Book Project: Secrecy and Oversight: U.S. Law and the Domestic Politics of Covert Action

Scholars have focused on the choice of covert action over overt use of military force. But covert action is not monolithic—it includes psychological, political, economic, and paramilitary activity—and current theories do not explain under what conditions each covert option is chosen. Congress regulates the CIA’s capacity to conduct covert action as part of its oversight of national security agencies. I categorize laws passed by Congress to regulate covert action along two dimensions: authority (prohibitive or enabling) and doctrine (substantive or procedural). Using logistic regression, I find that Congress is more likely to choose substantive laws when faced with recent intelligence scandals. Congress is less likely to create a prohibitive law as support by the president’s party and the opposition party in- creases. These results are supported by the legislative history of a sample of these laws. I argue that these laws lead to a shift in covert action components that can best be explained through the interaction of principal-agent dynamics and the variation that exists in the perception that Congress’s laws are binding on the behavior of the executive and the national security agencies. To test this theory, I engage in process tracing of primary source documents of decision-making related to U.S. involvement in the Angolan civil war (1974-1989) and the Yugoslavian civil war (1992-1995). The results suggest that covert actions are modified in response to legal regulations, but sometimes in unanticipated ways that attempt to evade these laws. Substantive, prohibitive laws can result in an escalation of the violence of covert action, contrary to the intent of Congress. But procedural, prohibitive laws can caution decision-makers and slow escalation. Third parties, especially those invested in but not actively participating in the conflict, serve as a vital conduit for this covert activity but ensure an attenuated relationship to the U.S.

• Chapter: No Instructions or Green Light? US and Iranian Arms in the Yugoslavian Civil War

To test my theory, I engage in process tracing of primary source documents of decision-making related to U.S. involvement in the Yugoslavian civil war (1992-1995), when the Bosnian government was suffering from a UN-imposed arms embargo against the participants in the Yugoslavian conflict. Iran helped to smuggle arms to the struggling Bosnian government, and this arms pipeline through Croatia into Bosnia received U.S. blessing. 

The results suggest that covert actions are modified in response to legal regulations, but sometimes in unanticipated ways that attempt to evade these laws. Particularly, covert operations are sometimes modified in ways that run counter to the legislative intent of these laws. Prohibitive, substantive laws can sometimes encourage escalations in the violence of covert activity. In the Yugoslavian civil war case, Congress intervenes with the Nunn-Mitchell Amendment to prohibit any U.S. government spending on enforcement of the UN arms embargo against the combatants in the Yugoslavian civil war. This prevented U.S. intelligence gathering under the guise of enforcing the embargo and so led to greater direct involvement in paramilitary covert activity in Bosnia, such as CIA inspection of missile shipments. 

But prohibitive, procedural laws have also been successful at making decision-makers cautious and slowing escalation. The Yugoslavian civil war case shows that the 1991 Intelligence Authorization Act’s regulation of communication between the CIA and Congress led the executive to be cautious in involving the CIA in encouraging the Iranian arms pipeline to the Bosnian government, and relying more heavily on political action rather than direct paramilitary involvement.

---

Works in Progress 

• A Rogue Elephant: Credibility and Strategic Disclosures by the Intelligence Community (with Keren Yarhi-Milo)

We have extensive knowledge on how the intelligence community (IC) informs policy decisions, and how intelligence products can be politicized to justify policies. We have less analysis, however, of how the IC can assert its agency in policy disputes with the executive. During a policy disagreement, how can the IC shape policy in a democratic society? We theorize that the IC chooses between four disclosure strategies: internal dissent/whistleblowing; declassification; subversion; and leaks. Two main considerations shape this decision: whether the politicization risk for the disclosure is high or low, and whether the intelligence community seeks a prospective change in policy or a retrospective rehabilitation of reputation. Several cases illustrate how these strategies are chosen: CIA whistleblower on Trump conversation with Ukraine (whistleblowing); CIA’s Family Jewels (declassification); 1920s efforts of UK intelligence service to remove the government (subversion); and the NSA’s cancelled THINTHREAD program (leaks). In conclusion, we consider the implications of this article for civilian-intelligence relations and democratic backsliding.

• Let Everybody Know: Public Intelligence Disclosure, Audience Costs, and Credible Threats (with Ofek Riemer)

Russia’s recent war against Ukraine has highlighted an increasingly prevalent phenomenon in international affairs: the purposeful disclosure of secret intelligence as an instrument of foreign policy. The question remains as to the effect of official public intelligence disclosure (OPID) in international bargaining. First, does OPID generate audience costs, making it a credible signal of resolve? Second, what factors shape public perceptions of OPID? For instance, is moving from simply exposing new information to illustrating it by using solid, tangible evidence likely to increase audience costs? Since OPID constitutes a more “benign” way of signaling than an explicit threat to use force, we hypothesize that leaders who disclose intelligence on a security threat but ultimately fail to remove it will incur a price in public opinion, although not as high as leaders who threaten to use force. If true, intelligence disclosure could serve as an intermediary step on the escalation ladder, between privately and publicly threatening the use of force. There also exists a spectrum of coercion through public intelligence disclosure, and so we hypothesize that leaders who disclose costlier intelligence, but ultimately fail to remove the threat, will incur a higher price in public opinion. To test these hypotheses, we plan to field a survey experiment to a random sample of Israeli and American citizens. In the Israeli survey, we modify the classic “repel an invader” experiment, to include an arms shipment from one adversarial country to another.

• A Matter of Proportionality? Customary International Law and Public Support for the Use of Force

Existing observational studies have been unable to adequately explain why states comply with international legal obligations, especially those deriving from nonbinding international legal principles. Customary international law is defined by custom—evidence of a general practice accepted as law. Some studies suggest that customary international law should have no different effect than treaty law on support for the use of force. This article sheds light on several fundamental questions related to the effect of international law. To what extent does customary international law change public preferences about the use of force? Does the type of commitment to international legal principles matter? Through what mechanisms does international law exert an effect? This article employs a preliminary survey experiment of 1,800 Amazon Mechanical Turk workers fielded in November 2017. This experiment presented respondents with a hypothetical situation in which the United States was considering bombing a building where terrorist leaders would be meeting. The vignette randomized information about the international legal principle of proportionality, which does not permit attacks when the expected harm incidental to such attacks would be excessive in relation to the military advantage anticipated to be gained. It also randomized whether the United States publicly committed to following this principle, and whether the UN found a similar previous bombing to violate those legal principles. After this vignette, the experiment asked whether the respondent would support a U.S. bombing of the targeted building. This article finds that there is a weak effect of customary international law restraining potentially violative uses of force abroad. Specifically, when information on customary international law is paired with statements that international organizations may deem the use of force as a violation, respondents significantly decreased their support for the use of force.

• Assassination, Covert Norm Violation, and Customary International Law

This article uses the lens of historical institutionalism to explore why assassination (and political killings more generally) have become increasingly common and how this change in state practice affects customary international law governing the use of force. I argue that covert norm violation can create precedent at critical junctures that is then used to justify further norm erosion in the future. I develop a theory of how states can use covert norm violation to create strategic norm erosion—thus changing customary international law in a way that suits their policy goals. I identify that the early 1960s was one critical juncture in the development of U.S. assassination policy. I test my theory using a comparative case study approach based on the 1975 Church Committee, which investigated cases where the U.S. implemented assassination operations against foreign leaders. I engage in process tracing of two incidents: killing of Patrice Lumumba, Prime Minister of the Congo, in 1961; and the AMLASH operation against Fidel Castro. I examine the Church Committee reports, declassified primary documents, and secondary histories and memoirs. I conclude with the repercussions of these assassination efforts reverberate into modern day targeted killing efforts.