Combining panel data and macro information: an application to the estimation of a participation model

When studying particular subgroups of a population, like for instance lone parents, the econometrician typically has few observations at hand. In such a situation, it is vital to take advantage of any valid complementary information that may be available. In this paper we illustrate, for the estimation of a participation model for lone mothers on data from the German Socio-Economic Panel 1984-1990, the relative benefits derived from using the panel structure of the data and from including macro information in the form of extra moments, as proposed by Imbens and Lancaster. The efficiency gains we find amount to having up to six times as many observations, 'and are shared almost equally between using the panel structure optimally and including macro information.