Missing data are a common challenge facing empirical researchers. This paper presents a general GMM framework and estimator for dealing with missing values of an explanatory variable in linear regression analysis. The GMM estimator is efficient under assumptions needed for consistency of linear-imputation methods. The estimator, which also allows for a specification test of the missingness assumptions, is compared to existing linear imputation, complete data, and dummy variable methods commonly used in empirical research. The dummy variable method is generally inconsistent even when data are missing completely at random, and the dummy variable method, when consistent, can be less efficient than the complete data method.

We examine inference in panel data when the number of groups is small, as is typically the case for difference-in-differences estimation and when some variables are fixed within groups. In this case, standard asymptotics based on the number of groups going to infinity provide a poor approximation to the finite sample distribution. We show that in some cases the t -statistic is distributed as t and propose simple two-step estimators for these cases. We apply our analysis to two well-known papers. We confirm our theoretical analysis with Monte Carlo simulations.