Uncertain Shield: The U.S. Intelligence System in the Throes of Reform

Intelligence in Recent Public Literature

Richard A. Posner. Lanham, MD: Rowman & Littlefield Publishers, Inc., 2006. 256 pages.

Reviewed by Stanley Moskowitz

If you are reading this review of Judge Richard Posner’s book Uncertain Shield, you will want to read the book, for it is more interesting and insightful than, alas, this review may convey. Posner brings a fresh and welcome perspective to hoary intelligence issues, drawing on mathematics, economics, and organizational theory, but it is Posner’s intelligence and common sense that keeps you reading.

Posner is a writer of books—some four dozen according to the fly jacket—a lecturer at the University of Chicago Law School and, in his spare time, a federal appellate judge. The reviewer especially likes the judge part, because, while Posner displays a judicial temperament in mostly modulated prose, his book is full of telling judgments about the process, the people, and the sheer ignorance that brought us the reorganization of the Intelligence Community in 2005. Judge Posner has made plain that he is not a supporter of the Intelligence Reform and Terrorism Prevention Act (IRTPA) of 2004 or the 9/11 and WMD commissions that conspired with Congress to bring it about.

Uncertain Shield is a critique of the mischievous works of the two commissions. Since they deal with every aspect of the intelligence enterprise, so too does Uncertain Shield, which ranges widely, if incompletely, over intelligence matters. Posner argues that US intelligence has been unfairly castigated for the surprise attacks of 9/11. In an earlier book, Preventing Surprise Attacks, he argued that the historical record establishes that prevention of surprise attacks is not possible and that the effort to create a perfect intelligence system, capable of achieving perfect results, is a fool’s errand.

Surely Posner is right about this, but if perfection is not possible, how about better performance? This gets you to the second chronic question of the intelligence game, how do we know better performance if it is not correlated with being right or wrong? (As a manager, I used to ask analysts, would you rather be right for the wrong reasons, or wrong but for the right reasons? They usually thought about it, which was why they were analysts. My answer was always “be right” because would-be critics are not much interested in your reasons.) If only we could make judging intelligence performance an arithmetic exercise, like baseball averages. A hitter who manages to get hits three out of 10 times at bat is Hall of Fame material. And an intelligence analyst? Because we do not know what is a good intelligence average, and it is unknowable, we default to expecting perfection. Posner says no one argues that a .300 hitter is a failure for not getting hits 70 percent of his times at bat. Why should we obsessively talk about intelligence “failures” when the batting average is less than 1.000? Posner is generous, but intelligence professionals know better. If the bases are loaded in the bottom of the ninth of the seventh game of the World Series, and our .300 hitter strikes out, you bet fans will say he failed…or worse.

Posner understands, although he does not like it, that reorganization of the Intelligence Community was inevitable after 9/11. In his first chapter he astutely, and bitingly, describes the political circumstances in 2004 that made the IRTPA inevitable. He has many unkind words for the 9/11 Commission.

My experience in studying the legislative response to the 9/11 Commission’s report has been a disillusioning one. The length and literary panache of the report and the public-relations blitz that accompanied it disguised its intellectual shallowness, in particular its indifference to the relevant scholarly literature, to foreign experience and to history. The disguise worked.

And of investigatory commissions generally, which

have structural problems…that impair the quality of their recommendations [which] include selecting members…on the basis of prominence and diversity rather than expertise, a [related] preference for bipartisan over nonpartisan composition, which tends to politicize…and the pressure for unanimity…with the effect of making the commission’s recommendations a product of political compromise.

Posner pronounces that the “history of government reorganizations teaches that most such reorganizations fail especially [his emphasis] those imposed on an agency from the outside.” And yet the main fault he finds with the 9/11 Commission is that it chose the wrong people to serve and did not have enough academics and other experts. But one might argue, that given the history he sums up, we might be better off if the 9/11 Commission had confined itself to finding fact and left it to the Bush administration and the people who would have to implement change to find the appropriate change. Given that the whole country was watching, reforms would have had to have been meaningful, but they would have been better attuned to the deficiencies of the Intelligence Community than the recommendations of the 9/11 Commission or the legislation it huckstered.

Posner does not address another issue: How did it happen that the 9/11 Commission issued its report in the middle of a presidential election? Why was the commission afforded the opportunity to push a sweeping and ill-considered reorganization of US intelligence?

Posner sees a clever, political commission fixing the blame on intelligence as a way of avoiding digging deeper into the performance of Democratic and Republican administrations (the political class taking care of its own). He sees a Congress avoiding blame and taking the easy way out which, of course, is reorganization. Changing the boxes is easy to explain and it does not cost much, at least in dollars. And when you have done it, you can see the fruits of your work. Titles are changed, offices are moved, new confirmable positions are devised! Lights, camera, action!

Posner’s most salient point about the reorganization is that “most intelligence failures are not failures of organization, and reorganization is unlikely to solve organizational problems.” This reviewer could not agree more, but it does leave begging the question of whether there were intelligence shortfalls that could be addressed in a meaningful way? What are the nonorganizational problems? The text of the WMD Commission report is full of interesting but questionable observations about collection and analysis. Posner touches on a few but you will need to wait for Posner’s next book to get more sustenance.

As an aficionado of the literature of organizational theory, Posner also does battle with the particulars of the legislation. He says the 9/11 Commission believed that inadequate sharing of information contributed to the failure to uncover the 9/11 plot and that the remedy was to centralize what had been a decentralized system. He argues that this is a flawed approach because centralization runs counter to the idea of competitive analysis and that a weak DNI will have little power to make centralization work. Posner notes, for example, the DNI’s inability to hire or fire across the community or to move resources except in a circumscribed way. Posner also argues that the DNI’s lack of progress in information-sharing illustrates the flaws in the reorganization.

There are problems with this line of analysis. For one thing, it underestimates the DNI’s political power to get things done that are not strictly provided for in legislation. If he is willing to fight for a policy or a vision, the DNI can be larger than the sum of his legislative powers. What’s more, he can argue for, and probably get, more legislative power. It counts that everyone in the White House and the Congress has been rooting for him to be strong and decisive.

But the more fundamental flaw in Posner’s analysis is that the Intelligence Community is less centralized today than it was under the DCI. For one thing, the Department of Defense’s long overdue effort to get its own intelligence house in order and to reassert itself into human source collection, an area long neglected by DoD, represents, whatever its many merits, a de facto challenge to the head of Intelligence Community. Add to this the fact that the secretary of defense has been playing a larger role in hiring and firing the leaders of his large collection agencies and giving them marching orders.

Consider also the roles of the head of the Intelligence Community. The DCI ‘s responsibilities were to run the Intelligence Community, to manage the CIA, (the community’s only “central organization”, i.e., the only one that reported to the president, not a cabinet officer), and to be the president’s intelligence adviser. The legislation, in the name of creating a stronger head of the community, split this unity into a duality and did so without strengthening the head of the community’s authorities. Posner believes the split was a good idea, since he agrees with the 9/11 Commission that one person cannot do all three jobs. (Former DCIs were not polled.[1]) He would do the split differently, arguing that the principal intelligence adviser should also run the CIA. The DNI’s responsibilities would be “managing” the “tense” relations between DoD and CIA, coordinating the defense collection agencies, improving domestic intelligence, improving information sharing and, finally, “evaluating the CIA.”

Since the DNI is already saddled with these responsibilities, it is not clear how taking away his access to the president helps him, especially in the absence of increased legislative power. In the unequal struggle between CIA and the DoD that Posner postulates, could the former expect much help from a DNI who did little more than oversee the Intelligence Community budget?

Posner’s discussion of CIA reiterates myths—“the bulk of CIA’s human and technical resources” were devoted to the Soviet Union—and creates new ones— that there was a “witch hunt” after the Ames exposure—but it has the seeds of what should be his next book, the decline of CIA and what it means for US intelligence.

In the end, Posner comments ruefully that with the DNI and his deputies acting as the “real bosses” of CIA, with the DNI siphoning off CIA’s analysts and missions, and with the Senate objecting even to CIA management of human intelligence, it may turn out that the reorganization will be little more than the renaming of the Central Intelligence Agency as the Office of the Director of National Intelligence. So, maybe CIA is not declining so much as morphing—like the Terminator. We should be so lucky.

Posner asserts that there is a “crisis” in domestic intelligence because the FBI is proving that it cannot discharge its new intelligence duties. His evidence is mostly anecdotal, but it is also supported by history. The FBI, according to Posner, is in its eighth year of reorganization, and, despite protestations that it has “got it” this time, Posner argues that as a law enforcement agency, the Bureau will never be able to “change its culture” and embrace the intelligence mission. Posner describes at length the many differences between law enforcement and intelligence: law enforcement performance is measurable, intelligence is not; law enforcement seeks to make (countable) arrests and (countable) convictions, while intelligence wants to surveil and analyze and understand. He gives real-life examples including the so-called Lackawanna six case. For Posner, the intelligence mission is diverting the FBI from its real calling.

The Bureau structure is inherently decentralized and the Bureau’s 56 field offices essentially work for the Department of Justice’s 96 US Attorney’s Offices, which are scattered across the country. Since most crime is local and prosecuted locally, the provincialism of the FBI’s field offices is rational. If you add to that the fear that FBI agents have, according to Posner, that others will steal their cases and the fact that the judicial system discourages commiting to paper or sharing data that can be discovered by defense attorneys, then you can understand that the nature of the FBI’s undertaking gives rise to cultural norms that are antithetical to intelligence work

There are other problems. The Bureau is accustomed to investigating crimes that have been committed, less so to the idea of gathering data or pursuing leads in an environment in which no law has been broken. (Presumably counterintelligence investigations are an exception).

The Bureau incorrectly believes that intelligence is a natural outgrowth of traditional criminal investigative practices. Criminal investigations are after-the-fact reactions to specific events, and their scope is limited.…The process is remote from the collection of bits of evidence quite speculative threats and from the interpretation of that evidence with the aid of...analytical techniques.

The Bureau is interested primarily in arrests, but such arrests not only potentially compromise greater understanding of the networks or leadership of terrorist enterprises but they can alienate US Muslim communities whose good will is key to penetrating terrorist conspiracies.

Posner adds to his case that the Bureau is unlikely to successfully undertake terrorist intelligence by rehearsing stories about the Bureau’s current problems. These includes accounts of troubles in starting up an intelligence shop; the rapid turnover of officers working in intelligence; the lack, presumably inevitable, of intelligence experience; the hiring of clericals to do analysis and the treatment of them as clericals; the so-so reputation of bureau-led joint terrorist task forces; the continuing complaints of local law enforcement; and the latter’s setting up of their own intelligence operations. He says the Bureau will be unable to change its culture, because it is still by mission a law enforcement agency and the two cultures are incompatible.

Posner wants an independent domestic intelligence agency under the Department of Homeland Defense. Its mission would be to develop contacts with the country’s local law enforcement apparatus with the purpose of developing intelligence on radical groups and potential terrorists. It would collect information, analyze it, and distribute it to the nation’s police forces, the FBI, the national Intelligence Community and to national security policymakers. It would create a “coordinated nationwide intelligence network that we need and don’t have.”

Posner understands the obstacles. The FBI is against it. DHS does not have the confidence of the Congress or the country. Influential lawyers like DHS chief Chertoff and WMD commissioner Silberman have a prejudice against intelligence unencumbered by the judicial process and sensibility. And, perhaps most important, civil libertarians are opposed.

If truth be told, civil libertarians are opposed to anything that involves intelligence. Posner deplores their refusal to allow that there are tradeoffs between security and unbridled freedom, but he argues that we now have increased domestic intelligence, whether the civil libertarians like it or not. He points to the domestic intelligence shop already in DHS and to DoD’s domestic intelligence efforts. He also notes that separating law enforcement and the power to arrest from intelligence (currently both vested in the FBI) should serve the interests of civil libertarians. “It is ironic to note” he writes, “that in reaction to Nazi Germany having merged police and domestic intelligence functions in the SS, a 1949 Allied directive required the new Federal Republic of Germany to place the two functions in separate agencies.” Is Posner right in concluding that only a separate domestic intelligence agency will be sufficient for the struggle against terrorism? I would find the case more persuasive if he were more the analyst and less the advocate.

He understates the problems of standing up a new agency. Where do the experienced people come from? How long would it take to get traction in its work? CIA now coordinates extensively with the Bureau. With a new domestic intelligence agency, that cooperation would be greatly complicated, and it’s not an easy undertaking now. The reporting chain of the Bureau’s national security branch is too cumbersome, but the new domestic intelligence agency’s ladder would hardly be better. Posner’s discussion of “culture” is as devoid of subtlety as those he quotes discussing it. We all know that agencies develop (in a more mysterious process than Posner and others acknowledge) subcultures that reflect the individual missions and the people who carry them out. Would not the Bureau’s national security branch follow suit?

Finally, Posner points approvingly to the experience of foreign governments which typically divide their domestic and foreign intelligence operations between different organizational agencies, without noting that between them tension, and even strife, is the norm. He seems to think that the FBI can keep its special branch doing intelligence without running afoul of his new agency. Similarly, he does not discuss how the conflicting interests of domestic intelligence and those of the FBI will be adjudicated.

Getting domestic intelligence right will not be easy, and it may be that the Bureau is still the best bet, problems notwithstanding. We get a strong argument on one side from Posner, but not a judicious balancing that might have been more persuasive.

It was one of the clichés of the 9/11 Commission that the failure to properly share information was a weakness that led ineluctably to 9/11. Without arguing the merits of this simplification—and Posner likely would not agree with the commission—I agree that the Intelligence Community would benefit greatly from a better flow of information. What if all imagery, whether collected by satellite or air-breather, were instantly available to analysts in the Intelligence Community and in the military commands? How might that change how business is done? The WMD Commission was right in saying that collection agencies must be compelled to surrender the idea that they “own” the data they collect. The main issue is not technology, but policy.

In the chapter “The Woes of Automation,” Posner mostly addresses the technical issues impeding information flow. The culprits are the inadequacy of federal procurement, burdened by poorly paid and underachieving officials, excessive paperwork, and a Congress that does not understand the value of turning over software every three, rather than 10, years. The Feds, according to Posner, are bedazzled by fast-talking contractors bent on maximizing profits by undervaluing commercial off-the-shelf computer hardware and software. The government’s chief information officers do not have enough power or influence to do their jobs, and they suffer from fear of failure and criticism. All of this is good stuff and should be read by the right people in Congress and the heads of agencies. It would be easy to be depressed by Posner’s bleak depiction, but there are exceptions that prove that the government can get automation right.

Congress is the dark specter that looms over the intelligence enterprise. To Posner, there is little Congress contributes except mischief. It misreads the Constitution and believes its role in the separation-of-powers fandango is to micromanage the executive branch. Through a careful parsing of Articles I and II of the Constitution and the early history of intelligence as a near exclusive executive function, Posner makes a powerful case that Congress overreaches. The absurdity of the detailed prescriptions in the IRTPA is on display in the chapter Posner devotes to destroying Congress’s pretensions in matters of intelligence. One might also add that the Senate staffers who wrote much of the legislation were decent and smart people. They were willing to take advice, as long as it met the needs of their senatorial masters. But they were also all but totally ignorant of matters of intelligence. How could it be that uninformed staffers ended up writing the basic writ of America’s intelligence establishment? Only in America.

Does Posner’s constitutional analysis matter? It seems unlikely it will have much effect, except to strengthen the hand of those in the executive branch who argue from time to time that this or that congressional mandate or legislative requirement is less than divine writ and can be safely ignored. This is a rare occurrence and, if truth be told, most managers of US intelligence like oversight precisely because it gives them advice about what to do or not to do. (Their faux protestations not withstanding.)

It has been almost 30 years since Congress got seriously into the intelligence oversight business. One can surmise that most Americans think that it is good that someone is watching over a secret and powerful enterprise like intelligence. But how effective has that oversight been? Posner postulates that conventional wisdom argues that it has been weak and a failure. He offers little evidence to support this judgment except to repeat the bromides of the 9/11 Commission. He does not agree that better oversight would have made any difference in preventing 9/11. “If Congress cannot enact a responsible reorganization of the intelligence system,” he writes tartly, “ it is hardly likely that its committees can discern threats to national security before the intelligence agencies themselves notice them. The commission, maybe because its membership included former members of Congress, displayed an extravagant regard for Congress’s potential contributions to intelligence.” Posner might well have noted that the three former members were congressional “authorizers,” who are used to seeking their place in the firmament by poaching on executive prerogatives and power.

Posner concludes on a sour note “The intelligence community will have to continue muddling through with very little help from Congress, and occasional hindrance.” This is too severe a judgment. The congressional oversight committees do some excellent work, almost all of it outside the view of the public. (One might observe that the quality of oversight decreases as press coverage increases.) More, the impact of oversight on the folkways and habits of the intelligence practitioners is very great indeed, although it is generally underappreciated, even in Congress. A negative example, which Posner does mention, is that congressional divisions mirror and empower poor coordination in the executive branch. Congress deplored the last DCI’s alleged indifference to managing the community, but it did nothing to help him in that undertaking. Indeed, it preferred dealing with individual intelligence agencies, untied to the DCI. It sent the unmistakable message that the DCI was not terribly relevant to the community’s parts. Still there are numerous examples of congressional oversight helping the agencies face up to problems and opportunities. Almost 30 years ago, for example, the oversight committees were urging CIA to relocate HUMINT collectors.

What is to be done? We could do worse than take to heart Posner’s throwaway line: “The initiative for redirection, reform, and reorganization ... must come from outside Congress, normally from the intelligence community itself.…” A hard-nosed Intelligence Community that understands and confronts its own short comings before it has its nose rubbed in them is the best antidote to the kind of pathologies and ill-considered nostrums so well described in Uncertain Shield.


[1]For a study of how each of the 19 DCIs carried out this function, see Douglas F. Garthoff, Directors of Central Intelligence as Leaders of the U.S. Intelligence Community, 1946–2005 (Washington, DC: Central Intelligence Agency, Center for the Study of Intelligence, 2005).


Stanley Moskowitz died on 29 June 2006, not long after he delivered the book review published in this issue. Mr. Moskowitz, who held numerous senior positions in CIA, had retired from active service but continued to work on intelligence issues, including work on a study of congressional approaches to intelligence reform.


All statements of fact, opinion, or analysis expressed in this article are those of the author. Nothing in the article should be construed as asserting or imploying US government endorsement of an article's factual statements and interpretations.

Historical Document
Posted: Apr 15, 2007 10:14 AM
Last Updated: Jun 26, 2008 07:26 AM