Energy-Saving Housing Improvements in Canada (1979–82): A Nested Logit Analysis

Abstract

The determinants of energy-saving housing improvements are analysed in the Canadian context. The data used represent the time in which the Canadian Home Insulation Program (CHIP) and the Canadian Oil Substitution Program (COSP) were operating at their peak. A nested multinomial logit model is estimated, which predicts the likelihood that individual households will carry out different retrofitting measures. In the final model, socioeconomic variables have an impact that is less than anticipated, an indication that the government programs succeeded in reducing socioeconomic barriers to retrofitting. Dwelling characteristics and fuel-type variables are of more importance and provincial dummy variables are often significant. It is concluded that the likelihood of a dwelling retrofit had less to do with the simple socioeconomic surrogates of the occupants than with the nature and location of the dwelling itself. The data reveal further that CHIP and COSP were less effective since they operated separately, and that the hypothesized nested decision structure is an appropriate one.