Abstract
We survey some of the key problems confronting empirical applications in labor economics and how panel data can be utilized to robustly estimate parameters of economic interest. Topics include whether and how to introduce heterogeneity in intercept and slope parameters; measurement errors in regressors; endogeneity bias and associated panel instrumental variables estimators; sample composition dynamics to control for selection on (un)observables; and model specification and selection issues such as a static or dynamic framework. The pedagogical framework for our discussion is the Mincer-type human-capital wage equation used in quantifying the value of statistical life (VSL) in labor market data. We show in our application that whether we control for latent intercept heterogeneity is more important than how, and conditional on latent heterogeneity, the robust range of VSL estimates is in the $6 million $10 million range when we additionally address measurement error, endogeneity, and dynamics.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
