Public Debt and the Political Economy of Reforms

Co-authors: Christoph Esslinger and Brian Roberson

We develop a two-period model of redistributive politics in which two politicians compete in an election in each period. In the first period, the politicians propose both whether to experiment with an efficient reform with uncertain benefits and choose the amount of public debt. Politicians also allocate pork-barrel spending to voters in each period. We show that allowing politicians to raise debt ensures that the reform is always implemented when the reform's ratio of private good to public good gains exceeds a threshold, i.e. the reform generates enough private good benefits. This is not the case when the reform's ratio of private good to public good gains is below this threshold. While both debt and spending limits reduce the success of the reform in the political process, debt limits are always at least as efficient as the corresponding spending limits.

Personal webpage

Unpaired Kidney Exchange: Overcoming Double Coincidence of Wants without Money

Co-authors: Mohammad Akbarpour, Yinghua He, Victor Hiller, Robert Shimer and Olivier Tercieux

We propose a new matching algorithm—Unpaired kidney exchange—to tackle the problem of double coincidence of wants without using money. The fundamental idea is that 'memory' can serve as a medium of exchange. In a dynamic matching model with heterogeneous agents, we prove that average waiting time under the Unpaired algorithm is close-to-optimal and substantially less than the standard pairwise and chain exchange algorithms. We evaluate this algorithm using a rich dataset of the kidney patients in France. Counterfactual simulations show that the Unpaired algorithm can match nearly 57% of the patients, with an average waiting time of 424 days (state-of-the-art algorithms match about 31% with an average waiting time of 675 days or more). The optimal algorithm performs only slightly better: it matches 58% of the patients and leads to an average waiting time of 410 days. The Unpaired algorithm confronts two incentive-related practical challenges. We address those challenges via a practical version of the Unpaired algorithm that employs kidneys from the deceased donors waiting list. The practical version can match nearly 87% of patient-donor pairs while reducing the average waiting time to about 141 days.

Personal webpage

Crime Entanglement, Deterrence, and Witness Credibility

Co-author: Bruno Strulovici

When a defendant is accused of multiple crimes, one may consider punishing him if the probability that he has committed at least one of these crimes is high. We show that entangling criminal accusations in this way can severely harm deterrence and reduce witnesses’ credibility. When conviction entails a large punishment, an individual’s decisions to commit distinct crimes are substitutes and witnesses’ reports are complements. This tension induces negatively correlated private information and a coordination motive among witnesses, which leads to uninformative reports, ineffective deterrence, and frequent crimes in equilibrium. We discuss various remedies to restore edibility and reduce crime.

Personal webpage

A behavioral approach to analyzing games with strategic complementarities

Co-authors: Wouter Kager and Alvaro Sandroni

Games with strategic complementarities have many applications but can be difficult to analyze, in large part because they generally have many equilibria. Building on research in psychology on perspective-taking, we develop a new equilibrium refinement and obtain novel comparative statics and welfare results. Applications are discussed.

Personal webpage

The Winner-Take-All Dilemma

Co-author: Kazuya Kikuchi

This paper considers collective decision-making when individuals are partitioned into groups (e.g., states or parties) endowed with voting weights. We study a game in which each group chooses an internal rule that specifies the allocation of its weight to the alternatives as a function of its members' preferences. We show that under quite general conditions, the game is a Prisoner's Dilemma: while the winner-take-all rule is a dominant strategy, the equilibrium is Pareto dominated. We also show asymptotic Pareto dominance of the proportional rule. Our numerical computation for the US Electoral College verifies the sensibility of the asymptotic results.

Personal webpage

A Welfare Analysis of a Steady-State Model of Observational Learning

Co-author: Eszter Kabos

TBA

Personal webpage

Macroeconomics with Learning and Misspecification

This paper explores a form of bounded rationality where agents learn about the economy with possibly misspecified models. I consider a recursive general-equilibrium framework that nests a large class of macroeconomic models. Misspecification is represented as a constraint on the set of beliefs agents can entertain. I introduce the solution concept of constrained-rational-expectations equilibrium (CREE), in which each agent selects the belief from her constrained set that is closest to the endogenous distribution of observables in the Kullback–Leibler divergence. If the set of permissible beliefs contains the rational-expectations equilibria (REE), then the REE are CREE; otherwise, they are not. I show that a CREE exists, that it arises naturally as the limit of adaptive and Bayesian learning, and that it incorporates a version of the Lucas critique. I then apply CREE to a particular novel form of bounded rationality where beliefs are constrained to factor models with a small number of endogenously chosen factors. Misspecification leads to amplification or dampening of shocks and history dependence. The calibrated economy exhibits hump-shaped impulse responses and co-movements in consumption, output, hours, and investment that resemble business-cycle fluctuations.

Personal webpage

Social Learning and Social Context

People gather information from their peers to improve their decisions in many situations. I investigate the impact of a communication friction on which social learning networks form and on the quality of information transferred over these networks. In my model, agents cannot perfectly communicate their beliefs because of agent-level idiosyncrasies in expression. Social context (knowing what information a peer receives from her peers) allows one to better understand these idiosyncrasies and so better understand their information. If agents have sufficient time to communicate, a directed cycle network allows all agents to learn all information in society despite the idiosyncrasies in expression. With more limited time to exchange information, agents must trade off the reach of their network (i.e., from how many agents they will receive some information) with the clarity of the announcements (i.e., how well they learn the signal of any one agent they observe). Social context acts as a substitute for familiarity and can result in novel non-monotonicities in overall welfare with varying underlying parameters.

Personal webpage

A solution to the two-person implementation problem

Co-authors: J-F Laslier and R Sanver

We propose a solution to the classical problem of Hurwicz and Schmeidler (1978) and Maskin (1999) according to which, in two-person societies, no Pareto efficient rule is Nash implementable. To this end, we consider implementation through mechanisms that are deterministic-in-equilibrium while lotteries are allowed off-equilibrium. Under a very weak condition for extending preferences over lotteries, we build simple veto mechanisms that Nash implement a class of Pareto efficient social choice rules called Pareto-and-veto rules. Moreover, under mild richness conditions on the domain of preference extensions over lotteries, any Pareto efficient Nash implementable rule is a Pareto-and-veto rule and hence is implementable through one of our simple veto mechanisms.

Personal webpage

Time Inconsistency in Stress Test Design

When publishing bank stress test results, the Central Bank (CB) faces a time inconsistency problem. Before the stress test is conducted, the CB wants to appear tough since the threat of letting banks fail the stress test incentivises prudent behaviour. After the stress test is conducted, the CB wants to act soft by releasing only partial information in order to increase the perception of bank health. The CB is particularly concerned about weak banks, which might otherwise be further destabilised by belief-driven phenomena such as runs. This time inconsistency problem creates a trade-off: Any communication framework for the results of stress tests trades off reassurance and incentives. We characterise the socially optimal communication framework and show that delegating stress tests to a sender with preferences that differ from welfare maximisation can be optimal, as delegation is a partial substitute for commitment. We show that a hurdle rate framework, where all banks are judged against a common threshold and which is used in practice, can be optimal but implies that stress tests themselves become an informational contagion channel as changes in one bank’s fundamental health affect the perceived health of rivals.

Personal webpage

Optimal income taxation in the presence of networks of altruism

This paper characterizes the optimal income tax in the presence of networks of altruism. Individuals are modelled as being part of networks of altruism when they take into account the private utilities of other agents into their own social utility. This materializes through monetary transfers that take place between group members. Agents then have two dimensions of choice: their productive effort and the transfers they make to other members of their group. I derive the structure of incomes and transfers when there is no governmental intervention, and show that under certain conditions on two parameters – the heterogeneity of skill levels and the extent of altruism within groups – redistribution takes place at the group level. This then enables me to characterize theoretically the optimal linear and non-linear income tax schemes. I find complementarity between government and informal redistribution so that optimal marginal tax rates should potentially be adjusted downward. This depends on the two aforementioned parameters, as well as on the information available to the government. These results can then be applied in a variety of settings, especially that of developing countries, where these networks of informal transfers are widespread.

Personal webpage

Robust Monopoly Regulation

Co-author: Yingni Guo

TBA

Personal webpage

Better alone? Evidence on the costs of intermunicipal cooperation

While central governments encourage intermunicipal cooperation to achieve economies of scale, municipalities are generally reluctant to integrate. This paper provides new evidence on the costs of integration by assessing the causal impact of integration on municipalities resisting cooperation. Exploiting a 2010 reform in France that forced non-integrated municipalities to enter an intermunicipal community, I show that the loss of autonomy over urban planning plays a key role in explaining their opposition. Using a difference-in-differences strategy, I find that municipalities forced to integrate experienced a 12.2 per cent increase in the number of building permits delivered. Additional results suggest that municipalities were resisting further urban development to avoid congestion rather than to exclude minorities or prevent housing prices to fall. Exploring the impact of integration on fiscal revenues and public services, I find that, while congestion costs appear key in understanding urban municipalities' resistance, rural municipalities' opposition seems mainly driven by the loss of access to local public services. Overall, these results shed new light on the tension between the potential aggregate benefits sought by governments and local negative externalities driving municipalities’ opposition.

Personal webpage

Subjective complexity and comprehension under uncertainty

Departures of choice behavior under uncertainty from the predictions of expected utility theory are widely observed. Complexity of the problem of choosing among uncertain acts seems to be a salient feature of many of the environments in which these phenomena occur. I propose and axiomatize a model of choice under uncertainty which relates complexity to the size of the partition with respect to which acts are measurable. I derive a representation of incomplete preferences in which acts that are subjectively complex from the perspective of the decision-maker are bracketed by simple acts to which they are related by uniform dominance. I also consider decision-makers whose behavior can be characterized by an incomplete preference relation and a ``completion'' thereof. I discuss the relationship between this model and models of ambiguity and regret aversion. I also show how the model can be used to understand observed behavior in contracting, choice of insurance plans, portfolio choice, and equilibrium asset pricing. I also develop general comparative statics techniques.

Personal webpage