Local Institutions and the Dynamics of Community Sorting

Abstract

This paper studies the dynamics by which populations with heterogeneous preferences for public good provision sort themselves into communities. I conduct laboratory experiments to consider which institutions best facilitate efficient self-organization when residents can move freely between locations. I find that institutions requiring all residents of a community to pay equal taxes enable subjects to sort into stable, homogeneous communities. Though sorted, residents often fail to attain the provision level best suited for them. When residents can vote for local tax policies with ballots, along with their feet, each community converges to the most efficient outcome for its population. (JEL C73, D72, H21, H41, H71, H73)

Choosing peers: Homophily and polarization in groups

“…In particular, our mutually optimal groups would be "stable" under any protocol specifying which groups would be willing to add which types of agents. 13 In fact, all our results would remain qualitatively similar were we to assume that agents have more optimistic beliefs, predicting an equilibrium will be selected uniformly at random from the set of all equilibria in the task-selection game. That way, coordination on whom should contribute is resolved by randomization.…”

“…Second, our analysis will show that requiring only extreme agents, those who care su¢ ciently much about one of the two tasks, rather than all agents in a group, to hold such pessimistic beliefs would generate an identical characterization to the one presented here. 13 This notion of mutual optimality is a natural equilibrium condition for the group-selection stage in an environment that allows individuals in large populations to connect in an unconstrained way. 14 For example, university students have a wide choice of which association to join, or to form together with their peers.…”

“…This allows us to derive predictions on the size of mutually optimal communities, and on the correlation between size and other group features such as taste heterogeneity. 7 Robbett (2013Robbett ( , 2014 provides observations from an array of experiments that allow individuals with heterogenous preferences to dynamically move between communities in which a public goods game is later played, considering di¤erent institutions that determine the level and enforcement of individual contributions. She illustrates how chronic movements between communities may persist, that high connection fees in a location attract individuals of a particular type, and that voting on public goods provision helps in terms of utilitarian e¢ ciency.…”

Community dynamics in the lab

“…This paper uses a linear public goods game, in line with the body of previous work; the experiment lasts for 65 periods, allowing for sufficient time for the cyclical nature of community dynamics to play out; and the paper studies the effect of congestion and entry fees on community stability. The finding in this paper that individuals often refuse to remain in a location where others are contributing less than they are, even when the presence of free-riders does not harm the payoffs of contributors, suggests that community stability may depend on a punishment mechanism to support contributions, as in Gürerk et al (2006), or a local tax that all members are required to pay, as in Robbett (2014). The latter experiment used non-linear payoff functions to generate single-peaked preferences over tax rates, finding that tax institutions could allow agents to sort by their preferred tax rate and, when taxes could be voted on, achieve optimal provision levels within sorted groups.…”

Voting with hands and feet: the requirements for optimal group formation

“…Robbett (2010Robbett ( , 2014 each find that frequent movement leads to suppressed earnings in linear and nonlinear pure public goods games with free mobility.A. Robbett…”