Improving Power Purchase Agreements

Some initial thoughts on how we might improve markets for renewable Power Purchase Agreements (PPAs).

While I’ve worked in the energy industry for many years, I’ve only recently started looking at PPAs, which serve a valuable role in increasing renewable energy supply. I’m planning on having a few conversations, so have written this piece as a way of gathering my thoughts. Needless to say, I’m confident my thoughts will evolve as I read, think and talk to more people about it, so please don’t take these thoughts as concrete conclusions.

I am mostly focussing on the market in Great Britain, although I’m sure most of this would apply to other European markets (and likely to other markets too).

I must say to begin that I am particularly indebted to two reports by the World Business Council for Sustainable Development (WBCSD). They’re both well worth a read if you’re interested in this area:

• Corporate Renewable Power Purchase Agreements: Scaling up globally (October 2016)
• Innovation in Power Purchase Agreement Structures (March 2018)

Introduction

Increasing wind and solar generation capacity is likely to part of the most cost effective solutions to reducing emissions from our growing electricity system. In order to get financing, developments need reliable revenue streams.

However, there is considerable uncertainty around how much electricity the solar and wind generators will produce in each individual half hour. As it happens, lenders aren’t so worried about this risk itself. To the extent that it is due to meteorological uncertainty, we’ve got historical data that gives us a good sense of how much it might vary, and it will largely average out over time. And to the extent it is due to operational issues at the particular generator, lenders can diversify, or the generator can get insurance.

The bigger problem with not knowing how much will be produced each half hour is that it makes it difficult to lock in a price. The expected price for each half hour is likely to move a lot in the time before the production is likely to be reliably known. And, to make matters worse, the scenarios where the production is higher than expected tend to be the scenarios where the price is lower than expected.

One way of helping wind and solar generators get financing has been for governments to give them a fixed price per MWh for their total production. The generator is still left with the volume risk, but as mentioned, lenders seem to be alright with that. However, this puts a lot of risk in the hands of governments, at least some of which would be better placed in the hands of electricity consumers or insurers. To be clear, I’d rather see governments offering these contracts, than the wind and solar generators not being built. I just think there are better ways of doing it. And anyway, a lot of governments appear to be getting less willing to offer fixed price contracts.

The other common approach has historically been the power purchase agreement (PPA). In these, a corporate electricity consumer or an electricity retailer agrees to pay the generator a fixed rate for a certain quantity of electricity each year, for a number of years. Unfortunately for the generator, not many companies are willing to buy a completely unknown quantity of electricity for a fixed price. As a result, the most common PPAs usually agree to sell a fixed quantity each year of the contract. The fixed quantity has to be set in a way that minimises the risk that the generator won’t produce it over a year, but also minimises the risk that the purchaser receives more electricity than they need in any given half hour. Unfortunately, that constraint is extremely limiting, both in terms of how much of a generator’s capacity it can sell under PPAs, and in terms of how much of a purchasing company’s demand it can buy under PPAs.

A further limiting factor for these PPAs is that many purchasing companies may not want to fix a price for too many years of consumption: a typical contract of 10 years is both a long time for the purchasing company, but short compared with the generator’s loan-term. In some cases, even where a purchasing company wants to commit to such a contract length, this will pose too much credit risk. For example, if the purchasing company goes out of business 5 years into the PPA, the generator will have to sell that electricity, potentially at considerably lower prices.

I am therefore proposing an alternative way of structuring PPAs, that I believe will allow a greater volume of solar and wind generation capacity to become viable.

Before I do, I want to provide a more technical description of the existing PPAs, so it will be easier to see how my proposal differs. I define a few variables:

• GQ is the quantity generated in the given half hour
• PQ is the quantity consumed by the purchaser
• CQ is the quantity supplied under the PPA
• f is the fixed price (in £/MWh)
• p is the market price (in £/MWh) for the given half hour, for example, that from the Great Britain imbalance market. (Note that distribution and transmission costs are typically passed on to the purchaser separately, via a utility — I have therefore ignored them in this discussion).

From the generator perspective:

• Generator receives GQ x p from the market
• Generator receives CQ x (f-p) from the purchaser

From the purchaser perspective:

• Purchaser pays PQ x p to the market
• Purchaser pays CQ x (f-p) to the generator

Both the generator and the purchasing company are therefore likely to be left with significant exposure to market price p, as a result of the fact that GQ > CQ > PQ.

(Note that I’ve assumed here one generator selling to one purchaser — in fact a generator could sell to many purchasers, or a purchaser could buy from a number of generators.)

An alternative approach

My alternative approach is based on the recognition that the quantity of electricity from a particular generator has a component of uncertainty that is the same for all generators of that type in that market (eg England solar) and a component that is generator specific. (This is similar to the stock market concepts of beta and alpha). By combining them, we make them harder to manage.

As a result, I propose creating an index quantity IQ, which will be the average generation per MW of solar capacity in England (we’d create other indices for wind, and for other regions). Then, the risk that the generated quantity GQ was greater or less than this could be insured, while the PPA would be based on the index quantity.

So, from the generator perspective, we have:

• Generator receives GQ x p from the market
• Generator receives (IQ-GQ) x p from an insurer
• Generator receives IQ x (f-p) from the the PPA purchaser

And from the purchaser perspective, we have:

• Purchaser pays PQ x p to the market
• Purchaser pays IQ x (f-p) to the generator

The generator will now have a fixed price for all their generation. It should even be possible to hedge IQ x f in the market, guaranteeing the ability to service debt.

The purchaser will still have exposure to the market price p, where their consumption PQ differs from IQ, however this will be easier to manage as it will be less dependent on a specific generator. I will discuss this further below.

Term risk

As I mentioned above, generators will often want to sell for much longer terms than purchasers want to (or are able to) commit. For example, a generator with a 20 year debt repayment term, may struggle to find a purchaser willing to buy for more than 10 of those years. In this case, I believe it should be possible to find another market participant prepared to pay IQ x (f-p) for the final 10 years. For example, I (and plenty of other people like me) know that I am likely to be purchasing electricity in England in 10 years time — I could buy the product, knowing that the case in which I lost on the product was the case in which I was paying less in my electricity. And in 10years time, I could always sell it to companies ready to lock in their next 10 years of consumption. Alternatively, the government may be able to hold the later years of exposure, releasing them to the market as they get closer to term.

Credit risk

There is always a risk that the purchaser of the PPA may default on the agreement, which will cause the generator to lose money if the expectation of p is less than f. To reduce this credit risk, the PPA purchaser could pay some money up front, reducing (perhaps even to zero) the fixed price f. In this case, it would be similar to the purchasing company paying up front to install solar panels, in return for free electricity later.

Volume risk for the purchaser

I mentioned above that where the purchasing company’s consumption PQ differs from the index quantity IQ, the purchasing company will have exposure to the market price p. This could be where the company consumes more than IQ, in which case they would need to buy from the market, or where the company consumes less so effectively sells the excess.

The purchasing company may have an agreement where their utility agrees to fix the price of the imbalance. I’m not sure how great a deal this would be (I would be charging quite a bit to supply when there was no solar/wind, and to take when there was excess). However, showing a willingness to pay for this level of responsiveness might be what was needed to support increased investment in batteries.

Alternatively, the purchasing company may choose not to hedge their volume risk, but to use the fact that they are only hedged up to the level of IQ, as an incentive to reduce consumption when there is less renewable generation and prices are high. I see this as a good thing, as it goes some way towards reducing the need for less-green resources for balancing the grid.

Legal/contractual risk

I want to acknowledge that my proposal creates some additional legal/contractual risk — that it is a long term arrangement, reliant on an Index Quantity IQ. There is a risk that in 10 years time, whoever was responsible for calculating IQ might stop doing so, or that rival calculations might emerge. In many ways, these risks are similar to those held by any contract that indexes to inflation. I believe they need to be taken into account, but I don’t see them as being insurmountable.

I hope this has provided some food for thought — if anyone is interested in discussing further, please do get in touch.