TODAY’S STUDY: WHY SOLAR PRICES VARY
Photovoltaic System Price Quotes from Selected States, 2014-2015
Erika H. Myers and Vazken Kassakhian, December 2015 (Solar Electric Power Association)
The Solar Electric Power Association (SEPA) routinely speaks with utilities and developers to gather information on average national installed costs for photovoltaic (PV) solar. However, our information previously has not included state- or region-specific costs, which can be a valuable tool for utilities, consumers, regulators, and other stakeholders. Most publicly available reports are also limited to national prices. Free state-level data sources are available online, but the data points are self-reported and potentially unreliable.
This report is intended to provide a more granular level of information on solar pricing and some of the underlying market drivers. The report is based on nearly 11,000 aggregated residential, nonresidential (including commercial and industrial), and utility-scale solar prices from selected states obtained through a partnership with Mercatus and EnergySage. While recognizing the limitations of this relatively small sample of pricing information, we believe it provides a snapshot of the broader solar market.
Based on the provided data, average national residential prices ranged from $3 per watt to $4 per watt, while average national prices for nonresidential and utility-scale projects ranged from $2 per watt to $3 per watt between January 2014 and July 2015.
The dynamic nature of the market lends itself to variation due to factors such as state policies and incentives, supply and demand within local markets, competition, labor costs, permitting costs, electricity prices, and transportation. SEPA tested a range of hypotheses to tease out any correlations between residential, nonresidential, and utility-scale solar prices and labor costs, state incentives, elextricity prices, and solar market potential.
Despite our initial findings, information about potential drivers for state and regional pricing variability remains largely anecdotal. One potential source of variability is limited price transparency, in which consumers may not have a complete understanding of competitive pricing information prior to making purchasing decisions.
While the pricing data allows us to make some useful market observations, SEPA is seeking additional data to answer more questions about individual solar markets within each state. SEPA has issued a call for data and looks forward to expanding this pricing research in 2016…
The Solar Electric Power Association (SEPA) routinely speaks with utilities and developers to gather information on average national installed costs for photovoltaic (PV) solar, as illustrated in Table 1. However, our information previously has not included state- or region-specific costs, which can be a valuable tool for utilities, consumers, regulators, and other stakeholders. Most publicly available reports are also limited to national prices. Free state level data sources are available online, but the data points are self-reported and potentially unreliable.
To provide a more granular level of information on solar pricing, SEPA partnered with Mercatus and EnergySage to analyze nearly 11,000 aggregated residential, nonresidential (including commercial and industrial), and utilityscale solar prices from selected states. While recognizing the limitations of this relatively small sample of pricing information, we believe it provides a snapshot of the broader solar market.
Based on the provided data, average national residential prices ranged from $3 per watt to $4 per watt, while average national prices for nonresidential and utility-scale projects ranged from $2 per watt to $3 per watt between January 2014 and July 2015. However, as discussed in sections 3 and 4, variation of pricing data between regions and states is likely due to limited price transparency, in which consumers may not have a complete understanding of competitive pricing information prior to making purchasing decisions.
Additionally, the dynamic nature of the U.S. solar market lends itself to variation due to factors such as state policies and incentives, supply and demand within local markets, competition, labor costs, permitting costs, electricity prices, and transportation.
SEPA analyzed the pricing data to determine if any of these factors had a strong influence on price. As shown in Table 2, we identified (1) a moderate to strong correlation between average commercial and industrial electricity prices and average price-per-watt for nonresidential systems and (2) a strong correlation between total incentive amount and total system cost for residential systems. For additional information about how these factors were calculated, please see Appendix A.
EnergySage provided aggregated residential solar pricing information from quotes generated through its online solar PV marketplace.2 While other solar-related websites have exclusive partnerships with one or two installers, EnergySage partners with hundreds of prescreened solar installers across the country to provide a diversity of pricing information.
Summary of Residential Pricing Data
Residential pricing in states is driven by a combination of electricity prices, incentives, labor cost, permitting cost, competition, and supply and demand. Average prices ranged from $3 per watt to $4 per watt; however, there were some outliers within certain states, as shown in Figure 1.
For example, according to EnergySage, Florida’s inexpensive electricity and lack of solar incentives have generally kept prices quite low. Among the data analyzed from the state, the average price was $2.51 per watt.
Washington state, on the other hand, has a solar power performance payment incentive, which is significantly higher for projects using panels manufactured in the state, resulting in the highest average price of $4.43 per watt.
According to EnergySage, New York and Massachusetts both have vibrant solar markets with some limited price transparency among consumers and between vendors — factors that pushed the average price of solar to $4.03 per watt and $4.20 per watt respectively…
Summary of System Size, Gross Cost, Incentive Value, and Mounting Details
The average reported size for residential systems ranged from 5 kW to 10 kW, according to EnergySage. However, the data within each state contained a large standard deviation, reflecting the broad diversity of household types — from small bungalows to mansions to multifamily residential properties — using the EnergySage marketplace.
The average gross purchase costs, defined as the price of the entire installed systems prior to incentives, ranged between $21,225 and $38,058 with the median price ranging from $18,360 to $30,500.
The value of the incentives — defined as a total of federal, state, and local incentives and rebates — for a given quote were also diverse, ranging among states from an average of $7,275 to $19,674. The median incentive ranged from $6,288 to $18,485.
Among residential systems, the most common mounting method was a roof-penetrating system, representing over 97 percent of the quotes, with just under 3 percent from ballast mounted systems. Further, over 95 percent of the systems were roof-mounted with the remaining 5 percent of quotes for ground-mount systems. These mounting systems and locations are consistent with current installation trends in the residential sector to date.
Nonresidential And Utility-Scale Pricing
Mercatus provided aggregated state pricing information based on outputs from its Energy Investment Management (EIM) software platform. The 2015 Mercatus data was predominately within the ranges depicted in Table 1, albeit on the higher side, particularly compared to large utility-scale. Unfortunately, as the data was aggregated by state, it was not possible to break out nonresidential and utility-scale projects to compare to SEPA pricing information.
Consistent with information from other pricing publications, average costs, according to Mercatus, declined between 2014 and 2015, with lower average costs for ground-mount than rooftop systems. Mercatus data also included costs for total engineering, procurement, and construction (EPC), interconnection, and development as shown in Figure 7. The bulk of costs were associated with EPC, although Mercatus noted that some users did not differentiate interconnection and development costs within the software, which may have ultimately skewed the final results. As reported by users, although interconnection and development costs comprise a very small portion of the total project cost (between 1 percent and 6 percent), interconnection costs were higher on average for ground-mount systems than for rooftop systems. The difference here may be a function of the larger average size of the ground-mount systems, 6-8 MW, compared to 1-3 MW for rooftop.
Summary of Nonresidential and Utility-Scale Rooftop Solar
As shown in Figure 8, the average cost for a rooftop system ranged from $2 per watt to $3 per watt for most states in 2014 and 2015. With the exception of Arizona, Florida, Hawaii, and Puerto Rico, almost every state and territory saw a slight downward trend in average prices. Outliers, such as the high rooftop price in Minnesota, may have been skewed by the small sample size. Overall, the range of prices was also far narrower compared to the quoted residential prices — possibly the result of procurement sophistication and acumen of large commercial consumers, compared to the average residential consumer…
Summary of Nonresidential and Utility-Scale Ground-mount Solar
Ground-mount prices on average are slightly more competitive than nonresidential and utility-scale rooftop, a trend which follows average national trends from most price reporting entities. Rooftop solar projects typically carry a cost premium due to the higher soft costs associated with safety protocols, logistics and material handling (e.g., cranes), additional engineering design costs, permitting, building codes, interconnection arrangements, and other factors. Further, the size of installations is typically limited for rooftop projects, which may reduce any economies of scale.
Similar to nonresidential and utility-scale rooftop systems, pricing for ground-mount systems shows less variability among states and regions, with the exception of Hawaii (see Figure 11). Nearly all of the prices were within the $2 per watt to $3 per watt range, with average prices trending slightly higher in the New England and Mid-Atlantic regions in 2014. However, average prices declined between 2014 and 2015, with the exception of Arizona, Georgia, New Jersey, and Puerto Rico, where prices increased.
SEPA tested a variety of hypotheses to tease out any correlations between residential, nonresidential, and utility-scale solar prices and labor costs, state incentives, average electricity prices, and solar market potential.
First, SEPA compared average state prices with a proprietary tool for ranking state-level residential solar markets, including composite factors of solar capacity, average electricity prices, incentives and policies.3 This analysis found little, to no, correlation between average prices and solar market potential.
Second, SEPA compared average state prices to labor costs, based on figures for solar installer mean hourly wages developed by the U.S. Bureau of Labor and Statistics.4 At best, we found a weak correlation between labor costs and high average residential, nonresidential, and utility-scale solar costs, but not sufficient to make any strong conclusions.
Third, SEPA compared states with richer incentives. Just based on incentive level, we could find no correlation to price-per-watt. However, we were able to determine a stronger correlation between higher total gross cost (i.e., total price of the system absent incentives) and higher total incentives (see Appendix A).
This trend would suggest that consumers are strongly motivated by incentives, and while that might not translate to lower per-watt pricing, it leads to a higher likelihood of making a larger PV investment in those states with rich incentives. Local permitting costs and regulatory barriers also differ across cities, states, and regions, resulting in increased price dispersion.
Finally, SEPA compared the average price to average electricity prices posted by the U.S. Energy Information Administration5 for both residential and commercial and industrial customers. After removing outliers (Florida and Washington), the regression plots in Appendix A reflect a weak positive relationship of average price and residential electricity prices. However, when comparing average price for nonresidential and utilityscale projects to average commercial and industrial electricity prices, we saw a moderate to strong positive correlation, for rooftop and ground-mount respectively.6 This trend may be linked to a few possible hypotheses, but SEPA believes that it is related to the sophistication of the commercial procurement process. Residential buyers often do not receive the same quote volume and have limited comparative levels of financial acumen.
Information about potential drivers for state and regional pricing variability is largely anecdotal. In states with fewer solar incentives, we see lower average prices. In areas with higher average prices, it could be the result of higher retail rates, supply and demand imbalances, and limited price transparency.
For example, when comparing EnergySage data to publicly available state data in California, Connecticut, Massachusetts, and New York, analysts found that EnergySage prices were approximately 20 percent lower than state averages.7 This wide gap is possibly due to comparison shopping online, which increases price transparency.
While the pricing data provided through EnergySage and Mercatus allow us to make some useful market observations, SEPA is seeking additional data that will allow us to answer more questions about individual solar markets within each state. For example:
• Are we seeing a selectivity bias based on the types of customers using the EnergySage website (e.g., those more comfortable with technology)?
• In states without incentives, are the only installed solar systems those with shorter payback periods?
• Do incentives offer the ability for wider deployment of projects that otherwise would not get built?
• Do other factors make projects cheaper, such as age of housing stock or common roof type?
• Are prices a function of market competitiveness within a certain state or region? SEPA looks forward to expanding this pricing research to investigate additional data trends in the future and potentially answer these, and other questions.
We have issued a call for data as outlined in Appendix B. If you are interested in participating in our 2016 pricing report, please contact us.