Millions of households have adopted rooftop solar to reduce electricity bills and participate in the clean energy transition. Yet rooftop solar remains out of reach for many households, especially multifamily building occupants, renters, and low- and moderate-income (LMI) households. Inequitable solar access reduces the potential contribution of rooftop solar to grid decarbonization and has generated opposition to rooftop solar policies. Stakeholders and policymakers are exploring ways to expand solar access in communities underserved by rooftop solar.

Community solar—where multiple customers buy solar output from a single shared system—is often proposed as a measure to expand solar access. Unlike rooftop solar, community solar poses no specific barriers to adoption for multifamily building occupants or renters. Community solar also generally entails lower costs, reducing barriers for LMI households. Community solar has become particularly common in the United States, where around 5 gigawatts of community solar capacity had been installed by the end of 2022, compared to around 29 gigawatts of residential rooftop solar (Davis et al. 2023).

In our paper published in Nature Energy (10.1038/s41560-024-01546-2), we empirically analyze whether community solar has, in fact, expanded solar access. We were uniquely positioned to implement that analysis due to our group’s years-long, ongoing efforts to compile household-level data on rooftop and community solar adopters, as well as some personal experience with rooftop and community solar adoption (Figure 1). Our study uses household-level rooftop solar data compiled as part of the Lawrence Berkeley National Laboratory’s Tracking the Sun project, an ongoing effort led by two of the authors that annually compiles data on around 80% of the U.S. rooftop solar market (Barbose et al. 2023). We combine the rooftop solar data with community solar data compiled under the National Renewable Energy Laboratory’s Sharing the Sun project (Xu et al. 2023), another ongoing effort led by the other two authors. Our study is the first, to our knowledge, to combine household-level rooftop and community solar data in this way.

We augmented both data sets with household-level demographic data for housing tenure (e.g., rent or own) and housing type (e.g., single or multifamily housing), with modeled household-level income estimates, and with an algorithmically generated prediction for household race/ethnicity. We then winnowed the data down to a study sample based on 11 states with at least 100 records of both adopter types with coverage of the key demographic variables.

Through comparative statistical analysis, we find that community solar adopters are significantly different from rooftop solar adopters (Figure 2). Community solar adopters were about 6.1 times more likely to live in multifamily buildings than rooftop solar adopters in our sample, 4.4 times more likely to rent, and earned about 23% less annual income. These results suggest that community solar expands solar access under the reasonable assumption that some—and likely many—community solar adopters could not or would not have adopted rooftop solar. The results were more equivocal in terms of race/ethnicity, for reasons we can only speculate on in our study.

Our next task was to determine what explains demographic differences between community and rooftop solar adopters. Those differences could be partly explained by how the community solar business model reduces barriers to adoption for multifamily building occupants, renters, and LMI households. Demographic differences may also be driven by differences in the rooftop and community solar policy landscapes. Many states have implemented policies to enable LMI community solar adoption, including all 11 states in our study subsample. LMI community solar policies are, arguably, more generous than policies to promote LMI rooftop solar adoption. For instance, many states require that minimum shares of community solar be reserved for LMI customers, while no such policy exists for rooftop solar.

We find evidence that business models and policy differences each partly explain the observed demographic differences. We demonstrate this by implementing analyses that isolate adopters that participated in LMI programs. Based on a subsample of three states, we estimate that policy explains around 67% of income differences between community and rooftop solar adopters, 43% of the differences in housing tenure, and 23% of the differences in housing type.

Overall, our results suggest that community solar has expanded solar access relative to rooftop solar. However, as illustrated in Figure 2, the demographic data suggest that communities underserved by rooftop solar are likewise underserved by community solar. In most states in the subsample, community solar adopters are less likely than the general population to rent, less likely to live in multifamily housing, and earn higher annual incomes. Further, the results for race/ethnicity comparisons generally suggest that community solar is not effectively reaching households that identify as racial minorities. More could be and is being done. U.S. states are exploring and implementing increasingly ambitious LMI community solar policies that could further enhance solar access in the future. We encourage future research to explore how effectively these policies augment the access benefits of community solar.

Banner photo credit: Dennis Schroeder, NREL