Forecasting electricity demand in developing countries: A study of household income and appliance ownership

Michael A. McNeil and Virginie E. Letschert, Lawrence Berkeley National Laboratory

Keywords

appliance ownership, developing countries, residential, refrigerator, saturation

Abstract

Increase in the rate of appliance ownership with rising incomes is a major driver of the growth of residential electricity consumption in developing countries. Macroeconomic parameters such as GDP growth, while useful in forecasting residential electricity use, may fail to take into account the determinants of appliance use at the household level, such as total market penetration of primary appliances among certain classes, subsequent ordering and climate dependency of secondary products, the dynamics of economic growth and income distribution, urbanization and rural electrification. A detailed, multi-country investigation of the relationship between household income and appliance ownership is therefore valuable for predicting future residential energy consumption, as well as evaluating strategies for mitigation of its environmental impacts.

This paper describes an analysis of recent survey data from six countries (Brazil, Mexico, Nicaragua, Panama, Peru and South Africa) yielding data on income and appliance ownership for a large sample of households. We use an econometric parameterisation to model the relationship between income and ownership for each country, and for each appliance for which data are available. This model accounts for economic variability in households in different countries due to variation in development of local markets, and makes a determination of the order of successive appliance purchases and corresponding income levels. In addition, the paper presents the results of using non-economic parameters (such as climate, degree of urbanization and electrification rate) as a means of explaining variability between countries. Finally, we evaluate the use of the household-level model along with macroeconomic and other parameters in order to predict national ownership rates.