Octopus Agile: A Winter’s Tale

Octopus Energy offers UK domestic customers like me a tariff, Agile, that varies each half hour based on the wholesale electricity price. I am a member of a Facebook group and a forum that regularly discuss this tariff, and I have found it fascinating watching the reaction to December 2020 / January 2021’s record high prices: many have found it a nasty shock after the low prices back in spring/summer 2020.

I decided to write this blog post offering some reflections from this period.

Background

The best comparator for Agile is Octopus Energy’s static time-of-use tariff, Go, which offers a cheap price of 5p/kWh from 00:30–04:30, and a more expensive price the rest of the day of between 12.92 and 14.71p/kWh, depending on the region. Go is extremely good value, as even the expensive times of the day are cheaper than Octopus’s standard fixed rate. Octopus also offer some variants on Go, labelled Go Faster, which allow customers to change their cheap hours (for more details on these, visit this page).

How much Agile has cost

In this first chart I show what Agile and Go would have cost two customers each day over the past year, one using electricity solely during the 4 cheapest hours of each day, and another using electricity solely during the 4 most expensive hours each day. It shows that customers who use a lot of electricity in the most cheapest periods of each day, say Electric Vehicle owners, will have been paying less on Agile for most of the year, and only a couple of pence more in Dec 2020 / Jan 2021. On the other hand, Agile customers who use a lot of electricity in the most expensive periods of the day will have been paying more for most of the year, and significantly more in Dec 2020 / Jan 2021.

The chart above considered only the cheapest four hours and the most expensive four hours, irrespective of when those hours occurred. We might instead split the day into three periods: 00:30–04:30 (the Go cheap period), 16:00–20:00 (the Agile expensive period, plus an extra hour that also tends to be expensive), and the rest of the day. The table below shows how these compared to Go prices for November, December and January (again for London). Customers that were dependent on 00:30–4:30 or 16:00–20:00 have fared worse on Agile, while those that use most of their electricity in the other 16 hours are much better off on Agile.

Next, some customers have additional flexibility allowing them to benefit some more from Agile. Even over the higher priced periods, some days have been considerably cheaper than others, sometimes even going negative. Customers that can take maximum advantage of these occasions are likely to lower their average price even further.

There is also some regional disparity between the different tariffs. In the chart below I show average Agile prices and Go prices for a customer that consumed evenly over the 18th of December. In London Agile prices were lower than Go, while in neighbouring South East England they were slightly higher. In North Scotland Agile prices were significantly higher.

Finally, cost depends not just on the average price but also on the total amount consumed. Some customers don’t consume much electricity, so higher prices in December/January won’t lead to a material impact on bills. Other customers, especially those with electric heating and electric vehicles, will have saved a substantial amount on Go relative to Agile.

Is it alright to switch?

Some Agile fans consider it cheating, or at least outside the spirit of Agile, to switch when prices are high. While I don’t think Agile customers should complain or feel cheated when prices are high, I don’t believe Octopus want customers to suffer unnecessarily. I believe a lot of customers that joined Agile in 2020, lured by the low prices, didn’t really understand the risk, and are probably better off on a tariff with more certainty, and I am glad that Octopus allows this.

What does worry me a little is the many customers switching to Go, with the intention of returning in the spring when Agile prices fall. Go is a 12 month contract, with prices most likely set so that customers underpay in winter and overpay in summer. By switching to Go for the winter, and Agile for the summer, customers get the best of both worlds, and undermine the sustainability of these tariffs. Octopus don’t seem too worried, so I probably shouldn’t be either.

That said, I see tariffs like Agile as essential to our transition to net zero, and it strikes me as suboptimal if many customers feel the need to switch off it for 3–6 months each year. Perhaps Octopus will alter its Go tariff to having different summer and winter prices, reducing the benefit of switching. Or perhaps they will come up with insurance add-ons that allow Agile customers to reduce their uncertainty or smooth out the seasonality without switching away.

Decision making under uncertainty

While I agree that we never know for certain what will happen, I think we can draw some reasonable expectations. For example, prices in summer are likely to be low than prices now. Short term prices are ‘noisy’ and do tend to mean revert. But some shifts are real and more persistent. It is difficult for a customer to observe the price movements, and identify which are short-term and which are likely to persist. I look at futures markets to see what more knowledgeable market participants think. For example, I have constructed the following chart showing what I’d expect average Agile prices to be over the next 14 months, derived from the EEX futures curve in late January. But I also think about the fundamentals, and am aware that there is plenty that could happen to push these prices up or down.

Conclusion

I don’t blame the customers for making this decision. But, given the important contribution of Agile to Britain’s transition to net zero, it would be great to see innovations that would allow customers to feel confident remaining on the tariff, without eliminating its benefits.

Fascinated by what makes societies and markets work, especially in sustainable energy. http://guylipman.com.