Mechanism Design in Law and Economics
Paper Session
Monday, Jan. 4, 2021 3:45 PM - 5:45 PM (EST)
- Chair: Yeon-Koo Che, Columbia University
Optimal Standards of Proof in Patent Litigation: Infringement and Non-Obviousness
Abstract
We build a model of innovation and patent adjudication under two forms of uncertainty; uncertainty regarding whether the original invention merits protection (non-obviousness), and uncertainty as to whether a particular competitor’s product should be barred (infringement). We find that when it is practical to increase the rewards from innovation by extending patent length, the standards of proof for non-obviousness should be high. The intuition for this is that patent length should be set so that the increase in innovation from extending patent length is balanced by the increase in deadweight loss from extending monopoly pricing. In this situation, the ex-ante cost of failing to protect a good patent is minimal, but there is substantial deadweight loss from protecting a bad patent. In contrast, if non-infringing competing inventions substantially decrease the original inventor’s profits, it might be desirable to have a very low standard of proof for infringement, since the deadweight loss from an incorrect finding of infringement is mostly balanced out by the increased ex-ante incentive to invent.Judicial Mechanism Design
Abstract
This paper studies the design of welfare-maximizing criminal judicial processes. We identify properties of the generically unique optimal processes for two notions of welfare distinguished by their treatment of deterrence. These properties shed new light on features of the criminal justice system in the United States, from the prevalence of extreme, binary verdicts in conjunction with plea bargains to the use of jury instructions and an adversarial system, all of which emerge as the result of informational, commitment, and incentive arguments.Settling Lawsuits with Pirates
Abstract
A firm licenses a product to overlapping generations of heterogeneous consumers. Consumers have the choice to purchase the product, pirate/steal it, or forego its use. Higher consumer types enjoy higher gross benefits from product use and, if they choose to pirate/steal the product, are caught at a higher rate. If caught stealing, consumers face legal sanctions. We show that the firm will commit to an out-of-court settlement policy that is "soft" on pirates, so high-type consumers purchase the product and low-type consumers pirate/steal it. This scheme facilitates price discrimination and gives the firm an additional stream of revenues. We show that firm profits are even higher if the firm can bundle a license agreement with the cash settlement. However, requiring pirates to sign license agreements as part of the settlement may deter the entry of more efficient competitors.Predictive Policing
Abstract
We study a model of dynamic, data-driven enforcement policy making, known as "predictive policing." The crucial aspect of predictive policing is a possible link and feedback loop between the allocation of enforcement resources and the collection of information/data. To incorporate these features, we adapt the continuous-time exponential bandit model of Keller, Rady and Cripps (KRC 2005). In our model, a policymaker (PM) allocates at each point in time a budget of unit enforcement resource between two neighborhood: non-risky one in which the (Poisson) arrival rate of crime opportunity is fixed, and risky one in which the rate is uncertain and evolves according to Markov chain. We first consider a model where the crime probability in both neighborhoods is exogenously fixed, and study the optimal policy. Next, we endogenize the crime probability so that criminals respond to the amount of enforcement resources allocated to their neighborhoods. In the resulting dynamic inspection game between criminals and law enforcement, there exists an equilibrium in which the equilibrium policy sacrifices exploitation to avoid exploration, Instead of sacrificing exploitation to enable exploration-the usual prescription from bandit literature. This "anti-learning bias" result stands in contrast to the standard "learning bias" wisdom from the literature and suggests that endogenizing the value of bandit arms could dramatically overturn the main tenet of the bandit literature.JEL Classifications
- K0 - General
- D8 - Information, Knowledge, and Uncertainty