Estimating Almost-ideal Demand Systems with Endogenous Regressors

Abstract

In this article, we present the new aidsills command for estimating almost-ideal demand systems and their quadratic extensions. In contrast with Poi's (2012, Stata Journal 12: 433–446) quaids command, which is based on the nonlinear nlsur command, aidsills uses the computationally attractive iterated linear least-squares estimator developed by Blundell and Robin (1999, Journal of Applied Econometrics 14: 209–232). The new command further allows one to account for endogenous prices and total expenditure by using instrumental-variable techniques. Elasticities and their standard errors can be obtained using the aidsills elas postestimation command.