In recent years the U.S. solar market has gone from a fledgling industry to a booming one. The Energy Information Administration expects solar to rank behind only natural gas in terms of new plant capacity added through 2050. Naturally, investors are flocking to the space.
But getting involved with an emerging technical asset class has its risks, particularly as the growing economic competitiveness of solar has pushed developers to get even more aggressive on pricing assumptions.
A new report, with contributions from firms including DNV GL, kWh Analytics, SunPower and PV Labs, aims to shed light on the financial uncertainties associated with solar production. Taken together, their data-driven conclusions show solar investments aren’t infallible.
Instead, analysts say potential investors should be ready to question the assumptions solar developers make about project dynamics like expected production and residual value when financing and building a project.
“Not every project necessarily falls within these accepted assumptions that the solar industry tends to follow when putting together financial models,” said Michelle Davis, a senior solar analyst at Wood Mackenzie Power & Renewables, which also contributed to the report.
“[It’s] a good lesson for folks to know as they’re looking to invest in the space. Not that the space is unreliable…it’s just not risk-free.”
The partners say they’ll publish the collaboration annually, in efforts to help the investment community ask the right questions of developers when they’re signing on to finance projects.
“The idea here was to bridge the gap between these two communities,” said Richard Matsui, CEO and co-founder of kWh Analytics, “to create a guide for investors who are interested in recognizing the importance of having data-based insights to inform the way they invest.”
The “wall of money”
Having a steady flow of cash ready to back solar has been a boon to the industry, but now the supply of projects is strained. As we've reported in the past, 2018 saw a general pullback in the supply of U.S. solar projects because of uncertainty tied to tax changes and tariffs.
With a "wall of money" chasing fewer projects, financiers are now open to lower-than-average returns in order to beat the competition.
Meanwhile, prices for power-purchase agreements continue reaching new lows. Davis said developers are, at times, manipulating assumptions to make projects pencil out economically at those prices.
To get the same returns with lower-priced PPAs, developers have begun to rely on exceedingly optimistic assumptions for post-PPA residual value, determined by unknown merchant power prices decades in the future.
And fudging the numbers on a system’s output puts projects in further danger of not delivering on those expected returns. Value assumptions, when combined with potential underperformance, create even more uncertainty on revenue.
“The focus has been a lot on reducing price — lower, lower, lower,” said Davis. “There needs to be a focus on price, but also on value and the production of the system.”
“A P99 isn’t actually a P99”
Data compiled by contributors showed that developer assumptions about solar performance are often overly optimistic. And lower performance means lower returns.
The value of a solar asset is primarily determined by modeled energy yield and the likelihood of achieving that yield, according to independent engineering and quality assurance firm DNV GL.
In an assessment of 39 projects totaling 1.2 gigawatts in capacity, DNV GL found that energy models overshot actual performance by a median of 3 percent. That means the main tool used to assess solar value and returns is regularly incorrect.
“Investors are already assuming that they’re just going to break even or even lose some money during the PPA [term]. Even after all these assumptions, they’re assuming that they’ll make 8 percent returns at the end of this process,” said Matsui at kWh Analytics. “But hold on — what if the plant actually, on average, underperforms by 3 percent?”
In its section of the report, kWh Analytics analyzed over 200,000 operating solar projects and found that the P99 estimate — which calculates a 1-in-100 worst case irradiance scenario — actually occurs 6 percent of the time.
Source: kWh Analytics
“In other words, a P99 isn’t actually a P99,” said Matsui. “We were able to show that this worst-case scenario is happening way more frequently.”
That means a seven-year solar loan would have a 37 percent chance of underperforming, which adds significant risk to a project defaulting on agreements.
“Banks are always worried about the downside scenario; that’s just in their DNA,” said Matsui. “When data is showing that the downside scenario is six times more likely to happen than what they currently believe. That means the banks need to take that step back and say, ‘Hey, do I really understand what risk I’m taking here on the solar production side?’”
Shrinking maintenance budgets
In addition to development assumptions, risks in managing projects can also endanger financial returns.
SunPower found that work left over by the engineering, procurement and construction team for the operations and maintenance (O&M) crew to handle reduced performance during the first year of a sample utility-scale project by 1.2 percent.
“In most utility-scale plants, most of the rework can be accomplished at night to minimize the impact to energy production,” SunPower reported in its section of the report. “However, this pulls the technicians away from being available during daylight production hours to respond to equipment outages, resulting in lost energy revenue and decreased performance metrics.”
Creating further strain: O&M budgets are shrinking while labor costs are rising. That means that projects are getting less maintenance attention overall.
Wood Mackenzie Power & Renewables notes that O&M pricing has dropped nearly 60 percent in the last several years.
“Vendors report that they have been driven toward rock-bottom prices, making it difficult to provide necessary services while remaining profitable,” wrote Wood Mackenzie in its portion of the report.
Potential downtime brought on by inadequate servicing is likely to drag down performance further.
“No one in the industry has a bird's-eye view to understand how much damage is being caused,” said Matsui.
Taken together, the aggressive assumptions and belt-tightening on portions of the value chain like O&M can depress project returns. Those factors suggest that solar projects shouldn’t only be assessed on their value when functioning flawlessly.
“If a PPA is incredibly low-priced, then there might be some risks that are not necessarily reflected in that price,” said Davis. “That’s what investors should be aware of.”