Apr 19, 2018
We examine how pension rule changes affected teacher retirement by estimating an option value retirement model using data from a large cohort of late career Missouri public school teachers. In so doing we offer potential solutions to several statistical challenges that often arise in estimating structural models of retirement using large panel data sets. The first challenge concerns modelling the formation of expectations of future pension rules on the part of late career teachers. The second challenge is bias induced by baseline sample selection: in baseline cohorts we only observe teachers who are still working. This is a bias that also evolves with pension rule changes. A third challenge arises from estimation using large panels of micro-data on individual teachers. The teacher-level data can be difficult to obtain and the likelihood of teacher-data is costly to compute. We address these challenges by incorporating policy expectations and sample selection directly into estimation of the likelihood function. We also show that the likelihood can be efficiently estimated by using teacher-data grouped by age and experience cells, which permits: a) estimating structural models of teacher retirement with data that are more widely available; b) use of longer panels, and c) dramatic reductions in computation cost. Counterfactual simulations of the estimated structural model suggest that Missouri's pension enhancements reduced the average retirement age by about 0.3 years for the 1994 cohort and by more than one year in a steady state.