,
25 tweets,
7 min read
Read on Twitter
My initial reactions to AEO 2019, with charts and data:
The most important high level point is that the US is poised to become a large net energy and fossil exporter post-2020. Note that the US will remain an OIL importer through 2050 in most scenarios, although the US may become a NET oil exporter 
A primary reason we are poised to become a net exporter in 2 years is because LNG exports are booming. 2019 is expected to see massive growth, with EIA projecting large (10%+ of domestic supply) LNG exports in all scenarios
In general, EIA's scenarios are worth looking through for the oil and gas sectors. The scenarios run provide a good mix of oil and gas prices
That said, EIA is projecting that US oil production peaks in the 2020s in most scenarios. Assuming climate policy, this may be true, but EIA isn't modelling that. A good reminder that even if shale is abundant, its has long term limits
Of course, shale remains key. Production in other areas is projected to decline in most (all?) scenarios. Note ANWR production is lost in the noise here...
Turning to electricity; the key result/assumption is that electricity demand growth is around 1% through 2050. Over 30 years, this increases load by a lot (more than 25%). As the chart shows, recent load growth has been near 0 despite robust economic growth
This 1% load growth assumption is thus critical for understanding the electric sector projections. Additional load means reduces resource competition within the model. If load growth is closer to 0%, the fuel mix for the electric grid would be very different
Across their scenarios (which all feature positive load growth), EIA finds a relatively surprising mix of fuel share outcomes. Not in particular the difference b/n renewables in the low and high O&G resource cases. Assuming no climate constraint these are reasonable(ish)
This -ish part comes when you dig into the details at the fuel/technology level. For a while, EIA's renewable numbers were a problem impacting their model results (shameless article plug: sciencedirect.com/science/articl…)
In AEO 2019, this is.... much improved for solar. Utility-scale capacity and generation across scenarios is 5-10 times higher than 2018 levels. While I'd argue these may still be a bit on the low side (assuming no climate constraint), they are reasonable
Wind, however, is projected to die an early and industry-shattering death. In only 1 scenario does wind capacity see any growth post 2020. EIA also projects 0 MW of off-shore wind. Both of these projections are in significant conflict with industry expectations
EIA scenarios constantly project this death of the wind industry. Historically, they have blamed tax credit expiration. However, the tax credit expiration is (more or less) final now and the wind industry still plans to exist in five years
IMO this is a model artifact. Modelling is an abstraction and EIA's model is unable to fully represent the calculus that goes into wind PPAs. Effectively, the real market has many other considerations (scalaiblity, price stability) that are hard to represent in a national model
But this brings me to another major critique of AEO 2019's results/method: its coal projections for the electric sector. If EIA expects wind to die a premature death (relative to industry expectations) it is because EIA expects coal to live (unlike industry expectations)
On a generation basis, EIA projections for coal are surprisingly resilient. In all scenarios, generation declines into the 2020s, but then holds flat at 15-20% of total US generation in all but one scenario
Considering that EIA assumes not climate constraint, these numbers might be reasonable. However, when looking at capacity, it doesn't make sense
EIA projects coal capacity to decline by 40-50% from *2018 levels* by 2030. This is almost 2/3 lower than the fleet's peak 10 years ago! This is a massive change and, assuming no climate policy, is reasonable with current market trends
However, the generation and capacity number differences create a large disconnect. In effect, EIA is saying that coal capacity is going to crash but the remaining plants are going to significantly increase utilization. Check this chart! =>
Historically, we've seen natural gas and coal capacity factors converge as NG replaces coal for baseload and coal becomes intermediate/peaking (its happening!). I don't know of any reason to expect that aging coal facilities would be able to reverse this trend
In the model, whats likely happening is that the model is finding a lot of capacity is not economic and so retires. The remaining capacity, however, is relatively efficient, and can thus use very cheap coal prices to maintain fleet generation levels. Its justifiable but wrong
Why is it wrong? Good question. I honestly expect EIA is trying to figure out this exact question right now - the disconnect b/n capacity and generation levels is just so striking
I'll be looking into the result in more detail. Before I go, two last points: transportation and emissions!
On transportation, I found this chart striking. Despite related data showing larges increases in transportation services (especially vehicle miles traveled), efficiency, a moderate shift to cars instead of trucks, and more alternative vehicles actually limits energy demand
Final point: none of EIA's projections indicate significant emissions reductions (note chart axis doesn't start at zero). Even with the wind/coal issue, we're way off track on emissions trajectories and need to deal with both oil and natural gas to address that (END)
