Preparing for AR8: managing uncertainty in CfD auctions

Allocation Round 8 (AR8) is an important round for delivering the government's Clean Power 2030 ambitions, with substantial new renewable capacity expected to be secured.

Projects participating in AR8 are now developing and, in many cases, finalising their bids. And they are having to do this in the face of significant uncertainty. 

Bidders do not know their own costs with certainty. They also don’t know their competitors’ costs, the government’s demand curve, or the likely competitive conditions and demand in future auction rounds. All these factors will matter, which means that working out what to bid can become extremely complicated. 

Our Contracts for Difference (CfD) auction simulator is designed to help you explore these trade-offs. Rather than producing a single forecast of the AR8 clearing price, it allows you to test different assumptions and see how they affect the auction outcome for your project. 

Using a few simple examples, we illustrate how different sources of uncertainty affect bidding decisions and demonstrate how our simulator can help bidders navigate them. We are publishing a simplified version of the tool alongside this article, so feel free to try for yourself. 

Let’s assume there are four competing projects, each with an uncertain minimum acceptable strike price. 

Let’s further assume an estimate of the government’s budget constraint at between £0.75bn and £1.25bn, with a CfD payment load factor of 60%, and an expectation that subsequent allocation rounds will be similar in price level to this one. 

Using this example, we can explore a few bidding strategies. 

What happens if we bid at our P50? 

If we assume the role of Project 1, a first question might be what happens if we bid our P50 minimum acceptable price. To answer this, we need to understand the possiblemerit order in the auction - Figure 1 illustrates a “static” merit order, where we assume all competitors bid their P50 minimum acceptable strike price. 

​​​Figure​ 1 Illustrative merit order if all competitors bid at their P50 minimum acceptable strike price 

Using our CfD auction tool, we can then introduce uncertainty about competitor bids and simulate the merit order across many thousands of simulated auctions.  

Figure 2 shows the bid stack distribution: in each auction simulation, projects are sorted from the lowest bid to the highest; the horizontal axis is cumulative capacity and the vertical axis is bid price. Darker cells are places where those simulated bid stacks pass more often. 

​​​Figure​ 2 Simulated bid stack distribution 

It is clear that based on our P50 bid, we stand a pretty decent chance of winning even with reasonable uncertainty around demand – around 75%. However, if we bid 10% above our P50, by contrast, our probability of winning declines to around 40%, but our expected headroom value actually increases from around £2.00 / MWh to £4.60 / MWh. The relationship between the probability of winning and expected value with a bid 10% above P50 is illustrated in ​Figure​ 3. In this case, the expected value curve also shows that the expected value would increase with a small further increase in bid.

​​​Figure​ 3 Award probability and expected headroom 

If uncertainty around our own costs grows, then the expected value of bidding our P50 worsens. Let’s suppose our P25 minimum acceptable cost remains at £90/MWh, but because of supply chain risk, our P75 value is actually £120/MWh (i.e. asymmetric risk). Bidding our P50 now has an expected value of -£5.40 / MWh. 

What happens if we are more optimistic about future rounds? 

Returning to our original specification, we can also see that what we think about future rounds matters. If we think that the P50 outturn for future rounds is 5% above this one (once we have factored in the costs of waiting), with a P75 at 10% above, then even if we do not shade up our bid, our expected value is slightly better, at £2.40 / MWh. If we do shade up our bid by 10%, then because our probability of losing in this auction is higher, what we think about the future matters more. Compared to an expectation of future rounds being similar in nature, being more optimistic increases the expected value from our bid to £6.10/MWh. 

What happens if we face keener competitors? 

We can also see what happens if we are facing competitors that are universally more optimistic (perhaps because they are less worried about negative prices). Suppose P50 and P75 minimum bids for competitors are 5% lower than our original specification (but that everything else stays the same, save that we mark up our P50 bid by 10%). 

​​​Figure​ 4 Simulated bid stack with keener competitors 

Project 1 is now much more marginal. Our probability of winning falls to around 30%, and our expected value even from a 10% addition to our P50 bid is now £3.30 / MWh. In this situation, if we think future prices might be 5-10% higher (as in our second example), then our expected value from our 10% addition is £ 4.90 / MWh. 

What happens if we are more uncertain about the government budget? 

The attitude of the government is clearly critical. If we are more bearish about the budget that the government will have, it will reduce both our probability of winning and our expected payoff from a given bid. Going back to our original specification (but keeping our 10% addition to our P50 bid), if we reduce our P25, P50 and P75 estimates of the government budget to £500m, £750m and £1.0bn respectively, then the impact is material. It reduces our probability of award to around 18% and our expected value to £2.6 / MWh. Again, a lower probability of award means what we think about future rounds matters more. 

All of the results above were produced using our CfD auction simulator. It is designed to help explore these trade-offs. Rather than producing a single forecast of the AR8 clearing price, it allows you to test different assumptions and see how they affect the auction outcome for your project. You can see a simplified version of the simulator below, to allow you to try your own assumptions (don’t worry: we don’t see any of the data). 

Try the interactive simulator below:

Playing with the simulator 

Start with your own project economics 

The starting point for any bid is the strike price your project needs. The simulator allows you to enter a range for your project's minimum acceptable strike price, rather than relying on a single point estimate. This makes it possible to see how uncertainty around your own economics affects the value of different bidding strategies. 

Then consider what other projects might bid 

Your own required strike price does not tell you whether you will win. That depends on how your project compares with the rest of the auction. 

The simulator allows you to build a view of the competing projects, their capacity and the range of strike prices at which they might be willing to bid.  

Add the government's auction constraint 

Even a competitive project will only win if the government procures enough capacity. The simulator allows you to test different assumptions about this constraint. You can represent it as a limited CfD budget, or a maximum strike price the government is prepared to accept. 

It then simulates many possible competitive bid stacks rather than assuming that every competitor's bid can be forecast precisely. This helps show where your project might sit in the auction, how your probability of success changes with your bid, and which individual competitors are most important to your outcome. 

Add what happens if you do not win  

Losing does not necessarily mean losing the project. A project may have another opportunity to participate in a future allocation round, although keeping the project alive may itself cost money. The simulator allows you to reflect a view of future auction conditions and bring this option value of waiting into the bidding decision. 

The simulator adds this to the current round competition, to make the trade-off through time explicit: a lower bid may increase the probability of winning, but at the cost of giving up value if you succeed. A higher bid preserves more value if successful but increases the risk of missing out. 

And if this whets your appetite… 

The simulator allows these uncertainties to be considered together, to stress test questions such as: 

• What happens if our own costs turn out higher or lower than expected? 

• What happens if the auction is more competitive than we expect? 

• How much does our bid depend on a particular competitor? 

• What if the government procures more or less capacity than expected? 

• How valuable is it to win AR8 rather than wait? 

• At what point does a more aggressive bid create enough additional chance of winning to justify accepting a lower margin? 

The online simulator is a simplified version of a more complex tool, which allows auction simulation to be combined with detailed project economics, competitor intelligence and scenario analysis to help develop and stress-test a bidding strategy. The important thing is to understand the objective. Simulation is not intended to tell you the right AR8 bid. Its value is in helping you understand why a particular bid might make sense, which assumptions are driving the result and where the key risks to that strategy lie.  

We hope you enjoy using it!