Specification issues and confidence intervals in unilateral price effects analysis

This paper contributes to the economics and econometrics literature on unilateral effects analysis. It introduces the Rotterdam demand system; considers the effect and importance of demand restrictions for minimizing the mean-squared error of price simulations; demonstrates that approximate price changes are often misleading indicators of exact price changes; and uses bootstrap techniques to determine confidence intervals and standard errors for simulated price changes, allowing determination not only of their economic significance but also of their statistical significance. (C) 2002 Elsevier Science B.V. All rights reserved.