Research

We provide a competitive equilibrium theory of urban segregation in a linear city. Households demand consumption and housing along the city and are exposed to neighborhood externalities. We show that equilibria that are robust to small coalitional deviations are segregated. Our results explain urban segregation in a standard neoclassical framework and shed new light on the difficulties faced by authorities to integrate cities.

Accessible knowledge and public information are of tremendous importance, yet acquiring and transmitting information usually takes place in environments full of frictions. There are three important tensions that characterize most situations in which public information is being produced and consumed: it conveys costs, it creates conflicts of interest and it is received by heterogeneous agents. I model the interaction between a unitary mass of receivers of a costly public signal and a sender who has private interests on the actions they take. More specifically, I take a binary state of the world version of the standard Bayesian persuasion framework with costs and extend it in two directions: allowing for heterogeneous priors and for hidden conflicts of interest. 

The first result is that the optimal strategy of the sender is to employ a cut-off rule. She can exploit the asymmetry of information by sending truthful messages when stakes are low and lying when stakes are high. This way she increases her expected returns conditional on sending false information, without decreasing the average reliability of the signal. This effect is particularly important when priors are heterogeneous. In the case of uniformly distributed beliefs, there can be no communication in equilibrium if conflicts of interest are observable. And more generally, in the case of arbitrary distributions, perfect information minimizes the sender's rents. On the contrary, when conflicts of interest are unobservable and sufficiently spread, the sender can get arbitrarily close to the revenue maximizing outcome, regardless of how priors are distributed. These findings shed light on the difficulties of regulating information, which can reduce the sender's incentives to invest and may end up completely eliminating communication, and it opens up the alternative of exploring possible ways of generating publicly funded information.

During the past decade there have been serious concerns over the reliability of scientific research and especially some features suggest there might be a large amount of false positives. Journals tend to significantly favor striking results and small p-values, researchers frequently repeat trials and conveniently analyze data, experiments and other tests are seldom reproduced. With the outbreak of information design in economics, several papers have tried to explain these characteristics as the result of the strategic interaction between a sender and a receiver of a signal. But what if some of these failures are simply the result of the professionalization of research and the strong individual incentives faced by scientists?

In this paper I apply the classic Spence (1973) signaling model to academic job market incentives. Just like in the traditional model, researchers with heterogeneous skills use publications as a channel to signal their types to employers. In this framework there is an additional group of agents, the journals, who act as intermediaries for the message between workers and firms. I allow for a large class of objective functions for journals, that may or may not be aligned with social welfare, and show that in the presence of competition research pursued in stable equilibria is independent of journals' goals. Researchers' signaling incentives outweigh journals in deciding what type of research is conducted, and they manage to separate in the cheapest way possible. Finally I show that this model can be applied to explain theoretically some observed features of scientific publications, like the omission of insignificant results, the striking result bias and formalism without connection to reality.