Efficient Estimations of Random Coefficients Model Based on Survey Data
M. Ishaq Bhatti, Bhatti, Muhammad Ishaq · AgEcon Search (University of Minnesota, USA) · 1992
Much applied work in econometrics and other social sciences involves regression analysis using survey data. Recently, King and Evans (1986) and Bhatti (1991) considered an equicorrelated error term model with fixed coefficients and proposed some diagnostic tests for detecting the equicorrelated parameter (p). This paper considers an efficient estimation procedure for the linear regression model with stochastic coefficients and equicorrelated disturbances across blocks. An algorithm for the maximum likelihood estimator is presented, after King (1986), for an error components model. Its application to model based on Bangladesh data is discussed.