There are still trade-offs in energy policy

How to make power cheaper

The problem with debates over Britain’s energy mix is that they’re not just about Britain’s energy mix. They have instead become proxy debates over the desirability of decarbonising our economy.

People in general do not like to accept that there are trade-offs. Boris Johnson’s most popular stance was his stance on cake: pro-having it and pro-eating it. Now, sometimes trade-offs are illusory. Building new homes near railway stations in commuting distance of the best jobs for instance is not just good for the environment because it means fewer trips in cars running on fossil fuels. It is good for growth too because it allows more people to access higher paying more productive jobs.

Yet trade-offs do exist on climate change. Airport expansion undoubtedly increases emissions, but it also promotes trade and growth. In fact, air travel is so important for economic growth (amongst other reasons) that it ought to be one of the last things we try to reduce on carbon grounds.

What about electricity? Is the shift to renewables like densifying our most productive cities or more like blocking airport expansion? Some argue the electricity debate has been solved and that electricity is more like the former than the latter.

On Question Time the other week, Octopus Energy’s CEO Greg Jackson told the audience: “We've crossed the rubicon. Clean energy is now cheaper than fossil fuels. Power from wind and solar is cheaper than power from fossil fuels.”

Our Energy Secretary Ed Miliband says something similar. We used to talk about “an ‘energy trilemma’ between security, affordability, and sustainability. But now the signs all point in the same direction - the sprint for clean power.”

Yet when the National Energy Systems Operator gave its assessment of Ed’s mission of “Clean Power by 2030” (defined as gas making up less than 5% of total power generation in a normal year), they found that while achievable (if everything goes right) it would put up costs unless gas prices spiked to around 100p per therm on top of a substantially higher carbon price (£148 per tonne).

There is a lot to recommend about setting an ambitious target. The last Government’s stretch targets for vaccinations and testing meant they didn’t accept ‘business as usual’ and instead eliminated unnecessary red tape like the multiple rounds of irrelevant e-training required to become a volunteer vaccinator.

Clean Power by 2030 necessitates radical reform to planning, habitats regulations, and judicial review. Already, Ed Miliband has been deciding on planning approvals at pace – approving three projects in his first week. Reforms that would be good whether the target was Clean Power by 2030 or Clean Power by 2050 are now firmly on the table.

But there are risks too.

1. Timing: The cost of building a wind turbine is itself a function of energy costs (all that steel requires power and heat) and of interest rates (unlike gas, the cost of building out renewables is almost all up-front). Both are expensive right now, but expected to fall in the long term. Go too fast and we risk locking in high prices.

2. Competition: Renewables are procured (primarily) through reverse auctions for Contracts for Difference. In essence, these contracts pay out at a fixed rate regardless of what the actual price for power is at that time. Historically, such auctions are oversubscribed forcing developers to bid low prices. However, the last auction wasn’t. Nine out of ten bids were accepted. This creates a risk of market power and efficient developers putting in high-ball bids knowing that there’s a good chance they’ll make it through anyway.

3. Delaying Reform: Reading NESO’s report I was struck by the sheer scale of everything that had to go right. Failing on any single aspect, and there are a staggering number of moving parts, will mean missing the target. This makes putting off difficult, but ultimately cost-saving reforms, the path of least resistance. Locational pricing, for example, can cut bills and end the perverse system of paying wind farms in Scotland to turn off because the grid can’t bring it down to England. But locational pricing might lead to some investment decisions being reconsidered and others pushed back, so a reform that’s necessary to cut the cost of a renewable grid in the long-term risks being pushed back to hit the 2030 target.

4. Deprioritizing Nuclear: Renewables have a big role to play in a net zero world, but they’re intermittent and they’re not well-suited to meeting the needs of the growth industry of the next decade: AI. Reliable nuclear baseload power will be necessary to underpin the grid and reduce our reliance on expensive (or dirty) flexible tech when the wind isn’t blowing and the sun isn’t shining.

5. Making the Harder Stuff even Harder: Greening our electricity mix is the easy bit of net zero. France already did it once with nuclear. We have a pretty good idea of what is required for a grid that’s 95% clean. The tricky bit is the next stage. Electrifying heat and transport. If we want people to voluntarily switch to heat pumps and EVs then we’ll need to make them the cheaper option. Costs will fall for both as we install more, which will help, but they need to be cheap to run too. If electricity costs are higher than necessary we risk discouraging consumers from switching over.

To circle back for a second, it is important to understand why costs might be higher under an 80% renewable system.

When advocates of a rapid renewable rollout say ‘renewables are cheaper than fossil fuels’ what they’re referring to is the levelised cost of energy.

This is a metric which looks at the total lifecycle costs of producing one megawatt hour of renewable energy. Renewables, like nuclear, have high up-front capital costs, but near zero marginal running costs.

There’s a problem. In practice, we don’t actually care about the levelized cost of energy. Energy demand varies throughout the day and year. When it’s very windy, wind power can produce more power than we can use. In such situations, we actually have to pay for wind turbines to switch off. All of those extra units of wind power are less than worthless. And the opposite is true too. When the wind isn’t blowing, we have to pay for expensive backup generation. Sometimes that’s cheap (and getting cheaper) battery power charged during those wind peaks, but other times it’s expensive gas power plants being run at inefficiently low capacities. In our Clean Power 2030 future, it will probably mean using more expensive and untested forms of Long-Duration Energy Storage (LDES) such as Green Hydrogen.

As an aside, LDES is absurdly important and yet woefully under-invested in. I often think about this stat from Institute for Progress’s Brian Potter “between 2001 and 2021, only $1.8 billion was privately invested in LDES startups. By comparison, more than $33 billion was invested in blockchain and crypto startups in 2021 alone.” 

Not only does LCOE fail to account for the fact that the value of energy is different at different times. It also fails to account for the system costs of renewable energy. 

1. When renewables are producing more power than we can feasibly use, they’re paid to switch off. These are known as constraint payments. NESO’s model has them reaching close to £7bn per year by 2030.

2. Because renewables often need to be topped up by alternative sources such as batteries or gas generation, you need to factor in the costs of building and maintaining this infrastructure. (And any efficiency loss associated with it.)

3. Unlike the old model of a single power station providing large amounts of power for a region, power in a mostly renewable grid is radically decentralised. This requires a massive increase, what’s known as the ‘great grid upgrade’, to get power to where it’s needed.

The problem with finding an alternative to LCOE, which to its credit is a decent way to look at how costs have fallen/increased for a given technology, is that system costs aren’t fixed.

The more we invest in transmission, the less we have to worry about constraints or backup. In theory, we could go fully solar powered if we built enough connections to somewhere like North Africa where it’s almost always sunny. This isn’t purely theoretical either. There’s actually a proposal, known as XLinks, to build a cable from Morocco to Britain. And we already have multiple interconnectors with France, Denmark, and Norway.

The ‘great grid upgrade’ itself is set to dramatically cut our constraint payments down. Now, building all those wires and towers from Scotland and across the East of England won’t be cheap. But it’s a cost that can be spread out over decades through transmission charges. 

Market arrangements too can change things. If the UK had a locational pricing model for power, it’d encourage demand to shift to where the power was and therefore reduce the amount of power that’s constrained and the amount of wires we need to build.

How flexible our demand is also matters too. On the same episode of Question Time, Greg Jackson made the point that an electric vehicle (EV) could hold enough power for the average household for the best part of a week. Like EVs, heat pumps can be programmed to shift demand to when power is cheap and conserve it when it's expensive. Though there are limits here – people will still want to drive their car during times when power is scarce.

And planning can affect things too. If NIMBY objectors to new power lines are listened to, then we may end up burying new transmission lines underground. This is not only more expensive in the first place, but also makes maintenance a nightmare - increasing costs over the long term. It’s not that environmentally friendly either.

The systems costs of renewables also depend on how much of the grid they make up. Put simply, when renewables are at very low shares, pretty much every unit they produce translates to less gas burnt. When they’re dominating the grid, we have to pay the costs of keeping gas backup online even though it’ll rarely be used.

You can look at the OBR’s own estimates of ‘Enhanced Levelised Costs of Energy’ below. But what’s striking is how many uncertain variables make up this figure and how quick they are to change. If they were to publish this graph again today based on recent CfD auctions (which gave a higher price for renewables than at previous auctions) and current lower gas prices, the figures look much more favourable to gas. 

Of course, that might change. Conflict between China and Taiwan or escalation in the Middle East could create another gas price spike. Likewise, most estimates of wind and solar prices have been made to look foolish in rapid time. Look at this OfGem graph of DESNZ (then BEIS) forecasts of CfD prices. Maybe solar and wind will defy the forecasts again and get even cheaper. Then again, maybe they won’t.

Lower than expected costs have translated to higher than expected investment in new renewable capacities. The International Energy Association has upgraded their forecasts of how much new renewable capacity will be installed pretty much every single year this millennium.

While it may shelter us from extremes like the spike in 2022 after Russia invaded Ukraine, there’s no guarantee that moving fast to a clean power system will reduce costs as some imply. But, what’s crucial to understand is that what determines the costs of a mostly renewable system is far from fixed.

We can bring those costs down by:

• Keeping CfD auctions competitive to avoid market power pushing up strike prices.

• Reforming energy markets (including CfDs) to reflect local constraints on our grid.

• Reforming our planning system so new infrastructure projects can be delivered faster and cheaper (e.g. without expensive and badly-targeted environmental mitigations.).

• Allowing investment in the grid to avoid network constraints leading to sky-high payments to wind turbines to switch off.

• Extending our existing nuclear power stations lifespan as much as feasible to provide baseload energy.

• Reforming nuclear regulation, investing in SMR development, and creating special compute zones to avoid AI demand swamping our grid.