From Executive Summary:
Although photovoltaic (PV) penetration in the United States is increasing rapidly, properly valuing homes with PV systems remains a barrier to PV deployment. Previous studies show that PV homes command sales price premiums. Still, some appraisers and other home valuers assign no value to a home’s PV system, and those who do often cannot find comparable home sales to help determine the PV premium. This has spurred the development of alternative methods of valuing PV homes, including the use of an income approach (based on the present value of PV energy produced over its useful lifetime) and the replacement cost approach (based on the present installed cost equivalent of the PV system). However, those approaches have just begun to have been validated against actual market premiums. Moreover, the drivers underlying PV home premiums are not well understood, which may deter some appraisers from assigning value to PV systems.
This study, which builds on a previous study conducted by the same authors (Hoen et al., 2011; 2013), helps fill both of those gaps by: 1) using regression analysis to examine actual PV home sales price premiums from a large dataset of California PV homes; 2) exploring the sensitivities of those estimated premiums to the size and age of the installed PV system at the time of home sale, and 3) comparing the actual premiums to predictions made with the income and cost approaches.
Our analysis offers clear support that a premium exists in the marketplace; thus, PV systems have value, and their contribution to home values must be assessed. We find that premiums in California are strongly correlated with PV system size and weakly correlated with PV system age, in other words larger systems garner larger premiums and older systems garner smaller premiums. We estimate that each 1-kW increase in size equates to a $5,911 higher Premium (p-value 0.000) and each year systems age equates to a $2,411 lower premium (p-value 0.087).
Additionally, the actual California premiums appear to erode over time (estimated to be approximately 9% per year), more quickly than either the income (approximately 0.5% per year) or cost approaches (5% per year) predict and thus the premiums for homes with older systems (e.g., between 6 and 10 years old) appear to be substantially smaller than predicted.
Further, premiums appear to be substantially larger than predicted using the income (42% of premiums when the average income estimate is used, p-value 0.000) and cost approaches (65% of premiums, p-value 0.000). There are a number of plausible explanations for this disparity including: premiums might be larger because buyers were willing to pay more for the PV system owing to its green cachet; there could be transaction costs that are avoided by purchasing a home with a PV system already installed that are not incorporated in the cost estimates; the average utility-specific California residential electricity retail rates, which are used for the income estimates, might be lower than they should be in CA where steeply tiered rates are commonplace; and, the market-based Premium estimates could contain effects from omitted variables and therefore potentially overestimate the actual premiums.
We conclude by proposing future research ideas to further improve understanding of the impact of PV systems on home values and therefore related barriers to deployment.