February 29, 2016
There’s a simple, compelling argument that the world ought to be building many more nuclear power plants. We’ll need vast amounts of carbon-free energy to stave off global warming. It’s not at all clear that renewables can do the job alone. And nuclear is a proven technology, already providing 11 percent of electricity globally.
So what’s the catch? Cost is a big one. More than safety or waste issues, cost is nuclear’s Achilles’ heel. Modern-day reactors have become jarringly expensive to build, going for $5 billion to $10 billion a pop. Worse, the price tag seems to be rising in many places. Back in the 1960s, new reactors in the US were one of the cheaper energy sources around. Two decades later, after a series of missteps, those costs had increased sixfold — a big reason we stopped building plants.
Ever since, experts have been debating whether or not nuclear’s cost problems are an intrinsic flaw that will doom the technology. Nuclear skeptics, such as Joe Romm,argue that soaring costs are an inevitable side effect of building massive concrete-and-steel structures that need layers of radiation safeguards.
But there’s also an optimistic story for nuclear — and one that I think is worth hearing out. A recent paper in the journal Energy Policy by Jessica Lovering, Arthur Yip, and Ted Nordhaus of the Breakthrough Institute looked at construction costs for hundreds of reactors built in the US, France, Canada, Japan, German, India, and South Korea between 1960 and 2010. Their data tells a more nuanced story.
Nuclear construction costs in the US did spiral out of control, especially after the Three Mile Island meltdown in 1979. But this wasn’t universal. Countries like France, Japan, and Canada kept costs fairly stable during this period. And South Korea actually drove nuclear costs down, at a rate similar to what you see for solar. Studying these countries can offer lessons for how to make nuclear cheaper — so that it can become a useful clean energy resource around the world.
“The biggest thing we found is that there’s nothing intrinsic to nuclear that leads to cost escalations,” Lovering told me. “It depends on what policies are in place, on the market dynamics. You get very different cases in different countries.”
So here’s a look at where America’s nuclear industry went awry — and how France and (especially) South Korea avoided those mishaps.
How nuclear costs soared in the US — and why it wasn’t inevitable
Before we dive into the US story, a note on numbers. The Energy Policy paper focuses on “overnight construction costs” for power plants. This is the price of parts, labor, engineering, and land. It doesn’t include fuel, operations, maintenance, or financing costs, but it’s the dominant component of lifetime costs. And it’s phrased in terms of dollars per kilowatt, so we can compare plants of different sizes.
For context, the EIA calculates overnight construction costs for new US power plants ordered in 2014. Advanced nuclear reactors are estimated to cost $5,366 for every kilowatt of capacity. That means a large 1-gigawatt reactor would cost around $5.4 billion to build, excluding financing costs. By contrast, a new wind farm costs just $1,980 per kilowatt. A new gas plant costs just $912 per kilowatt, or one-fifth as much. (This isn’t a perfect comparison, since reactors run at capacity more often than wind farms or gas plants. But even if you adjust for capacity factors, those construction costs make nuclear uncompetitive. High upfront costs can also scare off investors.)
So how did nuclear get so expensive?
The story starts in the 1950s, when the Atomic Energy Commission supported the first wave of commercial reactors. As with any nascent technology, early demonstration projects were pricey. Yet within a decade, companies were figuring out how to build bigger reactors, take advantage of economies of scale, and drive down the price.
A breakthrough came in 1963 with GE’s contract to build a low-cost light-water reactor at Oyster Creek, New Jersey. By the late 1960s, overnight construction costs for new reactors had dropped to $600 to $900/kW in today’s dollars — cheaper than modern gas plants. Atomic energy was on a roll.
But then … things got messy. As utilities ordered more reactors, supply chains for parts and skilled labor became stressed, causing delays and cost hikes. Meanwhile, both industry and environmentalists were finding new safety issues to deal with. Early core cooling systems had flaws and required upgrades. California’s reactors needed earthquake contingency plans. Most of these changes, Lovering says, were ultimately good things — they made the reactors safer.
But the process unfolded haphazardly. Rules and requirements sometimes changed midway through construction. That meant delays. And delays are crippling for any big, labor-intensive project. Idling workers and equipment can lead to massive budget overruns. By the early 1970s, nuclear construction costs had risen to $1,800 to $2,500/kW in today’s dollars — about the cost of modern wind farms.
Additional woes followed. In its 1971 Calvert Cliffs decision, the DC Circuit Court ordered nuclear regulators to change their rules to comply with the National Environmental Policy Act. That opened the door for citizen lawsuits to intervene in the licensing and construction process, sometimes causing further slowdowns.
Then nuclear suffered a mortal blow after the much-publicized (but nonfatal) meltdown at Three Mile Island in 1979. Every reactor still under construction at the time — 51 in total — suddenly faced major regulatory delays, changes in safety procedures, and new back-fit requirements. Construction times doubled, stretching out past 10 years. Costs went through the roof, past $7,000/kW for some reactors:
After that, nuclear power in the United States was moribund. Utilities, scared off by soaring costs and stagnating electricity demand, canceled more than 120 reactor orders. The wave of utility deregulation started in the 1970s disfavored large, expensive plants. Not a single new reactor began construction between 1978 and 2013. Instead, coal and natural gas dominated the grid — and CO2 emissions soared.
Today, nuclear power provides just one-fifth of America’s electricity, and there are only five reactors currently under construction. Because the industry is essentially starting afresh, these new reactors are quite pricey.
That’s the capsule history. As you’d imagine, people bicker over the details. Industry groups blame overzealous environmentalists. Opponents counter that the reactors were inherently unsafe and cost overruns were a natural consequence. Other analysts highlight problems in project management.
What’s notable, however, is that other countries had very different experiences with nuclear — they didn’t always see the same catastrophic cost escalations. So what did they do differently?
France, Canada, and Japan did a better job of keeping costs down
France started making a big push for nuclear power in the 1960s, starting with gas-cooled reactors and pressurized water reactor designs from the US and later, in the 1970s, developing its own PWR designs.
As Lovering, Yip, and Nordhaus show, overnight construction costs stayed relatively stable throughout this period, hovering around €1,400/kW ($1,500/kW). Nuclear kept expanding until it provided more than 75 percent of France’s electricity:
How did France pull this off? It helped that the country had only one utility (EDF) and one builder (Areva) working closely together. They settled on a few standard reactor designs and built them over and over again, often putting multiple reactors on a single site. That allowed them to standardize their processes and get better at finding efficiencies. Canada and Japan kept costs relatively stable with similar tactics.
Contrast this with the US, where our electricity sector is split up among dozens of different utilities and state regulators. As a result, US nuclear vendors had to develop dozens of variations on the light-water reactor to satisfy a variety of customers. That pushed up costs.
France’s regulatory process was also less adversarial than America’s — and, for better or worse, doesn’t allow legal intervention by outside groups once construction gets underway. After the Soviet Union’s Chernobyl disaster in 1986, the government tweaked safety rules, leading to some delays. But costs didn’t skyrocket like they did in the US after Three Mile Island.
To be fair, France hasn’t totally solved the cost problem. Areva’s newest generation of massive EPR reactors have recently been plagued by delays and budget overruns in France, Finland, and Britain. And some French politicians are now calling for a partial shift away from nuclear. More on those EPRs below.
South Korea actually lowered costs
But the most interesting story is South Korea, which has seen nuclear costs decline sharply since the 1970s:
South Korea had an advantage in that it didn’t start entirely from scratch. The country imported proven US, French, and Canadian designs in the 1970s and learned from other countries’ experiences before developing its own domestic reactors in 1989. It developed stable regulations, had a single utility overseeing construction, and built reactors in pairs at single sites.
The results were remarkable: overnight construction costs fell 50 percent between 1971 and 2008 as South Korea built 28 reactors in all.
In fact, the Energy Policy paper notes, the decline in South Korean nuclear power costs is comparable to the decline in solar power costs in Germany over the same time period. (Though solar has kept getting cheaper past 2008.) Analysts have marveled at how solar panel costs come down as companies get better at manufacturing them — a process known as “learning by doing.” South Korea’s experience suggests similar reductions are possible for nuclear.
Four broad lessons for making nuclear power cheaper
Right now, there are a bunch of companies working on advanced reactor designs that, they hope, could one day prove far safer and cheaper than the reactors of old. Some designs, for instance, include passive cooling systems to automatically avoid overheating in the event of an accident. Not only does this reduce the risk of meltdown, but it eliminates the need for costly containment and backup systems.
Early estimates suggest these futuristic reactors might initially cost around $2,000 to $3,000/kW to build — cheaper than current US reactors, but still not able to compete with gas or coal. The hope is that learning-by-doing will kick in and costs can come down over time.
It remains to be seen if any of these designs pan out and get approved by regulators. (The sluggishness of the US government’s licensing process could be a big obstacle here.) But if they do, here are a few lessons for how to drive nuclear costs down — rather than letting them careen out of control again:
1) Stable regulations are essential for nuclear power to thrive. Nuclear reactors will always need safety and environmental regulations. The key, says Lovering, is making sure these regulations are predictable. A rule that forces reactors to limit the temperature of its water discharge can be dealt with. But a requirement that changes midway through construction can be devastating. Predictability was a key difference between the US and places like France or South Korea.
2) Standardization of design helps. Another lesson from countries like France and South Korea is that standardization is invaluable. If a company can build the same reactor over and over under consistent conditions, then learning by doing is more likely to occur.
This is easier in some countries than in others. South Korea has a single state-owned utility in charge of all projects. France had EDF and Areva working together. By contrast, the balkanized US electricity market makes it much, much harder to coalesce around a single design and standard.
This may bode well for China, however, which has embarked on a big nuclear push to shift away from coal. The country has focused on a standard reactor design, the CPR-1000, and is trying to build a large number, hoping eventually to export abroad. If the nuclear industry is to scale up globally, Lovering notes, it may have to coalesce around just a few companies and designs, the way today’s commercial aircraft industry has centralized around Boeing and Airbus.
3) Build multiple reactors at the same site. France, Japan, Canada, and South Korea often built anywhere from two to eight reactors at a single power plant site. That led to major efficiencies: It saved on site-related costs (like evacuation plans), it’s easier to get specialized equipment and workers, and you can consolidate control rooms.
In the US, by contrast, we built a lot of power plants with just one or two reactors — another inefficiency. It’s notable that the five new US reactors under construction today are all occurring at existing plant sites.
4) Smaller reactors may be the way to go. In the 1960s, US nuclear vendors thought that building increasingly large reactors was the best way to create economies of scale and cut costs. Today, this looks less viable. Larger construction projects are trickier to secure upfront financing for, and costs escalate more rapidly when there are delays.
Areva, the French manufacturer, seems to be learning this the hard way in trying to build a new wave of gargantuan 1.6-gigawatt EPR reactors in France, Finland and Britain. These are the first reactors of their kind, so you’d expect some hiccups. But the sheer size and complexity of these projects has led to multiyear delays and multibillion-dollar overruns.
That explains why some companies are now focused on building smaller modular reactors— 300 megawatts or less. Ideally, you could mass-produce parts for these smaller reactors in factories and ship them by rail, allowing for more learning by doing in manufacturing. Smaller reactors should also be easier to finance; a utility could order one and then gradually scale up.
One company, NuScale Power has developed a 40 MW light-water reactor that it’s currently submitting to the US Nuclear Regulatory Commission for design certification. The hope is to have one up and running by 2024. We’ll see.
Can nuclear power overcome its cost problems?
It’s tough to say what the future holds. Nuclear power has been in sharp decline globally since the 1970s, and it seems like every three years the media hypes a nuclear “renaissance” that never pans out:
Bringing costs down would certainly help — but that’s also not the industry’s only headwind. Nuclear power remains deeply unpopular in some countries. Germany is phasing out its reactors by 2022, and there’s heavy opposition in Japan to restarting reactors shut down after Fukushima. Plus, there are thorny issues like how to dispose of radioactive waste and how to decommission old plants.
So … it’s hard to predict. In theory, nuclear is capable of providing reliable, carbon-free power at a reasonable price. In practice, who knows? There’s ample reason to be skeptical.
That said, there’s also an double standard in the way some climate hawks talk about nuclear versus how they talk about renewables.
As the Energy Policy paper notes, back during the 2000s, the price of building wind turbines was rising by 10 percent per year in the United States — the sort of frightful cost escalation now haunting nuclear in some countries. But analysts didn’t argue that wind power was doomed and should be abandoned. Instead, they tried to understand why costs had risen and what could be done to reverse that. And that effort paid off: Wind costs are now falling sharply.
Similar work should continue for nuclear. Advanced reactors may never pan out. It might turn out that a renewables-only grid is the way to go. But we don’t know that yet. And decarbonizing the global economy is a far too vast a problem to exclude any solutions right now. Many environmentalists call for a World War-II-style mobilization to fight climate change. In a situation like that, it’s odd to be picky about your weapons.
— Alan Nogee wrote a smart, worthwhile critique of this piece on Twitter after it went up. He’s skeptical the South Korea lesson applies neatly to the US, given South Korea’s advantage on wages and labor productivity. (“Compare to scattered rural labor force where US nukes would be built,” he notes, “[with] much less experience.”)
Nogee also argues (as I did) that the balkanized structure of the US electricity grid makes nuclear standardization much harder. Ultimately, he thinks some of the lessons here may be applicable to China and even to small modular reactors — but not to building bigger reactors in the US. His points are all worth reading.
— The International Energy Agency put out a huge report in 2015 on policy steps needed to deploy nuclear power more widely and bring down costs. And Jessica Lovering wrote a report for Breakthrough in 2014 titled “How to Make Nuclear Cheap.”
— Josh Freed wrote an in-depth piece for the Brookings Institution about companies working on the next generation of nuclear reactors. He warns that current rules around licensing could choke innovation — though there are ways to fix that.
— Finally, I should note that the Energy Policy paper by Lovering et al. is not the only one to describe the history of cost escalation in the US or France (though it’s the first to include Canada, Japan, South Korea, etc. as a point of comparison).
For other perspectives, see this 2007 paper by Koomey and Hultman and this 2015 paperby Escobar-Rangel and Lévêque, which are more pessimistic about the US and French experiences. Another 2013 paper by Koomey and Hultman is skeptical that Three Mile Island was the main reason US nuclear costs escalated. Lovering and her co-authors explain in detail why they disagree, but those papers are worth reading.