CMStatistics 2016: Start Registration
View Submission - CMStatistics
B1598
Title: Estimation of random coefficient models with fixed effects Authors:  Thi Thu Hien Pham - University of Leuven (Belgium) [presenting]
Abstract: The aggregate random coefficient model previously proposed is widely used for demand modeling where only aggregate market data is available. The model incorporates random coefficients to explain the heterogeneity of consumers' tastes for differentiated product characteristics and price. However, product characteristics which are invariant across markets lead to unobserved heterogeneity. We extend the model to include fixed effects. To solve the extended model, we make use of the limited information maximum likelihood method. Monte Carlo simulations show that the employed method produces consistent estimates of the random coefficients in the presence of fixed effects. The asymptotic properties of the limited information maximum likelihood estimator are also discussed.