The capacity mechanism: less esoteric than it might seem

In November 2015, the European Commission opened in-depth investigations into the French capacity mechanism.[1] It may seem like a highly technical subject of interest only to an audience of insiders and aficionados of the arcane workings of electricity markets. Far from it. The issues raised by the introduction of a capacity mechanism are simple and universal: they boil down to the effectiveness of government action to correct an apparent market failure, and solidarity between EU Member States. This post explores these two points.

1. What are capacity mechanisms?

Several EU Member States have introduced or are about to introduce capacity markets in the electricity industry. But what are they?[2] The electricity market is currently "energy-only", meaning that power stations are paid exclusively for the megawatt hours they generate. If, however, wholesale prices are very high at the time of production, a company that produces little electricity generates little revenue, and the result is often bankruptcy. On the face of it, there's no problem here. Why should generation units be kept afloat if they produce little or nothing at all? And yet, that is what the public authorities are trying to achieve in creating a system that remunerates megawatts of installed capacity, even when they generate nothing... but are in a position to.

2. Is the electricity market flawed?

Imperfect markets

Markets are very efficient ways of coordinating the actions and decisions of thousands of individuals and companies, but they are not perfect. Economic theory and common sense suggest that governments should intervene to correct market flaws.

For example, a substantial category of market flaws are caused by supply and demand anomalies. Major economies of scale result in the emergence of big companies that might abuse their dominant position or enter into anti-competitive agreements. Consumers may be insufficiently informed and/or in a position to act when such suppliers emerge. Citizens and small businesses expect governments to regulate companies that might abuse their market power and buyer weakness. History is littered with examples of antitrust behaviour resulting in structural[3] or financial sanctions.[4]

Two flaws are regularly cited to justify the need to pay for megawatts of installed capacity over and above the megawatt hours produced: first, the wholesale energy market does not in itself guarantee compliance with the authorities' reliability criterion; second, the "energy-only" electricity market generates price peaks which are regarded as suspect.

The reliability criterion

In 1949, Marcel Boiteux showed that electricity systems (there were no markets at the time) alternated between off-peak periods during which, demand being lower than installed capacity, the cost of producing a megawatt hour essentially equalled the cost of the fuel, and peak periods when, with demand equal to installed capacity, the cost of producing a megawatt hour included not only the cost of the fuel, but also the cost of building a generator. When there is a long-term balance between supply and demand, installed capacity adjusts until the profits made at peak times exactly cover the marginal capacity cost.

To implement this mechanism in a market, the electricity price must increase at peak times while demand falls to adjust to installed capacity. Until the very end of the 20th century, there was no spot electricity market, so no real time price signal[5]. Demand was barely sensitive to variations in the actual cost of electricity generation. The engineers who designed electricity systems then adopted another less economically efficient but pragmatic approach: they scaled the generation system to respond to maximum demand by applying a security of supply criterion. In France, for example, the authorities have set the average blackout duration at 3 hours per year[6], which translates (for example) in no blackouts during mild winters; just over ten hours during exceptionally harsh winters.

The first market flaw cited to justify a capacity mechanism is what happens when installed capacity is insufficient to meet this security of supply criterion. The flaw certainly exists, and for several reasons: investors do not want to take too many risks, they coordinate their actions poorly; they struggle to anticipate demand; and facilities cannot be split. In virtually all highly capitalist industries, installed capacity is not optimal. It is therefore unlikely that, in the electricity industry, installed capacity will spontaneously become optimal, and it is even less likely that it will meet the security of supply criterion.

Does this argument in itself justify a regulatory intervention, in this case the implementation of a mechanism regulating the electricity market? There are more caveats to a 'yes' answer today than there were in the 1980s: spot markets exist, and thanks to information and communication technologies, demand can respond more nimbly to variations in the spot price.  When wholesale prices are very high, it is likely that some (big) users will cut their consumption at peak times, and demand will naturally adjust to supply.

This is very likely to happen, but not certain. What if the equilibrium outcome was on average ten blackout hours per year, more than three times the current security of supply criterion? During particularly harsh winters, or when the generation system is down for long periods, the French system operator, RTE, would implement rolling blackouts: tens of thousands of consumers would be cut off for two hours on a handful days per year.  This might not cause economists to lose any sleep, but politicians would: they regard the reliability of the electricity system as their responsibility. For example, on 4 November 2015, the UK came very close to implementing rolling blackouts. Prime Minister David Cameron, who could hardly be accused of being an interventionist, promptly convened a ministerial meeting, promising to "do everything we can to keep the lights on".

As the capacity mechanism comes at a time when demand is becoming flexible, it can't be said to be a response to a market flaw. It is above all a response to the political need to ensure that the security of supply criterion is met for all but three hours per year, on average.

The question, therefore, is whether it is the responsibility of the public authorities to meet this criterion. France, the UK and the north-eastern states of the US say 'yes'; others, such as Texas and New Zealand, say 'no'. This is therefore not a purely technical debate for electricity system experts.  On the contrary, it is a simple issue of whether governments are responsible for how electricity markets are organised, and in particular for meeting the security of supply criterion.

Price peaks

The second flaw cited in support of capacity mechanisms is the fact that for several hours each year, the "energy-only" market gives rise to prices of several thousand euros per megawatt hour.[7] Here again, as long as it is possible to ensure that these price peaks are not caused by producers flexing their market power, an economist sees no reason for concern. However, the operation of the free market creates a political problem. These high prices, which come precisely when electricity is needed most, are perceived as an injustice; in particular, they pose a severe risk to the poorest households. Might they have to turn down their electricity-powered heating on very cold evenings to avoid taking too big a slice out of their budgets?

An economist would choose to protect the least well-off households with appropriate supply contracts, or better still, energy vouchers, rather than a capacity market. To protect residential customers during price peaks, governments could ask suppliers to offer at least one fixed-price contract. However, it would be better if retail prices reflected variations in wholesale prices at least partially and if struggling households received flat-rate financial aid.

The example of passenger transport

The passenger transport industry provides a good illustration of the problem of capacity and public intervention. Historically, taxis needed to park in precise locations to pick up their customers. So the authorities regulated capacity, i.e. the number of vehicles available, by limiting the parking rights granted.  As capacity was limited administratively, the authorities naturally regulated the prices that taxis could charge.

IT is making this industrial structure obsolete. Now, thanks to mobile phone apps, drivers and passengers can meet anywhere. It is no longer necessary to regulate the number of vehicles. Numbers will balance out naturally at the capacity needed to satisfy the market. In economic terms, the number of vehicles is in long-term balance when the marginal vehicle exactly breaks even.

Even better, the Uber platform has implemented its surge pricing system. When demand exceeds supply, the "normal" price is doubled or tripled. This system has exactly the effects Marcel Boiteux anticipated for electricity: customers cut their demand, i.e. wait a few minutes before booking their ride, and drivers head to areas where supply is low. The drivers interviewed by the authors of this blog confirmed that most of their profits, once they had covered their vehicle and fuel costs, are made at peak times.

3. The question of European electricity solidarity

While opinions differ on the pros and cons of the EU's energy policy, the construction of the single European electricity market has been a success. Today, the vast majority of the continent's generation systems form part of a single market. When the state-to-state transport network is not congested, a single power station sets the price for hundreds of millions of consumers. When the network is congested, the market coupling mechanism implemented gradually over the last decade or so ensures that the congestion is managed efficiently.

Therefore, the implementation of national capacity mechanisms poses a potential risk to these arrangements. Meeting a national security of supply criterion requires a country to hold back megawatt hours during crises, or in other words, close the borders. In practice, if France was in danger of running out of supply, RTE would have to limit exports to avoid blackouts in France, at the risk of causing blackouts in Belgium.

Here too, the question raised is not technical, but political: what does European solidarity mean?  Are we ready, in the name of our common European identity, to share a power failure? Recent crises (the euro, refugees) have shown the limits of European solidarity.

* * *

So, why did the European Commission launch an enquiry into the French capacity mechanism, as we mentioned at the start of this post? Because any intervention by a Member State government aimed at altering the market's natural balances is suspect. "The Commission has concerns that these plans to remunerate electricity capacity could, in the case of the country-wide capacity mechanism, favour certain companies over their competitors and hinder the entry of new players." The French authorities therefore have to prove that they are merely trying to protect consumers against the risk of power cuts, without skewing competition.




[2] Loyal readers of this blog will remember that we discussed capacity mechanisms in May 2015. See

[3] In 1914, Standard Oil was broken up into thirty-four independent companies:

[4] In May 2009, the European Commission imposed a €1.06 billion fine on Intel:

[5] The rollout of Effacement des Jours de Pointe (Peak Day Cancellation) tariffs by EDF is a way of communicating a scarcity signal in the context of a regulated tariff.  Unfortunately, this innovation has not really been adopted outside France.

[6] The risk of blackouts, the theoretical possibility that a consumer might find his or her supply cut off, is determined by ministerial ruling. This is not a question of flaws in the transmission system; on this subject, see

[7] On EpexSpot, on 8 February 2012, during a very cold snap, the price per megawatt hour for delivery the following day between 10 and 11 a.m. hit €1,938, up from the usual €100 to 200;

Add new comment