Solar prices are stabilizing at new lower levels, but causation remains elusive

The factors that drive price differences remain largely a mystery, but a new study provides some novel insights

A new industry report shows solar photovoltaic prices are moving toward a standard range nationally, and also reveals some valuable details about the persisting mystery of state and regional solar price variations. 

The purpose of the study — “Photovoltaic System Price Quotes from Selected States, 2014-2015” — is to reveal where there could be price flexibility, explained co-author Erika Myers, Sr. Research Manager with the Solar Electric Power Association (SEPA).

It asks where the price differentials are in residential, commercial-industrial rooftop, and ground-mounted utility-scale solar and what causes those differences, she added.

Utilities will want this information because they know residential and business customers increasingly want solar, Mysers said, and they are working toward becoming sources of solar information on behalf of those customers.

Utilities are increasingly offering services that provide access to solar price information as part of an effort to be a solar consultant to their customers, she pointed out. Georgia Power, California’s Sacramento Municipal Utility District, Avista in the Pacific Northwest, and Duke Energy are among those that tout solar information websites.

EnergySage, which provided quote data from its online solar marketplace for the SEPA assessment, is about to announce a similar deal with “a well-known investor-owned utility,” according to CEO Vikkram Aggarwal. "They are launching a new solar program and they will direct customers to the EnergySage platform through their marketing materials.”

EnergySage is also in talks with ten or more utilities, he added. “They know their customers are thinking about solar but they are not yet equipped to provide solar. They want to maintain the customer relationship and provide quality information.”

Correlations and lack of correlations

The EnergySage data, along with quotes on commercial-industrial and ground-mounted utility-scale installations from energy investment consultant Mercatus, were examined for correllations with a number of factors, according to SEPA Reseach Analyst and paper co-author Vazken Kassakhian.

The factors included labor cost, state incentives, average electricity prices, and an independently determined multifactorial variable SEPA calls the "solar market potential," he explained.

“We wanted to go down the list of metrics and test the things that would intuitively seem to contribute to price,” Kassakhian said. “We are a little less than excited to report we didn’t find any clear factor that, on its own, explained why prices vary from state to state.”

Some factors had some impact on price, but there was no clear correlation with any particular factor. For solar market potential, which combines capacity factor, installed capacity, electricity prices, incentives, and policies into a single rating, the “correlation was all over the map,” he said.

The impact of local and state incentives seemed “muted” and didn’t seem to move prices either way, the researchers found. That could be because the 30% federal investment tax credit (ITC) is a much more significant incentive.

There was a strong correlation in "incentive amount" with "total gross cost" even though the amount of the incentive did not correlate with the end price of a PV system.

“It is possible when the incentives are bigger, consumers are inclined to build bigger systems,” Kassakhian said.

The researchers found only found a weak correlation to labor costs, Myers said. “But it is not the only thing causing the price differential and there are other things happening that mute that correlation.”



SEPA’s findings were largely corroborated in "Tracking the Sun VIII," the annual survey of U.S. photovoltaic solar prices from the Lawrence Berkeley National Laboratory (LBNL). LBNL also incorporated EnergySage data.

“Our data reinforces the notion of tremendous variability in PV system pricing that is a reflection of many different underlying differences between systems, installers, geographies, data reporting conventions, and many other factors,” LBNL Research Scientist Galen Barbose told Utility Dive.

The variability is not to be underestimated and is important to understand, SEPA VP Bob Gibson said of LBNL’s findings. California’s dominance of solar is shrinking as solar-supportive policies and incentives in a wider range of states help overcome solar resource limitations.

To study variability more closely, LBNL focused on Arizona, California, Massachusetts, and New Jersey, four of the biggest markets. It found huge installer level pricing variations in each state, Barbose said. But that variation had little to do with how active the installer was or whether the installer was doing third party-owned or customer-owned solar, he explained.

There is no escaping the variability, but the explanation for it is not clear, Barbose said. “The inability to find a relationship is not necessarily proof the relationship doesn’t exist. It may be washed out by other factors.”

One other version of variability was that LBNL’s 2014 national median installed prices were higher than prices found in the SEPA data and in other studies. LBNL’s findings came out at a $4.30/watt cost for residential systems, $3.90/watt for non-residential systems smaller than 500 kW, and $2.80/watt for non-residential systems bigger than 500 kW.

Other nationally recognized price benchmarks surveyed by LBNL came out at $2.80 to $4.50/watt for residential systems and $1.70 to $4.10/watt for non-residential systems, ranges similar to SEPA's findings.


Price visibility

The SEPA research turned up one consistent factor in price variations.

There is “a laundry list” of things that could be impacting PV prices but “the lack of price visibility seems to be involved in all the state-to-state residential system price variability,” Myers said. In the larger system size market segments, there is more visibility and less price variation, she explained.

“There is not a lot of pricing visibility available to residential consumers,” she said. Online marketplaces often offer comparisons but they are typically limited to “a small subset of local contractors who know each other’s prices.”

But the EnergySage information used by SEPA is of a different sort, according to the report.

"While other solar-related websites have exclusive partnerships with one or two installers," it reads, "EnergySage partners with hundreds of prescreened solar installers across the country to provide a diversity of pricing information."

But beside the EnergySage data, there is even less visibility into other states' or other regions' prices.

Larger industrial customers and utilities have procurement experts and guidelines, competitive bidding processes, and a network of solar industry contacts to provide quotes, Myers said. A residential consumer’s access to price comparisons is much more limited.

Correlations between electricity prices and PV prices are also likely a product of the visibility factor, the SEPA researchers believe.

The electricity price moves only weakly with residential PV prices but more, though only moderately, with C&I PV prices. The electricity price, however, correlates strongly with utility-scale PV pricing.

“In places where there is a procurement expert, there might be more sophistication in understanding market prices,” Kassakhian said. “We don’t have a metric of visibility but we surmise that is what is going on.”

EnergySage's Aggarwal hesitantly agreed. “Based on the data we have, we know the reasons for pricing behaviors is a something of mystery, but we do not object to SEPA’s analysis and conclusions.”


The narrowing gaps

The SEPA study also turned up one overarching data-driven conclusion.

“Prices do cluster in the $3/watt to $4/watt range in residential and the $2/watt to $3/watt range in the larger market segments,” Kassakhian said.

Florida’s very low $2.51/watt price and Washington state’s very high $4.43/watt price are outliers, he added. “This research adds to others' findings, providing confidence that prices are starting to cluster in these affordable ranges.”

This SEPA conclusion was corroborated in LBNL’s research. “Variability has narrowed over time,” Barbose said. 

Since about 2010, the data shows a slow but steady narrowing in dispersion of prices, he observed. “One interpretation is that markets are maturing.”

Customers are becoming more informed and installers are becoming more competitive, Barbose added.  “These are things we would expect to see in a maturing market. But the data is not granular and specific enough to make that interpretation with certainty.”

Aggarwal agreed. “Right now the price gap is still substantial, about 20% to 25%, but when we started keeping track it was almost 50%, so it is narrowing. As transparency becomes the norm, as it is in products like travel and cars, the gap will narrow and the price will revert to the mean.”

A call for data

Appendix B of the SEPA paper provides the “laundry list” of possible price drivers that the researchers did not have adequate data on to fully assess, Kassakhian said.

"We seek partners willing to share cost component information for one or all of these individual costs,” Appendix B declares. It then lists:

  1. Hardware Costs, including modules, inverters, racking and other balance of system and installation equipment
  2. Soft costs, including customer acquisition, permitting and commissioning expenses, installation labor, financing, and supply chain expenses
  3. Other costs, including profit and other overhead costs, sales and other taxes, and siting expenses

Feb. 11 SEPA webinar will advance discussion of those topics and address some key takeaway questions from the report, including:

  • Average state/regional utility-scale PV pricing from 2014-2015
  • Regional/state impacts of pricing factors including: labor, incentives, policies, and rates
  • Multi-sector PV market trends
  • New applications of technology and data analytics for compliance and program design

“This work has in many ways raised more questions than answers and we hope the call for data in Appendix B will bring new information to help answer some of the questions,” Myers said.

Correction: An earlier version of this article implied that the EnergySage online marketplace was limited to a small subset of local installers, when it in fact partners with hundreds of installers across the country to provide fuller information on pricing. The original version also misstated the focus points of the Feb. 11 SEPA webinar.

Top image credit: GraphicStock