This is a question I’ve been asked a number of times this week, since the results of the UK’s latest offshore wind auctions were published (on Sept. 20th 2019).
If you’ve missed all the excitement, 15-year CfDs (‘Contracts-for-difference’) were awarded to a total of ten offshore wind projects, to start operating in four to five years’ time, at prices from £39.7 to £41.6 per MWh.
Given that just two years ago offshore wind auctions were awarding contracts at £57.5 per MWh, that’s quite a spectacular reduction.
Here are a couple of headlines following the announcement of the auction results:
Let’s read that last sub-heading at face value.
The implication is that offshore wind projects can now make a business case selling at prices below the market price.
So why don’t we see developers building ‘merchant’ projects instead, ones where they don’t bother competing for CfDs at all, and just sell into the electricity market? Surely they’d be selling at a price higher than would those projects holding CfDs?
But if that’s the case, then why - in one of my favourite energy quotes of all time - did Keith Anderson, CEO of Scottish Power (one of the UK’s ‘big six’ electricity firms) tell a conference earlier this year:
“If you think we would build a £2.5bn offshore windfarm at market risk then you are bonkers, completely bonkers.”
Perhaps we need to delve behind the headlines and understand exactly what ‘subsidy-free’ and 'market prices' mean…
Reference, Strike & Market Prices
In an article about the latest offshore wind auctions, here, Carbon Brief write that the CfD 'strike' prices…
“…are some £8-9/MWh below the government’s ‘reference price’, the level it expects to see for electricity on the open market in each year. If the market follows the government’s reference price expectations, then the new renewable schemes will pay nearly £600m towards consumer bills by 2027, instead of receiving a subsidy.”
This statement points out that not only is there the possibility that these projects will receive no subsidy, they may even pay money back towards consumers! Yay!
To understand this, we need a brief explanation of how a CfD contract works…
In a CfD auction, the winners receive a contract with a fixed “strike price”: the top horizontal line in the above diagram (so a strike price of about £70/MWh in the picture). Once the project is operating, a flow of payments depends on how this fixed strike price compares to the market price at the time electricity is generated.
If the market price is lower than the strike price, the generator earns the market price, plus is paid (as a top-up subsidy from the government, funded by consumers) the difference between the low market price and their higher strike price. However if the market price is higher than the strike price, this difference is paid back by the generator (to the government/consumers).
So let’s say one of our newly-auctioned offshore wind farms, with a CfD strike price of around £40/MWh, puts electricity into the grid at 4am one day, when the market price is £25/MWh. That wind farm will receive a top-up subsidy of £15 for each MWh of electricity it generates.
On the other hand, if that same wind farm puts electricity into the grid later that same day, say at 6pm when the price is £60/MWh, the wind farm now pays back £20 for each MWh it generates.
In the long run therefore, what matters to the question of subsidy is whether wind farms are selling most of their electricity at times when the market price is lower than their CfD strike price (in which case they’ll usually be receiving payments); or whether they’ll be selling mostly when the market price is higher than their CfD strike price (in which case they’ll usually be making payments).
That takes us back to the Carbon Brief quote above, in particular the importance of “the government’s reference price, the level it expects to see for electricity on the open market in each year”.
That reference price can be found in ‘Appendix 2’ of a government document with the exciting title of “Round 3 Allocation Framework 2019”.
To save you the bother, I can tell you that for offshore projects operating in four years’ time the government is assuming a reference price of £48.1/MWh (rising slightly to around £51-52/MWh for the three years after that).
If the government’s assumption is correct, the price that offshore wind electricity could typically get on the open market in four years’ time is £48.1/MWh. That is higher than the strike price awarded to the offshore wind farms that will be operating then (£39.7/MWh). Hence Carbon Brief and others can, quite correctly, conclude that those offshore wind farms will likely be making payments, not receiving subsidy.
But hang on a moment…
Variable 'reference prices'
When previously explaining how CfDs work (and looking at that last diagram), the market price of electricity isn’t some fixed ‘reference price’. It varies!
Indeed, here’s a chart showing an example of how it varies: the hourly price variation for two sample weeks this year, 2019, one in the winter (January) and one in the summer (July):
(source: Grey Cells Energy analysis, based on government data - see link below)
These aren’t any old prices I’ve looked up: they are what are termed the “intermittent market reference prices”. They can be downloaded from this website, entitled “Settlement Data for CfD Generators”.
They are the market prices against which offshore wind farm CfD strike prices are actually compared (and subsidy payments calculated). So when it comes to actual future payments, that single figure government reference price is irrelevant - what will matter is how these variable 'intermittent market reference prices' look at the time.
Let’s say these same market price patterns happen in four years’ time, when one of those offshore wind farms is operating, with its CfD strike price of just below £40/MWh. In January that wind farm will spend most of its time paying money back, since the market price is almost always above its contracted strike price. In summer, when the market price is often below its strike price, it will more likely receive payments.
Whether the offshore wind farm ends up being ‘subsidised’ overall, over the full course of its CfD contract, depends on whether cumulatively it pays out more than it gets paid over this 15-year period.
So the question of subsidy depends not just on the strike price a project has contracted for but – crucially - on how the real market prices vary over the CfD period.
If market prices fall, the balance of payments will shift in favour of the wind farm (more subsidy, or fewer payouts to consumers). If market prices rise, the wind farm will find it has to pay out more often, with the balance therefore shifting towards the consumer.
So it is very important to stress that the timing of generation matters: in particular, how a generator’s variable supply of electricity sits alongside that future price variability curve.
Timing is everything! (And more on this later...)
A CfD is an insurance policy against falling market prices
The net effect of the subsidy and payout process described above is that a CfD guarantees a known, steady ‘strike’ price for the electricity a power project generates, for a long period (15 years). A project therefore has revenue certainty; which is good, since it has to pay back a big dollop of fixed capital cost.
As described above, whether a project proves to be ‘subsidy-free’ depends on how market prices change over the duration of the CfD.
Looking at it today, a strike price may be less than the average price that a project could be projected to receive by selling directly on the market, as a ‘merchant’ project. If – at times when the project is generating most electricity - those market prices stay the same or even increase over the next fifteen years, then a CfD could look like a bad choice. In hindsight, the project could have made more by selling on a merchant basis. The government and electricity consumer will be pleased though, because not only has the project been unsubsidised; it’s actually paid into the system.
On the other hand if – at times when the project is generating most electricity - those market prices fall over the next fifteen years, the CfD could prove to have been a useful insurance policy. It doesn’t matter how low those prices fall, the project will always receive revenues based on its strike price: it will always be 'topped-up' to this level.
If market prices fall far enough, a project which today is expected to be ‘subsidy-free’ may turn out not to be!
In other words, the CfD transfers the risk of higher-than-expected market price drops away from the project and onto the government (and electricity consumers).
‘When the project is generating most electricity...’
You probably noticed that I’ve repeated this phrase a number of times in this article.
That’s because the question of how market prices may change in future – and hence whether having a CfD proves to be a shrewd bet for winning bidders – is not a question of averages. It’s very much a question of timing, combined with price volatility.
Future electricity market prices will depend on a complex interaction of factors: the changing generating capacity mix; changing electricity demand patterns (for example from electric vehicle adoption); changes to market operations; and plenty more besides.
Certainly there are too many factors to consider here, but perhaps one that does deserve a quick mention is ‘price cannibalisation’.
That’s the idea that, as we add more zero marginal cost capacity to the system – such as wind and solar – they will, when they are generating strongly, drive market prices down and reduce their own revenues (if relying on a merchant sale).
Targets for the growth of offshore wind in the UK are bullish: up from about 8.5GW of installed capacity as I write, to 30GW by 2030 (just six years after those recent auction winners are due to start operating). Beyond that, numbers as high as 75GW by 2050 have been suggested. Those recently-won CfDs will run until almost 2040. Perhaps there could be 50GW operating by then?
If that’s the case, and continued efficiencies keep UK electricity demand in check (nowadays, peak is typically just below 50GW), on windy days that much offshore capacity will surely have a big impact on prices.
When the wind is blowing, just the time when those CfD-holders will be selling, it’s not a stretch to imagine market prices heading towards zero – or even below. After all, even now, we already seeing occasional such pricing events (see figure below, showing March 24th 2019). At which point a £40/MWh strike price will start to look like a good deal...
Perhaps a CfD provides offshore wind projects with insurance against the rapid growth and downward price impact of their own industry!
Risk, reward and price
A final point that’s very important to make when thinking about the business cases of CfD vs. merchant offshore wind projects concerns financing costs.
From the analysis so far, it should be apparent that a CfD may not produce the highest potential revenues for an offshore wind project over its 15-year period, if the strike price is below expected market prices and the latter don’t substantially fall. However, by providing revenue certainty, a CfD certainly does reduce financial risk for project investors.
Indeed the question of ‘CfD vs. merchant’ is a classic risk-reward one.
Agreeing a CfD contract at below-market strike prices, as the recent bidders have, means accepting potentially lower revenues in return for building a lower-risk project. You don't have to worry about all those variables that will interact in potentially unexpected ways to influence future market prices.
By contrast, putting together a business plan for a merchant project would mean taking a much higher risk, by betting on electricity prices up to 20 years into the future (if we add development and construction time to a CfD duration). To justify that higher risk, those providing the money will need to see a business plan that projects higher returns: the costs of finance will have to rise.
Financing costs are an aspect of energy economics that are often overlooked. Much tends to be written about the cost of equipment and installation work. Less tends to be written about the cost of the money used to pay for that equipment and installation work.
As with future market prices, there are many variables that come into play. However to illustrate how big a difference financing costs can make, I ran some rough-and-ready numbers through a financial model.
Let’s say, based on those 15-year CfD prices of £40/MWh, market prices of £50/MWh and a bunch of other reasonable assumptions on energy production, installed costs, operating costs, debt and so on, a project produces returns of 7.5% (IRR) for its equity investors, over a 25-year project life/analysis period. (By the way, I've no idea if this is the kind of number last week's projects are hoping for, but that's not particularly important).
Now let’s keep all the non-financing cost inputs the same, but assume that equity investors will need to make 10% in order to take on the extra risk of the project selling on a merchant basis. If we also assume that debt will be hard to find in the absence of a contracted, fixed revenue stream, the financial model now requires market prices of £54/MWh to make the business case work (to achieve those higher returns).
Of course that price is now higher than the £50/MWh ‘reference’ market price that the government is assuming in future. In other words, those 10% returns depend on an assumption of higher-than-projected market prices.
If equity investors take an even dimmer view of merchant risk and want 12.5% returns in order to part with their money, then even in the unlikely event that debt lenders can be persuaded to lend the same amount at the same rate as in the CfD case, the project still needs market prices of £51/MWh. In the absence of debt, market prices of £62/MWh are needed in order to keep the equity investors happy!
Given the numerous different variables at play within a financial model, don’t focus on the specific numbers I’ve produced above! Focus instead on the takeaway message, which is this:
You cannot compare side-by-side the price required by a CfD-backed project with the price required by a merchant one. Even producing the same amount of energy and having the same capital and O&M costs, the merchant project costs more because its financing costs are higher. And if a project costs more, it needs a higher price in order to compete.
Furthermore, the increased financing costs that come with increased risk can easily tip a business case for merchant offshore wind into being reliant on persuading investors that there will be market price increases in future. With increasing proportions of zero-marginal-cost renewables entering the system and the prospect of price cannibalisation as described earlier, that’s likely to be a hard sell.
It’s this combination - of future market price uncertainty and its impact on financing costs - that means developers and investors will certainly continue to value the security of a CfD over merchant risk, even when strike prices appear to be below market ones. (If not CfDs, there are other fixed-price routes to revenue security, such as corporate PPAs).
The case for merchant projects may sound obvious in the context of quick headlines around ‘subsidy-free’ offshore wind.
But it’s the revenue risk that kills them.
It’s perhaps worth writing a little bit more about the government’s 'reference price' assumption; by which I mean the one mentioned in the Carbon Brief quote and published within the government's CfD auction guidance documents, as a fixed number (£48.1/MWh for those first offshore wind projects, due online in four years' time).
This number doesn’t exist as the basis for whether projects will be paid or be paying out relative to their strike price, once operating. Those payments are calculated at the time, based on the variable ‘intermittent market reference price’ charted earlier.
So why does it exist?
It's because each CfD auction has a budget. This budget is a limit on the total amount of money that the government is prepared to spend as a subsidy (where needed to top-up market prices to strike prices). So subsidies added up across all the contracts that are awarded in an auction round cannot exceed this budget.
To work out how much subsidy they might be committing to pay to awarded projects, the government needs an ‘expected market price’ figure against which they can compare bid strike prices. So this expected market price represents their forecast as to how much on average bidders will be able to sell their output for.
Where bid strike prices are higher than this average market price, the subsidy due to a project is simple to calculate:
= (strike - reference price) * energy output * contract period
The energy output can be calculated from a project capacity and its projected 'load factor'. It's notable that, for the latter, the government assumed a very high figure of 58.4% for offshore wind in those latest auctions.
Now, to decide how many projects can win a contract, it’s simply a case of stacking up the bidders (lowest price first) and adding up their calculated subsidy estimates until the total budget is used up.
As discussed previously, because this reference price is simply a forecast of future pricing, the government is taking on a risk that their subsidy calculation will be wrong - if their forecast price turns out to be higher than the actual future ones, projects will end up being 'topped-up' to their strike price more often.
[Of course in the most recent auction, where bid CfD prices fell below the reference price, meaning subsidies aren’t expected to be paid, this total budget issue isn’t really relevant].
Pedants beware: within this article, I’m certainly guilty of some simplification in terms of the CfD and its procedural details. But I consider that allowable in order to get across the key messages in a comprehensible manner!