Published Papers

​

Testing Models of Strategic Uncertainty: Equilibrium Selection in Repeated Games 

(with Emanuel Vespa, Taylor Weidman, and Alistair J. Wilson)

forthcoming at Journal of the European Economic Association

 

In repeated games, where both collusive and noncollusive outcomes can be supported as equilibria, it is crucial to understand the likelihood of selection for each type of equilibrium. Controlled experiments have empirically validated a selection criterion for the two-player repeated prisoner's dilemma: the basin of attraction for always defect. This prediction device uses the game primitives to measure the set of beliefs for which an agent would prefer to unconditionally defect rather than attempt conditional cooperation. This belief measure reflects strategic uncertainty over others' actions, where the prediction is for noncooperative outcomes when the basin measure is full, and cooperative outcomes when empty. We expand this selection notion to multi-player social dilemmas and experimentally test the predictions, manipulating both the total number of players and the payoff tensions. Our results affirm the model as a tool for predicting long-term cooperation, while also speaking to some limitations when dealing with first-time encounters.

​

Goals, Constraints, and Transparently Fair Assignments: A Field Study of Randomization Design in the UEFA Champions League  

(with Alistair J. Wilson) 

Management Science 2023, 69(9), p. 3473-349

Online Supplementary Material, Data/Code

​

We analyze the design of a randomization procedure in a field setting with high stakes and substantial public interest: matching sports teams in the Union of European Football Association Champions League. While striving for fairness in the chosen lottery— giving teams similar distributions over potential partners—the designers seek to balance two conflicting forces: (i) imposing a series of combinatorially complex constraints on the feasible matches; and (ii) designing an easy-to-understand and credible randomization. We document the tournament’s solution, which focuses on sequences of uniform draws over each element in the final match, assisted by a computer to form the support for each draw. We first show that the constraints’ effects within this procedure are substantial, with shifts in expected prizes of up to a million euro and large distortions in match likelihoods of otherwise comparable team pairs. However, examining all possible counterfactual lotteries over the feasible assignments, we show that the generated inequalities are, for the most part, unavoidable and that the tournament design is close to a constrained-best. In two extensions, we outline how substantially fairer randomizations are possible when the constraints are weakened, and how the developed procedure can be adopted to more-general settings.

​

Balanced Growth Approach to Tracking Recessions

(with Jean-François Richard)

Econometrics 2020, 8(14)

Online Supplementary Material, Data/Code

​

In this paper, we propose a hybrid version of Dynamic Stochastic General Equilibrium models with an emphasis on parameter invariance and tracking performance at times of rapid changes (recessions). We interpret hypothetical balanced growth ratios as moving targets for economic agents that rely upon an Error Correction Mechanism to adjust to changes in target ratios driven by an underlying state Vector AutoRegressive process. Our proposal is illustrated by an application to a pilot Real Business Cycle model for the US economy from 1948 to 2019. An extensive recursive validation exercise over the last 35 years, covering 3 recessions, is used to highlight its parameters invariance, tracking and one- to three-step ahead forecasting performance, outperforming those of an unconstrained benchmark Vector AutoRegressive model.

​

​

Working Papers

​

Lab to Algorithm: Predicting AIs with Humans, and Vice Versa

(with Emanuel Vespa and Alistair J. Wilson) 

​

A now mature literature on repeated prisoner's dilemma has outlined a number of regularities in how human subjects behave. In this literature a core task is to predict when the participants will collude on the jointly cooperative action, and when they will coordinate on the myopic solution: joint defection. Orthogonal to this, a new literature in industrial organization has begun to look at when Artificial Intelligence (AI) pricing agents collude in repeated settings. In this paper we begin to explore the extent to which the regularities that show-up in human subject behavior also manifest in the behavior of pricing agents. While there are similarities, that we document, there are also points of divergence. Moving forwards, the aim is to connect both literatures: Theoretical rules developed for human subjects can be predictive for AI agents, and thereby a useful tool for theoretic exercises in predicting AI in counterfactual settings. Conversely, AI agents can be used to develop insightful experiments to further refine and test our understanding of human behavior through experiments. As such, the tasks of predicting and understanding both human and AI behavior can be symbiotic. 

​

Scoring from Difficult Angles

(with Battista Severgnini)

​

The allocation of talent is a crucial factor in determining the efficiency, inequalities, and growth trajectories of economies. While theoretical models consistently suggest that one of the main drivers of self-selection into different job market positions is opportunity costs, extensively testing this hypothesis using data proves to be extremely challenging. This paper empirically tests whether the initial conditions at birth can explain self-selection into high-risk (and thus remunerative) tasks. We collect a rich dataset on football players of English nationality and link this information with a set of macro and micro measures of economic performance. Our econometric analysis suggests a negative and significant relationship between the economic condition of the birthplace of players and their future economic performance. Furthermore, these results remain consistent even when changes in opportunity costs are driven by a quasi-experiment based on sudden and significant changes in regional funding from the European Union.

​

Screen vs Scene: Impact of News and TV on Belief Formation

(with Natalia Khorunzhina)

​

This study examines the influence of news and television on belief formation. We analyze public beliefs about crime using data from an online survey of a nationally representative US sample, comparing the results with both current and historical news, and content from popular media. We focus on the influence of streaming movies and TV shows in the US to see how media shapes opinions. By explicitly modeling belief updating we further enhance our understanding of these processes. Our findings suggest that popular culture, including both fictional and non-fictional content, significantly influences people's views. This research highlights the significant effect media representations of issues like terrorism, COVID-19 pandemic, global warming and climate change, and international relations.

​

​

Works in Progress

​

Community-Wide Responses to Unexpected and Prolonged Shocks in Safety and Public Health

(with Natalia Khorunzhina)