Distributed solar deployment is often a game of off-taker credit. That is, behind-the-meter project success depends largely on whether the end-use customer has credit sufficient to support a long-term power contract or lease.
For residential customers, this often restricts deployment to those who own their home and have a relatively high credit rating, measured as their FICO score (FICO stands for Fair Isaac Corporation). Historically, residential solar adopters have been those with very high credit scores. For example, in the SolarCity securitization that transacted in November 2013 (which included systems installed from 2008 to 2013), residential customers in the portfolio had a minimum and average FICO of 680 and 762, respectively. SolarCity applies other underwriting criteria as well such as the customer has not filed for bankruptcy within the prior five years. As shown in Figure 1, residential FICO ranges from roughly 300 to 850; SolarCity average customer score of 762 is near the upper end of the scale.
Figure 1: Distribution of U.S. consumer FICO scores
Residential contracts represented 71 percent of the future cash flows aggregated through the SolarCity 2013 securitization, officially referred to Series 2013-1. Based on average cashflows for residential customers, I calculate there are about 4,525 residential systems in the securitization pool. There has been no disclosure on what proportion of the systems are leased vs. those that are on a power purchase agreement (PPA).
One primary concern raised by rating agencies in the securitization of solar cash flows is whether the contracted payment (either through a lease or PPA) would continue to be paid for if there are homeowner changes, e.g. through a home sale, or in stress-related events such as death, divorce, or foreclosure. Even with the relatively high FICO scores represented in Series 2013-1, Standard & Poor’s (S&P) — in it’s presale ratings report — raised the issue of limited “customer performance history,” suggesting customers may stop paying for their old solar systems if they have the opportunity.
However, the S&P ratings report offers some insight into this risk. According to the report, 99 percent of contracted cashflows were recovered during normal sale events and 91 percent during all other events. Normal sales represent 82 percent of all contract reassignments. Overall, 97 percent of cash flows from reassigned contracts were recovered.
But what will be the effect on asset recovery as solar developers sign lease or PPA contracts with residential customers who hold lower credit quality? Will the aggregated cash flows be more risky, and how will that impact the accessibility of capital market investment via securitization and other vehicles? In the SolarCity securitization referenced above, S&P modeled several highly stressed scenarios including one where assumed residential customer defaults increased from 25 percent to 50 percent. Even the reference assumption at 25 percent default rate is roughly 8x historical levels of default. Under the stressed case, S&P’s model indicated that the transaction could pay “timely interest and full principal by rated final maturity” of the debt securitization. In other words, there was enough extra cash in the deal to pay off the interest on schedule and the principal by the end of the debt securitization period even if every other residential customer defaulted.
Does that suggest that securitization offers a mechanism to expand the deployment of solar assets into customer segments with lower credit quality? It would appear there is sufficient integrity in the SolarCity securitization to withstand higher rates of default and continue full payment of the bonds. Clearly, the diversity of the pool offers a valuable credit enhancement to the overall repayment of the debt.
Unfortunately, investors and rating agencies apply “doomsday” scenarios on credit defaults, inverter replacements, electricity price trends, and other uncertainties wherever data may be lacking to prove otherwise. The NREL-led initiative known as the Solar Access to Public Capital (SAPC) working group is currently developing a dataset that illuminates the credit quality of solar as an investment asset class. Hopefully, the dataset will provide the necessary insight to attract a widening pool of investors.
Importantly, solar is unique among securitized asset classes as, unlike credit cards and auto leases, solar systems allow customers to offset other costs. According to Danny Abajian, Senior Director, Project Finance for Sunrun, solar companies apply additional indirect forms of underwriting by ensuring customers are projected to save money in year one of the contract, and that contract escalators do not exceed historical increases in utility rates. If the solar contract is more expensive than utility power, Abajian notes, “we do not underwrite that contract, or [we] require the customer to buy down their monthly rate by increasing their down payment. By doing this, we are creating a high likelihood of there being a system with an in-the-money contract regardless of the homeowner.”
For commercial customers, there is no universal score akin to FICO, and this makes it difficult to assess their ability to pay back the expected contract payments. Most commercial entities do not have rated credit — when they borrow, the loan is secured by the various assets of the business such as the machinery and inventory. Most small-scale businesses — and even some large ones — do not have a credit rating from a national rating agency such as Standard & Poor’s, Moody’s, or Fitch Ratings. Now consider the multitude of small-scale entities across our shopping centers, malls, restaurants, and other retail and wholesale enterprises, and it is easy to comprehend why solar has not proliferated broadly across the commercial sector.
Commercial property is generally distinguished into four components by the real estate industry:
- Multi-family housing
The problem with placing solar on these properties is several-fold. First, there is credit. As mentioned above, small businesses don’t have rated credit because they are not in the public markets for debt. The SolarCity securitization referenced above included only commercial and governmental customers with an investment-grade rating from a nationally recognized rating agency or it’s equivalent.
Second, there is long-term occupancy expectations. Businesses generally lease their space, and do so over relatively short time frames. The national average lease term for offices and industrial properties is 4 to 5 years, far shorter than the timeframe necessary for solar equipment cost recovery. Of course, this is true for real estate in general. But real estate is its own security — the asset can be resold with minimal loss in value. Solar property, however, is generally considered “unsecured” because much of the value is in the installation process. The conventional wisdom is that the cost to remove the panels and inverters and re-use them on another property is too high and thus not an economic alternative if systems are not paid for.
Third, a large percentage of commercial property owners do not pay electricity costs associated with their properties. Commercial properties are contracted under what are called “triple-net leases” — whereby the tenants pay for taxes, insurance and interior maintenance — or other contractual structures that pass the cost of power to the tenant. In these cases, solar installations only directly benefit the tenant and provide no direct incentive for the property owner. But tenants are not likely to purchase systems where they do not own the property and are uncertain of their tenancy over time.
So, if solar-adopting end-users are not expected to remain in the properties, does it really matter if they have rated credit? Or alternatively, is there a better metric to evaluate whether a solar system on a commercial rooftop will produce cash flow as predicted? And last, if a solar system is expected to reduce power bills for the tenant, is it reasonable to assume it will do the same for the next tenant so long as the general usage doesn’t change?
With those questions in mind, I researched national and regional vacancy rates for commercial properties. According to the National Association of Realtors, the following vacancy rates (shown in Table 1) are projected for 2014 and 2015 and a few top markets in each category:
Table 1: Projected Vacancy Rates by Commercial Property Type 2014 – 2015, with Vacancy Rates for Top Markets in Q1 2014
So here’s the $64 million question: can commercial solar portfolios be designed for securitization whereby low vacancy rates and high economic value of the solar energy represents a material credit enhancement to the security? And if real estate can be developed without tenancy certainty over the life of the asset, can a solar facility with near zero operating costs be afforded the same benefit?All categories of commercial real estate are improving, and default rates in the best geographic regions are frequently in the low single digits. Interestingly, some of the areas listed in Table 1 are also some of the most active solar development regions as well.
This article was originally published on NREL Renewable Energy Finance and was republished with permission.
Lead image: Solar installation via Shutterstock