Working Papers
Clark, Tom S., and John W. Patty. "Which Issues Should We Decide? Coalitions and Agenda Setting." (working paper)
It is well-known that the classic model of multidimensional policy choice generically has no core under majority rule. We show that this instability has a simple anatomy. Fixing the preferences of a group's members and a prevailing status quo, every proposal that a majority would adopt falls into exactly one of three categories, distinguished entirely by which of its narrowings the body would also adopt. Convenient vehicles are proposals every issue of which would be adopted on its own; bundling them is a convenience. Bridging vehicles are proposals that pass as a package but fail if any issue is removed—the spatial analogue of a logroll, sustained by a coalition of coalitions drawn from opposite sides of the policy space. Clown cars are the residual: proposals that pass and would continue to pass even if some of their issues were stripped away. We show that bridging vehicles are exactly the adopted proposals that move policy away from the dimension-by-dimension median on every issue (when there are two issues), that they exist generically whenever the status quo is Pareto efficient, and that their existence is necessary for majority-rule “chaos.” The typology also sharply separates positive from negative agenda control: division of the question threatens clown cars but never bridging vehicles. We develop all of this in the simplest possible setting—a single proposal and a single sincere vote—and indicate how it organizes the strategic analysis of agenda formation that we pursue in companion work.
Clark, Tom S. "A Model of Criminal Justice Discretion." (working paper)
Criminal prosecutors wield tremendous power in the American legal system. However, their ability to direct prosecution against criminal suspects relies on coordination with the police who investigate crimes and make arrests. That relationship has been made more prominent in the context of reform prosecutors who propose to shift traditional law-enforcement priorities. I propose a novel model of police-prosecutor interactions that focuses on the informational asymmetry between police and prosecutors. Police are endowed with private information about the true criminality of an arrested suspect but also have biases about prosecution and charges. I evaluate how prosecutors' preferences over leniency affect credible communication between police and prosecutors. Analysis of the model identifies empirical implications for the types of criminal suspects prosecuted, prosecutors' adoption of police charges, and the success rate for prosecutors who take cases to trial.
Clark, Tom S. "Coordinated Justice: The Politics of Prosecution and Policing." (working paper)
Criminal justice outcomes emerge from the coordinated actions of police and prosecutors, yet these institutions are typically studied in isolation. We lack a theoretical framework for understanding how the strategic incentives of police agencies and elected prosecutors interact to jointly determine enforcement patterns. This paper develops a game-theoretic model in which police choose evidence investment and enforcement allocation while prosecutors set evidentiary standards for charging. The analysis reveals a strategic complementarity between these choices that generates multiple equilibria, institutional path dependence, and electoral cycling in enforcement regimes. Identical formal policies can sustain sharply different enforcement environments depending on coordination history, and reforms that fail to account for this interdependence may produce null or heterogeneous effects. The framework yields direct implications for the empirical analysis of criminal justice reform, offering guidance for interpreting estimated reform effects.
Clark, Tom S. "Municipal Liability Insurance as Police Oversight: Evidence from Vallejo, California." (working paper)
Municipal liability routinely translates police misconduct into costs for local governments, yet we know little about whether liability's structure actually changes police behavior. This article studies one underappreciated mechanism: insurance. Private liability insurers can function as decentralized oversight because they price risk, impose deductibles, and provide risk-management services. When a city shifts from pooled risk-sharing to market-based coverage, the fiscal consequences of misconduct become sharper. I examine such a transition in Vallejo, California, which in 2018 left a municipal risk pool after high police liabilities and obtained market-based coverage. Using a generalized synthetic control design with monthly California data, I estimate that Vallejo's violent-crime clearances rose substantially, stabilizing at roughly 30 additional clearances per month relative to the synthetic counterfactual — on the order of a near doubling of the pre-treatment rate. This evidence suggests a shift in resources towards activities that increase the rate of crime solving, supporting the expectation that liability institutions shape ex ante incentives inside public bureaucracies, not merely compensate victims.
Clark, Tom S. "Politics, Elections, and Coercive Bureaucracy." (working paper)
When does an election change the bureaucracy beneath it? Elections may follow shocks that alter an agency's environment, making office-taking a poor date for attributing bureaucratic change. I study felony charging in the Cook County State's Attorney's Office around Kim Foxx's 2016 election. Linking arrests to review, charging, and disposition records, I distinguish changes in the distribution of stable reviewer types from common movement within reviewers. Adjusted felony-review approval fell 3.7 percentage points from 2014 to 2019, but reviewer composition explains little of the decline. Neither the Democratic primary nor Foxx's inauguration marks a detectable abrupt office-wide break; announced policies instead produce sharp, offense-specific changes. The broad decline began before Foxx took office and continued afterward. The evidence identifies the timing and organizational form of the movement, not whether scandal, anticipation, supervision, training, or other common pressures caused it. Elections may occur during bureaucratic change rather than initiate it.
Clark, Tom S., and Justin Grimmer. "Political Prosecution." (working paper)
We develop a model of political control inside a prosecutor's office. Assistant U.S. Attorneys differ in intrinsic partisan alignment, professional resistance to politically distorted charging, and the value they place on remaining in office. A politically responsive principal chooses the career pressure attached to resisting her preferred charging pattern, after which prosecutors comply, resist, or leave. Political pressure has two effects. It induces marginal prosecutors to take actions they would reject under neutral supervision, and it selectively drives more resistant prosecutors out of the office. The threat of exit constrains the principal because replacing experienced prosecutors sacrifices case-specific knowledge and ordinary prosecutorial capacity. Principal-induced political prosecution can nevertheless occur with no visible turnover when marginal prosecutors comply and remain. The model also yields two downstream implications of political charging: opponent overcharging produces weaker cases, and charging rules that respond less to conduct weaken deterrence. It also changes the empirical target. Cross-sectional partisan skew combines intrinsic preferences, induced behavior, and personnel selection; separating them requires variation across political principals while holding prosecutors fixed, together with evidence on exit and assignment. We develop these implications for a new dataset covering roughly 2.5 million federal criminal defendants.
Clark, Tom S., and John W. Patty. "Managing Intractable Problems." (working paper)
Intractable problems are inherent to governance and politics. A problem is "intractable" if it is not only difficult to agree on a solution, but also difficult to agree on what a potential solution will lead to. Intractable problems are characterized by broad and persistent low public satisfaction. We introduce two related, but distinct, notions of why reform might be pursued in a political system: complexity and complicatedness. Complexity centers on the difficulty of understanding the problem itself, and complicatedness focuses on the difficulty of making accurate decisions to mitigate the problem. We also consider how these notions are caused and/or ameliorated by having an "outsider" reconfigure the decision process, and how their presence affects the potential costs and benefits of reform by an outsider.