Scott Kelly University of Cambridge
With the growing prominence of climate change and the rising costs of domestic energy, the ability to predict domestic energy consumption and carbon emissions from the dwellings is growing in importance. Through the use of structural equation modelling (SEM) and maximum likelihood estimator (ML) the underlying structural relationships that cause energy consumption can be identified and quantified. Using the 1996 English House Condition Survey consisting of 2531 records it is found that there are a number of confounding variables that explain domestic energy consumption. The variables most shown to explain domestic energy consumption include household occupancy rates, household income, winter weekly heating
patterns, living room temperatures, dwelling floor area and dwelling energy efficiency (SAP). In this model the direct effects, indirect effects and total effects that each variable has on energy consumption can be calculated. More importantly it is shown that energy efficiency (SAP) has reciprocal causality on energy consumption. Logically, dwelling efficiency is shown to have a direct negative effect on dwelling energy consumption. However, the reverse is also true; homes that have a propensity to consume more energy are also more likely to have higher SAP ratings. Due to the non-recursive nature of SAP and
energy consumption a method for differentiating these two bidirectional effects on one another was thus required. Problems like this are commonly solved using path analysis and structural equation models. This paper therefore presents an application of this model to dwelling energy consumption.