PJM is at it again. The final version of the 2014 load forecast was published in February. PJM is forecasting a 4.4% load increase in 2017. The forecast covers the year three years in the future, because the load forecast serves as the basis for PJM’s capacity auctions held every May. The main auction involves generation capacity for delivery three years in advance. This year’s 2014 auction is for capacity to be provided in 2017.
As I pointed out in my comments about the draft forecast PJM published in December, PJM has never provided a load forecast that has proven to be accurate by actual experience. Every year, PJM predicts peak load growth, and every year, they revise projections downward. In past posts, I have shown samples of their load forecast graphs in which actual demand is flat or falling, yet PJM always projects future growth. As actual figures fail to meet PJM’s predictions, PJM forecasters revise the next years forecast downward, but insist on growth rates that are still too high. It’s been that way since I started covering PJM’s forecasts in 2010.
I guess it’s not surprising that PJM forecasters always err on the side of over-estimating future load. Their job is to make sure there is enough electricity to serve their region three years in advance. Before the capacity auctions every year, however, PJM adds a reserve margin to predicted load to ensure that level of caution. If PJM forecasters are also including a safety margin in their load forecast, they should be explicit about that. The RTO’s load forecasts are used for all kinds of planning beyond the capacity market. They should be as accurate as any forecasts can be.
A big part of PJM’s problem is the load forecast’s reliance on Moody’s economic forecasts. These forecasts of economic activity drive assumptions about predicted load. Moody’s forecast is always upbeat. A new upturn is always just around the corner, and the Moody’s forecast used in PJM’s 2014 predictions is no different, despite the stagnation that has characterized the US economy for the past 6 years. Perhaps PJM should rethink its reliance on a company that throughout the 2000s gladly slapped AAA ratings on bogus securities backed by shaky mortgage loans. But Moody’s rosy assumptions allow PJM to inflate load growth while maintaining the rationale that they were just listening to the “experts.”
This sentence, from the 2014 load forecast, has appeared in one form or another in every executive summary of every PJM load forecast I have read since 2010:
Revisions to historical economic data and the addition of another year of load experience to the model resulted in generally lower peak and energy forecasts in this year’s report, compared to the same year in last year’s report.
Yup. Another year, another downward revision of load forecast, and another admission that PJM’s forecasting has failed again.