Six Misconceptions and Six Hard Facts About the Solar Sector
Solarcity has bombed their earnings and Sunedison has been taken to the woodshed over the last few months. In that time it is remarkable how many bears have come out of the forest to trash the entire solar industry. Some of this is very justifiable and some of it is not. I thought it would be good to clear up some misconceptions and highlight some of the harsh realities of these businesses.
Misconception 1: Solar isn’t cost competitive! It is all driven by subsidies and will never work!
This is flat wrong — mostly. Solar’s viability is driven by geography and incumbent fuel costs more than regulations. Solar has radically different economics depending on where you are. If you are in Alaska the breakeven power price to get a particular internal rate of return is radically different to that in Nevada. The best tool to evaluate this is the National Renewable Energy Lab’s System Advisor Model. This is why you can get a 10% unlevered IRR at ~$45 per MWh in Nevada and $84 in Seattle. The economics vary by how many sun hours there are which comes down to cloud cover, latitude, humidity and other factors. Where solar “makes sense” is a very local calculation. By most measures it makes a lot of sense in the sunbelt unlevered without the investment tax credit. If you are in Maine, well, gas prices need to be higher to justify extensive solar installations. For the vast majority of Americans, solar makes a great deal of sense to say nothing of people in countries with higher gas prices, extensive sunshine and unreliable power grids — like India.
Misconception 2. The panels degrade and will be useless in a few years.
This is one of the more ridiculous arguments against solar. Extensive data from NREL indicates that panel degradation runs about 0.5% per year — so in 20 years, the life of a standard solar PPA, the plant will have 90% of its rated output at the date of construction. If my house or car had that rate of degradation I would be ecstatic but there is no way that is going to happen. To that end, solar should be looked at as a long life, low maintenance asset. Operation and maintenance data from NREL and others indicates that solar O&M runs at about $10,000 per MW of capacity per annum for utility scale installations and $20,000 for residential. If you take residential installation costs around $2.50 per Watt and utility scale of ~$1.50 per Watt, the running opex is about 1% or less of capital cost per annum. Solar degrades slower than a car park.
When it comes to residual values this math makes it pretty clear that there is life at the end of the initial power purchase agreement. Of the typical ~$1.50 for construction of a utility scale solar site only ~$0.50 is the cost of the panels. The land grading, clearing and steel frame racking does not suddenly vanish after twenty years and, assuming wage inflation does not go backwards over that time frame probably has more value in twenty years than it does today. That means that at the end of the PPA or even before it the plant can buy newer, cheaper panels at whatever the prevailing price is and continue running with a new PPA or as a merchant plant having depreciated down all its capital to zero. Those incremental returns are pretty amazing, and are an almost free option. For this reason there are good reasons to build now, get a good IRR and preserve the option to “repower” later.Think about it in real estate terms: even if a car park is twenty years old and depreciated to zero, is it worth zero today?
Misconception 3: The PPAs are risky and everyone is going to default
This is one of those things that quietly drives me insane. Getting out of bed is risky. Crossing the road is risky. A holiday in Syria is risky. Are they all the same level of risk? Of course not. PPAs with US utilities that are investment grade are a radically different proposition to subprime lending for residential solar. To that end there is a very big difference between a company like 8point3 Energy Partners LP which operates exclusively in the utility scale solar area, a Sunedison which is primarily in utility and commerical and industrial energy and Solarcity which is in pure residential solar. For example, American Electric Power has a senior unsecured credit spread of about 45bps for 10 years — that is the cost of insuring your unsecured PPA over a decade. If you are making an 8–9% unlevered IRR on that counterparty risk you are killing it, pure and simple. Similarly if your commercial and industrial customer is someone like Walmart and you are getting a similar IRR their CDS is 55bps for a decade which leads to much the same conclusion. For residential solar the math is quite different — five year auto loans are approximately 5.4% per annum and are quite clearly a different level of risk. Hurdle rates should be higher and the discount rate investors apply to Solarcity’s books should be materially different to that applied to First Solar, Sunpower and Sunedison.
The key message here is that different counter parties have different risk and should be priced appropriately by both the providers of solar solutions and Wall Street. Until Solarcity’s earnings last night it seemed that there was very little differentiation going on. Sunedison’s acquisition of Vivint was a strategic blunder and while there is a good business in taking a margin on residential solar installations these assets still need to be proven out in the minds of yield and fixed income investors and are not easy to hold on balance sheet or lever as Solarcity is finding out.
4. Why would anyone buy a yieldco? Aren’t they just stuffed with overpriced assets?
For investors who can source a PPA, build plants themselves and not need to access the securities market, sure. Similarly if you are shooting for 20% returns by investing in strong growth at a significant multiple it may be that yieldcos are of precisely zero interest to you. That would appear to be most of twitter and the tech momo crowd. That is fine. But for the vast majority of investors particularly those aged over 50 who do not want to stake their retirement savings on the future of Gopro or Tesla a 5–7% dividend yield with 5–10% growth is *very* appealing. There is a reason people buy utilities and staples, and there is a reason people buy investment grade bonds and it is not for excitement or fun.
When comparing yieldcos to other companies I find toll roads pretty useful: they have limited maintenance capital expenditures or price and margin risk and some growth. Transurban, an ASX listed toll road operator trades at a dividend yield of ~4%. While I find that a bit rich for my tastes the idea that a portfolio of similarly dull assets with long PPAs trading at ~10% yield seems odd especially in light of their long lives.
5. This is Enron math! What is with all these non-GAAP numbers in the presentations?
In the case of Solarcity, guilty as charged in my humble opinion. “Retained value” is an NPV of cash flows from solar installations. It is a measure of the net present value of excess spread between where you fund something and what you earn on it. If banks did this math it would be absurd and it is a silly metric for this industry. Until such time as Solarcity provides investors with a model which they can sensitize themselves for defaults, funding costs and so on the more they put this in presentations they more credibility they will lose.
In the case of the utility and C&I based businesses the focus on cashflow available for debt service makes a lot of sense. The “trick” of yieldcos is to maintain growth while having aggressive depreciation schedules such that you almost never end up paying tax. If you maintain growth, net income does not happen, more or less and until such time as the entire power grid is renewable it would be safe to say the scope for further growth is very clear. The problems come when the thing you build has no incremental demand for capital expenditure — something that coal MLPs like Alliance Resources know a thing or two about, and arguably the entire midstream MLP space. For solar yieldcos until we get close to a completely decarbonized economy the runway for growth is still clear.
6. There is no moat in these businesses and they just compete on price
Large, capital intensive engineering, procurement and construction businesses tend to be reasonably consolidated industries. Building a solar plant, ensuring it works and is delivered on time is not buying a coffee — if the builder, owner or operator is incompetent or deficient you cannot just walk three doors down and find anther to replace them. To that end this business should look at lot like EPC businesses like Fluor. Reasonable margins that are fairly stable for large, scale players. If nothing else people will pay a bit more — but not a lot more — to ensure that if something goes wrong they have someone they can sue. To that end players want to be in the top five players in this space because the diversification of projects, economies of scale and brand makes the business more profitable. This isn’t Alphabet or Facebook but it is not a coffeshop either.
That is enough for the misconceptions — solar still makes sense, yieldcos are a pretty sensible way to package these assets and if this sector has a fault it is that the underlying assets are incredibly boring and predictable. So what has gone wrong here?
Hard Fact 1: Credit Risk Matters
As alluded to above, the credit risk of the counterparties is not just a little important, it is *really* important. If this sounds like boring bank underwriting it should because it is. Hand wavy arguments from companies that the utility space is the same as residential is garbage.
Hard Fact 2: This business is very capital intensive and capital market confidence matters
If you build assets with stable revenues, almost no opex and a lot of capex you can do it two ways: at low returns with 100% equity or some leverage. Most people choose leverage. To run a business like that you need the confidence of your bankers. The utility and C&I space in solar appears to still have that funding support. After completely mismanaging market expectations and funding Sunedison managed to park most of its pipeline in warehouses for infrastructure investors happy to clip an 8–9% all in yield. The residential space is something less clear to me and others. Time will tell what default rates are and whether it makes sense.
Hard Fact 3: Residential leasing is dangerous because it is poor value for the consumer
Embedded yields in residential products are high — generally high single digits to low teens after one considers the mandatory step up in power price over time. This is not a virtue for the lender but a vice: eventually this business dies because good credits will buy outright as the cost of installation comes down and bad credits will lease and default. It is far better to do an installation, clip a 20% gross margin or better and get some servicing fee arrangement on the back end where you can hopefully up sell on new panels, a battery and so forth over time. Holding leases on balance sheet attracts all the risk for very low incremental return.
Hard Fact 4: Pure Play Yield Vehicles Are Preferred By Investors
If you look at how the yieldco space has traded it has become abundantly clear that vehicles full of various segments — utility, C&I, residential, some transmission lines — like Abengoa Yield have been hit hardest. Some investors want steak, some want ox tail soup, nobody wants sausage. To that end stuffing the Vivint portfolio into Terraform has a great deal to do with it being demolished in the marketplace and exiting from this portfolio would be wise. 8Point3 Energy has done a much better job here and the market reflects that.
Hard Fact 5: Most of the Problems for Sunedison are of their own making
Other developers and yieldcos like First Solar and Sunpower have traded back up since the end of September and have produced solid earnings, others like Solarcity have collapsed. Most of this is due to the differentiation between growth, underlying asset risks I have discussed and balance sheet management. Sunedison has got itself out of trouble but it is hard to imagine the market giving the company the same credit as Sunpower and First Solar until some of the architects of this mess have been removed. Carlos Domenech and Brian Wuebbels need to go. The aggressive push into acquisitions funded on the fly that are in areas outside the firm’s expertise have caused most of these issues and appear to have been driven by this pair. Removing them would send a good signal to the market about investing discipline and focus going forward. It would also make it clear that management is accountable. If nobody leaves after this much market cap is vaporized then it is abundantly clear that there is something deeply wrong with the culture of this firm: it is the kind of place where everything that is good happens because of their brilliance, and everything that is bad that happens is because of external forces. I don’t invest in companies like that, and many other investors would feel the same. Sunedison needs a CFO as even keeled as a CFO at a mortgage REIT or toll road operator and business development people need to wake up with a deep seated fear of destroying value, not the kind of deal junkie optimism that one gets from former investment bankers who have not had to take balance sheet risk before. Business development people need to be value focussed, not transaction focussed. In my observation you hire investors for those jobs not transaction focussed former bankers.
Hard Fact 6: Most of these problems can be fixed and the market has made it clear how to do so
The recovery in Sunpower and First Solar since September makes it clear that having a sensible, conservative plan and sticking to it works in this business even as the MLP and energy sector looks more and more like an dumpster fire. By pushing the Vivint lease portfolio outside of Terraform, de-emphasizing the leasing business for residential and moving back to the core C&I and utility business Sunedison can re-rate and earn the investor’s trust.