Plan ahead, they always say. Turns out that might not always be the best advice: thanks in part to a two-year delay, grid operator PJM just saved consumers in the mid-Atlantic $4.2 billion. That works out to around $100 per household in the 65-million-person region. The reason is simple: PJM usually arranges its power supply three years in advance, and always overestimates how much power the grid will need. But, because of some legal squabbles, the auction for 2022 took place just last week, two years later than planned. Waiting two years gave PJM a much more accurate view of what the grid needed, and so it purchased about 2,500 MW less. That 1.5% drop in demand led directly to the $4.2 billion drop in what it paid for guaranteed capacity.
The Sustainable FERC Project has been calling attention to PJM’s chronic overprocurement for years. In 2020, NRDC and Sierra Club commissioned economist James Wilson to take a close look at this problem, and he found that fixing overprocurement could initially save as much as $4.4 billion per year. Now, to be sure, there are a lot of moving parts in PJM’s markets, and this isn’t a pure test of that prediction. PJM also partially implemented another of Wilson’s money-saving recommendations and reduced the price it will pay for capacity to follow improvements in technology. Still, it’s hard to miss that an auction run with a more accurate load forecast and pricing nearly identically matches Wilson’s prediction.
These billions in unnecessary spending hurt more than peoples’ pocketbooks. PJM is rife with old coal-fired plants that are huge carbon emitters. A recent report by think-tank RMI (Rocky Mountain Institute) shows that many of these coal plants are both unneeded and hanging on by an economic thread. The billions of dollars in excess capacity payments help keep those plants alive. Thanks to the more accurate forecast in this most recent auction, more than 8,000 MW of coal plants didn’t clear. That’s about 10 good-sized power plants, which now face a greater chance of shuttering.
To permanently shut down these surplus coal plants, we need to make sure PJM doesn’t go back to the old “buy too much” approach. Over the next few years, PJM will work off the backlog in auctions and get back to running them three years in advance. As it does that, it must make sure that over procurement due to forecast errors doesn’t creep back in. There are a few things PJM can do to prevent this:
- Better forecasts. The simplest thing is to improve the load forecast. Of course, this is a bit like telling someone who lost a race that they need to run faster—everyone knows it, and we have to assume PJM’s planning staff is trying as hard as they can. Still, when your prediction misses one way every single year, it’s time to get more independent voices into the process. But improving the forecast can’t be enough. Forecasts will always be uncertain, and PJM should also work on ways to manage that uncertainty.
- Don’t buy everything up front. Right now, PJM’s procurement strategy is basically to buy what it thinks is needed three years in advance, then throw out what turns out to be excess. That’s tremendously inefficient and can easily be improved. For example, up until 2017, PJM bought 97.5% of expected need three years in advance, then waited to get the last bit. Nearly every time, it ended up not needing it, and this tiny holdback saved (again) billions of dollars. Reinstating some version of the holdback is a proven move that will end a huge unintended subsidy for some of the dirtiest power in the country. Some analysts have gone further and suggested that it’s better to hold the entire auction less than three years in advance. PJM should look carefully at these recommendations and commit to an auction design that more efficiently preserves reliability.
- Shift the risk away from consumers. As things stand now, the general public pays for forecast error—when PJM buys too much, the cost shows up in electricity bills. This is a primitive approach. The use of financial instruments to manage future risk is well established and is commonly used for nearly every other commodity in the world. The basic idea is simple: those most willing to take on risks do so and get paid for it. In the electricity world, this means a factory that can shut down production once in a while or a mothballed power plant that can come back into service on a few months’ notice could commit to doing so if needed and get paid for that commitment. If it turns out they’re not needed, they’re released from the commitment. If it turns out that they are needed, then they’re held to their commitments. Of course, derivatives markets get more complex fast, and need to be carefully managed. Still, allowing limited financial trading should both give us more accurate capacity prices and let flexible loads and suppliers manage forecast risk. This could save, you guessed it, billions every year, so it’s worth the effort to get it right.
Last week’s auction results show that PJM’s overprocurement is keeping surplus dirty power plants alive at enormous cost. Now that we know it’s possible to run a leaner, more efficient capacity market, it’s time to lock in those gains. PJM must act now to avoid going back to the excess capacity procurements we’ve seen for so many years.