Dynamic panel data models and causality: Applications to labor supply, health and insurance

P.C. Michaud

Research output: ThesisDoctoral Thesis

319 Downloads (Pure)

Abstract

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.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Tilburg University
Supervisors/Advisors
  • van Ours, Jan, Promotor
  • van Soest, Arthur, Promotor
Award date20 May 2005
Place of PublicationTilburg
Publisher
Print ISBNs9056681427
Publication statusPublished - 2005

Fingerprint Dive into the research topics of 'Dynamic panel data models and causality: Applications to labor supply, health and insurance'. Together they form a unique fingerprint.

Cite this