Econometric analyses of household fuel demands

This study develops econometric models of residential demands for electricity, natural gas, and petroleum products. Fuel demands per household are estimated as functions of fuel prices, per capita income, heating degree days, and mean July temperature. Both cross-sectional and dynamic models are developed using a large data base containing observations for each state and year from 1951 through 1974. Long-run own-price elasticities for all three fuels are greater than unity with natural gas showing the greatest sensitivity to own-price changes. Cross-price elasticities are all less than unity except for the elasticity of demand for oil with respect to the price of gas (which is even larger than the own-price elasticity of demand for oil). The cross-sectional models show considerable stability with respect to own-price elasticities but much instability with respect to the cross-price and income elasticities. Agreement between the cross-section and dynamic models is good for the own-price elasticities but inconclusive for the other coefficients.