Application of an econometric multiple discrete continuous fusion approach to link residential sector energy demand and travel infrastructure and usage

Abstract

The considerable body of earlier research on household level residential sector energy demand does not consider the impact of household residents’ travel infrastructure on energy consumption patterns. The absence of travel infrastructure elements in energy demand models can be attributed to the lack of data providing this information in energy surveys. In this study, a novel econometric fusion approach is utilized to combine the traditional energy dataset −Residential Energy Consumption Survey (RECS) data – with a transportation survey data − National Household Travel Survey (NHTS) data. The probabilistic fusion approach is employed to study energy consumption by end use type with a Multiple Discrete Continuous Extreme Value model framework. The framework will quantify the impact of travel infrastructure and usage related attributes on household end-use energy demand and remedy the over-estimation of the impact of socioeconomic attributes. The model results reveal the impact of several household socioeconomic attributes (i.e., household size, location and income) and travel infrastructure and usage related attributes (i.e. number of vehicles of different fuel and body types, household annual mileage and frequency of long-distance trips) on household end-use energy demand. The model estimation results are augmented with an elasticity analysis and policy analysis to highlight the implementation of the proposed framework. The elasticity results reveal that ignoring the influence of travel infrastructure and usage variables can contribute to errors for elasticity values for other independent variables such as household size (up to 1800%), number of adults (up to 50%) and residential location (up to 15%).