Who is participating in residential energy efficiency programs?

December 1, 2021

new Berkeley Lab report describes how 11 demographic and household characteristics including income, race and ethnicity, and education affect participation in residential utility customer-funded energy efficiency programs. The report compiles previous work on this topic and adds new primary analysis of four datasets with different levels of detail from the Residential Energy Consumption Survey (RECS), two New England states, and a Midwestern state. Our analysis includes single-variable statistical models, which describe the relationship between each factor and program participation, and multivariable models that disentangle the effects of individual factors from other factors they are correlated with (e.g., income and education). Parsing these factors suggests specific opportunities for programmatic intervention.

Numbers in the “Literature” rows indicate the count of studies that found a particular result.
* Racial and ethnic groups were not compared individually to the share of non-Latino White householders because of sample size. The share of non-Latino White heads of household in the zip code was positively correlated with the market-rate and eligible income-qualified participation rates but negatively correlated with the overall income-qualified participation rate.

Figure 1. Simplified summary of selected results

Key findings from the study include:

Participation rates were strongly associated with educational achievement and building type. The clearest associations with energy efficiency program participation were with education and building type. Households with more educated heads of household and households in single-family homes were more likely to participate in nearly all cases studied, and these relationships remained strong in multivariable analyses. The single-family results are in part structural – many programs are only available to single-family households – though our results suggest that these structural factors should be examined. The clear and consistent impact of education on participation across program types suggests that program administrators may wish to explore strategies to better engage households and locations with lower educational attainment.

Race and income had varying associations with participation rates. Both in our analysis and in the existing literature, relationships between program participation and particular racial and ethnic groups or income levels depend on the individual program. Still, patterns emerged that suggest inequities regarding these factors that program administrators may wish to address. In single-variable models, household participation in market-rate programs was positively correlated with income and negatively correlated with Black heads of household. These associations were not always significant in multivariable models and varied for other racial and ethnic groups.

Results depended on the participation rate studied. One of our datasets allows us to compare two different participation rates for the same income-qualified program – the overall participation rate (the share of total households who participated) and the eligible participation rate (the share of eligible households who participated). As Figure 2 shows, the results of our analysis differ between the two metrics. For example, higher income areas had a lower overall participation rate but a higher eligible participation rate for the income-qualified program – indicating that within the eligible low-income population, households in higher-income areas participated more. We observe a similar reversal of this relationship for education and some race/ethnicity variables.

The report authors will host a webinar to present key findings on December 9 at 10 am PT / 1 pm ET. Register for the webinar here: https://lbnl.zoom.us/webinar/register/WN_4q_7OnUOTBq0QHaytwWc5A

We appreciate funding support from the U.S. Department of Energy Strategic Analysis Office.



Signup for our Mailing List