Yield Cos - State Of Play: A roundtable discussion about yield cos at the Infocast projects & money conference in New Orleans in January

Six yield cos have been formed to date by large renewable energy developers in the United States. Many people are curious about how much yield cos reduce the cost of capital for the development companies that have formed them, what discount rates they use to acquire assets, how much room there is for additional yield cos, what it takes to be able to form one, and similar questions.

The following is an edited transcript from a well-attended roundtable discussion about yield cos at the Infocast projects & money conference in New Orleans in January. The panelists are Andy Redinger, managing director and group head of utilities, power and renewable energy at Keybanc Capital Markets and who contributed to the early buzz about yield cos by being one of the first investment bankers to write about them, Carl Weatherley-White, president of LightBeam Electric Company, former head of project finance at Lehman Brothers and Barclays Capital and another early proponent of yield cos, Alejandro Burgaleta, chief financial officer of Gestamp Wind, a Spanish wind developer with a global footprint that is considering forming a yield co, Hunter Armistead, executive vice president of Pattern Energy, a prominent US wind company that has reorganized itself as a yield co, and David Mcllhenny managing director of project finance of SunPower Corporation, a prominent solar developer that has been debating whether to form a yield co. The moderator is Keith Martin with Chadbourne in Washington.

MR. MARTIN: David Mcllhenny, what is a yield co?

MR. MCILHENNY: A yield co is a synthetic master limited partnership with equity investors who are looking for regular dividends and expect the dividend to grow significantly over time. One way to maximize how much cash flow is available for dividend payments is to minimize how much taxes the yield co pays by having the yield co hold projects that throw off as much in tax shelter as the taxable income they generate.

MR. MARTIN: That is a fairly sophisticated answer. Carl Weatherley-White, do you want to add to it?

MR. WEATHERLEY-WHITE: A yield co is a company with stable operating cash flows from long-term contracted assets, that expects dividends per share to grow over time and that taps into the public equity market to raise capital at a high multiple to earnings.

MR. MARTIN: Let me offer another definition. A yield co is a simple concept. A development company separates its operating projects from its development pipeline. It puts the operating projects in a separate corporation that lists on a stock exchange and is able to raise capital more cheaply because its projects are de-risked; they have operating histories.

Hunter Armistead, you take issue with these definitions. What is the problem with what all three of us said?

MR. ARMISTEAD: Anything that starts with the word synthetic makes me nervous. There are different flavors of yield cos. Is a yield co a separate company or is it a financing vehicle? Each of the yield cos that has been formed to date has had a strong sponsor standing behind the yield co with a healthy development pipeline that the yield co can buy as a path to future growth. However, Pattern is a different flavor of yield co than NRG Yield is. I do not believe all yield cos are the same.

MR. MARTIN: You asked whether a yield co is a company or a form of financing. By asking this, you are implying, I think, that there is not much more room in the market for yield cos that are merely roll ups of assets acquired from third parties; you need a real developer with a big pipeline of development assets to support a yield co. Is that what you meant to say?

MR. ARMISTEAD: Is this a negotiation? [Laughter.] Good lawyers put the question in a way that comes close to answering it. I think that the investor base that has been attracted to the existing yield cos has put a premium on strong sponsorship, a strong pipeline that allows for growth, stability and a track record running the operations side of the business. If those variables are there, then I think there is more room in the market. Maybe someone who cannot check off all these boxes will find another group of investors that is looking for something different.

MR. MARTIN: Andy Redinger, is there much room in the market for more yield cos?

MR. REDINGER: Absolutely. There is a lot of room. I think a yield co is a real estate company. It is an opportunity to invest in an asset class that looks a lot like real estate with characteristics that are much better than what is in existing REITs today. There is a huge volume of potential yield cos coming to market in 2015. Look at the REIT market. REITs have been around for more than 25 years and have a current market capitalization of more than $400 billion. I challenge anybody to compare REIT assets to yield co assets, and I will tell you all day long that yield cos are a better asset class.

MR. MARTIN: So you see a lot of pent-up demand for yield cos in 2015. How many more do you foresee, and will the next round of yield cos be roll ups of assets from third parties as opposed to the Pattern or NRG Yield model?

MR. REDINGER: Hunter Armistead is absolutely right. We will see a different type of yield co in 2015. The yield cos we have seen to date have all had strong sponsors backing them. It will be interesting to see whether the market differentiates between sponsor-backed yield cos versus the two of us getting together, going out and buying assets and saying we are a yield co.

MR. MARTIN: Can we do that?

MR. REDINGER: We absolutely can.

MR. MARTIN: So roll ups work.

MR. REDINGER: Well, I am not calling it a roll up. Otherwise, you would have to call the entire REIT industry a roll up. Yield cos are similar to what is already done in the REIT space.

MR. MARTIN: I did not hear a number of how many more yield cos will come to market in 2015. Can you give us a number?

MR. REDINGER: It is more than a dozen.

MR. MARTIN: Carl Weatherley-White, do roll ups work?

MR. WEATHERLEY-WHITE: Absolutely.

MR. MARTIN: Isn’t future growth less certain for a roll up? Yield cos are like vacuum cleaners. With so many yield cos competing for assets, eventually you run out of things to vacuum up.

MR. WEATHERLEY-WHITE: That’s a great question. I think the emergence of more yield cos is not only possible, but really required for this market to develop. Investors are looking for more opportunities to participate in the space. Depending on how you define “yield co,” there are five or six companies today. The market capitalization of the existing yield cos is about $12 billion. The market capitalization of the existing master limited partnerships is about $600 billion.

More than 50 times as much capital is looking for opportunities like yield cos than is invested in the existing yield cos.

The question is whether there are enough assets to support that type of growth, and I think there are. Many new energy and other infrastructure projects are being built worldwide.

Necessary Predicates

MR. MARTIN: Andy Redinger, what are the necessary predicates before you can have a yield co? For example, how much asset volume do you need to start?

MR. REDINGER: Two years ago, you needed at least $1 billion in equity value. Today, $500 million is probably the minimum with at least $30 million a year in distributable cash flow. Can you do something smaller? Sure, but I do not think it will trade as well. You will see all kinds of yield cos coming to market this year. I think you need a development pipeline. You need to show an ability to grow, and you need to show some geographic diversity.

MR. MARTIN: So the yield co must have at least $500 million asset value. How much capital must it come to market hoping to raise against that much asset value? At least $100 million?

MR. REDINGER: You usually see 30% to 40% of the company sold in the initial public offering, so that would be $150 to $200 million as the initial capital raise. The objective is to take as little public as possible out of the box, but enough to ensure good execution. This will let the sponsor benefit from any upward price appreciation on its remaining ownership percentage after the IPO. You come to market as small as the market will allow and leave a lot in the development pipeline to show the potential for growth.

MR. MARTIN: Hunter Armistead, you have been through the process. Do those numbers sound right?

MR. ARMISTEAD: They are the right numbers for when Pattern went to market. One of the real challenges today is asset valuations have stepped up. Delivering growth is not just a matter of adding new assets, but it is also growing dividends per share. Depending on the yield co’s cost of capital, that could be hard to do, given the sky-high asset valuations, unless the yield co has a strong development pipeline from an affiliated sponsor.

MR. MARTIN: This is where I was trying to pin you down before! So a yield co needs a sponsor with a healthy development pipeline?

MR. ARMISTEAD: There are two different kinds of yield cos.

MR. MARTIN: Let’s move to cash flow. Some assets put into yield cos are subject to debt or tax equity. What percentage of the operating cash flow should be available for distribution to shareholders, or put differently, what percentage of it can be used for debt service or be distributed to tax equity? 65%? Less?

MR. REDINGER: Roughly a third of the revenue should be available for distribution. However, the trend today is to de-lever the assets going into yield cos or to have an amortization profile that is better suited for a yield co, for example, by having less amortization early on and maybe more later or there may be no amortization.

MR. MARTIN: So you front load the cash available for distribution?

MR. REDINGER: Yes. Yield co investors tend to be more focused on the short term. Yield cos tend to be valued based on next year’s cash flows.

MR. MARTIN: Why would anyone be fooled by that? [Laughter.]

MR. WEATHERLEY-WHITE: It is important for management teams and bankers involved to have a strong handle on the cash flow projections. To miss a dividend and to fail to grow the dividend will lead to a failed company. Therefore, to value something based on first-year cash flow that trails off is setting the company up for a huge problem. The fancy vehicles, as Andy said, are designed to produce steady cash flows, less leverage, tax equity with less dramatic flips: those sort of things.

MR. MARTIN: Maybe one lesson for anyone doing tax equity is not to give 99% of the cash to the tax equity investor, but to work out a different sharing ratio, and not to have cash flow sweeps to pay indemnities. What cash sharing ratio should a sponsor aspire to if he wants to preserve the option to move a project later into a yield co?

MR. WEATHERLEY-WHITE: Just try for a fixed sharing ratio that leaves a significant amount of cash each year for the sponsor.

MR. MARTIN: I hope John Eber [head of tax equity investments for JPMorgan Capital Corporation], who is sitting in front of me, is listening to this.

MR. ARMISTEAD: John stuck me on his panel, so maybe this is my opportunity. It is an unnatural state when the tax equity is earning a higher return than the sponsors. [Laughter.]

MR. MARTIN: John asks, “What wrong with that?” [Laughter.]

MR. ARMISTEAD: The wind business has reached a stage of maturity where performance is much more reliable. This has allowed the cash sharing ratios in tax equity structures to move today to a better split that makes projects more suitable for yield cos.

MR. BURGALETA: You must send a message to your investors that you will not fool around with cash flow, and you have to be very careful about what you announce because you will have to keep to that every quarter.

MR. MARTIN: So you must be careful about the dividend you announce because you create an expectation.

How much should one expect to spend to put a yield co in place, and how long does the process take?

MR. ARMISTEAD: This is an interesting thing. We have a banker who loves every yield co.

MR. MARTIN: Andy Redinger, that is smack talk. Are you going to let that go unanswered? [Laughter.]

MR. REDINGER: The question is how costly, how hard and what does it mean when you get there?

MR. MARTIN: Give me some numbers.

MR. REDINGER: I don’t like to share them because somehow they will end up all going to you, Keith, as a lawyer.... [Laughter and talking over each other.] Four to $5 million in costs and nine months to a year to complete the offering.

MR. MARTIN:... which is not bad. Two data storage companies that converted recently to REITs said it cost them as much as $145 to $155 million in conversion costs.

MR. ARMISTEAD: I don’t think they included underwriter fees in that. Actually, I am certain of that. [Laughter and talking over each other.] Andy does it pro bono

Cost of Capital

MR. MARTIN: Yield cos pay low dividends on the order of 3%, 4%, 6% and yet most of the shares are held by institutional investors who have other places they can put their money that earn returns in the teens. The shareholders are looking for growth on top of dividends. What is the true cost of capital to yield cos in the current market?

MR. MCILHENNY: A rule of the thumb is that if you can project 15% dividend growth, then you can pay a current dividend yield of 3%.

MR. MARTIN: To what does that translate as a cost of capital?

MR. MCILHENNY: You can figure it out mathematically. The perception of growth is key to value. That perception can lower the cost of capital for the initial public offering. A high growth rate can lower the cost of capital for drop-down assets. Higher than expected growth can increase the value of the sponsor’s incentive distribution rights.

Growth is key, and that is a different metric for financing through a yield co than the project finance market traditionally looks at things. Project financiers look at the long-term cash flow and tax benefits and come up with an internal rate of return that they need on their investment. In contrast, a yield co investor looks for a current dividend yield, some years of comfort that the yield will be maintained, and a dream that it will be there for a long, long time. The quality of the yield co assets affects whether the dream will come true.

The bottom line is it is hard to say what the cost of capital is for a yield co. The cost is affected by a number of factors.

MR. MARTIN: Is it as simple as to add a 3% dividend yield to a 15% growth expectation, which equals 18%?

MR. MCILHENNY: No.

MR. MARTIN: So you solve for the internal rate of return that sets the present value of the dividend stream, starting at 3% and growing at 15% a year, equal to your investment.

Hunter Armistead, is a yield co a form of financing fora development company?

MR. ARMISTEAD: It is an interesting question that we faced when we were evaluating the right vehicle. Keith told us to keep the answers short and controversial. That was his only guidance, so I need to shorten it up. [Laughter.]

MR. MARTIN: We also need the controversial.

MR. ARMISTEAD: I am trying to avoid that. It is cost-effective capital. The cost is not stratospheric compared to what we would be looking for from a pension fund. The numbers that we saw when we valued our assets in anticipation of forming a yield co were not that different than when we ultimately did our trade on the public exchange. The reason for going public had nothing to do with monetization. It was funding our growth. If you are just buying current short-term dividend growth or capital yield, with a low IRR, then that will turn on you in the future. You have to keep an eye on both.

MR. MARTIN: Carl Weatherley-White, I was looking at the dividend yields for three yield cos. They are 4.87% for Pattern, 3.6% for Terra Form, 2.91% for NRG Yield. What accounts for the differences in dividend yields? Is it just growth expectations?

MR. WEATHERLEY-WHITE: The market is evolving. There is a differentiation among business models. Investors assign different values to different models. This is a natural evolution. As more companies come to the market, there will be more differentiation and ways for sponsors to play it.

MR. MARTIN: David Mcllhenny, you are nodding.

MR. MCILHENNY: I was just going to ask Carl a question. Why is NextEra’s yield so low compared to Pattern’s yield?

MR. WEATHERLEY-WHITE: NextEra has an enormous inventory of assets that it owns and operates, so investors can be more confident about the potential for growth. Pattern has an outstanding development team and a proven track record, and it will grow as well. You are seeing differentiation in yield, which I think is a function of growth expectation.

Growth

MR. MARTIN: Fair enough. Let’s focus on how yield cos grow. They distribute almost all their operating net cash flow. So in order to have cash to make acquisitions, they must either borrow or raise more equity, thereby diluting the existing shareholders. Why isn’t this a zero sum game? How does one get growth out of such a process? Alex Burgaleta.

MR. BURGALETA: The sponsor must drop down or contribute assets in an accretive way for the existing shareholders.

MR. MARTIN: What does it mean to contribute assets in an accretive way?

MR. BURGALETA: When you dilute the existing shareholders by issuing new shares, the amount of cash that the shareholders will be distributed per share is higher than before.

MR. MARTIN: Are there any other ways of expressing this?

MR. WEATHERLEY-WHITE: It is the same model that has existed in the REIT and MLP sectors for many years. Alex said it well. There must be a difference between the value at which assets are brought into the yield co and the cost of capital.

MR. MARTIN: You basically raise money at a higher price-to-earnings ratio and reinvest at a lower one. What are current spreads in the yield co market? Are they narrowing or widening? Andy Redinger.

MR. REDINGER: The spreads are definitely narrowing as the demand for assets heats up. We are seeing other players coming into the market and competing with yield cos for assets. Pension funds and infrastructure funds are two examples of other sources of capital that have become more aggressive recently.

MR. MARTIN: Spreads are tightening. Does that suggest a shortage of product? Hunter Armistead.

MR. ARMISTEAD: I think you are getting to one of the core challenges if your growth model is acquisitions. As the spread narrows, it becomes harder to do more acquisitions on an economic basis. If you can’t do any more deals, then you can’t grow. That is where it helps to have a deep development pipeline to have a captive group of assets that provides that accretive growth as opposed to having to rely on the spread.

MR. MARTIN: Hunter Armistead, your company has said that the market has become too frothy. Asset values are inflated —you are wincing as I say this — so you will focus on your own pipeline probably to the exclusion of buying assets from third parties.

MR. ARMISTEAD: We continue to evaluate third-party acquisitions, but at the core, we maintain a deep, strong development pipeline that can feed the growth. To the extent we can augment that with acquisitions from third parties, we will do it, but we don’t feel like we have to do it.

MR. REDINGER: These ebbs and flows are natural in the marketplace. The tightening will force yield cos to look other places for assets: international, residential solar. What ebbs today will flow again later. Smart companies are already thinking about this. You will see a broader mix of assets put into yield cos.

MR. WEATHERLEY-WHITE: That’s a great point. There is a natural evolution. I am also interested in what Hunter Armistead said because he is a developer. I see this as a search for a lower cost of capital that can then be applied to the ownership of assets, which then results in a more competitive electricity price. When we talk to developers about the opportunities of yield cos, they are working backwards from what electricity price works.

MR. ARMISTEAD: Hit me with the question again.

MR. WEATHERLEY-WHITE: When you think about what returns are necessary to support a power contract bid, you think about your cost of capital and do you consider yield co capital as giving you a competitive advantage?

MR. MARTIN: You are setting him up, because you are asking Hunter to admit that his yield co is a form of a financing. [Laughter.]

MR. ARMISTEAD: Thank you! [Laughter.]

MR. MARTIN: I’m a lawyer.

MR. ARMISTEAD: I think the answer to your question is yes. We were originally funded by Riverstone when Pattern was first formed. That’s a private equity fund. The returns we were searching for, which included some yield compression, were materially higher than what we see would deliver very solid growth to our public vehicle. So, yes, the yield co allows us to be more aggressive when bidding for power contracts.

MR. MARTIN: Alex Burgaleta, do you want to add to that?

MR. BURGALETA: We have to look at the cost of capital of yield cos as something that is not as steady. The cost depends on the rate at which the yield co will acquire assets and the market prices at which it will do so. You cannot look at the yield co on a stand-alone basis; it is a more dynamic process.

MR. MARTIN: Your cost of capital may be harder to predict.

MR. BURGALETA: Yes. A yield co has a window in time when it may be most efficient and aggressive. But you have to be careful because you have to deliver the growth, and then we will discuss how you do so. Do you already have the assets? Do you develop the assets? Do you buy the assets? How much value are you bringing to the shareholders by contributing those assets? These different paths to growth are not equal.

Discount Rates

MR. MARTIN: There are many strategics that felt, when yield cos first appeared, that yield cos would have the lowest cost of capital and would win all the auctions for assets. Has that in fact happened?

MR. ARMISTEAD: As a loser of a bunch of options, no. There have been some big transactions that were transformative for a couple of the other companies that pay dividends like us; you can see I am moving away from the term “yield co.” In our experience, the most aggressive bidders are buyers who want to form a new yield co and are trying to get to a critical mass of assets. We have seen bidders pay materially more than we thought justified. We feel like we have a good growth engine and do not have to stretch to grow.

MR. MARTIN: What discount rate do you think yield cos are using to win assets in the current market?

MR. REDINGER: I think 9% to 10% levered would get you right in the middle of the bell curve.

MR. MARTIN: What about unlevered?

MR. REDINGER: Three hundred basis points below that, so somewhere mid-6% to 7% unlevered may win you the deal, or at least get you into the second round.

MR. MARTIN: I would have thought 8% was enough to get you in.

MR. REDINGER: That is the tightening that we talked about earlier.

MR. ARMISTEAD: One of my favorite things is that it is not just about the return, it is also about the assumptions. What will it really cost to operate an asset? How well will it perform? Maybe you will not hit the flip points on schedule in the tax equity partnership and that changes the entire game and the internal rate of return. You may think you bought at 9% but you did not get it at 9%.

MR. MARTIN: I didn’t follow that.

MR. ARMISTEAD: The banker for the seller sends you a pro forma that was prepared by the sponsor. If you bid based on the pro forma numbers, sure you bid at 9%, but when you actually realize that production is different, the operating costs are higher, and you have to add more workers at the local site, things get a lot more tight.

MR. BURGALETA: You also have to consider the profile of the cash flow as it changes dramatically from one market to the next. We are looking at projects in Brazil, South Africa and other places with high inflation rates, where cash flows are growing by 6% or 7% a year. Compare that with markets like the US and Canada with really flat cash flows or cash flows that decrease over time.

MR. MARTIN: Yield cos are usually built around a renewables base, but does it work to have solely renewables? Don’t you need some tax capacity in the yield co itself? Andy Redinger, you are shaking your head no.

MR. REDINGER: No, because yield cos are vehicles that need tax credits and depreciation to shelter the income going forward. It is true that the yield co is not using the tax benefits efficiently, but the ability to roll the tax benefits forward and use them to shelter future income is what gives yield cos access to cheaper equity capital.

MR. MARTIN: If the yield co has a tax base, doesn’t it have a valuable asset that can be used to earn an additional return in the renewable energy market?

MR. REDINGER: That is another story entirely. One of the things we are always tracking closely is what is our carryforward. Yield cos like to be able to project enough tax shelter far enough into the future so that they do not have to worry about taxes reducing the cash available for distribution.

MR. MARTIN: You like to be able to say to the investors that you expect to operate tax-free for nine or 10 years based on the existing asset portfolio.

MR. REDINGER: Correct. At the same time, we are trying to figure out how to make more efficient use of tax equity.

Sensible to Form?

MR. MARTIN: David Mcllhenny, what are some pros and cons that a company that is already in business thinks about when deciding whether to form a yield co?

MR. MCILHENNY: On the pro side, the company should be able to reduce its cost of capital. A yield co is a form of financing future cash flows. It is also a way for a developer to retain long-term exposure to the value of the project rather than just selling it off, or doing some other type of financing where the upside is given to some other party. The long-term value is captured through retaining incentive distribution rights.

The cons are the yield co will divert management attention. It will make you think more like a finance company and less like a solar company or developer.

MR. MARTIN: Alex Burgaleta, do you want to add to that?

MR. BURGALETA: Yield cos are good examples of sensible risk and return allocation. Forming a yield co forces you to change the way you handle your business because the business is different. We have development companies that are also listed and are feeding yield cos. That changes how we think about the development business. If yield cos do a good job maintaining asset quality and see their dividend payments grow, then we will see even more yield cos as they offer something of value: access to cheaper capital.

MR. MARTIN: David Mcllhenny, some people say that yield cos do not work for the solar rooftop business because there is so little development risk. You are not getting much of a pop in moving to a yield co. Make sense?

MR. MCILHENNY: Residential assets have not been part of yield cos so far because yield cos are evolving, and they started off with the best assets, which are big utility-scale projects with long-term power contracts with creditworthy offtakers and little technology risk. As the yield co market has evolved, every new yield co has done something different: incentive distribution rights, distributed generation, projects that actually are not in service but count as IPO projects, foreign assets. Residential solar is an asset class that can be part of yield co, but it just has not been done yet.

Other Issues

MR. MARTIN: As an investor, are you better off investing in the development company or in the yield co?

MR. BURGALETA: What kind of car do you have? Do you have a Prius or a Corvette? [Laughter.]

MR. MARTIN: I’m the one asking the questions here! You don’t think it matters?

MR. ARMISTEAD: When you invest in a developer, you can lose everything or the investment can pay big dividends. A yield co should be a less risky investment.

MR. MARTIN: What happens when the market shifts and developers find a cheaper way to raise capital or they find other buyers willing to pay more for assets? What happens to the yield co at that point? Andy Redinger.

MR. REDINGER: They all get bought out or merged.

MR. MARTIN: Hunter Armistead, you were one of the early adopters of yield cos. What issues have come up in the first year and half of operation that you perhaps did not foresee?

MR. ARMISTEAD: We had a lot to learn. We have successfully managed and developed assets, but we have never managed a public company. Dealing with all of the reporting requirements and setting up the infrastructure so you can close your financials; if you are a private company and are a few days late on your financials, you call Riverstone and say, “We are a few days late.” There is no leeway with a yield co.

We went into the yield co with our eyes open and we assumed it was going to be tough, but it was tougher. Then there is the pressure to keep growing at the same time.

Another thing is Andy Redinger said it is $5 to $7 million dollars to stand up a yield co. It was more for us. Maybe we were not efficient at it the first time, but it is expensive to do. The infrastructure to support it is significant. If you start going into new jurisdictions, you have a lot more rules to master, like the need to close the books at the right time. It sounds really cool to do a deal in Chile, but you end up creating more stress at every turn with additional regulatory requirements. We have had to hire more people than we thought we were going to have to hire.

MR. MARTIN: That is a good bridge to my next question. What complications are created when the yield cos run out of assets to vacuum up in the US and start looking more widely overseas? You mentioned one. Does anything else come to mind? Andy Redinger.

MR. REDINGER: Currency risk. Obviously, there are also repatriation risk and the tax consequences of bringing the money back into the US.

MR. MARTIN: Alex Burgaleta, you deal with this all the time as your company operates in multiple countries.

MR. BURGALETA: I was going to say, if you are worried about Chile, wait for Brazil. [Laughter.]

MR. MARTIN: Energy Capital Partners packaged a group of gas-fired power plants and put them in a public vehicle. Compare that play to a yield co.

MR. MCILHENNY: Fossil fuel plants have fewer tax benefits, so the public company will be paying taxes, and there will be less cash for dividends.

MR. MARTIN: Where do you see the growth in that sort of play?

MR. MCILHENNY: Buying up old assets, I would think.

MR. WEATHERLEY-WHITE: But will those be contracted? That is a really interesting point because not all assets meet what you are trying to do here. Five- and 10-year contracts are tricky.