A Tale of Two Electricity Forecasts
Regional grid operator PJM’s capacity auction is front-page news in the mid-Atlantic region. Conducted annually, this auction helps procure enough generation to serve total projected peak electricity demand a few years into the future, with some added cushion. Broader public interest in the subject stems from recent auction price increases that, in some states, are passed directly to consumers rather than internalized by market participants. These increases result from forecasted demand that in some cases defies logic and outstrips available or anticipated supply.
As one might imagine, assumptions used in PJM’s load forecast directly affect its capacity auction. PJM’s forecast is based on economic measures like local industrial activity within defined geographic areas and the adoption of electric vehicles, but is recently most impacted by utilities’ near-term additions of new demand in the form of large loads—mostly data centers. Consumer groups and FERC commissioners have questioned these forecasting processes, and PJM implemented a new procedure last year seeking some minimum level of commitment from these large loads before including them in the short-term (next three years) load forecast used for the capacity auction.
Just yesterday, PJM issued a new load forecast that shows lower expected demand in the next few years than previously anticipated. The immediate takeaway is that their check on financial commitment is working. PJM’s process requires some level of activity or financial commitment of new large loads in order for them to include the projects in the forecast. If no activity or commitment is provided, then a project’s presumed in-service dates are pushed beyond the immediate three-year window. PJM’s summer load forecast for the 27/28 delivery year (June 1, 2027-May 31, 2028) was 4,400 megawatts (MWs) lower than last year’s forecast for the same period. This is notable for a number of reasons, one of which is that the initial auction for the 27/28 delivery year came up 6,600 MWs short of supply. While that auction focused largely on meeting winter peak, the decrease in the forecast will have a similar impact throughout the year.
The new forecast also has nearer-term implications. For instance, PJM’s new rules reduced the summer peak forecast for the next delivery year (starting June 1, 2026) by nearly 2,500 MWs. This reduction in peak demand—coupled with changes in other planning parameters—also served to increase available excess supply for the upcoming delivery year. If the same impacts from the new forecast and changed planning parameters similarly impact expectations for the 27/28 delivery year, it is not unreasonable to expect a previous deficit turning into a surplus by 2027. This calls into question those claiming a crisis.
Based on more and more certain data, these changing realities also change the prevailing narrative. While the new forecast certainly evidences load growth over the next few years, the timing and magnitude of that growth are becoming clearer. While many were worried about the ability of new supply to meet the near-term demand, more time to plan and complete that supply will be welcomed. These changes also provide fuel for more efforts to get better information. PJM should consider bolstering their checks on utilities’ large-load additions, and state and federal regulators should make a formal effort to engage each other on the topic and ensure they understand how to facilitate reasonable and dependable load forecasting. Finally, restructured utilities have pointed to recent capacity auction results as their primary argument for seeking state approval to start owning and building generation again. State policymakers will likely use these recent updated forecasts (which call past utility assumptions into question) to reconsider their support for such proposals.