Working Paper: CEPR ID: DP9026
Authors: Mathias Reynaert; Frank Verboven
Abstract: We shed new light on the performance of Berry, Levinsohn and Pakes' (1995) GMM estimator of the aggregate random coefficient logit model. Based on an extensive Monte Carlo study, we show that the use of Chamberlain's (1987) optimal instruments overcomes most of the problems that have recently been documented with standard, non-optimal instruments. Optimal instruments reduce small sample bias, but prove even more powerful in increasing the estimator's efficiency and stability. Other recent methodological advances (MPEC, polynomial-based integration of the market shares) greatly improve computational speed, but they are only successful in terms of bias and efficiency when combined with optimal instruments.
Keywords: Optimal instruments; Random coefficients demand model
JEL Codes: C36; L00
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Optimal instruments (C36) | Performance of GMM estimator (C51) |
| Chamberlain's optimal instruments (L64) | Small sample bias (C83) |
| Optimal instruments (C36) | Bias and efficiency of GMM estimator (C51) |
| Optimal instruments (C36) | Variance of the random coefficient (C29) |
| Optimal instruments (C36) | GMM objective function flatness (C51) |
| Weak instruments (C26) | Performance of GMM estimator (C51) |