Initial Conditions Problem in Event History Analysis: An Indirect Inference Procedure
In event history analysis, the problem of initial conditions is associated with the first spells being left-hand censored (LHC). Currently there are two ways of treating the LHC spells: the Bayesian approach and the direct likelihood approach. Both require strong assumptions about functional forms. This paper proposes an indirect inference procedure which uses simple simulation to correct the discrepancy of a maximum pseudo likelihood estimator.
KeywordsTransition Intensity Event History Analysis Indirect Inference Maximum Pseudo Likelihood Initial Condition Problem
- An, M. Y. (1992): “Econometric Analysis of Sequential Discrete Choice models,” CAE Working Paper x #92–03, Cornell University.Google Scholar
- Gallant, A.R. and G.E. Tauchen (1992): “Which Moment to Match?” Manuscript, Duke University.Google Scholar
- Heckman, J.J. and B. Singer (1986): “Econometric Analysis of Longitudinal Data,” in Z. Griliches and M.D. Intriligator (eds.) Handbook of Econometrics Vol. III, 1917–1978.Google Scholar