ABSTRACT
This chapter discusses consequences and solutions of unit- and time-specific effects in panel data models and model transformation to eliminate the unit-specific effects. It describes dynamic panel data models and unit-specific effects and fixed effects models vs. random effects models. The chapter points out that previous attempts at estimating the economic model of crime with aggregate data (aggregated over time, so one observation for each unit) relied heavily on cross-section econometric techniques, and therefore do not control for unobserved unit-specific heterogeneity. It examines several explanations of the persistent wage differentials between industries. One competitive-market explanation is that the inter-industry wage differentials are due to differences across workers in “unobserved” ability (unobserved to the researchers). The chapter also examines the international tourism demand to the Canary Islands by estimating a dynamic panel data model. The model is a form of log-linear function and its coefficients measure the demand elasticities.
