Estimation of a Nested Logit Model for Appliance Holdings

Abstract

This paper estimates a discrete choice model for room air conditioning, central air conditioning, space heating, and water heating, using data from two recent surveys of energy consumption by households—the 1978 National Interim Energy Consumption Survey (NIECS) and the 1980 Pacific Northwest Energy Survey (PNW). Estimation for these two data sets proceeds in parallel so that results based on the national level survey may be compared with those derived from Pacific Northwest regional data. We are thus able to address the important issue of model transferability. The estimated structure involves a ten alternative logit model of space heat/air-conditioning system choice. We first match a time path of operating costs to each household using historical state level energy prices and then analyze the role of price expectation formation in the choice of heating and cooling equipment for single family owner occupied dwellings. We compare a basic static expectation model with two alternative models: perfect foresight over a limited planning horizon and adaptive expectation formation. Finally we consider alternative conservation policies and alternative scenarios for the prices of electricity, natural gas, and fuel oil in order to predict the path of durable saturations from present to the year 2000.

Files

Additional details