Why the capacity market favours distributed energy

The UK energy market is amid a once-in-a-generation shift. Our system is rapidly moving away from traditional, large thermal plants and towards flexible and distributed sources of energy and storage. But even when we can see change coming, the pace and full repercussions often surprise. Nowhere is this clearer than the recent T-1 and T-4 capacity market auctions.

The capacity market was created to incentivise security of supply and to set loss of load expectations and capacity margins at sensible levels. Implicit in the creation of this market was the idea that it would ultimately deliver CCGTs in a timely and orderly fashion.

While the capacity market has delivered capacity at relatively low cost, CCGTs were not the winners. Gas and diesel reciprocating engines (recips) triumphed at a scale few analysts foresaw, along with batteries and demand-side response (DSR).

In the 2016 T-4 auction, the emergence of 1.4GW of unproven DSR, in particular, caught some analysts off-guard, but in retrospect was less surprising given the policy set-up. DSR is not affected by the Triad or supplier obligation reforms, and testing is only required after the contract has been awarded. Whether the full 1.4GW is delivered is now an open question.

The surprises continued in the most recent 2017 T-1 auction. The final clearing price of £6.95/kW was lower than many analysts forecast with new build recips bidding in lower than £10/kW, despite embedded benefit reforms.

And so why have flexible technologies and, in particular, recips won?

First, recips won because at a very fundamental level the future GB system requires significant additional capacity that is cost competitive at low load factors. If the current pipelines of nuclear, interconnectors and renewables materialise, the residual load duration curve implies space for 18GW for technologies with load factors above 15-20 per cent (i.e., CCGTs) in the late 2020s. With 22.5GW of currently contracted CCGTs and 7.5GW of expected CCGT retirements, this implies approximately 3GW of space for new CCGTs to 2030. Conversely, those same load duration curves indicate that about 10GW of new capacity is required that will run below 15-20 per cent load factors (e.g. recips).

Second, Aurora analysis indicates recips won because a sizeable chunk of market participants bet that significant capacity promised in previous market auctions will not be delivered – that previous capacity market contracts were treated as options, not as obligations. If that is the case, there will still be sufficient scarcity in the system to re-create the significant price spikes in the wholesale electricity and balancing markets seen over the last few winters. Recips can benefit by more than £10/kW from scarcity-driven price spikes and so can then bid at lower levels into the capacity market.

Given the rise of recips, what role is there for new build CCGTs in the future?

The Department for Business, Energy and Industrial Strategy (BEIS) has a range of policies available to both improve the overall efficiency of the Capacity Market and incentivise the entry of CCGTs. Four policies, in particular, are most frequently cited as increasing the likelihood of CCGT delivery: increasing the procurement target; removal of embedded benefits; removal of EU-ETS exemption for recips; and changes to reliability standards for storage.

To use the reliability standards for storage as an example, under current rules all batteries are de-rated at 96 per cent, regardless of duration. Consequently, a very short duration battery (say 30 minutes) is treated the same as a much longer duration battery (say 4 hours). A short duration battery – which are well-suited to frequency response services – can secure a capacity market contract at a 96 per cent de-rating level when any individual unit can only deliver power for 30 minutes. Making storage de-rating dependant on duration and deployment assures a like-for-like substitutability with other generation technologies and reduces batteries’ competitiveness, making space for an additional 1GW of CCGT.

But such is the underlying need for flexibility, Aurora modelling indicates that, even if the four potential policies available to BEIS were implemented, we would still see 11GW of some mix of recips, DSR and batteries enter to 2030, and only 5GW of CCGT. As recips and other flexible technologies continue to win, Aurora forecasts relatively low long-term prices in the capacity market through the 2020s with price spikes in the mid-2020s as coal retires. Canute cannot turn back the tide, it would appear.

There are potential non-policy market developments that may induce additional CCGT capacity: less interconnector development than is currently forecast; slower battery cost reductions than Elon Musk and his acolytes forecast; and a failure to develop any nuclear after Hinkley. Each of these outcomes creates space for additional CCGTs to enter via the capacity market – in the case of nuclear, up to 5GW of additional CCGT capacity by 2030.

The repercussions of the shift to flexible and distributed generation will continue to surprise. It will create discontinuities in ancillary markets, change the shape of the wholesale price curve, profoundly alter the way we charge for transmission, and make life hard for big players who are slow to adapt to a new eco-system.