Consumption Risk-Sharing in Social Networks

Abstract

We develop a model in which connections between individuals serve as social collateral to enforce informal insurance payments. We show that: (i) The degree of insurance is governed by the expansiveness of the network, measured with the per capita number of connections that groups have with the rest of the community. “Two-dimensional” networks—like real-world networks in Peruvian villages—are sufficiently expansive to allow very good risk-sharing. (ii) In second-best arrangements, insurance is local: agents fully share shocks within, but imperfectly between endogenously emerging risk-sharing groups. We also discuss how endogenous social collateral affects our results. (JEL D85, G22, O15, O17, Z13)

Structural Experimentation to Distinguish between Models of Risk Sharing with Frictions in Rural Paraguay

“…Ligon et al (2002) and Laczó (2015) show that observed changes in households' income and consumption flows are consistent with a model of limited commitment. Ambrus et al (2014) construct a model of risk-sharing in social networks with limited commitment and show that networks and transfers in Peruvian villages fit the predictions of this model. On the other hand, Ligon (1998) provides evidence that in at least some Indian villages private information may be important, while Ambrus et al (2017) construct a model of risk-sharing in social networks with hidden information and Milán (2016) then shows that bilateral transfers in Bolivia are in accord with that hidden information model.…”

The Formation of Partnerships in Social Networks

“…This characterization of the optimal strategy (which is obtained by induction on the size of the graph) relies on the identification of all maximum matchings of a graph, and of the players who belong to all maximum matchings. 3 It immediately implies that a maximum matching is obtained in equilibrium, as players will never break a link which reduces the number of matches in the graph. In this baseline model, the maximum number of pairs is formed because agents delay the formation of the partnership until the social network is such that there is a risk that they will not be able to find a partner in the long run.…”

“…Ambrus, Möbius and Szeidl (2014) are all papers which share the same basic structure as our model, where agents request favors or transfers at different points in time, and favors or transfers are enforced through reciprocation in the future. The paper in the literature on favor exchange which is closest to our model is the paper by Jackson, Rodriguez-Barraquer and Tan (2012).…”

Economic analysis of resilience: A framework for local policy response based on new case studies

“…Indeed, we see this reflected in the Clifton, 2011, case study in this project, where family groups based on "trust and obligation" form the basis of the insurance provided to the society's members, although this study also indicates that the local gift exchange system has existed side-by-side with wider trading relationships at some points in history. Ambrus et al, 2010, propose village elders as natural intermediaries for their own analysis and could perhaps be thought of as a good place to start in developing more complete insurance systems.…”

“…Ambrus et al, 2010, pose the question of whether local arrangements can, in aggregate, supply good global risk-sharing where threats are in the form of idiosyncratic income shocks. Note that this implies that resource constraints are not an issue in their paper so that the advantage of large insurance networks of pooling large amounts of resources is not present.…”