Selection Bias and the Demand for Electricity

Abstract

This article looks first at the issue of policy-motivated electricity rate designs with special attention to the situation of master-metered households. It then discusses the implicit bias (both substantive and statistical) in demand functions and other regression-related analyses when these consumers in master-metered households are excluded from such research. Using econometric techniques to adjust for selection bias, it estimates and compares two demand functions, one of which corrects for selection bias. While most effects are comparable, some important differences do result.