JPSM Presentation : Rare Events Across Time and Space

This research compares the effects of family breakups, unemployment, becoming ill, and aging between Germany and the United States, and across time. The research question of the project requires the use the American Panel Study of Income Dynamics (PSID) and the German GSOEP for the analyses. Several features of the two data sets, however, raise doubts on the comparability of the measured effects of four key life events. The doubts arise from the different target populations, tracking concepts, and the nonresponse adjustment of the two panel studies. The doubts are intensified by the necessity of using an unbalanced panel design for the analysis. It is shown that not all problems can be solved with one research strategy. The paper, however, proposes a compromise that involves an unbalanced panel design, with fine tuning of target populations, estimating non-coverage bias, and a revised set of longitudinal weights optimized for comparability.