Saturday, April 18, 2020

Designing Future Electricity Markets: Evolution Rather than Revolution

We have a new working paper on designing wholesale electricity markets with high levels of zero marginal cost intermittent energy sources. The paper reports on the findings of a session on this topic at the Future of Electricity Markets Summit that was held in November last year in Sydney. Gordon Leslie led the writing with contributions from myself, Akshay Shanker, and Mike Hogan. The session itself featured papers by Paul Simshauser, Mike Hogan, Chandra Krishnamurthy (a paper coauthored with Akshay and myself), and Simon Wilkie.

Imagine the future where fossil fuel generation has disappeared and generation consists of intermittent sources like solar and wind. The cost of producing more electricity with these technologies is essentially zero – the marginal cost of production is zero – once we have invested in the solar panels or wind turbines. But that investment is costly. How would markets for electricity work in this situation? The naive view is that introductory economics tells us that the price in competitive markets, like many wholesale electricity markets (e.g. Australia's National Electricity Market – the "NEM"), is equal to marginal cost. If renewables have zero marginal cost, then the price would be zero. So, an "energy only market" where generators are paid for the electricity they provide to the grid won't be viable and we need a total redesign.


But the consensus in our session was that zero marginal cost generation doesn't mean that energy only markets can't work. Marginal cost of production isn't as important as marginal opportunity cost. Hydropower plants have a near zero marginal cost of production but if they release water when the price of electricity is low, they forego generating electricity when the price is high. Releasing water has an opportunity cost. Mike Hogan pointed out that the same is true of reserves more generally. There will also be electricity storage including pumped hydropower as well as batteries and other technologies in the future. The Krishnamurthy paper focused in particular on the role of storage. The price these receive should also be determined by opportunity cost. Even without storage and reserves there should be a market equilibrium with a non-zero price as long as there is sufficient flexibility of demand for electricity:

Intermittent renewables supply to the market however much power, Rt, that they can generate. They can't supply unlimited power at zero marginal cost... The market clearing price is p, where the demand curve, Dt, crosses the vertical supply curve. Price does not equal marginal cost of production. Of course, the quantity and price is continually fluctuating. But, if, on average, the revenues are insufficient to cover the costs of investment, generation capacity will shut down and exit the market until long-run equilibrium is restored. Storage has the effect of making the supply curve more elastic like the St supply curve in this graph:

In this example, storage lowers the price and increases the quantity of electricity consumed. This is when storage is discharging electricity. When it is charging, it will raise the price relative to the case with no storage. Storage has the effect of reducing volatility.

This is of course very simplified. The real world is more complicated. The paper shows that increasing intermittent generator penetration increases the importance of adequately pricing scarcity and all network constraints and services. Wind generation tends to colocate in the best wind resource areas overloading the gird. Locational marginal pricing pricing is required to deliver investment incentives for the right technologies to locate at the right locations to efficiently maintain a stable and reliable electrical network as discussed by Katzen and Leslie (2020). The NEM currently only has a spot market that schedules generation for each 5 minute interval. A day-ahead market could help in valuing the provision of flexible generation and storage. The current charges for transmission costs in the NEM also disincentivize grid-scale storage. Electricity users pay for these costs. This makes sense when the user is an end-user. But storage is treated as a customer and also pays these charges.  So "fresh electricity" where only one set of charges is paid has a cost advantage over stored electricity.

In conclusion, electricity markets need to evolve to provide the correct incentives to generation and storage. A total rethink isn't needed.