What's happened in the US power system in the last decade?

Electricity generation in the US has changed a lot in the last decade (from 2010 to 2020). The country has seen a natural gas boom, due the rise of fracking, and a corresponding decrease in electricity generated by coal. Wind and solar have grown from almost nothing to generate a significant portion of US electricity, due to dramatically decreasing prices.

Electricity Usage Has Flattened

The first major change is that electricity generation leveled off around 2007. After 50 years of near-continuous growth, electricity usage staying flat is a huge change. If you measure per-capita it’s even slightly lower - electricity use per person is now at mid-1990s levels. The Great Recession was an initial contributor, but the trend has remained flat since - with energy efficiency playing a role. Demand didn’t level off because the price increased though – residential electricity prices have remained fairly flat. This change isn’t a simple triumph of green conservation - the US continues to use much more electricity per-capita than almost any other country (only Canada and some tiny petro-states use more). It may be that - along with more efficient appliances and lighting - many people in the US have finally found a limit to how much electricity a person can possibly use.

Generation Shifted from Coal to Fracked Natural Gas

The way the US produces energy changed dramatically. Natural gas generation sharply increased, and coal generation corresponding declined. Renewables rose from generating a negligible percentage of energy in 2010 to become a significant source of electricity within a decade.

Here’s the percent of electricity generation by source, and how it changed in the decade from 2010 to 2019:

source

2010

2019

percent change

coal

44.9%

23.5%

-48%

natural gas

24.0%

38.5%

60%

nuclear

19.6%

19.7%

1%

hydro

6.3%

6.7%

6%

wind

2.3%

7.3%

217%

solar

0.0%

1.8%

Here are those same sources, looking further back over the 50 year period from 1970 to 2019:

decade-in-review: generation by type

The replacement of coal for natural gas is even more dramatic if you look at the last 50 years. Coal peaked in the 1980s and has been on the decline ever since. That decline has been accelerating since the mid 2000s, with the rise of natural gas. In 2016, natural gas plants produced more electricity than coal for the first time ever.

Renewables have become a significant part of the electric grid. In the mid 2000s, renewable power was negligible and even in 2010 renewables generated only 2% of US power, essentially all of it from wind. Since then, wind power more than tripled, surpassing hydroelectric power in electricity generated for the first time in 2019. For the first time ever, solar is also contributing a non-negligible percentage of electricity generated, reaching almost 2% in 2019. The numbers above only include electricity produced by power companies – but about 40% of recent solar capacity has been added to residential and commercial buildings. If you included this distributed solar in the chart above, solar power would be even higher.

Nuclear power has produced an essentially constant amount of electricity for the past 30 years, when the last nuclear plants were built in the US. Hydroelectric power generation has been similarly flat over the past decade. These two electricity sources are not a significant part of the story of the changes of the past decade.

Carbon emissions are down, but not enough

US total carbon emissions are down about 5% over the decade. Carbon emission reductions in the electricity sector are a big reason for this larger change - which makes sense, because electricity generation is the largest single contributor to emissions in the economy. Emissions from electricity generation are down about 20% due to shifting to natural gas and increased renewables. Just the displacement of coal by natural gas accounts from about one third of the overall US emissions change - more than any other factor.

We’ll talk more about climate policy and electricity generation in a different post, but the 5% reduction over the last decade is insufficient to meet the challenge of climate change. A serious commitment to avoiding catastrophic climate change requires much faster and deeper emissions cuts - on the order of 7% per year, every year for the next decade.

Other emissions decreased too

The emissions of other harmful pollutants into the air are also way down - both in measurements of emissions from power plants and measurements of overall US air quality. Air pollution kills over 100,000 people and causes $1 trillion in health impacts in the US annually - mostly from respiratory problems and heart disease. Burning coal in power plants is responsible for about 10% of US pollution deaths - mainly due to sulfur dioxide, which also causes acid rain. These health impacts are so significant that in 2012 the EPA estimated that for every dollar spent on pollution controls would yield up to $170 in health benefits. More recent science estimates that the health impacts are even worse than we thought, so much so that “it would be worth freeing ourselves from fossil fuels even if global warming didn’t exist”.

Pollution is down significantly since 2010, largely because of the reduced use of coal power. Sulfur dioxide is down 70%, with nitrogen and particulate matters also down. Mercury pollution is also down about 80% since 2010, in part because of additional EPA regulation that required coal power plants to install additional scrubbing equipment (which also drove down particulate matter). These are big improvements, but regulating emissions and closing coal plants is still important, because the harm from these emissions is still quite high and disproportionately hurts Black and Brown people.

What caused such big shifts these last 10 years?

These shifts are huge, compared to previous decades: coal generation has been cut in half along with reductions in its many harmful emissions and renewables are now starting to produce significant amounts of energy. Has the US been successfully carrying out a plan toward a low-carbon future? No. In the US, these changes stem mostly from business-as-usual economics, a lack of climate policy, and a nearly unregulated new form of natural gas extraction called “fracking”. To the extent that renewables have grown it is primarily because of dramatically reduced costs (while the renewable industry does receive some subsidies, they are dwarfed by the subsidies to fossil fuels).

Over the past decade, Republicans have blocked and continue to block any progress on climate change. Republican obstruction in Congress has blocked any discussion of climate policy since 2010, while hampering investment in renewables. Rule-changes from the Trump administration (EPA, DOE, FERC, NHTSA, etc) have favored coal and fracking by any means necessary and taken pot-shots at renewables. The ship of US climate policy has moved between having no rudder and being steered the wrong way.

Trends from the last few years suggest that in the absence of climate policy the future of power generation will continue to be driven by the economics we’ve seen in the past decade. We will continue to use less coal and more renewables. However, as we talked about in the carbon section above, the changes under this business-as-usual scenario won’t be enough to avert catastrophic climate change. We will need to policy to drive additional changes. Before we talk about policy and where we need to go in the next decade, let’s talk about the economics of power systems and how the changes we’ve seen in the past decade have come about. It boils down to just two things: cost to operate and cost to build.

Costs to operate

Cost to operate is how much money it takes to run the power plant: fuel costs, operations, and maintenance (aka O&M). For coal and gas plants, fuel is about 75% of the cost to operate. Renewables have a clear advantage, because their “fuel” is free.

The big change in the last decade was the quantity of natural gas produced went way up, causing fuel prices to drop for natural gas based generation (see the next section). Coal prices stayed more or less flat in comparison, so coal consumption and production dropped. Coal plants are slightly more expensive to operate and maintain, so as natural gas got cheaper, the cost to run a natural gas plant dipped below the cost to run a coal plant in 2015, and it has continued to fall.

2010

2019

% change

units

quantity of natural gas

27

41

52%

trillion cubic feet

quantity of coal

957

550

-43%

million tons

price of natural gas

$5.96

$3.55

-40%

$/MBtu

price of coal

$2.66

$2.06

-22%

$/MBtu

total cost to operate natural gas

$57.03

$32.43

-43%

$/MWh

total cost to operate coal

$41.84

$35.86

-14%

$/MWh

total cost to operate wind

$9-20

$9-20

0%

$/MWh

total cost to operate solar

$7

$5.60

-20%

$/MWh

It’s difficult to get exact numbers for the cost to operate solar and wind farms, in part because many are so new. For wind farms, technology had matured by 2010 and costs to operate leveled off. Solar farms continue to find new ways to reduce O&M costs - sometimes dramatically - as the business matures and solar technology continues to improve. For both wind and solar, technology improvements have led to modern plants producing more energy for the same size plant (aka capacity factor) - which helps these economics.

When it comes to fossil fuels - even with cheaper gas - it’s hard to compete with the renewables. Coal has another disadvantage to gas here: coal has essentially one customer: US electric electricity generation uses about 80% of US mined coal (two thirds of the remainder is exported and industry uses the rest). Natural gas has more uses, with only 40% of gas going to electricity generation. So coal usage has fallen almost in lockstep with falling natural gas prices.

Fracking

Fracking is the main player in the increase in US natural gas production in the last decade. Fracking involves pushing liquid down a well at high pressures to “fracture” rock formations deep underground, allowing oil or gas to flow through the cracks and back to the surface. The basic The idea of fracking to increase gas (or oil) production has been around for decades, but fracking shale formations didn’t start to take off in earnest in the mid-2000s. By the end of the decade, shale gas production was over 60% of US natural gas production - a large share of it from the Marcellus region of northern Appalachia. Shale fracking has boomed enough to drive down conventional gas production:

decade-in-review: natural gas source

Fracking, from an ethical and environmental perspective, is horrible: notably it’s caused ground-water pollution (to the point where people can taste and smell pollutants in their drinking water), dumping of waste into streams, earthquakes, and significant methane emissions. Most of the fracking in the US is concentrated in poor, rural areas - like the Marcellus shale in PA/OH/WV and the several shale formations in TX/LA/OK/AR. The US is essentially the only country in the world fracking shale (Canada does a bit too). Other countries have shale reserves but have regulated or banned fracking.

Recent research suggests that methane emissions from gas leaks may make fracked natural gas as bad for the climate as coal. In response, the Trump administration has proposed that we just stop monitoring these methane emissions. Despite the boom in production, fracking companies have been dubiously profitable, have nearly continuously sought additional investment, and sometimes resemble ponzi schemes - raising questions about fiscal as well as ecological sustainability. One of the largest fracking companies filed for bankruptcy in April 2020. Unfortunately, gas infrastructure spending and shale reserves (15 years worth of production and about 70% of US gas reserves) indicate that if we continue on our current path (unregulated industry and unchecked emissions), fracking may be with us for a long time.

Cost to build

The cost to build a new power plant is the other major component of power system economics. Here we’ve seen renewables get dramatically cheaper to build - which is the primary driver behind their growth. The table below was compiled in 2018 by a large investment bank (the Dept. of Energy has similar estimates) and represents the combined cost (in 2018 dollars per MWh) to build and operate a plant.

cost to build

2010

2018

percent change

natural gas peaker

$243

$179

-26%

nuclear

$125

$151

21%

coal

$130

$102

-21%

natural gas

$96

$58

-40%

solar

$290

$43

-85%

wind

$145

$42

-71%

solar + storage

$124

And these costs are before any renewable subsidies (but do include the sneakier subsidies to coal and gas, which are significant). For a power company, it’s now cheaper to build new wind and solar than any other technology before renewable subsidies.

Current costs mean that energy storage co-located with solar farms will start to be built instead of new natural gas “peaker” plants. Peakers are essentially fast moving generators designed to balance out unexpected fluctuations in supply and demand - they are essential to ensuring the power stays on. At the start of the decade, storage wasn’t an option - the grid had essentially zero storage and building it would have been fantastically expensive. Now, we’re starting to see power companies choose storage over natural gas because it's the cheapest option. Storage is only projected to get cheaper - with some analysts predicting storage will be consistently cheaper to build than natural gas peakers by 2025.

Additional estimates suggest that in the next decade building new renewable energy will be cheaper than continuing to operate existing coal plants. This implies we’ll start to see coal plants shut down even before the end of their lifetimes because they can’t compete on costs.

Nuclear doesn’t currently seem like a competitive option due to a combination of high prices, long times to build (think decades), and public concerns about waste and safety. There have been just three nuclear plants in the pipeline since the 1990s. The last US nuclear plant opened in 2016, after starting construction in 1973! There was one nuclear plant planned for SC in 2009, but plans were abandoned in 2017. There is only one US nuclear plant currently under construction, in Georgia - it’s already 10 years into construction, billions over budget, and has no certain completion date in sight. The supposed “nuclear renaissance” of the 21st century is pretty much officially dead.

Power plant additions and retirements

It’s good to see where this might be going, but how’s this played out so far? We can look at US electricity generating capacity over the past decades, and how it’s changed:

energy source

2010

2020

percent change

natural gas

38.2%

43.7%

22%

coal

31.9%

21.4%

-28%

nuclear

10.3%

9.0%

-7%

hydro

9.4%

8.8%

-1%

wind

3.8%

9.6%

168%

solar

0.1%

3.6%

4313%

storage

0.0%

0.1%

3421%

These numbers mostly agree with what we’ve seen in the electricity generation data - a decline in coal, paired with an increase in natural gas, wind, and solar. But changes in capacity represent much more permanent change. Once a coal plant closes, it’s not reopening.

Aging Fossil Fuel Plants

With energy use flat over the decade, the cost of building new plants has been mostly to replace old plants. Cost to build and operate were part of the reason the US retired so much of its coal fleet, but another reason is that those coal plants are getting old! We can look at when US plants were built – in these graphs the lines show capacity additions by year:

decade-in-review: generation capacity

Coal plants were mostly built in the 1970s and 1980s - with no new coal capacity added after 2015. The average age for a coal plant is 40 years old. Most coal plants are retired after about 55 years, so in the next 15 years, much of the existing US coal fleet will shut down due to old age – and when utilities look to replace that capacity, building renewables will be the cheapest option.

Wind farms started being built around 2005 (with big swings from year to year as Congress let federal tax credits for wind expire and then renewed them) and solar around 2010 (with a dip in 2018 as Trump added largely ineffective tariffs). Renewable capacity additions are on a shaky but steep trajectory upward over the past decade, and are likely to continue to grow.

Natural gas is a bit more complicated - there was a giant boom in building gas plants from 2000-2004, a period when power companies were restructuring and there was cheap natural gas and cheap interest rates. We’re now in the midst of a second building boom in natural gas plants. These natural gas plants are likely to become stranded assets as renewables become cheaper – and even more so if the US starts to take climate change even a little bit seriously.

Conclusion

The changes in the US power system in the last decade have been significant, with renewables and natural gas replacing coal. We talked through the economics behind this transition - costs to build and operate power plants. We also looked at the effects on US emissions, air quality, and the generation fleet. The next post will be about how the power system might change in the next decade, based on the several recent climate policy proposals. I’ll link to it here when it’s ready.


Technical notes

In 2010, I started work on my PhD dissertation in power systems. After finishing in 2013, I took a job in another field. This blog post has been part of my process of poking around again to see what's changed in the industry, what's likely to happen next, and how it all relates to the rest of the world.

The EIA provided much of the data in this post. The charts and tables here were compiled by me. Source code to create the charts can be found here and the tables can be found here (along with notes on the sources behind the numbers).

Thanks to Amelia Greenhall for lots of editorial work on this post.


To get an email when the next post comes out, sign up here!