Working Papers
Homophily and Specialization in Networks, with Luca Paolo Merlino [Supplemental Appendix]
Revise & Resubmit at the Economic Journal
In this paper, we study a game where players contribute to two local public goods for which they have different tastes and sponsor costly links to enjoy the provision of others. We characterize the game’s equilibria and show that linking costs influence who contributes vs. who free rides. Lower linking costs can reduce welfare and increase polarization. When the available budget is small, a social planner should subsidize players who are already providing significant amounts of public goods. With a larger budget, subsidies should target a single player who provides the public good bundle that is most beneficial to society.
Revealing Information - or not - in a Social Network of Traders, with Paolo Pin and Fernando Vega-Redondo
We propose a simple micro-founded model of trading with ex-ante asymmetric information similar to one proposed by Kyle (1985) in which the equilibrium price is fully revealing under rational expectations. We analyze under which conditions a privately informed trader may want to share her information with other traders for free. Despite the strictly competitive setup and conventional wisdom, we show that there is a unique separating equilibrium in which the informed trader reveals some signals and conceals others. A consequence of this is that the price need not be fully revealing of the aggregate information in the market (even if traders are risk neutral), which in turn has welfare implications on the distribution of the social surplus at equilibrium. We establish these results for a context where the pattern of communication among traders is restricted by a given social network, studying as well what network arises when links are established endogenously.
Ideology and Tolerance in Networks
Agents with heterogeneous ideologies select an interval of tolerable types and invest effort to form connections within this range. The resulting weighted network classifies agents as allies, opponents, and defines the strengths of alliances. Agents outside each other’s tolerance intervals are opponents, generating benefits that increase with (i) a player's relative strength and (ii) network effects, which reflect the number of shared allies disputing the same opponent. When network effects are strong, players form cliques, whereas weaker network effects lead to opponents sharing mutual allies. Effective policy interventions to reduce polarization depend critically on the initial structure of the network.
Socially Disadvantaged Ethnic Groups and Distributive Politics, with Vangjel Bita and Aldo Elizalde
This study examines whether socially disadvantaged ethnic groups practise favouritism when they reach national power. We use the case of Bolivia, where Evo Morales, a member of the traditionally disadvantaged Indigenous Aymara, became the first Indigenous president in 2005. We develop a theoretical framework to uncover the necessary conditions for favouritism. The incentives to practise favouritism vary according to the income and size of an ethnic group. In municipalities with higher incomes or relatively few Aymaras, favouritism is more likely. We investigate whether different forms of favouritism occur using a fuzzy regression discontinuity design and a first difference estimator. Our empirical results show no evidence of favouritism towards Aymaras at the national level. However, we do find evidence for favouritism in high-income areas or where few Aymara live. Ethnic groups thus differ in how much they tend to practise favouritism and how much they suffer when others do so.