We present a method that accommodates missing data in longitudinal datasets of the type usually encountered in economic and social applications. The technique uses various extensions of missing at random' assumptions that we customize for dynamic models. Our method, applicable to longitudinal...

We use longitudinal individual wage and employment data in France and the United States to investigate the effect of changes in the real minimum wage rate on an individual's employment status. We focus on workers employed at wages close enough to the minimum in a reference year as to be illegal...