Reversing scorched earth policies in the electricity sector: Part 3 [Emissions Trading]

In my previous article in this series, I referred to the impact that abolishing the Emission Trading System (ETS) would have on the net cost of dispensing with solar and wind generation. The ETS is perhaps the most important but least well-understood component of the support structure for renewable generation, not only in the UK but in the EU. Variants of the ETS have been adopted in many other countries. In all cases, governments and economists rely upon a standard piece of economic analysis which, unfortunately, is completely irrelevant to the electricity market.

The analysis of emission taxes and trading in the electricity market is a classic example of how the simple analytical models of economic theory get misapplied when transferred to a setting for which the implicit assumptions are simply wrong.[1] The idea goes back to a brilliant economist who taught at Cambridge in the 1930s – A C Pigou. He suggested that certain kinds of damaging externalities - actions by one individual that affect the welfare of other people – could be dealt with by imposing corrective taxes (known as Pigouvian taxes) on the activities that give rise to the externalities.

In the case of environmental externalities, and in particular CO2, the standard theory suggests that a tax equal to the marginal damage per tonne of CO2 (tCO2) would reduce CO2 emissions in an efficient manner. Note that the theory does not refer to eliminating all emissions of CO2, just to balancing the external costs of emitting CO2 against the benefits using coal or gas to generate electricity. This analysis gives rise to a large but very controversial literature focusing on calculating the social cost of carbon, i.e. the external damage per tCO2 emitted.

Estimates of the social cost of carbon tend to rely on use of models that attempt to capture the interactions between economic and environmental variables looking forward to 2100 or 2200. The results are extremely sensitive to small variations in assumptions and have become extremely politicised. Estimate vary from $10 to $200 or more per tCO2. Anyone who reads the Wikipedia article on the social cost of carbon will get a clear sense that this is an area in which prior convictions outweigh all other considerations.

The standard theory implies that we should use an environmental tax – a fixed amount per tCO2 – to deal with environmental externalities. This argument was modified by Martin Weitzman, who pointed out that we might not have any good idea of the damage caused by an externality, but we might be willing to say that we want to reduce emissions by, say, 80%. In such a case, he argued that a permit trading system would be more efficient than a tax. The total number of permits issued would be equal to 20% of existing emissions and the price of traded permits would reveal the tax or penalty per tonne of emissions.

While this is an elegant answer – and the idea has been adopted in many countries – it does not, in fact, deal with the original problem. It is little more than handwaving. How do we know that 80% is the “right” amount of reduction in emissions. Why not 60%, 70% or 90%? This comes back to what we think about the costs and benefits of reducing emissions by different amounts. Are we genuinely willing to pay, say, £500 per tCO2 to reduce CO2 emissions? Or what about only £5 per tCO2?

This brings us to a variant of emissions trading in which the number of permits is adjusted at regular intervals to keep the permit price within a range defined by a floor price and a cap – perhaps between £20 and £80 per tCO2. The UK adopted a CO2 floor price in 2013. Under the arrangement a top-up tax - the Carbon Price Support (CPS) - was charged to businesses required to buy EU-ETS permits. The CPS was equal, in principle, to the difference (if greater than zero) between the EU-ETS price and the carbon floor price.

The CPS top-up started at £4.94 per tCO2 for 2013-14, nearly doubled to £9.55 per tCO2 in 2014-15 and again to £18.08 per tCO2 in 2015-16. At this point, the complaints from industries affected became too loud along with threats to close businesses, so the CPS top-up has been frozen at £18 per tCO2 since 2015.[2][3]

Over the 17 years since the EU-ETS was established in 2008, the annual average permit cost in the UK at 2025 prices, including the CPS, has varied over a range from £10.6 per tCO2 in 2012 £128.3 per tCO2 in 2022. In 2025 the annual average cost of permits was £77.3 per tCO2.

Under no coherent view of the world has the external cost of CO2 emissions varied by more than 12 times over a period of a decade. Even if we leave out the period 2021-2023 as affected by exceptional factors, the increase in real terms from 2012 to 2025 was more than 7 times. Such variations undermine any notion that the ETS provides a reasonable way of signalling the cost of the environmental externality associated with using either coal or gas for electricity generation. Instead, the EU-ETS and now the UK-ETS are simply a shambolic and dishonest tax on electricity use.

On top of these practical considerations, there are two key reasons why the ETS is fundamentally flawed. The first is that the economic theory of Pigouvian taxation outlined earlier is valid if and only if the corrective tax is the sole intervention designed to correct the externality. In other words, there should be no support for renewable generation. Patently that assumption does not hold. Energy policymakers believe that if one intervention to encourage low carbon generation is good, two interventions are better and five interventions are better still.

With an incoherent and ever-changing set of policies that are supposed to support the transition from reliance on fossil fuels to low carbon alternatives, the standard economic case for either a carbon tax or emission permit trading is simply wrong. That conclusion does not rule out the possibility that a, probably low, carbon tax might improve welfare, but the case has never been made and would be contingent on a whole range of empirical circumstances. As in the classic anecdote, the most likely conclusion is not to start from here.

The second reason is rather more subtle and explains why I described the ETS as a dishonest tax on electricity use. To explain the point, I will use a highly simplified example. Think of electricity system with 20 GW of gas plants and 20 GW of solar capacity. The solar plants receive a fixed annual payment and incur no variable operating costs. The gas plants vary in their heat rates, which is a measure of the amount of gas used per MWh of electricity generation. Their operating costs vary directly with their heat rate – see the Technical Note. The most efficient plants use 1.75 MWh (HHV) of gas per MWh of electricity, while the least efficient use 2.5 MWh (HHV) of gas per MWh of electricity.[4]

Gas plants must buy emission permits amounting to 0.18 tCO2 per MWh of gas used, so buying emission permits increases the effective cost of using gas to generate electricity. If the market price of gas is £30 per MWh and the market cost of emission permits is £60 per tCO2, the effective cost of gas is £30+0.18*60 = £40.8 per MWh. The operating cost for a plant with a heat rate of 2 is £81.6 per MWh of electricity. If the market cost of emission permits increases to £100 per tCO2, the effective cost of gas rises to £48 per MWh and the operating cost for the plant would increase to £96 per MWh of electricity.

If total demand is 10 GW and solar output is 11 GW, the system will rely entirely on solar generation. The market price will be zero, since any higher price would lead to solar plants wanting to supply more power than the system can use. On the other hand, if total demand is 10 GW and solar output is only 9 GW, the system will require output of 1 GW from gas plants. The market price will be equal to the operating cost of the gas plant whose efficiency ranking means that 1 GW of gas plants have lower operating costs and 9 GW of gas plants have higher operating costs. If that plant has a heat rate of 2, the market price of electricity must be £81.6 per MWh if the market cost of emission permits is £60 per tCO2 or £96 per MWh if the market cost of emission permits is £100 per MWh.

The point of this simple example is that the cost of buying emission permits is passed directly through to electricity users whenever it is necessary to use gas plants to meet demand for electricity. More generally, the price of CO2 permits only affects the market price if gas plants must run to meet demand, but it has no effect on either their output or the quantity of CO2 emissions. This is completely counter to the assumptions on which the model of taxing environmental externalities relies. If output and emissions are independent of the level of the CO2 permit price, neither a carbon tax nor an emission trading system serve any purpose other than to collect tax revenue.

While this example is very simple, it captures a critical feature of the GB electricity market. Low carbon modes of generation always run first, while their output is constrained by capacity and the availability of solar power, wind power or hydro resources. The price of ETS permits is directly passed through to market prices. It has no impact on the overall balance between low carbon generation and fossil fuel generation.[5]

Under such conditions the ETS permit system is, in effect, a pure tax on electricity use which has, at most, a minimal impact on the level of CO2 emissions in the short and medium term.[6] What effect it might have in the longer run depends on whether investment in low carbon generation and fossil fuel plants is affected by the expected level of the ETS permit price. Such an effect is likely to be very small. Almost all investments in low carbon generation rely on one or more of the various support mechanisms for renewable generation. What matters is getting a contract, not the expected future permit price.

The same is broadly true for investments in dispatchable generation, though it is the role of the capacity market that is critical in such cases. A somewhat baroque argument might be constructed to argue that the expected level of the ETS price could affect investment decisions for battery storage plants, but again these rely primarily on capacity market contracts. Given the path of ETS prices over the last decade, few investors would put much weight on guesses about what the annual average ETS price will be in 2030.

In summary, the current ETS is just a way of taxing electricity users. Variations in the number of permits made available and external circumstances lead to huge variations in the market price of permits. The price has a minimal effect on the balance between low carbon generation and fossil fuel generation. Both in theory and in practice, the argument that carbon taxes and emission trading offer a good way of promoting the transition from fossil fuels in the energy sector reflects a failure to understand the economics and inherent features of electricity markets.[7]

The case for abolishing the ETS as well as the Climate Change Levy is extremely strong. Economists and lobbyists may enjoy experimenting with toys that they do not understand, but this is not a good reason for keeping an arrangement that has a major impact on energy costs and economic activity without having any practical benefit in terms of reducing CO2 emissions.

[1] Anyone interested in the standard theory of economic externalities should refer to David Newbery’s chapter in a book that I co-edited with Geoff Heal – Public Policy and the Tax System, which as re-published as an e-book by Taylor & Francis in 2025.

[2] Most emission permits are auctioned fortnightly through each year. Some businesses, though not electricity generators, receive a free allocation of permits but free allocations are being phased out. The current government plans to reduce the total number of permits issued each year sharply up to 2030. In 2024 the total value of permits auctioned was £2.56 billion with an average auction price of £37.2 per tCO2. The revenue from auctioning permits accrues to the government along with the revenue from the CPS.

[3] As if the system were not complicated enough, there is a separate Climate Change Levy (CCL) which is charged to business users of energy other than electricity generators at a rate of £8.01 per MWh of electricity and gas. This is presented as a carbon tax, but it is simply a stealth tax on energy use designed to raise revenue.

[4] This is a simple way of capturing the idea of a merit order of generating plants by which the most efficient plants (with the lowest heat rates and low operating costs) have priority over less efficient plants (with higher heat rates and higher operating costs). Because of this advantage, plants with low heat rates will operate for more hours in the year than other plants with higher heat rates.

[5] The story would be more complicated if there were coal plants that compete with gas plants. That is not the case now in the GB market and many other European countries. When coal plants were still operating in the UK, they had much lower operating costs than gas plants, so that the cost of emission permits would have had to be extraordinarily high to change their ranking in the merit order.

[6] To be precise, the ETS is tax on electricity use when total use exceeds low carbon generation. In the GB market this is almost all the time. Since the amount of low carbon generation is essentially random, the incentive to transfer consumption from periods when low carbon generation is low to periods when it is high is minimal. In addition, since most suppliers rely on period average prices backed by complex hedging contracts, they must assume that demand will be concentrated in periods when gas generation is required.

[7] The increasing reliance on imports from Europe complicates the way in which the GB electricity market functions. However, it does not affect the essential point that the ETS increases the market price of electricity in periods when gas generation is required to meet total demand.

Technical Note: When using standard indicators of gas prices and the efficiency of gas power plants it is important to bear in mind the distinction between HHV (High Heat Value, based on Gross Calorific Value) and LHV (Low Heat Value, based on Net Calorific Value) measures of the heat content of natural gas. The difference between the two is the latent heat of the water vapour produced when gas is burned. The HHV assumes that water vapour is condensed to a liquid, while the LHV assumes that water vapour remains as vapour. The HHV of natural gas is usually assumed to be 10.8% higher than the LHV.

Gas prices in £ or € per MWh or $ per million Btu or pence per therm are almost always quoted on an HHV basis to reflect the total heat content of the gas. However, the conversion efficiency of gas power plants, turbines, and engines is usually quoted on an LHV basis as these systems tend to vent hot water vapour rather than condensing it. When the efficiency of a CCGT plant is quoted as 60%, that is on an LHV basis. To convert that to a heat rate which can used with the market price of gas as above, it is necessary to divide the efficiency by 1.108 (= 0.54) to convert to HHV and then take the reciprocal (= 1.85).

US sources tend to use heat rates expressed as Btu per kWh, because US gas prices are quoted in $ per million Btu. As an illustration, a heat rate of 7,000 Btu per kWh in US units translates to 2.05 MWh per MWh since 1 MWh = 3.412 million Btu.