Nathan Alexander Sales · February 2010
78 GEO. WASH. L. REV. 774 (2010)
By Nathan Alexander Sales
Information sharing is the one counterterrorism initiative virtually everyone supports. Yet no one seems to have any idea how to make it happen.
Like the American public as a whole, the academy remains divided on the wisdom and legality of many measures undertaken in the name of national security since the terrorist attacks of September 11, 2001. Sharp disputes persist over lengthy detentions of suspected terrorists outside the criminal-justice system, coercive interrogations of captured al Qaeda leaders and other detainees, and eavesdropping on Americans’ international communications without court orders. Yet the need for more effective information sharing remains a rare area of agreement, both within academia and without. Of course there are exceptions. There always are. Still, the consensus in favor of more information sharing has proven surprisingly broad and durable.
Egged on by the commentariat, Congress and the executive branch have enacted a series of measures intended both to eliminate legal restrictions on information sharing and to promote data exchange among national security players. For example, in 2001, Congress allowed federal prosecutors to share information uncovered in grand-jury investigations and through wiretaps with their counterparts in the intelligence community. A year later, Congress directed federal agencies to exchange “homeland security information” with one another. And in 2004, the government established an “information sharing environment” to encourage the free flow of national security data.
Despite a decade of effort, however, there is a widespread sense that information sharing is “going nowhere.” Why?
Because policymakers have failed to account for the iron law of agency self-interest. Reformers have repealed a number of legal impediments to data exchange, and they have exhorted intelligence agencies to do a better job of sharing. But they have done little to eliminate the natural bureaucratic incentives that dissuade agencies from cooperating with one another. Nor have policymakers fostered new incentives to encourage agencies to exchange data. No matter how many bills are passed or executive orders signed, intelligence agencies won’t connect the dots unless it’s in their interest to do so. It’s not enough simply to tear down the wall. Agencies must be given reasons to climb over the rubble.
Hence this article, the first comprehensive analysis of why intelligence agencies fail to share information and what may be done about it.
Part I recounts a number of recent information-sharing failures. Some are notorious, others obscure. Some are operational missteps—such as CIA’s refusal in 2001 to alert other agencies that an al Qaeda member (and eventual 9/11 hijacker) had entered the United States. Others are failures of policy—for example, the Justice Department’s decision to erect a “wall” between intelligence officials at the FBI and their law-enforcement counterparts. Part I then describes the major post-9/11 legislative and executive initiatives to correct these problems.
Part II consults public-choice principles and insights from other legal disciplines to explain why intelligence agencies tend to hoard information. It begins by asking: What do agencies maximize? The answer, I suggest, is twofold. Intelligence agencies seek to maximize their influence over senior policymakers in the executive branch, as well as their autonomy—i.e., the ability to pursue agency priorities without outside interference. Information sharing can undermine both goods. Data exchange can lead to free riding and, with it, a loss of relative influence; if the FBI shares information with CIA, policymakers might give credit to CIA for any resulting intelligence breakthroughs. Agencies also fear that information sharing will enable their rivals to muscle in on their turf, such as by seizing control of an ongoing operation.
Part II then proposes a series of analytical frameworks, or lenses, through which agencies’ tendency to hoard might be understood. First, data exchange sometimes resembles an intellectual-property (“IP”) problem. Agencies treat their intelligence information like private firms treat trade secrets. As in the private sector, sensitive information is valuable to an agency only to the extent it is able to shield that data from competitors. If the FBI free rides on CIA data, it might enhance its own influence and cause CIA’s to wane. Antitrust law supplies a second lens. An agency’s reluctance to share information with rivals can be thought of as the bureaucratic equivalent of a refusal to deal. Intelligence agencies also engage in familiar forms of rent seeking, such as lobbying the President to grant them monopoly rights in various intelligence submarkets. And they form cartels to enforce market-division arrangements, insulating themselves against competition in their respective market niches. Organizational theory is a third lens through which to view the hoarding problem. Intelligence officials are conditioned by an institutional culture of risk aversion, which stems from the asymmetries between their career rewards and penalties. They almost always stand to lose more from bold and independent action than they stand to gain, and so they have strong incentives to keep their information to themselves. Analysts are well aware that, as an FBI supervisor warned in the 1990s, information sharing can be a “‘career stopper.’”
Part III considers how to adjust the incentive structure to encourage agencies to share—and to do so in a way that doesn’t weaken their existing incentives to collect information in the first place. In particular, it examines whether intellectual-property, antitrust, and organizational-theory solutions could be adapted for the information-sharing context. Policymakers might replace the trade-secrets regime that currently governs intelligence products with a system based on patent and copyright principles. That could be accomplished by requiring agencies to somehow publish their intelligence data as a precondition of receiving IP-type protections, or by subjecting them to a copyright-style compulsory licensing scheme. To mitigate free-riding concerns, originating agencies could be offered various kinds of compensation when rivals use their products. Policymakers also might look to antitrust law’s suite of enforcement tools. They could establish a central regulator with the power to impose sanctions (monetary and otherwise) on agencies that refuse to share. They could create a mechanism for individual intelligence agencies to litigate challenges to their rivals’ efforts to hoard. And they could harness market forces to promote information sharing, such as by relaxing the intelligence market’s severe entry barriers. Finally, policymakers could mitigate intelligence agencies’ cultural risk aversion by creating new, pro-sharing incentives. They could increase the benefits an individual analyst could expect to gain from sharing (such as by offering cash bounties) while decreasing the expected costs of doing so (such as by eliminating legal ambiguities about which kinds of information may be shared and which may not).
