Pierre-Carl Michaud
| Date of Ph.D. defense: | May 20, 2005 |
| Title of thesis: | Dynamic Panel Data Models and Causality: Applications to Labor Supply, Health and Insurance |
| ISBN: | 90 5668 142 7 |
| Promotores: | Prof. dr. Arthur van Soest, Prof. dr. Jan van Ours |
Abstract: Full text
Dynamic Panel Data Models and Causality: Applications to Labor Supply, Health and Insurance is a collection of essays demonstrating how dynamics in panel data can help uncover causal relationships that are helpful to understand microeconomic behavior. These methods are applied to the analysis of labor supply decisions of elderly couples, the process determining wealth accumulation and health trajectories of elderly couples and finally, to the analysis of asymmetric information in dynamic insurance contract arrangements.
One of the main findings concerns the importance of common persistent factors, or unobserved traits of respondents, in order to study dynamic relationships between two variables of interest using panel data. The "hand of the past" can reinforce existent causal relationships, or blur their effect, potentially leading the analyst to misleading conclusions. In applications using elderly respondents, one must be aware that life-cycle aspects of their decisions and persistent unobserved traits of respondents reinforce the association between a multitude of socio-economic and health variables within households. In insurance markets, the dynamics introduced by experience rating of insurance premiums potentially mask the existence of moral hazard, a short-term response to changes in insurance coverage. In the different applications considered, it is shown how dynamic panel data models can be used to identify causal relationships of interest.
Global / English