Identification and Estimation of Continuous Time Dynamic Games

Jason R. Blevins.
The Ohio State University, Department of Economics
Unpublished Manuscript.

Versions and Availability:

Abstract. We consider the theoretical and econometric properties of a recently proposed class of continuous time dynamic games with discrete choices. This class of models has computationally-desirable properties and includes widely-used entry, exit, and investment games that underlie many applications in empirical industrial organization. First, we establish a linear representation of the value functions in the model which assists in both identification and estimation. Next, we examine conditions for nonparametric identification of the intensity matrix and the structural primitives of the model when only discrete-time data are available. Finally, we propose a two-step pseudo maximum likelihood estimator based on a nonparametric estimator for the intensity matrix in the first step and an inverse conditional choice probability mapping used for pseudo maximum likelihood estimation of the structural parameters in the second step.

Keywords: Microeconometrics, continuous time, Markov decision processes, dynamic discrete choice, dynamic games, matrix exponential, identification, pseudo maximum likelihood.

JEL Classification: C13, C35, C62, C73.

BibTeX Record:

  author       = {Jason R. Blevins},
  title        = {Identification and Estimation of Continuous Time
                  Dynamic Games},
  institution  = {Ohio State University},
  year         = 2014,
  type         = {Unpublished Manuscript}

Note: Many results from the current draft this paper have been incorporated into the following paper: Estimation of Dynamic Discrete Choice Models in Continuous Time with an Application to Retail Competition.