Late on 10 December the second capacity market auction closed its books, with 46.4GW of capacity procured and a clearing price of £18/kW. Although slightly higher than our prediction of £14kW-16/kW in our pre-auction comment in Energy Spectrum, it is still being universally characterised as a "low price". Cue the predictable government fanfare about the delivery of energy security at low costs to the consumer.

We would agree, in one respect, that the two capacity market auctions have achieved their basic objectives: procuring the required security volume at the lowest price in the context of a technology-neutral auction. But in our view, whilst the price is low, the wider costs are high.

In each of the first two auctions we have seen pay-outs to coal plant--against which government has hardened its position and ultimately now wishes to legislatively close. We have seen additional incentives for new small scale peaking plant that National Grid never expected to be built. Whilst the clearing prices have been low they have been well above the exit bids of the vast majority of plants in each supply curve, not only solving the "missing money" problem for generators but giving them enhanced rents to boot. Considering at the same time the moribund prospectus for new large build gas that the government explicitly wants, and that demand side has failed to provide more than 1% of the cleared capacity in the last auction (well shy of its innate potential), and the picture is even more disappointing.

The issue is that capacity market in its current format has only limited strategic value. At its heart, it fails to provide any compelling answers to what will replace closing old coal and old gas capacity at a scale we think is consistent with our organic and evolving system security requirements, our increasing requirements for flexibility and our hardening decarbonisation objectives. With a forecast widening shortfall in the Fourth Carbon Budget, the challenge of the recommended Fifth Carbon Budget, and a potential tightening of national level ambitions following COP21, it is hard to see where the capacity market, as currently designed, fits in as part of the solution to the trilemma. This is more acutely the case when combined simultaneously with a savage reduction in money available to support new renewables deployment.

Relying on a '"cheapest megawatt", single delivery year-focused capacity market to deliver complex long term strategic system objectives was always going to be akin to a man standing in a bucket and trying to lift himself up by the handle. This could perhaps be tolerated if the wider system impacts were neutral; but more worryingly, the capacity market is in our view now actively undermining efforts to address the trilemma.

It is becoming ever clearer that before the end of the decade we will most probably witness tighter margins as plant that are disadvantaged relative to their competitors in failing to win capacity market agreements shut down or mothball. It is a strange system that aims for medium-term security of supply but that compounds short-term system frailty as a consequence. In less benign commodity markets we would already be seeing the projected tightening margins feeding through into wholesale prices on the forward curve. The consumer is being spared this cost only because of oil and gas being at historic lows, fickle variables that the government cannot claim credit for, or control in the future.

As a result of anticipated further plant closures, system security margins will tighten further in the next couple of winters. It is highly likely that we will see National Grid coming under pressure to expand Supplemental Balancing Reserve (SBR), the interim measure before capacity market delivery years commence. SBR is already currently subject to a "minded-to" approval for extension to 2018 from Ofgem.

Some plant that otherwise would close following unsuccessful bids into the auctions will compete hard to fall into the envelope and so provide economically valuable services outside of the wholesale market. Those that can't get into SBR face a difficult future. But for those that do, we shouldn't be under the illusion that this is a cost-free route to keeping the lights on. National Grid is expected to allow marginal wholesale market plant to deliver before invoking SBR, so SBR itself does not dilute the impact of tighter margins on wholesale prices. In addition, the costs of the service are recovered through balancing charges, and any triggering of the use of "last-resort" SBR is priced into the imbalance price at the Value of Lost Load (£3000/MWh).

There is also a risk that SBR may need to become a more long-standing feature of the market, even once we begin to overlap with capacity market delivery years. This is because National Grid may harbour legitimate fears that the mix of procured plant in capacity market auctions does not guarantee the same level of system service as procured through SBR, particularly during prolonged periods of high demand in winter.

SBR requires that plant: can be available between 6am and 8pm on weekdays between November and February; have either a gas connection or sufficient fuel stocks to run during the availability period for five consecutive contracted service windows at a level of output equal to the offered capability of the unit; and have necessary control and monitoring facilities installed to enable it to be deployed by the system operator. By contrast, the capacity market in its current format looks likely to procure several gigawatts of small-scale embedded plant in its volume for any given delivery year, plus a couple of gigawatts of interconnectors. That sort of capacity may, but cannot be guaranteed to deliver the same level of system service as those within the SBR envelope. This disparity in degrees of confidence in outcomes matters for the system operator.

Of course, peaking plant like small gas and diesel does have a role to play in the energy mix to meet system peaks--more so in a system with more intermittent renewables. But with low loads and high emissions for diesel in particular, allied to very high prices required to run, it should only be used at the margins of the merit order on an "as and when" basis. National grid already procures for this in balancing services like STOR. It is not obvious we need to be buying gigawatts more of this plant across the capacity market auctions, particularly when by doing so it is leaving no room for new build larger scale gas, which the government has now said it wants.

Aside from the interim role of SBR, there is also a medium to longer term risk of increased cost in the wholesale markets too. In its current format the capacity market looks set to deliver a generation mix that is overweight in new "invisible" (to National Grid) embedded peaking plant and underweight in new transmission connected mid merit and baseload plant. Without some level of like-for-like substitution the expected attrition in existing mid merit and baseload plant is likely to drive increased reliance on peaking capacity in periods of higher than average demand when compared to today. This will potentially drive up peak prices and increase their frequency of applicability, particularly when we consider that for decarbonisation we should see a continued proliferation of must-run, but less controllable and more variable renewables.

Building some new build, highly efficient CCGT would help manage this risk. CCGT has versatility to run to suit a variety of system conditions. It is also a lower carbon option. We may not need a huge amount of new capacity given the growth of renewables, interconnectors and the surviving existing fleet of gas plant, but we will need more than we are getting today where--Carrington aside--we are bereft of any emerging capacity of this type.

Incentivising this will be more expensive than £18/kW. In fact, a price of £50/kW would not be unrealistic. Bearing in mind that under the current pay-as-clear design such a payment would apply to plant that the supply curves have shown only need £5/kW or less to exist, the costs of procuring this plant could be very high if the same auction design is adopted as a delivery model.

The costs may be further increased if we fail to reform the auctions and the new build signal comes in one-go--as could happen in 2016 as a result of attrition. A panic buy of new-build capacity, concentrated in a single auction, would see far less competitive tension than if new capacity had been procured incrementally over several years.

To save the consumer on both counts, the reforms to the auction will be necessary to deliver new build gas, and will be needed very urgently. But, as we have alluded to in our previous coverage of the capacity market, EU state aid considerations could act as a constraint here in both scope and timing of reform.

In our view it is time we started to think less about the detailed features of instruments of delivery. We should instead take a step back to profoundly reconsider how to deliver on the trilemma in light of much changed, and at times incompatible, environments in domestic and international energy and climate policy.

COP21 shows that, internationally at least, writing the obituary of the trilemma would be premature. Even if the Treasury would prefer decarbonisation to be an afterthought of security of supply at the lowest cost, in rhetoric at least, Paris propelled it back to parity with cost and keeping the lights on.

Taking both domestic cost pressures and continued low-carbon ambitions into account, a more comprehensive strategy is now required. Such a strategy needs to set out a pathway into the next decade, not from year to year, but across the whole period. It needs to take into account the system margin and emissions impacts of the expected scale of plant attrition. It needs to examine the already "locked-in" new-build capacity in renewable and thermal generation through the same lens. At its core it needs to construct a realistic ambition around the role of disruptive demand side and storage technologies. It needs to determine the limit of our ambition on interconnection. Only then will it be possible to determine the incremental need for new capacity, and the types of technology required to finely balance the trilemma. This is a far cry from the myopic and low-ambition policy delivery frameworks we have today.

Of course, developing this strategic approach would necessitate another period of painful transition to new incentives and structures so quickly after we have concluded Electricity Market Reform. But if the alternative is the continuation of a false economy that damages rather than supports meeting the challenges of the trilemma then this would appear to us to be a price well worth paying.