The US energy information body's projections show us how policy makes better futures

In September 2022, the United States Energy Information Administration (EIA) released its 'retrospective ' analysis. This takes the 'reference' or baseline projections of various different variables in the US energy system (like price, production, or electricity generation) from each year's report (going back to 1993 for some, and 2006 for others), and compares them to what actually happened.

I love this stuff. I think comparing forecasts from the past to actuals tells a fascinating story of how our assumptions have changed over time, and perhaps can provide some insight into how today's visions of the future might be wrong, or right.

So - here's a very quick little dive into the data, which I've visualised by comparing the forecasts to what actually happened. Unfortunately, they don't include the post-2022 future forecasts in the dataset, so we can't really see what's being forecast. But as a look back into the past's visions of the future present, it's interesting.

Things that were underestimated

For quite some time, the EIA was predicted in their 'reference' cases that without any policy intervention, wind, solar and gas generation would level off. Interestingly enough, after 2016, their forecasts have actually been relatively accurate and sometimes even over-estimating wind and solar deployment.

While this often gets interpreted as crotchety old forecasters skeptical of wind and solar, such as the case with International Energy Agency forecasts, I think it simply shows that without the intervention of governments, the status quo will indeed remain.

However, the era of cheap gas and fossil fuel subsidies has also resulted in a significant growth of gas-fired power generation in the US.

As a consequence, the burning of fossil gas for power on the US' grids produces more carbon dioxide emissions than burning coal. While it has a lower emissions intensity than coal, that benefit is lost when you rapidly expand reliance on gas. And of course, 'intensity' doesn't count for much when the atmosphere itself responds to the absolute volume of added greenhouse gases.

It's a good illustration of how locking in long-term reliance on fossil fuels for a short-term hit on emissions reductions is a terrible pathway for climate action.

Also, as is highlighted in the 'Net Zero America' study by Jenkins et al, you don't actually need a very high volume of gas usage at all to help balance large volumes of variable wind and solar on the power grid - that share decreases rapidly over the coming years as wind and solar expand.

It's also interesting to note that oil production in the US was frequently forecast to level off, but saw a huge rise. The forecasts corrected but...then, COVID happened.

Generally, electricity prices have been underestimated, though the extent to which depends on whether you use a 'constant' or 'nominal' approach.

Things that were overestimated

Coal power and production were both constantly overestimated, even after the implementation of renewable tax credits. 2021 was the first year in a long time where forecasts of a return came true, of course a symptom of the COVID dip.

Both total energy consumption and energy intensity were both consistently over-estimated in the EIA's energy forecasts. I'm not sure why, but it's interesting. It's worth noting that total energy consumption has been somewhat growing since the 2010s, and hasn't been overestimated like it was in the past.