In this paper, we investigate the use of interactive effect or linear factor models in regional policy evaluation. We contrast treatment effect estimates obtained using Bai (2009) with those obtained using difference in differences and synthetic controls (Abadie and coauthors). We show that difference in differences are generically biased, and we derive support conditions for synthetic controls. We construct Monte Carlo experiments to compare these estimation methods in small samples. As an empirical illustration, we provide an evaluation of the impact on local unemployment of an enterprise zone policy implemented in France in the 1990s.

In this paper, we estimate a transition model that allows for measurement errors in the data. The measurement errors arise because the survey design is partly retrospective, so that individuals sometimes forget or misclassify their past labor market transitions. The observed data are adjusted for errors via a measurement-error mechanism. The parameters of the distribution of the true data, and those of the measurement-error mechanism are estimated by a two-stage method. The results, based on the 1990-1992 French labor force survey, show that neglecting measurement errors leads to an underestimation of the average durations spent in labor market states. The estimates of some important transition probabilities between states are also biased by the measurement errors.