Developing a Pre-Retrofit Energy Consumption Metric to Model Post-Retrofit Energy Savings: Phase One of a Three-Phase Research Initiative

This paper details the process and results from the first step of a three-step research process. This first step looks to identify the most predictive pre-retrofit metric of energy consumption to utilize in a model to predict the energy savings post retrofit. The ultimate goal of this research is to predict candidacy for retrofit using only a combination of demographic and home-characteristics data that is available for the entirety of the U.S. residential housing stock. This is important, as utility data is almost always protected for privacy and thus unavailable to assist in targeting where energy efficiency retrofits will be successful. It is found that the best metric is the simplest, total energy consumption divided by total floor area. In addition to evaluating which pre-use metric is most indicative of post retrofit savings, the paper evaluates the endogenous component of pre-use to post use and a potential method to alleviate this endogeneity. The research finds that by removing the year that is used to calculate the savings as the baseline pre-use year removes a portion of the endogeneity. It is also found that one year before the savings base year is the best year to utilize as the base.