Work in Progress

**Informal Risk Sharing and Bargaining Power**The standard models of risk sharing focus on the testable implications if constraint efficient informal insurance allocations of output, labor and consumption. These allocations may be different among otherwise homogeneous households, an asymmetry typically measured by the estimated Pareto weights of the observed allocation. However, these are endogenous to the environment itself, and may not be good measures of the implied bargaining power amongst households. Our first contribution is to provide testable implications on the implied Pareto Weights of different models of cooperative bargaining solutions. Specifically, we consider the ex-ante and ex-post bargaining solutions of Nash (1950) and Kalai and Smorodinsky (1975), which specify how relative risk preferences, productivity and income streams of households determine their shares in the informal risk sharing agreement.

Our second contribution consists on defining measures of bargaining power that are theoretically orthogonal to the risk sharing environment description. These are obtained by fitting the observed data to the asymmetric versions of the cooperative bargaining solutions studied, thus delivering relative bargaining weights implied by the observed allocation. Under the hypothesized bargaining solution concept, these measures should be invariant to changes in the physical environment, and hence may be interpreted as household specific bargaining power. This could be related to some other observable characteristics, like political power and relative network centrality.

**A note on Rationalizability in Infinite, Dynamic games**I extend the main results of Battigalli (2003) and Battigalli and Siniscalchi (1999, 2003) to environments with compact topological spaces of actions and payoff parameters, with potentially infinitely lived agents. This is necessary to be able to deal with continuum of action and payoff types, which are widespread in the literature on repeated oligopoly competition, risk sharing, dynamic contracting and dynamic mechanism design in general, among other applications. Under continuity at infinity and topological compactness assumptions, we provide an analog to their Universal Type Space Theorem, and prove basic topological properties of the sets of Weak and Strong Rationalizable Strategies