Capacity market details and demand side response
Catherine Mitchell, IGov Team, 1st August, 2014
Back in 2010, the UK Government announced an Electricity Market Reform (EMR) which was to transform the GB electricity system into one fit for the 21st Century. One of the four main planks of EMR is a capacity market (CM). It was argued that such a CM was necessary (1) in order to stimulate sufficient investment to ensure security of supply, and the Government’s original preferred option was a targeted strategic reserve mechanism (page 100, para 69). Since 2010, GB has moved from a preferred targeted strategic reserve mechanism (see below for explanation) to a market-wide auction, and we now know that the first auction will be for 53.3 GW in December 2014.
How did we in GB end up with a market-wide CM, despite the Government’s original preference for a targeted mechanism? The answer relates to the incredibly useful dimension of a market-wide CM: its complexity of operation. The blow by blow means of how that happened is a brilliant, albeit very depressing, example of how decisions are made in GB. In the end, GB has a CM which is a retrenchment, not a reform, of the electricity market; which is great for the incumbents and their plant; which is bad for the environment, because it will keep old coal going and is very conservative on the demand side; which is unnecessarily expensive; and which is bad for GB innovation.
The market-wide CM necessitates a whole range of assessments, which if calculated poorly can mean that far more capacity is paid for – or to put it another way, the market-wide CM can be very easily rigged to make it do what the operator wants. If a primary goal of the CM from the point of view of the Government is (i) to keep any old capacity on the system – irrespective of its characteristics (an incredibly old-fashioned approach) – but also (ii) to assuage the incumbents who have been irritated by the pro-nuclear EMR process and who kept threatening to shut down their plants – then a market-wide mechanism suits your purpose very well.
Lessons from America
As ever, when it comes to the details Britain is going it alone. Lessons could, and should, have been learnt from the US.
The US is at the global forefront of capacity mechanisms and markets, and has long experience of how to include demand side response. For example Pennsylvania Massachusetts and New Jersey electricity market (PJM) has incorporated it since 2002. Over time, slowly but surely, improvements have been made to the CMs so that different capabilities (including DSR) are eligible and can be paid for, thereby leading to increasingly efficient and cost-effective operation of the system. The Federal Energy Regulating Commission (FERC) publishes an annual Staff Report which sets out what US State or market has done in the last year to their CMs, and why. Over 10% of peak capacity is provided by DSR in PJM, while DSR averages 6% of peak across the US. This translates to between 5-8% reduction in wholesale prices averaged across the year, although this is a whopping 90% cut for the peak time price. FERC also charged the North American Electricity Reliability Corporation to analyse all resources of DSR, and it has found that there is still a lot more DS resource which can be incorporated. FERC wants this DSR used before new generation is built to keep costs down. Given all this experience, why is it we in GB have not used the opportunity to implement a forward thinking CM which will help to incorporate the demand side, keep prices down, fit with Europe and generally make our electricity system more secure and efficient?
How the CM will work in GB
The Implementing document sets out how GB’s CM will work and how it will be paid for. Capacity will be paid for by its availability – whether or not it is used. There is also a payment for the energy component of that capacity via the market when a capacity event is triggered. The price paid for the capacity will be the marginal bid (i.e. all those which bid in below the final marginal bid will receive the higher marginal bid) via a decreasing clock auction. There are two interlinked auctions: one, to be held 4 years ahead of delivery (the first take places in December 2014 for capacity in 2018-9, for a period of up to 15 years depending on the contract agreed); and then a demand auction, a year ahead of delivery (ie for demand response to be available a year later, for a year).
Targeted strategic reserve versus capacity-wide market
A targeted strategic reserve mechanism is when the System Operator (SO) decides that there is a need for a strategic reserve of capacity and then puts out a tender for that amount of capacity to be available by a certain time, for a certain time – and this capacity can be in the form of supply from a power plant, but it could also be capacity via an interconnector, demand side response, a particular type of generation (ie gas) or for a particular energy system capability (for example, fast ramping) and so on. The SO is therefore able to tender for exactly what they need / want for operating the electricity system in the way they deem most efficient. The targeted CM is therefore
- a flexible mechanism.
- It is much simpler than the market-wide mechanism to operate (as explained below);
- much less likely to pay for capacity that is not really needed or the type of capacity that is not wanted (for example, old coal plant that UK decarbonising policy and targets says it wants to get rid of);
- cheaper, because it only supports what is needed for at that time in the electricity system. We know that the electricity system can change rapidly (ie via rapid deployment of solar electricity in Britain (up to 3.6 GW mainly over 3 years) or via interconnectors), and such a targeted mechanism is able to keep up, and work, with that change.
- It also does not affect electricity market prices, thereby complementing electricity availability in Europe. If capacity is scarce then prices will rise and attract electricity from Europe, thereby helping to reduce the capacity problem. This is the opposite of a market wide CM, which does effect electricity prices by depressing them. In a European wide scarcity situation, GB prices will be relatively lower and electricity will flow to Europe, thereby making the scarcity problem worse in GB.
In brief, a market-wide CM is basically the opposite of a targeted CM – it is inflexible; likely to pay for more capacity than is needed; does not differentiate between what is wanted by the system and what is not wanted; is more supply orientated; and does not fit with Europe. So, if a CM is needed, then a targeted mechanism is the way to go.
The cost of choosing the market-wide CM versus the strategic reserve is not trivial. As said above, the overall amount of capacity to be auctioned and therefore paid for is 53.3 GW. The Impact Assessment is based on a cost of £39/kW. E3G has estimated a 2GW coal plant such as Fiddlers Ferry could receive £234m over 3 years (or up to £1bn over 15 years). Of this 53.3GW, E3G estimates about 10GW could be old coal. This means customers could pay around £5bn or so over 15 years to keep coal going, when the whole point of EMR was to make the electricity system fit for the 21st century and to move into low carbon. Others have estimated greater amounts of coal or higher amounts of payments.
Whatever the actual amount, a targeted reserve would have been far cheaper and preferable.
Rigging the Market
In a market-wide CM, it is necessary to work out how much capacity to auction to maintain a reliability standard. There are a number of steps along the way to do this – all of them eminently riggable.
- (i) The reliability standard has to be set. A reliability standard is a combination of the agreed amount of time that electricity should be available and an acceptable cost of ensuring that. It has been decided that electricity can go off for 3 hours a year (sometimes known as the loss of load (LOLE)). This reliability standard then has to be converted into a capacity requirement and GB has been criticised for its methodology, including by the Technical Experts Panel. If the methodology leads to too much capacity being supported (the IET estimates this at 10% more capacity than necessary) then the marginal cost of that extra capacity will be higher than necessary (e.g. see here and here). This higher marginal price is then paid to all available capacity, thereby unnecessarily increasing the total cost of the CM.
- The amount of capacity to be supported has to take account of (ii) what supply there is expected to be available without a CM; and (iii) what capacity demand there will be. This should lead to a gap between the two which equals the amount of capacity the CM should be paying for to meet the reliability standard. Within these assessments, there are numerous complicating factors, such as de-rating factors (ie how to assess the reliable capacity of any given type of capacity); availability of generation from interconnectors, pre-qualification processes etc.
- Finally, (iv) the criteria to call a capacity event has to be agreed. Again, NGs methodology has also been criticised as leading to more capacity being paid for unnecessarily.
All of these assessments and arrangements can very easily be got wrong – and in fact, because it is projecting the future and because the energy system is in a time of huge change, one can expect it to be wrong. This could therefore lead to too much capacity being paid for – bad enough, but completely bonkers if that capacity is undermining other policy goals (e.g. here, here and here). Furthermore, much of this uncertainty would have been significantly reduced, if not removed, with a targeted strategic reserve CM. Sadly, the European Commission approved this mechanism on 23 July 2014.
At the moment, the GB mechanism does not take account of interconnections, but will do so in the second auction (again potentially supporting capacity which will then become unnecessary); and is also building up to introducing demand side response, which, as said above, will bid in to a separate 1 year ahead auction. Some of the assessed capacity need for the main 4 year ahead auction will be kept back to allow DSR to bid in to the demand year ahead auction. Discussion of this is between 400-900 MW – or about 1% of peak demand. Given that the average in the US is 6% and PJM reduces the peak by 10% this is very conservative. Moreover, as the Implementation document explains, if the actual capacity need falls, then at least 50% of the kept back capacity for the demand year ahead auction will be maintained. However, this is a cross-incentive. If we in GB manage to reduce total demand more than expected, DSR will suffer. This is the wrong way around. What should happen is that the main 4 year ahead auction should be reduced to follow on from what DSR is available.
The GB capacity market is another example of poor energy policy decision-making. It suits the interests of companies who are the losers in the move to a sustainable future whilst again being unsupportive to those actors within the energy world who are trying to be innovative. And of course, it is the customers who will pay considerable amounts for something they should not be paying for.
(1) This blog will not enter into the discussion of whether a CM is necessary or not in GB, but point those interested to a paper which sets out the key questions to be asked before a CM is put in place. If any of those questions have a negative answer then it would seem that a CM is not necessary, and most of the answers in the GB case are negative