Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
This paper examines the implications of uncertainty in gravity models of trade due to the violation of the Law of Large Numbers (LLN) that we document in the data. When the number of available technologies is finite and the LLN does not hold, the variance of the stochastic component in gravity models is large, which leads to the poor goodness of fit of gravity models and high uncertainty in comparative statics results. We offer a procedure that specifies counterfactual predictions in terms of distributions rather than point estimates and helps to account for such uncertainty.