Cost of capital: 2016 outlook
More than 2,000 people listened in January as a group of project finance industry veterans talked about the current cost of capital in the tax equity, bank debt, term loan B and project bond markets and what they foresee for the year ahead.
The panelists are John Eber, managing director and head of energy investments at J.P.Morgan, Jack Cargas, managing director in renewable energy at Bank of America Merrill Lynch, Thomas Emmons, managing director and head of renewable energy finance for the Americas at Dutch bank Rabobank, Jean-Pierre Boudrias, managing director and head of project finance at Goldman Sachs, and Jerry Hanrahan, vice president and team leader, power and infrastructure, North American corporate finance at John Hancock. The moderator is Keith Martin with Chadbourne in Washington.
MR. MARTIN: John Eber, what was the tax equity volume in 2015, and how did it break down among wind, utility-scale solar and rooftop solar?
MR. EBER: We estimate that about $11.5 billion in new wind and solar deals were mandated in 2015. Of that amount, about $6.4 billion was wind. There were 40 wind projects with an aggregate capacity of 5,700 megawatts, which is the same number of projects in the previous year, but an increase of about $700 million in tax equity raised in 2015 over 2014.
There were three leading sponsors in the wind sector that did about $1 billion each. They accounted for about 47% of the total wind tax equity raised.
The solar market is not as transparent as wind, but we estimate that about $2.6 billion in tax equity was raised in the residential rooftop market by the three leading residential rooftop companies. That is up from about $1.9 billion in 2014. The top three sponsors account for about 90% of the residential market.
Another $2.5 billion in tax equity was raised for utility-scale and commercial and industrial solar projects. The solar figures are rough estimates.
MR. MARTIN: So there was only a modest increase in tax equity volume in 2015. In 2013, the total volume was $6.5 billion. In 2014, it was $10.1 billion. Why was there a slowing down in the rate of increase?
MR. EBER: The increase was still significant in absolute numbers because it comes off a huge year in 2014. These are historically significant numbers.
MR. MARTIN: How many active tax equity investors are there currently?
MR. EBER: We estimate about 20 in wind and 27 or 28 in solar. However, of the 20 in wind, we identified only 17 who actually entered into mandates in 2015.
MR. MARTIN: There is overlap between the two. How many total investors were there in 2015?
MR. EBER: No more than 30.
MR. MARTIN: Jack Cargas, what do you expect in the year ahead: deal volume, number of investors, mix of transactions, mix of structures?
MR. CARGAS: The volumes are likely to be similar to what we have seen in the last couple years.
In terms of transaction types and structures, the partnership flip structure will continue to be the leading application. There are lots of variations in the basic structure.
We expect to see similar volumes as in 2015 in wind and residential solar and a possible increase in the volume of utility-scale solar as there are some significant utility-scale solar projects that are expected to come to the market in 2016.
It would not be terribly surprising to see a further increase in the number of tax equity investors. The number increased in each of 2014 and 2015. The tax credit extensions in December may bring other investors into the market.
MR. EBER: I agree. The market should remain very active in 2016. Wind developers will feel more pressure to get projects underway in 2016 to qualify for full tax credits. Wind projects that start construction in the next three years after 2016 still qualify for credits, but at reduced levels. Solar developers may not feel as urgent a need to act in 2016 as they have until December 2019 to start construction and qualify for full tax credits and another two years after that to qualify for reduced credits.
MR. MARTIN: Many people expect that there will be a reduction in 2016 deal volume due to the tax credit extension since there will be less pressure to complete projects in 2016. It does not sound like either of you agrees with that.
MR. EBER: A lot of the commitments by tax equity investors are made well before a project is completed. Developers who plan to begin constructing projects and investing money in them will be in the market seeking tax equity commitments even though a project may not be completed until 2017. That will keep us busy this year.
MR. CARGAS: Sponsors will not have to commit unnatural acts to complete projects in 2016. Some of what would have been completed in 2016 before the extension will now be spread into 2017.
MR. MARTIN: A lot of deals carried over into the start of 2015 as tax equity and due diligence shops had too little capacity in late 2014 to handle the volume. Did you see the same thing happen at the end of 2015?
MR. EBER: Not so much. It helped that more firms were offering engineering services. Many of us were better staffed last year in anticipation of the need. We were very busy all through the year and especially in the fourth quarter, but we managed to get everything processed that our clients wanted to close in 2015.
MR. MARTIN: Jack Cargas, tax equity yields were trending down last year after several years of remaining stable. Do you think there is still downward pressure on yields?
MR. CARGAS: We are cautious about discussing pricing in public forums. Tax equity is unlike debt where you can look up accurate prices with a couple keystrokes.
There is also a more important point. Yields were a function for a long time of supply and demand. But as the deals themselves become more complicated, many more things bear on yield. You have projects with different cost mechanics, in different states, with different cash distributions and different credit buckets. There can be different offtake arrangements: corporate PPAs, PPAs with sponsor affiliates, puts, hedges, merchant tails. In the wind market, there can be various kinds of pay-go arrangements.
The point is that it has become much more difficult to generalize about yields. They are a function of a more sophisticated credit analysis based on numerous interactive risk and reward parameters.
MR. MARTIN: John Eber, let me ask the question differently. Falling yields are usually a sign of a buyer’s market. Are you seeing more deals put out for bid among two or more tax equity investors?
MR. EBER: Yes. That was not uncommon before the financial crisis. It became less common after the financial crisis. Everything is settling down now. There are more investors. It is not unusual today to see sponsors ask for proposals from more than one tax equity investor.
I agree with what Jack said. Yields have trended down somewhat if you look at a mainstream wind deal with a 20-year power purchase agreement with a solid utility using well-known wind turbines in a stable part of the country.
However, as Jack pointed out, there is more variation today in deals. Only a small subset of the potential tax equity market may be interested in a deal with some merchant characteristics or unusual turbines. We are seeing a lot more corporate PPAs.
MR. MARTIN: How much of a yield spread is there among the principal asset classes: wind, utility-scale PV, residential solar, C&I rooftop?
MR. CARGAS: Offering generalized data points for each asset class is difficult for the reasons John and I just mentioned.
MR. MARTIN: John Eber, same answer?
MR. EBER: Pretty much. I think you can say utility-scale wind and utility-scale solar look alike from a return standpoint, but rooftop solar is a different market.
MR. MARTIN: Jack Cargas, what percentage of the capital structure is covered by tax equity in the typical wind or solar deal?
MR. CARGAS: For wind, a third to two thirds of the capital stack is tax equity. For solar, the figure is probably a third to half.
MR. MARTIN: John Eber, are there any other noteworthy trends in the market as we enter 2016?
MR. EBER: The most noteworthy is the significant increase in the number of commercial PPAs for wind projects. The commercial and industrial solar market has always relied on them. The fact that they are showing up in wind is changing the dynamics of the marketplace.
It is great for wind. It is causing more megawatts to get built. However, we have to analyze a different kind of risk when underwriting deals. Bloomberg reported that more than 3,000 megawatts of commercial PPAs were signed in 2015.
MR. MARTIN: The commercial PPA market basically doubled last year from the year before. People expect further growth this year. Jack Cargas, is it harder to raise tax equity for a project with a commercial PPA than for one with a utility PPA?
MR. CARGAS: Yes, there is a difference. It has not been a standard offtake arrangement. We need to analyze the contract terms and the credit issues early in the process.
MR. MARTIN: Are there any other noteworthy trends as we enter 2016?
MR. CARGAS: We are keeping a close eye on what the states are doing on net metering. Nevada reduced the bill credits available to homeowners who send excess electricity from their rooftop solar systems back to the grid. Nevada homeowners are losing the financial incentive to install solar.
There are some pretty draconian predictions about what that does to the Nevada solar market, and other states are also evaluating changing their net metering rate structures.
It may be too early to call it a trend. It may never become a trend. However, if it were to become a trend, it could put a significant chill on the residential solar market in the affected states.
MR. MARTIN: Let’s move to bank debt. Tom Emmons, what was the North American project finance market in 2015 compared to 2014?
MR. EMMONS: Overall, 2015 was a very strong year, but there were winners and losers. The bank market was up 25% in dollar volume over 2014. Deal volume was $56 billion in 2015 compared to $45 billion in 2014. The increase is on top of a 65% increase in bank project finance lending from 2013 to 2014. The project finance bank market basically doubled in the last two years.
Those are the headline numbers, but what is going on within subsectors is more interesting.
In 2015, renewables were up 70% to $17 billion. Big growth occurred in both wind and solar, which was unlike 2014 when solar dominated. Both wind and solar were strong last year.
Oil and gas was flat at $20 billion, but within that number was a 50% jump in LNG loans to $17 billion in the first half of 2015 combined with a total collapse in upstream oil and gas from $4 billion in 2014 to almost nothing in 2015. Lending to finance gas-fired power projects was down 25% to $10 billion.
In summary, 2015 was the year that renewables moved into first place in power generation volume.
MR. MARTIN: How many active banks were there in 2015, and how many do you expect in 2016?
MR. EMMONS: There were 104 banks who were active in 2015, up about 10% from 2014 and up 50% from 2013.
But it is not the number of active banks that is telling; it is the volume of loans that the biggest players are making.
In 2015, 20 banks lent more than $1 billion each compared to 12 who lent more than $1 billion in 2014. The largest lender in 2015 committed almost $5 billion. The big players are doing more and more.
So as I anticipated a year ago, the depth of capacity in the bank market is coming from the bigger players doing more, and not from new banks coming into the market. Having said that, I expect in 2016 still more banks will enter the market.
MR. MARTIN: 2015 was an odd year. It got off to a slow start in terms of deals and picked up speed as the year went on. In late July to early August, share prices for TerraForm Power and NRG Yield and their affiliated sponsors crashed, and the market began withdrawing liquidity for buying operating assets. Until then, people had talked about a market awash in liquidity.
Did the market remain awash in liquidity to build new projects, and has lending recovered to buy operating assets?
MR. EMMONS: Banks always have an interest in new projects, but they also like operating assets, including portfolios, that can be seen as lower risk. My assessment of last year was that if there was a cooling of interest in operating assets, it was due to the sponsorship of those loans by yield cos and similar vehicles. The cooling off was not due to the asset class itself.
In terms of any recovery to buy operating assets, as I said, I think banks are always interested in operating assets, but they will remain selective and cautious as to the sponsorship for those operating asset loans until the yield co market stabilizes.
MR. MARTIN: So the market remained awash in liquidity all year, but some types of borrowers had a harder time as the year went on.
MR. EMMONS: That’s right.
MR. MARTIN: What is the current spread above LIBOR for senior bank debt, and what does that translate into as a coupon rate?
MR. EMMONS: Of course, there is a range depending on a lot of factors, but I will try to generalize. For short-term construction debt, the spread is typically 1.5% to 1.75% over LIBOR. For term debt, maybe add a quarter percent with step ups over time. If that term debt is back leveraged, then add some more depending on the risks and structure of the back leverage.
As a coupon, you would add those spreads to a base rate of about 2.4%, which is the current 10-year swapped LIBOR rate. All in, long-term rates are in the low 4% range for term debt, which I think is very attractive.
MR. MARTIN: You are speaking about renewables primarily?
MR. EMMONS: Yes. I don’t think the rates on loans to finance gas-fired power plants are all that far off, and gas-fired loans tend to have shorter tenors. Renewables, thanks to the long-dated PPAs, tend to be longer-term loans.
MR. MARTIN: The construction loan rate 1.5% to 1.75% above LIBOR. Add a quarter percent for term, and then you have a spread above that for back leverage. I think over the summer you said the spread for back leverage above term debt is 50 to 100 basis points.
MR. EMMONS: A big factor in the spread is deal size, because the number of back-leverage lenders is limited. Other facts are loan tenor, whether there is a hedge or a PPA, and then, of course, the size and shape of cash distributions to the sponsor over time.
The less lender-friendly these factors are, the shorter, pricier and tighter a back-leverage loan might be. For a back-leverage loan that is less lender friendly, there could be a premium of as much as 100 basis points.
MR. MARTIN: Back leverage tends to be found more commonly in the solar rooftop market. Is there is a difference in rates between residential solar and commercial and industrial solar?
MR. EMMONS: The more complex the deal, the fewer players there will be and the higher the premium for back leverage. For a straight, single project, utility-scale wind or solar deal, the spread will be smaller. For back leverage on a hedged wind deal, the spread will be larger, and the spread on a portfolio C&I solar deal will be larger still because of the complexity.
MR. MARTIN: What are current loan tenors? Start with senior debt and then move to back leverage.
MR. EMMONS: Loan tenors, with a few exceptions, are staying under 10 years. Banks have higher capital and liquidity costs and can be more competitive for shorter tenors, so they try to keep tenors short. Mini-perms are a common technique to do that. Much of the demand in the renewable energy sector is for tax equity bridge loans, and they are typically under one year.
MR. MARTIN: What are the current debt-service-coverage ratios for wind, utility-scale solar, rooftop solar, natural gas projects?
MR. EMMONS: Generalizing again, wind is 1.45x, and solar is 1.35x. Those are P50 numbers.
MR. MARTIN: Is that the coverage ratio for utility-scale or rooftop solar?
MR. EMMONS: There could be a premium for rooftop, but not necessarily if there is good diversification of credit risks. It really depends on the individual case. Those are P50 numbers. For P99, the coverage ratio would be 1:0x or 1.1x for deals with power hedges. For natural gas, it could be as low as 1.4x for a project with a fully-contracted revenue stream, or higher if there is more revenue risk in the equation.
MR. MARTIN: What are current advance rates on construction loans?
MR. EMMONS: A lot of construction loans are tax equity bridge loans, and those attract a 95% or even a 100% advance rate. They typically have shortfall indemnities from the sponsor in case tax equity does not cover the full loan at the end of construction. Otherwise the advance rates can be as high as 85% to 90% for strong projects.
MR. MARTIN: A tax equity bridge loan is a construction loan that is expected to be repaid out of the capital a tax equity investor contributes upon coming into the deal.
There was downward pressure on interest rates last year. The number of banks increased, which meant more banks were chasing the same number of deals. The US central bank increased the overnight federal funds rate by a quarter percent in December. Many analysts expect as many as three or four rate increases this year.
To what extent are the rates you quote correlated to the federal funds rate? Do they move up at the same time?
MR. EMMONS: It depends on whether the loan is swapped or not. Short-term loans like construction loans typically are not swapped and so, therefore, they are pegged to LIBOR, typically three-month LIBOR, and the three-month LIBOR rate moved up in step with the increase in the federal funds rate.
So for unhedged short-term debt, there is a direct effect. Long-term loans typically are swapped against a Treasury rate commensurate with the average life of the deal. Those rates are not very well correlated to the federal funds rate and, in fact, some forecasts show the 10-year swapped LIBOR rate even coming down. The good news for term project financing is that the rate increases have had, and are expected to have, a very moderate effect on the effective cost of long-term borrowing.
MR. MARTIN: Have you been asked to lend to any merchant solar projects?
MR. EMMONS: If “merchant” means a project with no PPA or hedge, then the answer is very rarely. One example outside the United States is Chile, where several merchant solar deals have closed.
A few developers who are determined to keep some price upside potential do not want to hedge and may be able to borrow at lower advance rates and higher cost in deep power markets, but they are the exception.
MR. MARTIN: What would the lower advance rate be: 40%, 50%, higher?
MR. EMMONS: It would depend on electricity price forecasts and how believable the forecasts are. The advance rate would be set by discounting cash flow against future revenues, not just as a percentage of cost.
Term Loan B
MR. MARTIN: Jean-Pierre Boudrias, what was the term loan B volume in the North American power sector in 2015, and how did that volume compare to 2014?
MR. BOUDRIAS: In 2015, there were 10 transactions for a total borrowing of $3.3 billion. That was down from 2014 when the deal volume was around $9 billion, and down from 2013 when the deal volume was around $11 billion.
The market tends to be a good place finance acquisitions. The market also has had in the past couple years a large number of refinancings. There were few of either type of transaction last year. There was also strong competition from banks to finance new projects. These factors help explain the drop in volume.
MR. MARTIN: That is almost a 50% drop from last year. It sounds like Tom Emmons is stealing your lunch.
MR. BOUDRIAS: The drop is due to a number of things. On the M&A side, acquirers have been using more corporate balance sheet finance, so that has removed some volume from the market. We saw the bulk of refinancings in 2013 and early 2014. In terms of new assets, the banks have financed the quasi-merchant gas plants to the tune of about $1.6 billion last year with only one deal in the term loan B market, which was a bank and term loan B deal for Panda Hummel, of which only $460 million was raised in the term loan B market.
MR. MARTIN: What percentage of the 2015 deals were merchant gas-fired power projects? You mentioned one.
MR. BOUDRIAS: It looks like approximately 60% of deals had a significant merchant component.
MR. MARTIN: They were all gas-fired power plants?
MR. BOUDRIAS: Yes.
MR. MARTIN: Were all of those projects in PJM or ERCOT? There was talk last year about merchant deals in New England.
MR. BOUDRIAS: The only new-build project that was financed in our market was in PJM. One project in New England was financed last year in the bank market.
MR. MARTIN: Has the market basically closed at this point to more PJM merchant gas deals? Are people feeling flush with that risk?
MR. BOUDRIAS: It requires a case-by-case determination. Obviously the investor community has a lot of exposure to certain sponsors. There may be more appetite for new sponsors at this point.
The broader theme across the market has been the continued retreat from energy stocks as oil prices fall and gas prices remain low. That has affected the fundamentals of certain power markets, ERCOT in particular. If we look at a sample of deals in ERCOT, for instance, a year ago all these loans were quoted around 99% or 99¾% of face value. The same transactions today are quoted in the low 80s — for example, 82 — so that is an increase of 500 basis points in terms of yield to worse. Loan value is moving in the same direction in PJM, but not by as much. For example, the same portfolio around the same bid level of 99¾ in PJM is down to 95, which is about 75 basis points more in yield equivalent, which is a little bit more in line with the broader market.
MR. MARTIN: Some listeners may not be familiar with a term loan B loan. What is it?
MR. BOUDRIAS: It is a loan documented largely like a bank loan, but that is placed with institutional lenders. The documents are more institutionally focused. By that, I mean that sponsors will generally have a greater degree of flexibility and freedom. Because it is harder to get a consent from investors than from a bank, the documents tend to be a little more sponsor-friendly by giving the sponsor more running room before amendments are required to the loan documents versus what you would see in the bank market.
MR. MARTIN: So it is basically the same bank debt that Tom Emmons is offering, but sold in then institutional market and perhaps a little more borrower-friendly.
MR. BOUDRIAS: The documents tend to be geared toward projects with higher risk: for example, for a merchant project or for holdco debt.
MR. MARTIN: Pricing a year ago for strong BB credits was around 350 basis points over LIBOR with a 1% LIBOR floor and 1% original issue discount, and single B credits were 500 basis points over LIBOR. Where do you see rates today?
MR. BOUDRIAS: We are probably 75 to 100 basis points higher than these levels. Some of it just reflects the broader malaise that we have seen across the leveraged finance markets overall. Some of it has been energy driven like the aversion to E&P companies that are directly affected by the downturn in oil and gas prices. It was widespread across most sectors of the market in 2015, so borrowing costs are higher than they were a year ago.
MR. MARTIN: So rates are continuing to go up. Are tenors and required coverage ratios the same as in the bank market?
MR. BOUDRIAS: Coverage has never been a good metric for term loan B debt because most term loan B’s sweep excess cash to pay debt service. As a result, people will try to understand under a variety of scenarios how certain the loan is to be repaid by the end of the term. You would expect a base case to show the loan being paid off and, in the downside cases, 50% of the principal, or perhaps less, paid off. As far as tenor goes, the market has been pretty consistent. B loans tend to have a seven-year tenor. We do not expect that to change.
MR. MARTIN: How large a transaction must one have to make it worth the trouble to do a B loan?
MR. BOUDRIAS: Anything less than $250 million is probably not worth the trouble. Obviously some slightly smaller transactions were done last year. There were two transactions that were done in the low $200 million range, but it is difficult to justify the expenses for a loan of less than $250 million.
MR. MARTIN: How long should a transaction take?
MR. BOUDRIAS: Generally speaking, three months from beginning to end. Most of it goes into producing the material required to go through rating agencies. Once the loan is in the market, things move quickly. There is usually a two-week period to closing after the rating is received.
MR. MARTIN: Jerry Hanrahan, the project bond market does not do well when the bank and term loan B markets are wide open and looking for product. We have heard the bank market is alive and well, and the term loan B was down last year.
There were no large syndicated project bond transactions in 2014. You and others did a few transactions on a direct, relationship basis. How many deals were there in 2015?
MR. HANRAHAN: There were probably a half a dozen or so deals last year in the investment-grade project bond market. There was one gas-fired deal early in the year that was well received, and the remaining deals were renewables, primarily solar, brought by corporate sponsors to the private placement or 144A market.
MR. MARTIN: How many active institutional investors were there?
MR. HANRAHAN: That is always difficult to gauge, but there is no shortage of liquidity. Everybody who can participate is participating.
There are probably somewhere around 25 institutional investors like ourselves who participate in these deals. The market is probably led by a group of eight or 10 of us who tend to be the anchor investors with larger teams and resources to bring to bear on a transaction.
MR. MARTIN: Last year at this time, you said there was one deal in the pipeline. How many are there in the pipeline today?
MR. HANRAHAN: Not many. It is hard to get visibility at this time of year as to what is coming, but there is one deal about which we are aware that went out as an RFP in the fourth quarter last year with the sponsor indicating that it prefers a bond structure. I expect that deal, if it ends up as a bond structure, to come to market in the first quarter this year.
Beyond that, we do not have a lot of visibility. Given what has just been said about the bank and term loan B markets, we will probably see a similar year this year to what we did last year.
MR. MARTIN: Project bonds are fixed-rate debt, and they tend to be longer term than bank loans and term loan B debt. Both of those products are floating rate debt. Project bonds need a spark, like a fear of rising interest rates, before that market gets traction. Are we at such a point today?
MR. HANRAHAN: It is possible. It all depends on one’s view of inflation and how much interest rates might increase. The advantages that we can offer are the longer tenor and the fixed-rate nature, and there tend to be little or no fees associated with bonds. So those are the pluses that we can use to attract borrowers.
If we are in a rising rate environment, then bonds will look more attractive. We do well then. We also do well when there is some dislocation in the market that people can arbitrage against; for example, if there is a larger spread than normal between the LIBOR swap spreads used to price bank debt and Treasuries.
MR. MARTIN: What is the current spread above treasuries, and what does that translate into as a coupon rate?
MR. HANRAHAN: It is a range like everyone else has answered. A typical investment-grade project could expect to pay a rate in the mid-200 basis point level above average life Treasuries, plus or minus depending upon the particular features and quality of the deal. It could be anywhere from the low 200s to high 200s or even 300 basis points above average-life Treasuries. We tend to do longer-term deals, so we are usually pricing off something in the area of a 10-year Treasury, which today is around 215 basis points, so you end up with coupons in the low-to-mid 4%, maybe 4.5%, range.
MR. MARTIN: That is not much different than the current rate on bank debt.
MR. HANRAHAN: Not much.
MR. MARTIN: Shouldn’t this market start to get traction when people think we are at the bottom of the interest rate cycle; it is a chance to lock in an historically low rate.
MR. HANRAHAN: Right.
MR. MARTIN: Another key difference between project bonds and term loan B debt is project bonds generally have the same tenor as the power purchase agreement, correct?
MR. HANRAHAN: Yes.
MR. MARTIN: There is no upfront fee like these other two products because the economics are fully baked into the spread. When are ratings required?
MR. HANRAHAN: Ratings are generally not required, at least by us and the other large insurance companies. Sponsors tend to get ratings if they are worried about execution or they want to attract some of the smaller players who are more comfortable in rated deals, or if they opt to go the 144A route, in which case ratings would be required.
We don’t require them ourselves, but we will structure a deal to a BBB investment-grade rating internally.
MR. MARTIN: Of course, the big difference between bank and term loan B debt and project bonds is project bonds require a make-whole payment if the bonds are repaid ahead of schedule. How is such a payment calculated?
MR. HANRAHAN: It is calculated off a spread to Treasuries when the bond is repaid. It is less likely in a rising rate environment that there will be a make-whole payment if rates are higher on the early payment date than at time of original issuance. You probably have no incentive to refinance in such a case in any event. You locked in a lower rate than you can get by refinancing.
MR. MARTIN: How long does it take to do a project bond deal?
MR. HANRAHAN: If the deal comes to us in the syndicated private placement market, then you are usually talking two to four weeks. A direct-placement deal requires a couple months to complete the due diligence and documentation.
MR. MARTIN: How large a transaction does one need to make it worthwhile?
MR. HANRAHAN: For the direct placement deals that we do, the transaction size is usually $30 to $50 million. A syndicated deal should be $100 million.