The Representative Consumer Approximation Bias in Discrete Choice Welfare Analysis

Abstract

Discrete choice welfare analysis is essential in non-market valuation to accompany studies using choice experiments and recreation choice models. McFadden (in: Melvin, Moore, Riezman (eds) Trade, theory and econometrics: essays in honor of John S. Chipman, Routledge, London, 1999) shows that discrete choice welfare measures derived by the common representative consumer approach can be biased under nonlinear income effects, and the percentage bias increases monotonically with the size of quality improvement. We present contrary results that percentage measurement errors in such welfare measures can go either direction and provide conditions under which the direction of bias can be identified. We also show that price and non-price variables play different roles in discrete choice welfare analysis.