When SunPower (SPWR) and First Solar's (FSLR) YieldCo, 8point3 Energy Partners (CAFD), went public two years ago, I used the financial nerd joke in 8point3's ticker symbol as a launching point to explain what "cash available for distribution," or CAFD, means.
In that article, I cautioned against the risks of using a short-term cash flow measure for long-term investing decisions. That risk is becoming more and more real for investors in 8point3 because the YieldCo is using short-term, interest-only financing to fund its long-term investments.
All of 8point3's debt matures in 2020, and refinancing that debt will reduce its ability to pay dividends for two reasons. First, interest rates are rising, which will lead to higher interest payments. Second, if the YieldCo is unable to secure interest-only debt, it will have to refinance with amortizing debt. The principal payments from amortizing debt will further reduce CAFD.
First Solar and SunPower are also considering a sale of the company. The better-capitalized YieldCo NextEra Energy Partners (NEP) has been mentioned as a possible buyer.
If another YieldCo were to buy 8point3, it would do so at a price that allowed it to raise its own dividend. This is the same yardstick YieldCos use when evaluating the effects of buying renewable energy projects from their sponsors or third parties. If a transaction will not increase a YieldCo's CAFD per share (preferably significantly), it will not do the transaction. This rule applies to both individual projects and large transactions like purchasing another YieldCo like 8point3.
A somewhat naive version of this rule would be that a YieldCo with a low dividend should be able to purchase a YieldCo with a higher dividend. Unfortunately, it's more complicated than that. Here's why.
1. Declared dividends are not the same as a YieldCo's ability to pay dividends. Different YieldCos have different payout ratios (the proportion of CAFD that they distribute as dividends). Naturally, an acquirer will keep its own dividend policy when acquiring another YieldCo, and so the value of the acquired YieldCo will depend on its CAFD, not the proportion of CAFD it has chosen to pay as dividends before the acquisition.
2. CAFD is what is called a non-GAAP measure, meaning that it is not defined by generally accepted accounting principles (GAAP). As such, different YieldCos do not use the same definition of CAFD, and some are quite aggressive in how they define it. This can lead to declared CAFD that is higher than the YieldCo's ability to pay dividends in the medium to long term. Naturally, a buyer will want to use a conservative measure of CAFD that is more indicative of the purchased YieldCo's assets to support medium- and long-term dividend increases.
3. 8point3's debt is interest-only and does not include principal payments. Most of this debt was issued when interest rates were lower than they are today, and all is due by the end of 2020. No other YieldCo uses this capital structure; most limit their interest-only debt to about one-third of total debt. A buyer would want to refinance most of this with amortizing debt to match its current capital structure. The increased principal and interest payments will reduce CAFD significantly.
4. As I pointed out in my article two years ago, CAFD exaggerates the value of projects that are likely to have less value at the end of their existing power-purchase agreements (PPAs). This will be dictated by the ability and cost to develop competing projects in the future. Solar projects can be built almost anywhere, and solar prices are falling rapidly. This means that solar projects are likely to have very little value at the end of their PPAs. Wind will be somewhat more valuable, while hydropower and geothermal are likely to be the most valuable renewable energy projects in the long term.
8point3's portfolio is entirely solar, meaning that using CAFD exaggerates 8point3's valuation compared to YieldCos that include other types of assets. All other YieldCos contain non-solar assets, and most have less than 50 percent solar in their portfolios.
5. IDRs, or incentive distribution rights, redirect a portion of a YieldCo's dividend to its parent company. For an acquirer with an IDR, any increase in per-share cash flow from a possible acquisition of 8point3 would have to pay for the IDR as well as a dividend increase to the buyer's common shareholders. In the case of NextEra Energy Partners, 25 percent of any increase in distributions goes to its parent as an IDR payment, while 75 percent goes to the YieldCo's shareholders.
Others have attempted to value 8point3 in a potential buyout by NextEra Energy Partners, but I have yet to see an analysis that takes most -- if not all -- of the above factors into account. Below, I try to do just that, evaluating what the other YieldCos (not just NextEra) might be willing to pay for 8point3 Energy Partners if they were to buy it.
Comparable, sustainable CAFD
The meaning of "cash available for distribution" should be clear. It's cash that is available to be distributed to shareholders. Actual definitions vary.
One would expect that CAFD should include all cash produced by the company's operations, investments, and financing that is not needed to maintain (but not replace) the company's investments.
For investors' purposes, these cash flows should not include one-off payments and cash flows that are likely to reverse in future periods. This is because investors are interested in a company's ability to both pay a dividend today and maintain that dividend for the long term.
The more conservative a YieldCo's definition of CAFD (i.e., the lower the final number), the more confident investors can be that expected future CAFD is a good measure of expected future ability to pay dividends.
The definitions of CAFD that I find most realistic are the ones used by Pattern Energy (PEGI) and NRG Yield (NYLD, NYLD-A). They start with the cash flow from operations (CFO), a measure defined by GAAP and hence comparable across YieldCos.
They also adjust for changes in operating assets and liabilities (which should reverse in later periods), subtract operations and maintenance capital expenditures (necessary to maintain equipment), add distributions from minority investments, subtract payments to the minority investors in its projects, and subtract payments of principal made from operating cash flows. Note that interest payments are considered an operational expense, and thus are already subtracted when calculating CFO.
The one adjustment that does not fit my model is the addition of one-time cash flows such as "network upgrade refunds." Several other YieldCos include these in their definitions of CAFD as well. These refunds are likely to be one-off for any particular project, and so they should be excluded from any attempt to estimate a YieldCo's ability to pay dividends in the medium to long term. In order to be useful to investors, the current CAFD should only include cash flows that are likely to repeat.
I attempted to build my own version of CAFD that would be comparable across all YieldCos and only include sustainable cash flows. I found that most YieldCos do not disclose sufficient information to complete this calculation.
The major barriers I found were different types of disclosure around minority investments in affiliates. These affiliates own the energy projects, and the YieldCo owns all or part of the affiliates. The lack of consistent disclosure meant that I could not be confident that my numbers were comparable across YieldCos. A related barrier to comparability is company structure. Most YieldCos have complex financial and legal structures, and these structures vary from one YieldCo to the next.
Below, I outline the results of my survey of definitions for CAFD.
More conservative definitions of GAAP
In my subjective analysis, I found that Pattern Energy, Atlantica Yield, and TransAlta Renewables had the most conservative CAFD definitions. Perhaps not coincidentally, all three of these YieldCos report results using international financial reporting standards (IFRS) instead of (or in addition to) U.S. GAAP.
IFRS is more of a principle-based system than GAAP, and so places more emphasis on intent than GAAP. GAAP is a rules-based system where the focus is more on the letter of the law.
Neither GAAP nor IFRS define cash available for distribution. GAAP gives a company nearly free rein to make up its own CAFD definition. IFRS principles provide somewhat more guidance, and seem to have resulted in more conservative definitions of CAFD.
The relatively conservative definitions of CAFD employed by Pattern, Atlantica and TransAlta are all defined with reference to cash flow measures. Pattern and Atlantica both treat their minority stakes in affiliates (renewable energy installations that the YieldCo does not own outright) conservatively by limiting the CAFD impact of these investments to actual cash received.
TransAlta instead includes a proportionate share of adjusted funds from operations (AFFO) for minority stakes in affiliates, but also subtracts a proportionate share of AFFO for minority interest in its affiliates when it has a majority stake. This is conservative because TransAlta has more majority-owned than minority-owned affiliates.
Treatment of minority/majority stakes in affiliates
The common factor in the two methods for dealing with minority and majority stakes in affiliates is that the three YieldCos discussed above treat the CAFD from their minority stakes in affiliates as they do the CAFD associated with the minority stakes of others when they have the majority share. In other words, they apply the rules consistently.
Hence TransAlta Renewables distributes the AFFO pro rata between the owners of each affiliate, while Pattern and Atlantica attribute only the cash payment to the minority investor, regardless of whether that minority investor is the YieldCo or some other party.
In contrast, the other three YieldCos seem to have chosen their definition of CAFD to maximize the number they report, rather than for internal consistency. NRG Yield, NextEra Energy Partners and 8point3 Energy Partners treat their own minority investments in affiliates differently than they treat other investors' minority investments in their majority-owned affiliates.
When these YieldCos own 25 percent of a project (a minority stake), they claim 25 percent of the project's CAFD. When they own 75 percent of the project (a majority stake), they claim more than 75 percent of the project's CAFD. These choices lead to a higher value for CAFD than what it would have been if they treated both types of ownership the same way.
Just how much this differential treatment of minority- and majority-owned investment affects CAFD varies based on the relative size of each YieldCo's investments.
8point3 did not respond to my request of a reconciliation of CAFD to CFO, which I had hoped would have allowed me to confidently isolate this effect. Lacking that information, my best guess (based on the slide below) is that 8point3's CAFD guidance for 2017 would be reduced from a range of $91.5 million to $101.0 million, to a range of $86.8 million to $93.4 million. With 79 million shares outstanding, this is a difference of between 6 and 10 cents per share.
Reimbursement of network upgrade costs
Among the one-off cash flows many YieldCos include in CAFD, the largest is usually refunds of network upgrade costs. These costs are incurred when a new solar or wind farm is built in order to ensure that the grid can handle the additional power production. The YieldCo can only be reimbursed for these costs once, and hence they are non-recurring.
8point3 Energy Partners expects to receive $13.2 million in network upgrade refunds in 2017. Since the YieldCo is making far fewer purchases of solar projects in 2017 than it was in 2016, we can expect network upgrade refunds to fall dramatically next year.
If we remove these one-off refunds from 8point3's 2017 CAFD guidance, it is reduced to $73.6 million to $80.2 million, or $0.93 to $1.02 per share. This is below 8point3's current annual dividend ($1.09 per share).
Refinancing 8point3's debt
Elsewhere, I have attempted to calculate the effects of refinancing 8point3's debt before it is due in 2020. My estimates range from a best-case scenario where all the debt is refinanced with interest-only debt at current interest rates, to one where two-thirds of debt is replaced by amortizing debt.
Today's higher interest rates would reduce annual CAFD by $11 million in the most optimistic case. Replacing two-thirds of 8point3's debt with amortizing debt would reduce annual CAFD by approximately $20 million to $25 million. I use $20 million in the calculations below.
8point3's sustainable CAFD
Putting these adjustments together, we find that, based on the company's 2017 guidance, its sustainable annual CAFD after replacing two-thirds of its debt with amortizing debt would be $54 million to $60 million, or $0.68 to $0.76 per share.
What another Yieldco should pay for 8point3
Most YieldCos have a target payout ratio (the percentage of CAFD paid to shareholders) of 80 percent to 90 percent. The residual value of solar farms after the end of their PPAs will be lower than most other types of renewable generation, as explained in the beginning of this article. For that reason, I assume that another YieldCo would only want to buy 8point3 if the purchase would both allow it to increase its dividend, while maintaining an 80 percent or lower payout ratio.
If the YieldCo were to pay for 8point3's shares with shares of its own, it would then be paying at most 75 percent of 8point3's per share CAFD, or $0.51 to $0.57 on each of the 8point3 shares. This amount would have to be further reduced for NextEra Energy Partners, which pays 25 percent of any dividend increases to its parent in the form of incentive distribution rights.
An investor expecting NextEra's stock to rise further should buy it. Barring that, investors hoping for another YieldCo to buy 8point3 above $13 will be disappointed. 8point3's current share price of almost $15 seems due more to investors chasing yield than a careful valuation of the company in a buyout.
Why does 8point3 continue to raise its dividend unsustainably? Most likely, it hopes the high yield will drive up the share price. While this has succeeded to a limited extent, it is not close to a level (over $20) where 8point3 might be able to issue new shares and grow its way out of its current problems.
The final victim of the YieldCo bust?
Another possibility is that 8point3 and its sponsors don't have a strategy at all. Before the YieldCo bust, it seemed to many that YieldCos could keep issuing new shares to buy more projects at increasingly higher prices.
Rising dividends, rising share price and rapid growth worked together in a virtuous cycle. If a YieldCo raised its dividend a little faster than it should have, the only consequence was an opportunity to sell more shares at high prices and use the proceeds to paper over past mistakes.
Then share prices stopped rising. The weakest YieldCo sponsors, SunEdison and Abengoa, had been relying too heavily on cheap capital from their YieldCos and other sources. They both fell into bankruptcy.
SunEdison's YieldCos, TerraForm Power and TerraForm Global are in the final stages of being sold to Brookfield Asset Management and Brookfield Renewable Partners at one-third and one-half of their IPO prices, respectively. Atlantica Yield is in the final stages of establishing itself as an independent company, but it also is down a third from its IPO.
First Solar and SunPower never relied heavily on 8point3, but the solar market is in a downturn, and all solar manufacturers are being squeezed. Neither can afford to support 8point3 if it is unable to stand on its own. Like TerraForm Global, 8point3 went public just months before the YieldCo bust and never had a chance to issue new shares at higher prices.
I think it's unlikely that 8point3 will share TerraForm Global's fate and be sold off to the only bidder at a third of its IPO price ($7 a share). Its sponsors are far less desperate than the bankrupt SunEdison. Using the same formula as in the table above, Brookfield Renewable Energy Partners could buy 8point3 for as much as $8 a share, but I doubt First Solar and SunPower would find that offer attractive.
More likely, the sponsors will simply fail to find a buyer at a price that satisfies them. In a few months to a couple of years, 8point3 will cut its dividend. The resulting stock price collapse may allow its sponsors to take the YieldCo private, selling off its assets piecemeal to fund the buyback of public shares.
The bottom line
Most of the above numbers are estimates. The most striking thing about the disclosures surrounding CAFD at most YieldCos is just how difficult it is to decipher the accounting underlying those numbers.
The bottom line is that 8point3's cash distribution outlook does not "look right." It's easy to get lost in the accounting, so in April I built a chart comparing these YieldCos' basic profitability and efficiency metrics in 2016.
For other YieldCos, adjusted EBITDA (orange bars) ranges from 60 percent to 85 percent of sales. For 8point3 (CAFD), adjusted EBITDA was 125 percent of sales in 2016.
For other YieldCos, CAFD (yellow bars) ranges from 15 percent to 38 percent of sales. For 8point3, CAFD was 120 percent of sales in 2016.
How can a company offer more cash for distribution than it is taking in sales? I asked. The folks on 8point3's investor relations team told me that it is because revenue from minority-owned projects is not included in total revenue. With this adjustment, 8point3's 2016 CAFD falls to 92 percent of sales plus distributions from minority-owned projects. This is more than double the unadjusted CAFD/sales for other YieldCos.
I keep coming back to this from different angles. Each time, I reach a similar conclusion: 8point3's CAFD is less reliable than that of other YieldCos.
***
Tom Konrad is the manager of the Green Global Equity Income Portfolio. Disclosure: Long PEGI, ABY, RNW.TO, NYLD/A, TERP, GLBL, BEP, FSLR. Short CAFD.