12:00

Paul Muller (VU University Amsterdam)

Methods to evaluate the Effects of Job Search Assistance
In March 2010, budgetary problems at the Dutch UI administration caused a sharp drop in the availability of job search assistance programs. We discuss various micro-econometric methods exploiting this discontinuity in the treatment rate and compare the results to non-experimental methods. This comparison allows an assessment of the plausibility of the assumptions made in non-experimental methods. In the empirical analysis, we use a large and rich administrative data set provided by the UI administration. We estimate treatment effects using regression discontinuity, difference-in-differences, matching and timing-of-events. Field: applied micro

*******

12:45

Bastian Ravesteijn (Erasmus University Rotterdam)

Good, Bad, and Ugly Controls: On Intervening Control Variables
Many studies in economics include an intervening control variable in an OLS regression model to estimate direct effects or partial associations. I investigate the size and the sign of the bias of the resulting estimators if the intervening control variable is correlated with the error term due to omitted variable bias or measurement error. The problem of endogenous intervening control variables can be solved if sources of exogenous variation in the intervening variable are available. I will use the effect of schooling on health as a running example, and illustrate the consequences of using endogenous intervening controls such as income, using data from the National Health Interview Survey. Field: applied economics