Standard economic models have long been applied to choices over private consumption goods, but have recently been extended to incorporate social situations as well. We challenge the applicability of standard decision theoretic models to social settings. We argue that, in economically relevant social settings, agents may choose to randomize over any of the deterministic outcomes in a way that clashes with standard decision theory axioms that require lotteries not to be valued strictly above the best deterministic outcome. Thus, mainstream economic analysis is unable to fully accommodate such choices. We find little evidence of such deviations in non-social settings.
On the Difference Between Social and Private Goods
We document a revealed preference for randomization for “social goods”, while such non-standard behavior is not present for private consumption goods.
International Economic Review, 2011, 52(1), pp 85-104. With M. Galenianos and G. Virag. [technical appendix] In directed search with a finite population, minimum wages improve employment but reduce output and efficiency, and reverse for unemployment benefits. Go to paper
Econometrica. 2018 86(1): 85-132. With Jan Eeckhout. When heterogeneous firms can choose both how many and which workers to hire, we illustrate consequences for firm-size and wage inequality. Note a correction for the condition with capital: corrigendum. Go to paper
Journal of Political Economy, 2017, 124(1), 224-264. With G. Grossman & E. Helpman. (simulations, matlab). We introduce two-sided heterogeneity into a Hecksher-Ohlin-style trade model to study factor reallocation and wage inequality within and across sectors. Go to paper
Journal of Economic Theory, 2010/145, 1354-1385. With Jan Eeckhout. Search affects competing mechanisms: if meetings with low types reduce those of high types, price posting and market separation replace auctions. Go to paper
Economic Journal, 2021 131: 713-744. With Ericson, Spinnewijn &, Starc Demand for insurance can be driven by high risk aversion or high risk, and we show how to separate the two using observed market shares. Go to paper