It's the Regulation, Stupid

Revisiting Lovering, Yip, and Nordhaus, 2016

It’s not that surprising, I suppose, that both pro-nuclear and anti-nuclear progressives are in full agreement that the current nuclear regulatory regime is necessary and requires no substantial reform. But the result of that shared and underlying ideological commitment is that it has turned the usual orientation of nuclear supporters and opponents toward the cost of regulation on its head. In the upside down world of progressive nuclear politics, the advocates argue that nuclear regulation hasn’t appreciably increased the cost of nuclear while the opponents argue that it has.

How did this happen? The origins of this strange inversion date to a time when progressives of virtually all stripes opposed nuclear energy. Progressive opponents of nuclear energy argued that the technology could not be utilized both safely and cheaply. Rather, they claimed that nuclear energy was inherently expensive because it was inherently unsafe. The increasing cost of nuclear energy over the last 50 years was a function of increasingly strict regulations that were necessary to assure safe operation. Nuclear energy was a dangerous and unnecessary technology that was rightly being regulated into economic obsolescence.

This view, which until recently was held virtually monolithically on the Left, has understandably created a fair amount of cognitive dissonance for the newer progressive nuclear supporters, who over the last 15 years or so have come to believe that nuclear energy is a safe technology that is necessary to address climate change. What is the point of being a progressive, after all, if you are not going to defend government regulation of corporations to protect the health of the public?

Of course, progressive nuclear advocates know, and frequently argue, that the public health risks associated with nuclear accidents and exposure to low dose radiation have been wildly overstated. But to follow that uncomfortable knowledge to its logical conclusion, that significant reform of nuclear regulation is needed, would smack far too much of Reagan era deregulation for most progressive sensibilities.

The evolution of the nuclear debate in recent years from an abstract one about the merits of nuclear power to actual policy-making aimed at building new nuclear plants has brought that cognitive dissonance to the fore. Progressive nuclear supporters have responded by simply denying that nuclear regulation has had any significant impact on the cost of building or operating nuclear reactors. If regulation wasn’t the source of rising nuclear costs, after all, there would be no need for reform even if the public health risks we are regulating are negligible.

Instead, progressive advocates claim that the real reason that nuclear costs have escalated so much over the last fifty years is due to a lack of standardization. Legacy companies like General Electric and Westinghouse kept changing their designs. Every utility wanted a different reactor. None of this would have happened if we’d just built a single type of reactor, like France and South Korea did.

There is much to be said for standardization, especially when it comes to the potential for a new generation of smaller reactors that can be produced modularly or even fully manufactured. But the mantra that the lack of standardization, not over regulation, is what explains the spiraling cost of nuclear reactors elides far more than it explains, particularly in the context of contemporary efforts to revitalize the US nuclear sector.

Socialization Not Standardization

The case for standardization has frequently been both overstated and entirely decontextualized. Insofar as it has been proven, it has been proven in places like France, South Korea, and China, where centralized authorities own and operate both the electrical system and the nuclear industry. Standardization, in other words, has been the byproduct of state ownership and planning.

As I noted recently in Foreign Policy, we have a lot of evidence from France, South Korea, China, and elsewhere that a state-owned and publicly-financed nuclear industry, building standardized large light water reactors for state-owned electric utilities, can hold the line on nuclear construction costs. Better yet, over the 60 to 80 years that these plants will operate, they will provide enormous amounts of affordable low carbon electricity.

My colleague Matt Wald made a similar point about private, vertically integrated, monopoly utilities in regulated markets in the US in the Breakthrough Journal last week. Progressives and leftists who are sincere in the belief that tackling climate change demands an alternative to neoliberalism would be well served to spend a lot more time advocating for the French nuclear model and a lot less time handing out huge public subsidies to private corporations.

But reducing these facts to technological standardization is deeply misleading. Many places that did deploy standardized designs, such as Great Britain and even France, saw significant escalation of nuclear construction costs nonetheless, just not as rapidly as the US did. Moreover, the differences between the US and those other places are less significant than is generally acknowledged. The US actually deployed significant multiples of a small number of reactor designs, the vast majority of them deployed by two firms, Westinghouse and GE, that were iterating on two basic designs - Westinghouse’s pressurized water reactor and GE’s boiling water reactor.

Meanwhile, many places held up as beacons of standardization actually deployed multiple reactor types. France imported a Westinghouse reactor and then indigenized it, building 3 different series of reactors of its own design. Even South Korea, which has the best documented record of managing cost through standardized design built 9 imported reactors of three entirely different designs before it indigenized and standardized its design, at a cost that was comparable to much of the standardized fleet it subsequently built.

What is clear is that irrespective of the level of standardization, large nuclear reactors are better thought of as public works projects and long lived infrastructure than as discrete technologies like a gas turbine or a solar panel. And whether built by state-owned enterprises or regulated and vertically integrated private utilities, they will prove over the long term to be a reasonably good deal for the public. France has among the lowest electricity prices in Europe. The same is generally true of utilities that still own and operate nuclear plants in the United States, even though most of those plants were built during an era of rapidly escalating nuclear construction costs.

But neither the lack of standardization nor the lack of socialization can explain why the cost of building nuclear plants has escalated so consistently in so many different places. Consider that with each new class of reactor that France deployed, construction costs rose. Its latest design, the EPR, will cost far more than prior designs. Some of that is due to the fact that it is effectively a first of a kind reactor, with the first two deployed in Olkiluoto, Finland and Flamanville, France. But there is almost certainly no amount of standardized deployment that will allow these reactors to ever approach the cost of earlier French reactors. The EPR features not one but two steel and concrete containment systems and four independent backup cooling systems. It will require somewhere in the neighborhood of twice as much steel and cement per megawatt of electricity capacity as any commercial reactor ever built.

The cost of nuclear regulation is often reduced to the time and cost of paperwork and navigating endless licensing proceedings that can drag on for years. But the real costs are literally embedded in the designs of reactors themselves, which over decades have become ever more complex and unwieldy in an effort to avoid infinitesimal and unobservable public health risks associated with low dose exposure to ionizing radiation.

The Regulated Stairway to High Nuclear Construction Costs

In the published literature on nuclear costs, the term of art for the way that regulation contributes to rising costs is “regulatory ratcheting.” The idea is that a drip, drip, drip of new regulations and requirements over time slowly escalates costs. But that doesn’t really describe how nuclear costs escalated over the course of the 1960s, 70s, and 80s, when virtually the entirety of the US commercial nuclear reactor fleet was built. Rather, there were two distinct step changes in the cost of US reactors. The first occurred in the early 1970s, the second in the early 1980s. Both coincided with major revisions to regulatory requirements for nuclear reactors.

The second is better known. In the aftermath of the Three Mile Island accident, the NRC imposed an emergency moratorium on licensing of all nuclear reactors. The agency enacted sweeping new nuclear safety requirements, applicable not only to plants that were seeking licenses but those already operating or under construction. These included hundreds of new regulations covering everything from containment to fire protection to backup cooling to systems for shutting plants down.

There is a reasonable case that some of these changes improved nuclear safety. But it is absurd to suggest that these sweeping new requirements, taken as a whole, had no significant impact upon the cost of building nuclear reactors. And indeed, a dramatic increase in nuclear construction costs after the Three Mile Island accident is exactly what Jessica Lovering, Arthur Yip, and I find in “Historical construction costs of global nuclear power reactors,” which is the most widely cited source on trends in nuclear construction costs around the world. That analysis documents a step change in costs for US plants that were completed after the accident, which cost, on average, three times more than those completed in the eight years before the accident. We conclude:

When the full cost experience of US nuclear power is shown with construction duration experience, we observe distinctive trends that change after the Three Mile Island accident… [For R]eactors that were under construction during Three Mile Island and eventually completed afterwards… median costs are 2.8 times higher than pre-TMI costs and median durations are 2.2 times higher than pre-TMI durations. Post-TMI, overnight costs rise with construction duration, even though OCC excludes the costs of interest during construction. This suggests that other duration-related issues such as licensing, regulatory delays, or back-fit requirements are a significant contributor to the rising OCC trend… These results suggest that the Three Mile Island accident in 1979 did uniquely affect the nuclear industry in terms of overnight construction cost.

That still leaves the question of what was driving the escalation of nuclear construction costs prior to the Three Mile Island accident. And while a gradual accumulation of non-regulatory cost pressures from things like labor and material costs definitely played a role, as well as a series of new regulatory requirements in the late 1960s and early 70’s addressing things like thermal impacts from the release of cooling water into waterways and new emergency core cooling requirements to address fears that a reactor core might melt through the concrete foundation of a nuclear plant (the controversial “China syndrome” hypothesis of anti-nuclear lore), there was also a very pronounced escalation of construction duration and construction cost in the aftermath of the Supreme Court’s landmark 1971 Calvert Cliffs decision, which established that US licensing and regulation of nuclear reactors, then under the auspices of the Atomic Energy Commission, was subject to expansive interpretation of the newly enacted National Environmental Policy Act.

The Calvert Cliffs decision had far-reaching consequences for US environmental law and policy, many of which are now being debated in the ongoing permitting reform effort at the federal level. But its consequences for the US nuclear industry were dramatic and immediate. Licensing of nuclear reactors in the United States ground to a halt for two years. As was the case when the NRC paused licensing eight years later in the wake of the Three Mile Island accident, the AEC completely revamped its regulatory standards in order to bolster the agency’s licensing determinations from anticipated legal challenges, even for reactors that were already licensed and operating.

New requirements for nuclear reactors included human factor engineering, control room indicators, shift management and operator training, and emergency preparedness requirements.

As after the Three Mile Island accident, a clear discontinuity in nuclear costs and construction times ensued. On average, reactors completed after the new rules were established took more than two years longer to complete and cost about 25% more than reactors completed prior to the Calvert Cliffs decision.

Progressive Uber Alles

Nonetheless, many progressives, including journalists, continue to downplay the role that regulation has had in the long-term escalation of nuclear costs, deploying a “just so” argument that because the cost of building nuclear reactors saw significant escalation before the Three Mile Island accident, regulation can’t be the reason that nuclear plants became so much more costly to build. “The claim that excessive regulation single-handedly ruined the American nuclear industry, however, doesn’t hold up,” Rogé Karma argues in a recent Atlantic article, to take one example, because “the cost of building new nuclear plants was already rising before Three Mile Island.” Karma, ironically, cites our analysis, as well as an outdated 2013 paper from Nathan Hultman and Jon Koomey, as the basis for the claim.

Karma then claims that the lack of standardization is the real reason that costs rose so much, quoting several sources, including Lovering, to the effect that standardization of nuclear designs and building practices, not regulatory reform, is the key requirement for building nuclear reactors affordably. A year earlier, Bloomberg’s David Fickling made a similar argument, pointing the finger instead at electricity sector deregulation. Fickling observed that there is simply no history of successful nuclear deployment absent vertically integrated, monopoly utility electricity systems. Utility regulation, not NRC reform, Fickling argued, holds the key to a nuclear future.

Both claims, to be clear, are not entirely without merit. The drivers of nuclear cost escalation are multifactorial. In the United States, the lack of standardization definitely played a role. The role of the decentralized and privatized monopoly utility sector, meanwhile, was double edged. As Fickling and many others have noted, it is the only context in which the United States has ever deployed commercial nuclear reactors. But the cost-of-service monopoly utility regulatory model also incentivized indifference among both utilities and vendors to escalating costs. Other factors also contributed to rising construction costs, notably rising labor and material costs during the inflationary decade of the 1970s, a dynamic that has more recently come back to haunt the sector in the post-pandemic run-up of commodity prices.

But it is also clearly the case that the dramatic expansion of regulatory requirements by first the AEC and then the NRC, in an effort to eliminate virtually all possible public health risk associated with nuclear energy, has undeniably played a very significant role in increasing the cost of building and operating nuclear power plants. The growth of regulatory requirements, procedures, and paperwork necessary to license and build a nuclear power plant is well documented. Study after study has also shown how those regulations increased nuclear costs. And the history of construction cost escalation tracks closely with these two watershed events in the regulatory history of nuclear energy.

Insofar as this multistep escalation of nuclear regulation has improved nuclear safety, it has done so at the margins of a technology that was already extraordinarily safe. Consider that the Three Mile Island accident was in no way a public health disaster despite the absence of all of those new safety systems and requirements put in place in its wake. A small amount of ionizing radiation escaped the plant’s containment. According to the NRC, about 2 million people living within 50 miles of the plant received an average dose of about 2 millirem, about one sixth the exposure one would receive from a chest x-ray. For a person standing on the other side of the Susquehanna River in the direction of maximum exposure for 6 days and taking no protective action, the dose would have been less than 100 millirem, less than the annual background radiation that someone living in the area would receive and about 20% of the background radiation that an individual living in a high altitude location such as Denver, Colorado would receive annually. Multiple long-term epidemiological studies conclude that no detectable public health effects resulted from the accident.

The miniscule radiological dose exposure that the NRC attempts to eliminate today is even more extreme. In virtually all cases, the NRC enforces health and safety standards predicated upon avoiding public health consequences from nuclear operations and accidents that are entirely theoretical, meaning that the impacts of those exposures are so low as to be unobservable epidemiologically even when tracking a large exposed population over a long period of time.

So while there is much else that will need to be addressed in order to launch a globally competitive nuclear energy sector in the United States, there is little reason to think that a new generation of nuclear reactors will be significantly less expensive to build than the current generation without some serious reconsideration of what level of safety and exposure to low doses of radiation is reasonable and necessary to assure public health. As such, the determination of progressives to point the finger at every conceivable cause of nuclear cost escalation except regulation is neither defensible nor a serious strategy for deploying a new generation of nuclear reactors at costs that policy-makers, utilities, or other customers are likely to be willing to bear.