Presidents, however, lack the capabilities to make the changes they desire. They have little time, limited political capital, few formal powers, and packed political agendas. Presidents therefore almost always prefer to focus their efforts on policy issues that directly concern (and benefit) voters, rather than on the arcane details of organizational design and operation. Tax cuts and Social Security lockboxes win votes, but no president ever won a landslide election by changing the CIA's personnel system. Moreover, presidents are especially reluctant to push for agency reforms in the absence of a crisis or

81. Herbert Simon, "Public Administration in Today's World of Organizations and Markets," John Gaus lecture, annual meeting of the American Political Science Association, Washington, D.C., September 1, 2000, reprinted in PS: Political Science and Politics, Vol. 33, No. 4 (December 2000), pp. 752-754; and Charles Perrow, Normal Accidents: Living with High-Risk Technologies (Princeton, N.J.: Princeton University Press, 1999). 82. Terry M. Moe, "The Politicized Presidency," in John E. Chubb and Paul E. Peterson, eds., The New Direction in American Politics (Washington, D.C.: Brookings, 1985).

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 98

in the presence of anticipated resistance. Presidents are thus loath to reform existing agencies through executive action or legislation. Although dozens of investigations, commissions, and experts identified shortcomings in the U.S. intelligence community between 1947 (when the CIA was created) and the September 11 terrorist attacks, no president attempted major intelligence reform.83

Self-interest leads most legislators to avoid tackling intelligence reform alto- gether or seek to block it. Legislators, like presidents, have little incentive to delve into the bureaucratic intricacies of intelligence agency design because doing so does not provide tangible benefits to voters back home.84 Indeed, the weak electoral connection is one of the reasons congressional intelligence over- sight committees until 2004 had term limits for their members, despite the fact that such limits hinder the development of expertise and despite repeated calls by commissions to abolish them. When crises do arise, intelligence committee members are rewarded more for holding hearings than taking corrective ac- tion. The Bay of Pigs, the congressional investigations into CIA abuses during the 1970s, the Iran-Contra scandal, and the Aldrich Ames spy case all triggered major investigations, but none produced fundamental change in the intelli- gence community. In addition, members of Congress care about maintaining the power of the institution. Generally, this means that legislators prefer execu- tive arrangements that diffuse authority and capabilities; the more agencies in the executive branch, the more power bases can accrue in Congress to oversee them.

Finally, national security agency bureaucrats have their own interests at stake and powerful means to protect them. Whereas most domestic policy agencies operate in relatively autonomous spheres-the U.S. Forest Service, for example, has no reason to think about the design or operation of the Inter- nal Revenue Service-U.S. national security agencies are more tightly con- nected. Policymaking inevitably crosses bureaucratic boundaries, involving diplomacy, the use of force, economic policy, and intelligence. In such a com- plex web, national security bureaucrats see reform as a zero-sum battle

83. For a historical overview of intelligence reform efforts, see Richard A. Best Jr., Proposals for In- telligence Reorganization, 1949-2004 (Washington, D.C.: Congressional Research Service, 2004). 84. David Mayhew, Congress: The Electoral Connection (New Haven, Conn.: Yale University Press, 1974).

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 199

for agency autonomy and power. No national security agency wants to yield authority or discretion to another.85

THE PROBLEMS OF DECENTRALIZED DEMOCRACY. Rational self-interest makes

reform difficult; self-interest coupled with the decentralized structure of the U.S. federal government makes it more so. Some of the cherished features of American democracy impede effective agency design and raise obstacles to re- form. Separation of powers, the congressional committee system, and majority rule have created a system that invites compromise and makes legislation hard to pass. This has two consequences for government agencies. First, political compromise allows opponents to sabotage the creation of any new agency from the start. As Terry Moe writes, "In the political system, public bureaucra- cies are designed ... by participants who explicitly want them to fail."8 Politi- cal compromise unavoidably leads to suboptimal initial agency design, even for national security agencies such as the CIA.87 Indeed, critics who contend that the CIA is poorly suited to meeting the needs of the post-Cold War world are only partially right: the agency was not particularly well designed to meet the United States' Cold War needs. Opposed from the outset to the CIA's cre- ation in 1947, existing intelligence agencies in the FBI, State Department, and military services succeeded in stripping the agency of any strong centraliza- tion powers. When the CIA was created in 1947, it was flawed by design.8"

Second, the decentralized structure of American democracy means that the worst agency problems usually are the hardest to fix. Although agencies can make some changes on their own and can also be altered by unilateral presi- dential action, the most far-reaching reforms usually require new legislation. But legislative success is difficult even under the best of circumstances because it demands multiple majorities in both houses of Congress. As Philip Zelikow, executive director of the 9/11 commission, put it, "The most powerful interest group in Washington is the status quo.""'

SUMMARY. These three enduring realities-the nature of organizations, ratio- nal self-interest, and the fragmented federal government-provide a basic

85. Amy Zegart, Flawed by Design: The Evolution of the CIA, JCS, and NSC (Stanford, Calif.: Stanford University Press, 1999). 86. Terry M. Moe, "The Politics of Bureaucratic Structure," in John E. Chubb and Paul E. Peterson, eds., Can the Government Govern? (Washington, D.C.: Brookings, 1989), p. 326. 87. Zegart, Flawed by Design. 88. Ibid.

89. Interview, Los Angeles, California, October 18, 2004.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 1100

model for understanding why U.S. intelligence agencies failed to adapt to the pre-September 11 terrorist threat. U.S. government agencies are not built to change with the times. Because reform does not generally arise from within, it must be imposed from the outside. But even this rarely happens because all or- ganizational changes create winners and losers, and because the political sys- tem allows losers many opportunities to fight back. Indeed, the greater the proposed change, the stronger the resistance will be. As a result, winners can never win completely. Organizational adaptation almost always meets with defeat, becomes watered down, or is postponed until the next crisis erupts.

The CIA before September 11

Efforts to reform the Central Intelligence Agency in the post-Cold War period illustrate the three enduring realities. No sooner had the Soviet Union col- lapsed than calls were heard demanding an end to business as usual in the CIA. At first, these came from Senator Daniel Patrick Moynihan and other leg- islators who saw the CIA as a Cold War relic that should be abolished. But fo-

cus soon turned toward changing, not casting aside, the agency. Between 1992 and 1996, Congress twice attempted, but ultimately failed, to pass major re- form of the intelligence community. During the same period, twelve public re- ports issued by blue-ribbon commissions, governmental reviews, and think tank task forces discussed shortcomings in the U.S. intelligence community, the FBI, and U.S. counterterrorism efforts more broadly.

COMMON FINDINGS

Although the reports covered a variety of issues, they all identified four major problems within the U.S. intelligence community. The first was its lack of co- herence or "corporateness." As the Council on Foreign Relations study noted, the organization and leadership of the intelligence community was a "struc- tural oddity," with fourteen major agencies and no one in charge of them all.9" Technically, the DCI was supposed to set broad strategies and coordinate ef- forts across these agencies (as well as run the CIA), but in reality the DCI held direct control over only 15 percent of the intelligence budget (the secretary of defense controlled the rest) and had weak management authority for allo- cating money, people, and programs to every agency outside the CIA. The

90. Council on Foreign Relations Independent Task Force, Making Intelligence Smarter, p. 25.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 1101

reports' specific recommendations varied, but all offered ways to enhance in- telligence community integration and coordination.

Second, ten out of twelve of these studies found that intelligence officials and policymakers did not devote enough attention to setting intelligence pri- orities."9 To provide useful information, intelligence agencies require guidance from policymakers about which puzzles, people, and places rank higher on the priority list than others; about which surprises U.S. policymakers can live with; and about which ones they cannot. A robust mechanism for establishing and updating intelligence priorities, however, had never developed during the Cold War. And the Soviet Union's collapse, along with the emergence of new transnational threats, exacerbated these weaknesses. As a result, during the 1990s intelligence agencies had to cover more issues with fewer resources and little guidance about where to focus their efforts. In 1993 the president's National Performance Review found the system for establishing intelligence collection and analysis priorities to be a "jumble of loosely connected pro- cesses" that did not satisfy the needs of policymakers.92 The 1996 House Intel- ligence Committee's staff study agreed, calling the prioritization process "one of the most vexing aspects of intelligence management" and the need for fixing it "fundamental and urgent."9' Five years later the Hart-Rudman com- mission warned that the continued absence of an effective process for setting intelligence priorities was creating "dangerous tradeoffs between coverage of important countries, regions, and functional challenges."94

A third finding was the need to revitalize human intelligence capabilities. Nine of the twelve reports called for more aggressive human intelligence efforts to combat terrorism; two did not address the issue"9; and only one, the Twentieth Century Fund report, advocated downgrading collection from hu- man sources. Most frequently mentioned was the need to revise the CIA's 1995 guidelines that required prior approval from CIA headquarters before an indi- vidual suspected of human rights violations could be recruited as an asset- guidelines that had come to be known as the "scrub order" because they had led to the removal of hundreds of assets from the CIA's payroll.96 Many of the

91. The two exceptions were the Webster and Bremer commissions. 92. National Performance Review, "The Intelligence Community." 93. House Permanent Select Committee on Intelligence, IC21, "Overview and Summary," p. 26. 94. U.S. Commission on National Security/21st Century, Road Map for National Security, p. 82. 95. These were the Webster and Deutch commissions.

96. The Bremer and Gilmore commissions recommended abolishing the guidelines. The Council on Foreign Relations task force and the Hart-Rudman commission urged the CIA to review the guidelines and make recruiting clandestine assets one of the intelligence community's top priori-

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 102

reports also advocated improving the intelligence budgeting process so that resources could be more effectively matched against priorities-improvements that would have had the likely effect of redistributing some of the vast resources dedicated to technical intelligence systems to human intelligence activities.97 In addition, the Aspin-Brown commission and the House Intelli- gence Committee staff reports recommended the revision of personnel incen- tives and restructuring of the intelligence community to ensure that human intelligence efforts could be more effectively and efficiently deployed against hard targets such as rogue states and transnational terrorist groups that are difficult to penetrate by other means.9" Perhaps most radical, the National In- stitute for Public Policy report recommended stripping analysis functions from the CIA so that the agency could focus exclusively on human intelligence col- lection and even raised the possibility of disbanding the CIA's clandestine Di- rectorate of Operations and replacing it with an entirely new clandestine service in order to address the directorate's long-standing cultural and man- agement problems.99

Finally, the reports ranked personnel issues high on the list of concerns. As the House Intelligence Committee staff study bluntly concluded, "[The intelli- gence community] continues to face a major personnel crisis that it has, thus far, not addressed in any coherent way." "" Although the specifics varied widely, two common themes emerged. First, the intelligence community lacked employees with the requisite skills to confront growing threats such as foreign terrorism.'" Despite being technically exempt from a number of Civil Service regulations, intelligence agencies rarely fired poor performers. In addi- tion, the Aspin-Brown commission noted that even when confronted with mandatory reductions in personnel in the early 1990s, intelligence agencies

ties. For effects of the "scrub order," see Seymour M. Hersh, "Annals of National Security: What Went Wrong," New Yorker, December 18, 2001, p. 40. 97. The National Performance Review, "The Intelligence Community." 98. In a 2004 report, the House Intelligence Committee revealed that it had repeatedly criticized the CIA's human intelligence efforts and recommended corrective action in classified annexes to its annual authorization bills years before September 11. The committee noted, "After years of try- ing to convince, suggest, urge, entice, cajole, and pressure CIA to make wide-reaching changes in the way it conducts its HUMINT [human intelligence] mission ... CIA, in the committee's view, continues down a road leading over a proverbial cliff." House Permanent Select Committee on In- telligence, Intelligence Authorization Act for FY 2005, H. Report 108-558, 108th Cong., 2d sess., June 21, 2004, p. 24. 99. National Institute for Public Policy, Modernizing Intelligence, pp. 85-98. 100. House Permanent Select Committee on Intelligence, IC21, "Intelligence Community Manage- ment," p. 21. 101. The Hart-Rudman commission had a broader focus, advocating a massive new educational initiative to improve the scientific and mathematics base for future national security needs.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies ? 103

reached targets through attrition and voluntary retirement rather than through strategically focused cuts to keep the best talent and those with the most needed areas of expertise for a post-Cold War threat environment.'12 Second, intelligence officers too often stayed in their home agencies rather than build- ing institutional bridges to other policymaking and intelligence agencies through temporary rotations. Three studies-the Aspin-Brown commission, the Council on Foreign Relations task force, and the House Intelligence Com- mittee staff study-recommended that rotations to other agencies be required for intelligence officers to be promoted to senior ranks. Another three-the Na- tional Performance Review reports, the Twentieth Century Fund task force, and the FBI's 1998 Strategic Plan-urged the establishment of vigorous rota- tional assignments without requiring them for promotion. The need to realign the personnel skill mix and improve coordination through temporary tours of duty in other agencies received major attention in all but two of the reports.'""

COMMON FAILINGS

In all of these areas, the studies' recommendations went unheeded. In 2001 the

intelligence community was 50 percent bigger than it was when the CIA was created in 1947, but the DCI had only slightly more power to oversee it. As the congressional Joint Inquiry into the September 11 attacks darkly concluded, "The inability to realign Intelligence Community resources to combat the threat posed by Usama Bin Ladin is a relatively direct consequence of the lim- ited authority of the DCI over major portions of the Intelligence Commu- nity."'"4 The 9/11 commission agreed, noting that the intelligence community "struggle[d] to collect on and analyze ... transnational terrorism in the mid- to late 1990s" in large part because the community was a set of "loosely associ- ated agencies and departmental offices that lacked the incentives to cooperate, collaborate, and share information."'"" Indeed, the intelligence community was so fragmented before September 11 that even Tenet's 1998 declaration of war against bin Laden and al-Qaida did not seem to have gotten much atten- tion beyond the CIA's walls. "' The Joint Inquiry and 9/11 commission also found major deficiencies in the

102. Commission on the Roles and Capabilities of the United States Intelligence Community, Pre- paring for the 21st Century, pp. 2-7. 103. The two exceptions were the Gilmore and Hart-Rudman commissions. 104. Joint Inquiry Final Report, p. 43. 105. The 9/11 Commission, "The Performance of the Intelligence Community," staff statement No. 11, April 14, 2004, p. 12. 106. The 9/11 Commission Report, p. 357; and Joint Inquiry Final Report, pp. 236-237.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 104

intelligence community's system for prioritizing collection and analysis. The Joint Inquiry noted that intelligence officials found the process "confusing" and "so broad as to be meaningless," with more than 1,500 formal priorities for the National Security Agency alone by September 11.107 The 9/11 commission concluded that the setting of clear intelligence collection priorities "did not oc- cur" before the September 11 attacks."'8 Even some of those responsible for set- ting priorities agreed. Former National Counterterrorism Coordinator Richard Clarke noted that the White House "never really gave good systematic, timely guidance to the Intelligence Community about what priorities were at the national level."109

Despite calls to vastly upgrade human intelligence efforts, the CIA's clan- destine Directorate of Operations continued to languish. In 1995 only twenty- five trainees became clandestine officers."" By the late 1990s, the Directorate of Operations had cut by nearly one-third the number of its personnel deployed overseas."' Today the CIA still does not have enough qualified case officers to staff many of its stations around the world."112 In addition, the 1995 guide- lines restricting recruitment of foreign assets remained in place until after September 11. Nor did funding priorities shift from technical intelligence sys- tems to any significant degree. As one senior CIA official put it, "I'm cynical, but I think the reason people wanted to keep the [intelligence] budget secret was not to protect spies in Moscow, but because they didn't want people to know that 99 percent [of the budget] was stuck in some satellite.""3

Personnel problems also continued. In 2001 only 20 percent of the graduat- ing class of clandestine case officers were fluent in non-Romance languages.114 Robert Baer, a veteran CIA clandestine case officer, noted that even after the

1998 U.S. embassy bombings, the CIA employed not one case officer who spoke Pashto, the dialect of the major ethnic group in Afghanistan, and still had none as of 2002.11' The Joint Inquiry's findings are consistent with these assessments. The congressional panel concluded that before September 11, the intelligence community "was not prepared to handle the challenge it faced in

107. Joint Inquiry Final Report, p. 49. 108. The 9/11 Commission, "The Performance of the Intelligence Community," p. 9. 109. Joint Inquiry Final Report, p. 49. 110. The 9/11 Commission Report, p. 90. 111. Joint Inquiry Final Report, p. 265. 112. Hersh, "Annals of National Security." 113. Telephone interview with senior intelligence official, June 15, 2004. 114. "Intelligence Gaps: America's Spy Network Needs a Quick Fix," editorial, San Diego Union Tribune, April 22, 2002, p. B6. 115. Powers, "The Trouble with the CIA."

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 105

translating the volumes of foreign language counterterrorism intelligence it collected. Agencies ... experienced backlogs in material awaiting transla- tion ... and a readiness level of only 30% in the most critical terrorism-related languages used by terrorists."116

Nor have temporary rotations been commonly practiced. Although Director Tenet declared in the late 1990s that all intelligence officials were required to do a tour of duty in another intelligence agency before being promoted to the senior ranks, every agency, including the CIA, ignored him. Instead, intelli- gence agencies usually have filled these rotational positions with poor per- formers rather than rising stars. As one senior intelligence official lamented, "I often think of writing a vacancy notice [for temporary transferees to his agency] that says, 'only stupid people doing unimportant work need apply,' or 'send us your tired, your sluggish, your marginally brain dead."''7

THE ADAPTATION FAILURE MODEL REVISITED

The adaptation failure model helps explain why recommendations to improve the four critical intelligence deficiencies described above were not imple- mented before September 11. In 1992 and 1996, congressional intelligence committees attempted to en-

hance the intelligence community's corporateness by introducing legislation that would have dramatically strengthened the DCI's powers over the entire intelligence community. As expected, however, bureaucrats within the Depart- ment of Defense resisted these changes because they would have diminished the Pentagon's control over budgets, personnel, and the operation of key de- fense intelligence agencies. At the same time, both legislative proposals drew heavy fire from the House and Senate Armed Services Committees, whose own interests and power were threatened. And neither proposal had the strong support of President Clinton who, like his predecessors, was reluctant to press the issue against the vehement objections of the Defense Department and was more interested in pressing policy agendas on issues closer to home that promised more direct political benefits. Indeed, Clinton's focus on domes- tic issues was so strong that when a mentally unstable pilot crashed a Cessna airplane onto the White House lawn in 1994, aides joked that it was DCI James Woolsey trying to get the president's attention."8

116. Joint Inquiry Final Report, p. xvi. 117. Telephone interview with senior intelligence official, January 30, 2004. 118. Spencer Ackerman and John B. Judis, "The Operator," New Republic Online, September 22, 2003, http:/ /www.tnr.com.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 106

The fragmented federal government ensured that reform opponents would succeed. In 1996 the pro-defense House National Security Committee stripped the Intelligence Committee's reform bill of virtually every measure designed to increase the DCI's power over military intelligence agencies. House Intelli- gence Committee Chairman Larry Combest chose not to send the bill to the floor in the face of such opposition. In the Senate, conflict between the Intelli- gence and Armed Services Committees proved even more acrimonious, with the same result. Unable to garner a majority in the Armed Services Committee for radical reform, the Intelligence Authorization Act of 1997 produced only minor changes across the intelligence community."9

Changes in the intelligence prioritization process required executive branch action. Here too reforms failed for reasons the model would expect. Although President Clinton issued a presidential decision directive in 1995 to establish intelligence priorities, the list was long and only grew longer with time. De- spite provisions calling for an annual review, new priorities were added but old ones were never removed. The intelligence community was required to cover most issues and most places in the world. From the president's perspec- tive, this made political sense. Held uniquely responsible for U.S. foreign pol- icy, the president (and his top advisers) had strong incentives to gather information on every conceivable international problem because for them, no surprise was acceptable. Paradoxically, these incentives only made surprises more likely, by stretching the intelligence community too thin.

Finally, organizational factors help explain why the reports' recommenda- tions for improvements in human intelligence and personnel systems were not made through internal reforms. Natural resistance to change, the persistence of old routines, and cultural obstacles to implementing new ideas ran deep. Although many intelligence officials, policymakers, and experts saw the need for ramping up human intelligence efforts in the 1990s, the CIA seemed unable to escape from the 1970s, when congressional investigations into CIA covert operations pushed the agency to reign in its clandestine efforts. Baer and other veteran CIA clandestine operators have described a culture of excessive risk aversion that, together with the Cold War's end, led to a crippling of human intelligence capabilities during the 1990s.12" Between 1992 and 2002, for exam-

119. Congress and the Nation, Vol. 9 (1993-1996): A Review of Government and Politics (Washington, D.C.: Congressional Quarterly Press, 1998), pp. 242-246. 120. Robert Baer, See No Evil: The True Story of a Ground Soldier in the CIA's War on Terrorism (New York: Crown, 2002); and Powers, "The Trouble with the CIA."

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies 107

ple, only 35 percent of the CIA's clandestine case officers spent more than three years overseas recruiting assets.121 Meanwhile, old Cold War spending pat- terns continued to favor technical intelligence, which was well suited for de- tecting the location of Soviet warheads, over human intelligence efforts better suited for penetrating terrorist groups on the ground.'22

As for other personnel issues, the CIA failed to realign its personnel to meet the terrorist threat in large part because doing so would have violated the age- old expectation that those who volunteered their service to the intelligence community were guaranteed lifetime employment in return.123 Decades of mistrust and compartmentalization between (and even within) intelligence agencies fed the widespread resistance to assuming temporary duty outside one's home agency.124 And because a successful temporary rotation program required the cooperation of every agency in the community, it was especially difficult to implement. Director Tenet's rotation directive was easy to ignore.

In sum, despite a high degree of consensus and a range of potential fixes, critical deficiencies in the CIA's structure, prioritization process, human intelli- gence capabilities, and personnel systems were never remedied before the September 11 terrorist attacks. Reforms from within ran into resistance from entrenched routines and a firmly established culture. At the same time, exter- nal reforms generated opposition or indifference from key political players.

Conclusion

Since the September 11 attacks, members of Congress, the 9/11 commission, executive branch officials, and intelligence experts have offered numerous pro- posals for reform of the U.S. intelligence community. These efforts, and their limited success, underscore three implications of the preceding analysis. First, it appears that the nature of organizations, rational self-interest, and the frag- mented federal government hinder adaptation even after catastrophic failure. Efforts to reform the U.S. intelligence community sputtered for nearly three years after the worst terrorist attacks in U.S. history, succeeded only after the

121. Mike DeWine, "Additional Comments," Joint Inquiry Final Report, p. 11. 122. See, for example, Robert David Steele, The New Craft of Intelligence: Personal, Public, and Politi- cal (Oakton, Va.: OSS International Press, 2002). 123. John Gannon, a senior intelligence official who moved to the private sector after September 11, said that he had fired more people in eight months in the private sector than during his entire career in the intelligence community. "Special Report: Time for a Rethink-America's Intelligence Services," Economist, April 20, 2002, pp. 23-25. 124. Telephone interview with senior intelligence official, February 5, 2004.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 ? 108

harmonic convergence of an extraordinary set of factors, and then produced only modest changes.

On at least three occasions after September 11, opportunities for intelligence reform arose but did not produce results. The first came in early 2002, when Brent Scowcroft, chairman of the President's Foreign Intelligence Advisory Board, delivered a classified report to President George W. Bush urging a radi- cal overhaul of the U.S. intelligence community. The proposal included trans- ferring control over the three largest defense intelligence collection agencies to the director of central intelligence. Defense Department objections were so strong, however, that the president took no action.125 Second, in November 2002, legislation established a new body, the Department of Homeland Secu- rity, but skirted the issue of intelligence reform.126 The third opportunity came in December 2002, when the Congressional Joint Inquiry issued its final report investigating the September 11 attacks. The report's nineteen recommenda- tions included establishing a powerful new director of national intelligence, re- vamping the intelligence priority process, and considering whether a new domestic intelligence agency should replace the FBI. But as Richard Falkenrath notes, the report's recommendations "were ignored."127

Only in July 2004 did three factors converge to make modest changes possi- ble: CIA Director George Tenet resigned; the Senate Intelligence Committee is- sued a scathing report criticizing the U.S. intelligence community's prewar assessments of Iraqi biological, nuclear, and chemical weapons capabilities; and the 9/11 commission issued its final report. Tenet's resignation removed the U.S. intelligence community's staunchest defender and thus weakened bu- reaucratic resistance to reform. The Senate Intelligence Committee's Iraq re- port galvanized committee members and shifted the electoral incentives for all members of Congress.128 Issued during the height of the congressional and presidential campaign season, the report linked intelligence problems to a powerful domestic political issue: whether legislators voted to authorize the Iraq war based on flawed intelligence assessments. Finally, the 9/11 commis-

125. Walter Pincus, "Rumsfeld Casts Doubt On Intelligence Reform; Changes Suggested by Presi- dential Panel," Washington Post, April 9, 2002, p. A17; and Spencer Ackerman, "Small Change," New Republic, December 13, 2004, p. 12. 126. The Homeland Security Act of 2002, Public Law 107-296, 107th Cong., 2d sess., November 25, 2002.

127. Richard A. Falkenrath, "The 9/11 Commission Report: A Review Essay," International Security, Vol. 29, No. 3 (Winter 2004/05), p. 172. 128. Senate Select Committee on Intelligence, U.S. Intelligence Community's Prewar Intelligence Assessments on Iraq, 108th Cong., 2d sess., July 9, 2004.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies i109

sion created a media sensation when it released its report and launched a cam- paign, supported by several of the 9/11 victims' family groups, to have Congress enact its recommendations. Between July and passage of the Intelli- gence Reform and Prevention Act in December 2004, the commission received greater national television news coverage than the war in Iraq.129 Even with all of these factors, however, the intelligence reform bill was nearly derailed and was ultimately diluted by vigorous opposition from House Armed Services Committee Chairman Duncan Hunter, House Judiciary Committee Chairman James Sensenbrenner, objections from the Defense Department, and tepid pub- lic support from President Bush.

Reform efforts since September 11 also suggest that changes to U.S. intelli- gence agencies are likely to continue lagging behind external environmental demands. The constraints that hinder adaptation in U.S. intelligence agencies can be mitigated by extraordinary events, but these constraints can never be eliminated entirely. Organizations, even those under duress like the CIA and FBI, often find it easier to resist change than embrace it. Conflict and self- interest are ever-present in politics. And the United States' separation of pow- ers system will always provide opportunities for opponents of reform to water down legislative proposals.

All of these forces can be seen hindering intelligence reform efforts since the September 11 attacks. The FBI's aversion to technology is so deeply en- trenched, for example-one veteran agent described the prevailing culture as "real men don't type"'13""that the agency has failed to develop a functioning electronic database for its case files, even though FBI Director Robert Mueller declared the initiative a top priority after September 11 and spent $170 million on it.1"' As for legislative changes, the Intelligence Reform and Prevention Act created a new director of national intelligence, as the 9/11 commission recom- mended. To win passage of the bill, however, reformers had to dilute the DNI's budgetary, personnel, and management authorities, leaving many to wonder whether the position can function effectively. Moreover, the commission

129. Analysis from LexisNexis using the search terms "9/11 Commission" and "Iraq War" be- tween July 22, 2004, and December 10, 2004. 130. Quoted in Eric Lichtblau and Charles Piller, "Without a Clue: How the FBI Lost Its Way," Mil- waukee Journal Sentinel, August 11, 2002, p. 01J. 131. Richard B. Schmitt and Charles Piller, "FBI Draws Heat Over Pricey Software Trouble," Los Angeles Times, January 14, 2005, p. A12; and Glenn A. Fine, "The Federal Bureau of Investigation's Trilogy Technology Modernization Project," testimony before the Senate Committee on Appropri- ations Subcommittee on Commerce, Justice, State, and the Judiciary, 109th Cong., 1st sess., Febru- ary 3, 2005, http://www.usdoj.gov/oig/testimony/0502/final.pdf.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

International Security 29:4 1110

avoided recommending more radical changes-such as removing intelligence agencies completely from Pentagon control-that could have more compre- hensively addressed the "deep institutional failings" that it found, but would have been harder to achieve.132

Finally, intelligence reform efforts since September 11 reveal a tendency to view organizational deficiencies as structural problems. The two core recom- mendations of the 9/11 commission and the Intelligence Reform and Preven- tion Act-creating a single head of the U.S. intelligence community and new centers to integrate collection, analysis, and operational planning on issues such as counterterrorism-focus entirely on structure. Although it is true that U.S. intelligence agencies have been hobbled for decades by a dysfunctional structure, this article suggests that the inability of U.S. intelligence agencies to adapt to the terrorist threat also stems from organizational routines and cul- tures that are highly resistant to change.

Successful reform must target these forces as well. Operating deep within U.S. intelligence agencies, they are the silent killers of innovation. No intelli- gence structure can dramatically improve U.S. intelligence capabilities so long as intelligence officials view the world through old lenses and cling to old ways. For example, the reluctance to share information across agency lines is deeply engrained, based more on the values and habits that have developed inside U.S. intelligence agencies and the policies that have reinforced them than on official organization charts. The fifteen agencies of the U.S. intelligence community have not functioned as a coherent whole mostly because their employees have never seen themselves as part of the same team. Current policies still provide weak incentives for intelligence officials to discard these traditional parochial perspectives; most intelligence professionals can spend twenty years or more cloistered in their home agencies without a single communitywide training experience or assignment to another intelligence organization.

Fixing these problems is difficult but not impossible. New legislation could provide incentives and opportunities to establish informal networks and build trust between intelligence officials across agencies in two ways: by requiring the establishment of community-wide training programs early in officials' ca- reers, before they become attached to parochial views, and by requiring rota- tional assignments to other agencies for promotion to the senior ranks.

132. The 9/11 Commission Report, p. 265.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms

Adaptation Failure of U.S. Intelligence Agencies ? 111

Strangely, the 9/11 commission noted how a rotational requirement in the De- fense Department "transformed the collective mind-set" of the military ser- vices and dramatically improved integration between them, but did not recommend such a requirement for U.S. intelligence officials.'33

A few months after September 11, Richard Betts wrote, "The awful truth is that even the best intelligence systems will have big failures."'134 Evidence suggests, however, that the U.S. intelligence community was nowhere close to being the best before the September 11 attacks. Improving the future perfor- mance of U.S. intelligence agencies requires understanding better the sources of organizational weaknesses, the barriers to organizational adaptation, and ways to overcome them.

133. Ibid., p. 409. 134. Richard K. Betts, "Fixing Intelligence," Foreign Affairs, Vol. 81, No. 1 (January/February 2002), p. 44.

This content downloaded from �����������63.236.101.18 on Fri, 06 Oct 2023 17:33:11 +00:00������������

All use subject to https://about.jstor.org/terms