The solar industry continues to grow, even in these unprecedented times. In addition to being a source of nearly 250,000 jobs in America and a systemic response to climate change, solar is also a growing, maturing asset class.
In this special edition of #Solar100, solar expert Jigar Shah and data expert Van Skilling meet to discuss the evolution of solar as an asset class.
As the founder and CEO of SunEdison, Jigar Shah unlocked a multibillion-dollar solar market. He has led a number of the industry’s firsts, including hiring the first independent engineer for solar and pioneering no-money-down solar. He’s known as both an expert and an influencer of solar’s history.
As the former CEO of Experian and Chairman of CoreLogic, Van Skilling has overseen the growth of industries ranging from consumer credit to home mortgages. He’s become the expert on all things pertaining to asset class evolution.
Bringing their respective areas of expertise to bear, Shah and Skilling contextualize solar’s past and present, identify solar’s parallels with other investment asset classes and forecast the future of the solar industry.
Looking back: The early days of solar as an asset class
Richard Matsui: I am excited to be joined today by Jigar Shah and Van Skilling, both titans in their respective industries. Our company happens to be at the intersection of both asset classes, solar as well as consumer credit, and in this interview, we’re inviting Jigar and Van to contextualize and lend insight into our industry’s past, present, and future.
Jigar, this first question on solar’s history is for you. We, alongside 10 other solar firms, published this year’s Solar Risk Assessment to share quantitative data on qualitative trends we’re seeing in the industry. Contributors included industry leaders Wood Mackenzie Power & Renewables and NextEra. A big headline-grabber this year: DNV GL reported that solar assets nowadays are underperforming by 5.4 percent, on average, even after adjusting for weather. In the early days of solar, solar was constantly overperforming estimates. Is this a trend you’re seeing in the industry as well, and can you help contextualize this?
Jigar Shah: Yes, and it’s a natural phenomenon. When you think about where solar was in 2010, people were getting loan guarantees or being forced to sell their projects to Warren Buffett at highway-robbery costs. Then in 2012, people started saying, “How about we make this more efficient?” Then the Chinese solar tariffs came in, and people began to ask, “Where do I flex to make my numbers work?”
Solar developers started shopping around for independent engineer (IE) reports, in an attempt to get a better fair market valuation and more tax equity. Every variable adjusted in the model was a couple of basis points, then a couple more basis points, and so on.
Now that time has passed, I think people are going back through the data and realizing, “Hey, these estimates weren’t true. This valuation wasn’t real, and now it’s time for a correction.” I don’t think this course correction is going to kill anybody, though I think some of the returns will be lower for some of the equity holders.
The bottom line is, as an industry, I think that we can do better and we should do better.
RM: Let’s turn back the clock to the first deals you were financing with SunEdison — did this industry always use IEs? How did technical questions get solved in early deals?
JS: At SunEdison, we hired the first IE for solar. I remember we signed a deal with Goldman Sachs in 2005, and Goldman asked us, “Who are you going to use for an IE?” So I went out to the most skeptical solar people that I knew, and I hired them. And they gave me a crappy report that said it would generate 8 percent less than we all thought. Goldman and SunEdison financed the project using the IE’s estimates.
We were extremely conservative in the early days. We sold a lot of those projects to Wells Fargo in 2007. Today, most of those Wells Fargo projects outperform by about 6 to 8 percent. In terms of production, we swept the excess cash so it was a good deal all around. Terms got tighter and tighter after that, but we were forced to be pretty conservative in the early days of solar.
RM: You’ve summarized it well — any incentive structure guides behavior. Lenders, tax equity and the entire financing edifice rely on IEs for production estimates. These production estimates come from IEs that are hired by either the sell-side or buy-side developers who have a profit incentive to see a certain number from these IEs.
When you talk to the IEs with the mics off at a bar at Solar Power International, they’ll tell you that they’re all competing with 30 other IEs for business and that there’s a very clear outcome that their client is trying to achieve. A prominent solar developer told me that for every large project they sell, they’ll hire five different IEs to provide production estimates, and then they’re going to pick the highest number. From a pure cost-benefit analysis, you can’t refute the strategy to pay $10,000 for each IE estimate, because they can get a higher P50 [production estimate] that translates to millions in the asset sale price. And over time, there’s a resulting drift in P50s, as you said, with “a couple of basis points here, a couple of basis points there; turning the soiling knob here, turning the shading knob there.”
It seems IEs haven’t seen an end in sight because the industry just continues to push further and further in this direction. How does that line up with you?
JS: The buy-side is into this trend, too. A lot of players on the buy-side actually really wanted the volume and were willing to compete for it. Remember, these were funds making fees. They didn’t really care about what the [limited partner] returns were.
It is the same in solar; the whole dynamic with IEs was basically the blind leading the blind. But as an industry, we’ve got another trillion dollars of solar to put to work over the next six years. It’s time to course-correct, address the problem of incentivized over-estimates and make sure that everyone gets a more fair deal going forward.
RM: Right. All that behavior between IEs and developers is rational, but it’s fundamentally changed the role IEs play; in practice, IEs are now hired to help the buyer or seller to achieve a better asset valuation. This reminds me of the role that lawyers play in the American legal system: Both sides hire lawyers not to find an objective truth but to argue their respective strongest case. This is sensible in the subjective world of justice, but this is engineering, not philosophy. What do you see as the trajectory of this market incentive structure and its implications?
JS: Well, we’re going to fix it now. When a report like the Solar Risk Assessment comes out, and the bosses of the people who had to put money out the door for projects read it, the bosses are going to confront this behavior and say, “Hey, let’s stop doing that crap.” As we go into the next phase of growth, we have to do better.
The reason our industry is in a good place is because it’s inevitable that the industry begins using market data to shine a spotlight on the excesses and biases. With companies like kWh Analytics publishing these insights, we can then fix those estimates.
It’s also an opportunity to realign the role of market participants and allow them to focus on their areas of expertise: IEs providing technical assessments and not solely relied on for financial estimates.
Van Skilling: Jigar, that realignment reminds me of how home inspectors and appraisers operate in tandem in real estate. Home valuations still leverage technical expertise, but appraisers and now databases like Zillow can help supplement with valuations data using market comps.
Moving forward: The maturation of solar as an asset class
RM: Solar is still a relatively young industry. Van, when you look at the evolution of other asset classes, do you see any parallels?
VS: I certainly do. Let’s look at an industry I’m very familiar with: consumer credit. Today, consumer debt drives the U.S. economy. And the fuel for that consumer debt is the consumer credit data, which allows the debt to be incurred and repaid.
Ten years from now, I both hope and expect solar, as a growing source of clean energy, [will] be our largest source of energy. So having the data to support that growth and to allow it, I think, is very important.
Credit has been around for thousands of years, but it’s really a relatively new business. Credit as we know it now was actually started in the 1930s by Sears, which was the Amazon of their time. A substantial number of houses in America had a Sears catalog, and you ordered whatever you needed from the Sears catalog. And so Sears kept credit records of their own customers.
As an industry, credit really didn’t take off until 1950 when Diners Club came up with a credit card. With that success came other credit card companies: American Express, Visa and MasterCard. These were credit companies, but they maintained their information on pieces of paper in these high-tech devices called Rolodexes, and they communicated between their offices by telephone.
In late 1960, one of the founders of TRW, which became Experian, said, “There’s going to be a cashless society that captures data on computers. And we know more about using computers for data than anybody else. So we have to get into this business.”
As a result, TRW, now called Experian, introduced computerization and data aggregation into a business that improved accuracy and efficiency in credit underwriting. This evolution in credit was inevitable given the availability of data in the market. And I think you’re going to see a similar evolution in solar, where the use of data is going to make solar financing better, easier, faster and more reliable.
RM: You’ve pointed out an interesting parallel. To expand on that, today everyone knows what a FICO score is — a universal market comp for lenders and consumers to assess credit risk. Obviously, there’s a time when FICO scores didn’t even exist. What was the inflection point that enabled TRW/Experian to gain market adoption and create that standard of comparables to the rest of the market?
VS: For context, FICO depends on Experian’s data for their model — FICO’s model relies on consumer credit database. FICO scores are universally accepted today, but that wasn’t always the case. What precipitated that change was that all credit investors found that it was in their best interest to contribute their information to the credit bureaus. The more information there was about the individual, the better credit records they could maintain. This enabled a source of truth for market participants to rely on and improved underwriting.
Prior to recognizing the value of contributing data, the consumer credit industry looked similar to the solar industry today, in the sense that each investor largely kept their own records of their assets in-house. If they shared those records, it was only shared on a local or regional basis. So each investor was dependent on a very narrow database to be making your decisions. Consumer credit hit an inflection point in the ’90s when stakeholders realized they could improve efficiency and accuracy by granting data access to a database that they could use for finding new customers as well as maintaining their own customer base. Combined with computerization, the database became global. Now virtually everybody uses a FICO score, which is based on consumer credit data.
Jigar, going back to your earlier point about the use of data, it sounds like you think solar will undergo a similar inflection point to use market data to fix inefficiencies.
JS: Right, I think data will inevitably get reported, the truth will come out and smart money will demand better. That ball has already started rolling because there’s actually a dataset out there. kWh Analytics has a dataset. And there are others that have a dataset. So at some point, the use of objective market data will be fully normalized. And to the extent that estimates deviate from that normalization, the buyer side will expose it very quickly, because they’re going to say, “Well, the last three deals done nearby had x in their IE report, so how is it that you think that your system’s going to produce 6 percent more?” That level of transparency will force accuracy.
Then the fight will be over the next generation of tracking software or panel technology improvement. And frankly, we want to keep encouraging that level of innovation.
It’s constantly a cat-and-mouse game, but my sense is that the deviation will be lower over time. So instead of a 5 percent deviation, as we found this time, my sense is that deviation in the future will be 1 percent or a half percent. And I think that that means that there’ll be a little bit more comfort within the financial soundness of the system.
The people getting screwed now are very sophisticated equity investors who should know better and [should know more about] what data they’re using. So if they are not reading your market reports or listening to the Currents podcast or figuring out how to educate themselves, then shame on them.
RM: Fascinating. You’re right — as an industry, we have the tools to course-correct and guide the evolution of solar. It is up to us to take that step forward.