A retail market for short-term electricity in the UK

Guy Lipman
8 min readMay 1, 2019

How retail markets could allow customers to optimise their consumption, reducing strain on the network.

When I talk to people about how the way we relate to electricity needs to change, I often describe the example of electric cars. In one possible scenario, everyone would get home at 6pm, and plug in their cars, creating massive strain on the network and raising prices for everyone. In a much nicer scenario, everyone would get home at 6pm, and plug their cars into a smart charger. Each car would then charge at some point over the night. This would not just avoid strain on the network but would actively counterbalance the variation in other supply and demand, reducing the need for extra batteries and fossil-fuel generators. A big part of my research involves thinking about ways to ensure we achieve the second scenario and not the first.

One option is to collect everyone’s preferences, and to optimise centrally. This is extremely challenging. Everyone will have different preferences, and they won’t all be defined in the same way. For example, one person might want the car to be fully charged as soon as possible, while another might be happy with it being 25% full, knowing that they weren’t planning on driving far until the weekend. People need the ability to change their minds as they go. And we need a way to charge fairly, without giving them an incentive to game the system by submitting tactical preferences.

I would therefore prefer a market to determine who gets electricity when. Customers (or perhaps their devices) could observe the prices for different times, and decide when it was optimal for them to consume. People that were prepared to pay more could get electricity at their choice of time, but they’d have to pay more (effectively subsidising everyone else), which I think is fair.

Octopus Agile: a first step

Octopus’s time-of-use tariff Agile is a first attempt at such an offering. Every day at 4pm, its customers are informed of the price for each half hour the following day. Customers then schedule their electric vehicle charging (or in my case, hot water heating), deciding how much and when to consume. Since moving to this deal (which I wrote about here), I estimate that I’ve almost halved my bills, and my demand profile changes dramatically from day to day to take advantage of the prices. Octopus has subsequently gone a step further, allowing customers with solar power to sell their excess electricity at the time-of-use tariff, which I see as a one of the signs of a liquid price.

The reason I call this a first attempt is its relatively simple structure — prices are fixed at a single point in time (4pm the day ahead), and customers have an option to use as much or as little as they like at that price. The two weaknesses of this to the customer are:

  • customers have no ability to lock in prices before that point (or even to see what prices are expected to be)
  • customers have no incentive or ability to gain from changes to their consumption should market prices subsequently change

I must acknowledge that Octopus don’t benefit from these weaknesses, and in fact take a risk as customers can use any amount of power at that price.

Could retailers offer more dynamic pricing?

A better model might be if instead, a retailer provided a listing, with a price for every half hour for the next week. These prices would update in real time, perhaps even up to the start of the half hour in question. At any point, I could lock in prices for specific volumes of consumption. And at any later point, I could decide not to use the power in that half hour (i.e. selling it back) and instead locking in the price for a different half hour.

So, imagine I own an electric vehicle, and I need 6 half-hours of charging by 8am (a total of 24kWh). I could get home at 6pm, and look at the prices for each half hour. Let’s say they were as in the left column below. I would pick the cheapest 6 half hours (highlighted yellow), giving me an average cost of 9.35p/kWh, and a total cost of £2.24.

But suppose by midnight, prices for the remaining hours had changed, such that they matched the column on the right. I could just leave my schedule unchanged, and I’d still pay the same £2.24. But I could reschedule and save myself some money. By selling back 1am-3am at an average cost of 11.55p/kWh, and buying 3–3:30, 4–4:30, 5–6 at an average cost of 11.10p/kWh, I would earn back £0.07. This isn’t a massive amount of money, but over time it adds up, and there’s no reason why a smart charger couldn’t automate this process.

Alternatively, imagine that at midnight you received a phone call, and you needed the car to be fully charged as soon as possible. You could reschedule for it to charge from midnight to 3am, paying a bit more but able to go where you needed.

You might also have locked in to pay 9.35p/kWh, based on the prices at 6pm, but not think it is worth paying anything over 10p/kWh. In that case, when you saw the higher prices at midnight, you might decide to sell back the electricity. In this case, you would sell back the electricity for £2.66, making yourself 41p profit.

Also, going a step further, once you could see the prices for the full week, you would be able to make sensible choices over the next week. If prices one night rose, you might decide to charge a bit less that night, and a bit more the following night. The power would be in your hands!

Finally, I should make clear that I don’t expect most customers to optimise in this way for most of their consumption — for example, planning the best time to switch on the bathroom light in the middle of the night. My guess is that people will mainly optimise for electric vehicle charging, water and space heating, and maybe their washing machines. For anything else, they’d prefer to accept a fixed price. But I believe that even having just these devices coordinating according to the market prices would make for a more stable electricity market.

What about a Retail Exchange?

In the scenario above, I imagined a retailer offering regularly updating prices for each half hour up to a week ahead. However, I’m not sure that retailers would necessarily want the administrative burden and risk of offering this kind of product. They might charge a transaction fee, perhaps a bid/offer spread to recoup some of their costs. But any retailer innovative to take this on might have too few customers to justify the cost. In particular, I don’t underestimate the effort in getting devices able to read prices and ‘trade’ with the retailer.

I therefore thought about how we could create an exchange that customers from any retailer could use. The solution I came up with was a retail exchange. Whereas in the retailer example above, the retailer was providing bids and offers for customers to accept, here we would allow multiple market-makers to list bids/offers. For example, a battery owner might be happy to buy power for below 6p/kWh, or sell it for above 8p/kWh. Customers (or their devices) could then see the best bids and offers, and make decisions on that basis.

I suspect it would become extremely messy if your counterparty on this exchange was responsible for supplying you with your electricity. For example, if I bought electricity 6am from ABC and later sold it back to DEF, the grid settlement agent (Elexon in the UK) would need to track both of these. Perhaps if we knew we had the same retailer and distribution companies, it might work, but that seems limiting.

Instead, I believe it would be simpler to make this exchange offer purely financial transactions, with buyers paying the agreed price and receiving a floating index price. This is sometimes called a Contract For Difference (CFD). Customers could then buy their electricity from their own retailer at that floating index price, meaning that the net cost to them was the price they had agreed on the exchange.

The floating price that makes most sense to me is the Elexon imbalance price, which is what retailers are ultimately charged for their customers’ electricity (plus distribution charges and environmental levies). So, if my retailer agree to charge me for electric vehicle charging and my heating at the Elexon imbalance price plus a fixed margin, I could effectively optimise my consumption using the retail exchange. I must note however that the imbalance price is quite volatile, ranging last year from -15p/kWh to 99p/kWh! (For more information about the imbalance price, see this post.)

Some additional details

Credit risk

Whenever you have counterparties trading, you have to worry about whether they’ll meet their end of the bargain. On an exchange, that credit risk is usually handled centrally, avoiding the need for participants to care who they’re trading against.

On this exchange, credit risk is reduced by the fact that participants would only be able to trade up to a week ahead, but there would still be quite a bit of risk due to the volatility of the imbalance prices. This could be managed by requiring participants to post collateral. There might also be a way in which retailers provided guarantees for their customers’ trading.

Linking to wholesale markets

UK markets offer the opportunity for wholesale market participants to buy or sell electricity in wholesale markets, which ultimately settles against the imbalance market. The most traded of these is the 11am day ahead auction, however EPEX SPOT also offers the ability to trade individual half-hours at any point between the day ahead auction and about an hour ahead of the half hour in question. Finally, there are some opportunities to trade blocks of hours a few days out (though this is not often traded).

If the retail market became out of line with the wholesale markets, there would be opportunities for wholesale participants to trade in the two markets, essentially bringing them back into line.

Linked transactions

In its simplest version, the exchange would allow bids and offers for single half hours, and customers would accept or reject them. A challenge of exchanges is to avoid partial execution of trades. For example, consider the example above, where the electric vehicle owner changed their charging schedule based on prices at midnight. Customers interacting with an exchange might worry about their sell trades being accepted, but then the market moving before the buy trades were cleared. There would need to be a way to ensure that the trades could all be executed before the markets moved.

In addition, I believe there would be considerable value in allowing compound products. For example, a battery owner might offer someone the ability to buy only one half hour of several, or even to simultaneously buy one half hour and sell another. I recognise that offering these products would add complexity, but I believe they would be worthwhile additions.


As well as Contract for Difference transactions, there would be merit in allowing customers to purchase options — for example, ensuring that they were compensated for prices above a certain level. Octopus Agile include a cap above which customers are protected from price rises — this seems a good idea. Clearly there are challenges in determining how options affect the overall level of collateral (especially for the seller of the option), but these products could prove popular.

I hope this gives some food for thought, for electricity retailers, market participants, and even device manufacturers — I do believe that something along these lines makes sense and is ultimately inevitable. As always, if anyone has had thoughts on this topic, I’d be very happy to discuss.



Guy Lipman

Fascinated by what makes societies and markets work, especially in sustainable energy. http://guylipman.com. Views not necessarily reflect those of my employer.