Of Solar Bond-age
Wherein I ponder whether SolarCity’s recent antics presage the end of solar securitization or its rebirth.
When I stopped being an editor @ijglobal, I imagined that I’d have lots of time to write self-indulgent pieces on Medium about such diverse subjects as forestry, renewables, the New Wave of British Psych Blues (hint: it doesn’t exist. Yet.) and British politics. I went into the woods instead.
But now I’d like to write for a little while about capital markets financing for solar assets. I’ve probably got as much of a track record writing about bond financings for renewables as anyone, and have reported on the first widely-distributed bond financings for onshore wind, ground-mounted solar, distributed solar, and (probably) geothermal. There were probably a bunch of super-quiet private placements that preceded those deals that I did not cover, but since a lot of what I have to say is about distribution we can safely ignore those.
Space-age structured financing
This subject is interesting because one Elon Musk-chaired company, SpaceX, recently decided to purchase short-term bonds issued by another Elon Musk-chaired company, SolarCity. I’m really glad that the transaction was that way round, because I know next to nothing about spacecraft financing, and SpaceX is much more closely-held than SolarCity.
You can find the most eloquent expression of disbelief at this transaction here. The author, Craig Pirrong, does not exactly come out and say “wait, this is the sort of thing you’d expect from a 1990s South-East Asian conglomerate or a 1920s business trust.” Because he’s a serious academic. He says simply: “When money is taken out of the left pocket to put into the right pocket, eyebrows should be raised.”
Both Pirrong and the FT’s estimable Izabella Kaminska focused a lot of their attention on Tesla, another Elon Musk-chaired entity, when looking for signs of stress in Musk’s empire. This is understandable, because Tesla has a more unusual business model, is listed, and is burning through cash at an awesome rate.
SolarCity is Nasdaq-listed, but has several competitors with reasonably similar business models, and its founders were cousins of Musk, Peter and Lyndon Rive, rather than Musk himself. So it is a slightly semi-autonomous part of the Musk empire with a more prosaic mission.
Still, it’s telling that this rather generous-minded look at the transaction in Fortune extracts an explanation for why SpaceX bought bonds from SolarCity out of SolarCity rather than SpaceX:
“The bonds offered SpaceX an attractive rate of return for a one-year investment compared to other investment options out there. SpaceX carriers a fair amount of cash at times, noted Newell, and the company wanted to put that cash to work in the short term with a high degree of reliability.”
Still, while all three Musk-connected companies are extremely capital-intensive, SolarCity’s competitive position is much more dependent on cost of capital than the other two, if only because SpaceX and Tesla have fewer direct competitors.
SolarCity claims to have two big competitive advantages — combining customer financing and installation in-house, and using the most up-to-date financing techniques available.
Much of SolarCity’s expansion has taken place against the backdrop of a huge oversupply in solar panel manufacturing capacity. Chinese manufacturers developed this capacity to serve a European utility-scale solar market that boomed and bust very quickly on the back of the provision, then the withdrawal, then the retroactive cutting, of generous feed-in tariffs.
SolarCity is now acquiring its own panel-manufacturing capacity, and may be able to start competing on panel efficiency against installers using more basic panels. But if its upfront costs are higher, its financing terms will have to be more competitive. You can read some more about potential changes to the competitive landscape in this Motley Fool article.
Why, yes, I will show you the money
I’m most interested in the financing angle, however. There are, broadly, three ways to sell solar power systems to consumers (I’ll leave utility-scale to one side for now), and three external sources of financing for installers. Firms can lease systems to consumers, retaining solar systems (and their tax benefits), they can own the systems, and sell power to the hosts/customers, or, increasingly, they can simply lend them the money.
Firms’ sources of external financing include tax equity funds (most common), wholesale and project finance banks (increasingly common) and the asset-backed securitisation market (still pretty rare).
There are some differences between the tax treatment of the three different sales models, and this can affect which financing model the seller/installer uses, but the three financings all have a similar credit profile. A typical financing for distributed/residential/roof-mounted solar (the first term is becoming most popular) involves aggregating the revenues from hundreds of systems into a single transaction.
Each of the three financing markets can claim to have some familiarity with some of the concepts used in distributed solar financing.
The tax funds involve pooling commitments from investors (including Google! Exciting!) that can take advantage of the investment tax credits available for installing solar panels. These investors overlap with the people who provide tax equity for wind farms (though the wind farms sometimes receive different types of tax credit), and with the people who invest in low-income housing projects. Both disciplines, which involve, respectively, understanding the credit risk attached to people renting or buying housing units, and understanding the performance of renewable generation systems, are applicable to distributed solar. These funds have been in existence since 2008, often with investment banks as managers and/or investors, and they understand residential solar. The government’s restrictions on who can use these tax credits make this kind of capital artificially scarce, so this financing is expensive. Lots of background at Greentech Media. The tax credits underpinning this market are set to expire at the end of 2017.
The project finance banks have been doing distributed solar since about 2013. They’re familiar with the performance risk associated with solar panels from their work on utility-scale solar, and may have financed some portfolios of larger solar farms, giving them some idea of how to do the portfolio modelling. Why are they doing it? Shortage of high-yielding opportunities elsewhere, mostly. It’s easy to undercut the funds (though tax investors may still be present in the bank-financed structure in lesser quantities), and still earn more than doing a large gas-fired power plant. It’s something of a niche business for the banks, and not all of them do it.
Securitization Until June this year, when Sunrun closed its long-awaited first transaction, this market had an issuer universe of one: SolarCity. SolarCity closed three securitisations in 2013 and 2014, raising about $325 million, and at the end of last week raised another $123.5 million. But it had been thinking about the structure for much longer. In 2009 I remember speaking to a (now-departed) SolarCity finance official about his hopes for a securitization market. “If you can do it with credit card receivables, you must be able do it in this market,” was his reasoning. I won’t name him because I was discussing starting a conference (which much later became this) rather than perpetrating journalism.
In 2009 it was quite reasonable to squint at him sceptically, because securitization techniques had only recently been the cause of a financial nuclear winter. But SolarCity had a few people on board who were quite good at banker arm-twisting. SolarCity’s former CFO (not the official I spoke to, I should stress), Bob Kelly, had previously been CFO at Calpine, where between 1999 and 2001 he somehow persuaded US power project finance lenders to abandon a raft of the protections that had girded their business since the late 1970s. The results in the near-term were not pretty, though Calpine is now prospering.
Solar securitization has now become respectable, if not yet immensely popular, in part because a number of market participants (investment banks, ratings agencies, and law firms) have spent a lot of time making it respectable. I’d also suggest that the success that Berkshire Hathaway and NextEra had in bond-financing utility-scale system nudged solar towards the bond mainstream.
Securitization can now meet the public
But solar securitization now exists, and these bonds pool a variety of credits with various customer FICO scores and contract lengths, extract a low triple-B rating from the agencies, and clear the market at coupons of 4.4% (weighted av), 4.5% (weighted av), 4.3% (weighted av), 4.59% and 4.8% (single tranche private placement).
The main structural development as these deals have evolved has been in the use of a subordinate tranche, but beyond that the structuring, including the use of a master lease agreement, is fairly uniform. And Asset Securitization Report informs us that Kroll gave the SunRun bonds a single-A rating, so there’s that.
That makes five deals in three years, and a pretty reasonable consistency in terms of coupon. Yes, I know I should work out the yield on each one, but all I have the energy to say is that the seven-year treasury has hovered between about 1.5% and 2.5% over the period that solar securitizations have been a thing.
OK, fine, I need to make a simplistic comparison between two transactions, so I’ll work out two of the spreads. The August SolarCity priced at 243bp over, and the June SunRun also at 243bp. I should allow for the fact that I might be a few days out on the prices, but I think we’ll call that a wash (Asset Securitization Report says the deal priced wide, but doesn’t mention the spread, so I may be wrong here). Given that SunRun uses third-party installers, I’m struggling to find evidence that markets are giving SolarCity’s integrated model much credit.
But the August deal from SolarCity at least dispels the notion that it had tired of the structure. It may, in fact, represent SolarCity’s admission that retail bonds will not offer much in the way of cheap financing. And it again positions securitization as almost the best source of third party financing.
SpaceX bought $165 million of these retail bonds, which are subordinate to all other SolarCity indebtedness and carry a coupon of 2% and one-year maturity. SolarCity made much of the fact that the SpaceX CFO was able to buy the bonds online from his desk, even though he was one of probably less than a dozen people who could have done the transaction just as quickly in person.
The Solar Bonds site blocks non-US IP addresses, and I’m now back to being a full-time Limey, but I consulted SolarCity’s 10-Q for the coupons, and chanced upon this delightful description of how an investor in one of its apparently third-party funds is compensated:
“As of June 30, 2015 the Company was contractually required to make payments to a fund investor in order to ensure the investor is projected to achieve a specified minimum return annually.”
The above has nothing to do with solar securitization, I’m just including it for lulz.
I’d love to know whether by registering the bonds as retail investments it makes it possible for SpaceX to account for them more generously than less liquid assets, but without reading the prospectus it’s hard to tell, and SpaceX is some way from going public.
Anyhow, what we’ve got there is the use of short-term debt (at least $165 million of the $200 million) going towards the financing of fairly long-lived assets. We don’t know what the interest rate on the other $35 million is, though SolarCity classifies $28 million of it as long-term, and specifies an interest rate range of 1.6% to 5.8%.
We can safely assume that SolarCity is some way off achieving the 5.75% for 15-year money that was its headline pitch to retail bond investors. But we can comfort ourselves with the fact that while there is very real refinancing risk, and the bonds are indeed subordinated, SolarCity has a real enough asset base that it is unlikely that investors will suffer the fate of some of the participants in the UK’s short-lived retail solar bond fad.
Lease worse option
I hope by now that I’ve convinced you that SolarCity’s enthusiasm for using the capital markets to fuel its growth is genuine, and that it deserves some credit for making this market happen. I’ve also suggested that it is still a compelling option for raising funding, though agencies insist that issuers have a decent geographical spread of customers so larger installers have the best chance of pulling it off.
But I was fibbing when I said above that there were three main external sources of growth for distributed solar developers. There’s a fourth source — creating a yieldco — that could be more important than the other three combined.
You can find explanations of what yieldcos do all over the place, but simplest explanation involves developers spinning off operational assets into a separate listed (but usually majority-owned) vehicle that promises to pay such a high proportion of its free cash flow out as dividends that investors are prepared to pay generous multiples of earnings for the yieldcos’ shares, and they, in turn, pay generously for new assets.
There is a lively debate about how sustainable yieldcos are, including how sensitive they are to interest rate rises, and whether they can keep paying such nice multiples for assets. But they will buy distributed solar, as SunEdison did for Vivint, and their presence ensures a strong bid for SolarCity’s assets if anything were to go wrong.
The Asset Securitization Report article I link to above (and whose pricing I quibbled) says simply that SunEdison has not done a securitization, and that the Vivint deal takes a potential issuer off the market. But there may be a bigger implication here.
Yieldcos can and do issue corporate debt, much of its subordinate to asset-level project financings. They can and do overextend themselves, as you can read in this excellent report on Abengoa Yield (one of many from a former colleague on Abengoa).
But what they can do is leaven the rather opaque buckets of residential systems with some big old utility-scale plants with big old corporate names as customers/offtakers. And they can do it quickly. More importantly, yieldcos, which are usually controlled and seeded with assets by their founders, are the related-party transactions it’s OK to like.
SolarCity has been mentioned as a candidate to use a yieldco before (it’s a reasonably common parlour game in US power finance), which leads me to wonder whether SolarCity has some kind of aversion to the perceptions of a loss of control of, or connection with, assets, that a yieldco can encourage.
If SolarCity’s selling points to customers (OK, fine, its brand) include control of as much of the residential procurement chain (panels, installation and financing) as possible, this might be a reasonably strong aversion, and a compelling case to stick with securitization. But given that yieldcos are creating (at least for now) a fairly solid floor price for assets, it’s difficult to see SolarCity as the weak spot in Musk’s empire.