Estimating the effect of a variable in a high-dimensional regression model

A problem encountered in some empirical research, e.g. growth empirics, is that the potential number of explanatory variables is large compared to the number of observations. This makes it infeasible to condition on all variables in order to determine whether a particular variable has an effect. We assume that the effect is identified in a high-dimensional linear model specified by unconditional moment restrictions. We consider properties of the following methods, which rely on lowdimensional models to infer the effect: Extreme bounds analysis, the minimum t-statistic over models, Sala-i-Martin’s method, BACE, BIC, AIC and general-tospecific. We propose a new method and show that it is well behaved compared to existing methods.