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 ...