Skip to content

Too Big to Fail?

(This post is co-authored by Catie Hausman)

The San Onofre Nuclear Generating Station (SONGS) was closed abruptly in February 2012. During the previous decade, SONGS had produced about 8% of the electricity generated in California, so its closure had a pronounced impact on California’s wholesale electricity market, requiring large and immediate increases in generation from other sources.

In a new EI@Haas Working Paper titled, “The Value of Transmission in Electricity Markets: Evidence from a Nuclear Power Plant Closure”, we use publicly available data to examine the impact of the closure on economic and environmental outcomes. Because of the plant’s size and prominence, the closure provides a valuable natural experiment for learning about firm behavior in electricity markets.


We find that the SONGS closure increased the cost of electricity generation by $370 million during the first twelve months. This is a large change, equivalent to a 15% increase in total generation costs. The SONGS closure also had important implications for the environment, increasing carbon dioxide emissions by 9.2 million tons over the same period. Valued at $35 per ton (IWG 2013), this is $330 million worth of emissions, the equivalent of putting more than 2 million additional cars on the road.

The closure was particularly challenging because of SONGS’ location in a load pocket between Los Angeles and San Diego. Transmission constraints and other physical limitations of the grid mean that a substantial portion of Southern California’s generation must be met locally. When SONGS closed, these constraints began to bind, essentially segmenting the California market. The figure below shows the price difference at 3 p.m. on weekdays between Southern and Northern California. After the closure there were many more days with positive differentials, including a small number of days in which prices in the South exceeded prices in the North by more than $40 per megawatt hour.


These binding transmission constraints meant that it was not always possible to meet the lost output from SONGS using the lowest cost available generating resources. Southern plants were used too much, and Northern plants weren’t used enough. Of the $370 million in increased generation costs, we attribute about $40 million to transmission constraints and other physical limitations of the grid. This number is less precisely estimated than the overall impact, but is particularly interesting in that it provides a measure of the value of transmission.

The paper provides all the gory details about how we made these calculations. It turns out to be more difficult than a simple before-and-after comparison because during this period the California market was also experiencing a whole set of simultaneous changes to hydroelectric resources, renewables, demand, and fuel prices. What is helpful, however, is that transmission constraints were rarely binding prior to the closure. This means that observed behavior during the pre-period provides a good sense of how firms would have behaved during the post-period had there not been transmission constraints.

Our findings provide empirical support for long-held views about the importance of transmission constraints in electricity markets (Bushnell 1999; Borenstein, Bushnell and Stoft 2000; Joskow and Tirole 2000), and contribute to a growing broader literature on the economic impacts of infrastructure investments (Jensen 2007, Banerjee, Duflo and Qian 2012, Borenstein and Kellogg 2014).

The episode also illustrates the challenges of designing deregulated electricity markets. A new book chapter by Frank Wolak (here) argues that while competition may improve efficiency, it also introduces cost in the form of greater complexity and need for monitoring. Transmission constraints add an additional layer to this complexity, by implicitly shrinking the size of the market. Constraints increase the scope for non-competitive behavior, but only for certain plants during certain high-demand periods, so understanding and mitigating market power in these contexts is difficult and requires a sophisticated system operator.

Lucas Davis View All

Lucas Davis is an Associate Professor of Economic Analysis and Policy at the Haas School of Business at the University of California, Berkeley. His research focuses on energy and environmental markets, and in particular, on electricity and natural gas regulation, pricing in competitive and non-competitive markets, and the economic and business impacts of environmental policy.

17 thoughts on “Too Big to Fail? Leave a comment

  1. It really is a shame they shut it down. My dad was currently working there as a mechanic supervisor when it was shut down. Yes there was a big impact by shutting it down that they are now feeling the affects of and from what I have heard, much of the decisions involved in closing the plant were political.

  2. An important point to note is that the operator of SONGS decided to retire the plant for economic reason that were just a fraction of what the shutdown will end up costing to society. The mean reason for the shut down was the economical cost of the uncertainty about when the NRC would allow the restart after what was a minor problem (authorized releases of coal plants, including those that are now running at full power due to the shut down, are many times more dangerous that the very minor amount of tritium that could have been released, some people have ingested tens of gigabequerels of tritium without any visible consequence on their health even 10 years later).

    • Sorry I should have clarified. Feel free to ignore the vertical bar labeled “Placebo Removal”. This indicates October 1, 2012 — exactly 12 months prior to the day that the import constraint was removed. We included this line because we wanted to use this as a placebo test, or control group i.e. to check visually for an impact on prices. If you saw a price change at both October 1st days, this would make you think that it was coming from some type of annual pattern or day of the year effect rather than from the removal of the import constraint.

  3. I haven’t read the paper yet, perhaps you considered one of the N-S import constraints that was eased by the CAISO in September 2013.

    The SCE branch group % import constraint was in place before 1998, and the CAISO kept enforcing it. In Sept, there was a decision made that the constraint didn’t make sense anymore – to the objection of some generators in southern California – and stopped enforcing it. It would be interesting to take a close look at price separation N to S before and after.

    • Our analyses use data through January 2013, so the removal of this constraint is not in our sample. Nonetheless, the removal of this import constraint is very interesting. We did a bit of additional data work and produced this figure which does seem to indicate a decrease in the N-S differential around the time the constraint was removed.
      Henry Fig

  4. The analysis in the detailed working paper includes marginal cost for a nuclear plant in Figure 6; this marginal cost consists of the sum of EIA estimates of nuclear power plant fuel and variable O&M costs.

    The use of these costs in the working paper suggests that “marginal costs” is used to refer to short-run marginal cost (SRMC).

    Fuel costs for a batch refueling light water reactor like SONGS should not be considered as a component of (SRMC). Fuel cost is fixed at or before nuclear fuel is loaded and the refueling outage schedule (i.e., between refueling) does not change as a result of small differences in output over the 12-24 month operating cycle. Under the classic definition of SRMC (a small change in output for a short time – 1 MW for one hour, for example), fuel costs for such a batch-refueling nuclear power plant will not change.

    A typical regulatory approach to recovering nuclear fuel cost is to amortize the total cost of fuel loaded over the expected output of the nuclear power plant in the future operating period. This provides an estimate of nuclear fuel cost per MWh, with a future true-up if the estimated output is lower or higher than actual output. This regulatory amortization approach does not, despite the $/MWh units, reflect marginal nuclear fuel costs.

    It is less clear whether “variable” O&M costs are a part of SRMC.

    • Thanks for the comment – very interesting. As you point out, available estimates of variable costs (in your example, nuclear fuel, but also operations and maintenance) do not always match the economist’s definition of a marginal cost. For our calculations, this distinction for fuel costs shouldn’t matter. First, because we assume that nuclear runs all the time except for refueling outages – that is, we’re not allowing nuclear to ramp up and down. And second, because we’re using the cost number to calculate the increase in generation costs over twelve months. So we’re not using the fuel costs to calculate a marginal change in generation within one hour, for instance.

  5. This all points out how essential it is for us to get energy use under control. Utility energy efficiency programs are helpful, but essentially leave the vast majority of residential customers out “in the heat”. Most residential customers do not have the disposable income to participate in the elaborate processes currently in use. To reach the majority of the residences we will need to rethink how we do things. Direct install programs, neighborhood by neighborhood are one viable approach, but given the current definitions of cost effectiveness, they will not be implemented.

    • Most ‘efficiency’ initiatives do not tackle behavioral change … ie a willingness to slightly higher summer [and lower winter] temperatures in the house. Driving less, rather than getting a more efficient higher mpg car. As long as we ‘feel entitled’ to ‘do what we please’ in our ‘inalienable right of the pursuit of happiness’ we will NOT be able to put a listing dent in our environmental impact and our energy consumption.

      • Behavioral change is nice, but as you mention very difficult to obtain. On the other hand buildings can be made much more energy efficient, even as retrofits. Cars, trucks and busses can be more efficient. The building changes can be concentrated in areas impacted by events like the closure of SONGS. If we had the will to do it we could do it. On the other hand “willing” other people to change their thermostat settings is not within our reach. In addition a savings of a few percent is significantly different from saving 50 to 75% of a buildings heating and cooling.

    • Small modular nuclear reactors with a safer design and waste with a short half life and smaller volume are on the horizon, particularly in countries like China, India, Norway, Canada and others. These reactors can be located near where power is needed and used, making transmission costs less of a factor.

      We need to streamline our US regulations to prepare to take advantage of these improvements in nuclear power. We also need to educate our public who have been victim to peddlers of scare tactics and a great deal of distortion and misinformation about nuclear accidents, current reactor safety and actual radiation dangers.

      We have a small “industry” of irresponsible speakers and writers making a nice profit off lectures and books that use scare tactics to make money. They are debunked by places like Popular Mechanics Magazine and the University of Texas environmental departments, but these revelations are not reaching to public. Making things worse, major news organizations with little knowledge of nuclear power and new generation developments pick up these distorted statements and use them in stories going out to a very wide public audience. All this puts the US at a disadvantage in keeping up with other nations in new energy advancements.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: