There has been ample discussion in recent years of a "nuclear renaissance," and many politicians and energy analysts believe that a meaningful response to climate change must include a new fleet of nuclear plants in the United States. The long-term planning studies that routinely come out of utilities, advocacy groups, and the Department of Energy now commonly include new nuclear units. However, many of these studies use nuclear and utility industry cost estimates for new nuclear plants, rather than estimates based on the actual experiences of companies currently trying to build nuclear power plants. Given the dollars and the environmental impacts at stake, it is critical that planners make resource decisions using the best information available.
Cost overruns: the rule not the exception. History has shown that the estimated costs of major energy projects can differ dramatically from actual costs, and seldom has the difference been greater than in the nuclear power industry. In the last round of nuclear power plant construction in the United States, the actual costs of 75 plants (excluding financing costs during construction) exceeded the initial cost estimates for those units by an average of more than 200 percent in inflation-adjusted dollars. In other words, the plants cost about three times as much as anticipated. Expressed in nominal dollars (dollars not adjusted for inflation) and including the costs of financing during construction, the overruns were far worse. For example, the cost of the two-unit Vogtle plant in Georgia increased from an estimated $660 million in the early 1970s to an actual completed cost of $8.7 billion -- a 1,200 percent overrun.
Delays, cost overruns, and expensive never-completed projects were common during the 1970s and 1980s. Then came a long period of inactivity in nuclear power plant construction in the United States.
This respite ended in the early 2000s, when a series of low-cost estimates by vendors and academics launched the so-called "nuclear renaissance." Proponents of new nuclear units pointed to "standardized designs" that they argued would facilitate regulatory approval and be less prone to cost overruns. Wall Street was less sanguine about new nuclear plants than these proponents. An interesting comparison of the cost estimates made by government, the power industry, and the financial industry appears in a 2009 report by Mark Cooper, senior research fellow for economic analysis at the Vermont Law School.
Institutional lenders made the extent of their skepticism clear when several new nuclear projects in the United States were launched and sought financing. Wall Street would not go near them. In 2005 Congress acted, authorizing the Energy Department (after consulting with the Treasury Department) to provide loan guarantees -- subsidies intended to enable projects to proceed with construction.
History repeating itself? The experience to date with these projects suggests that the pencils on Wall Street are indeed sharper than those in government and the power industry. A few recent examples:
- Progress Energy's cost estimate for a proposed nuclear plant north of Tampa Bay, originally $17 billion, has risen to $22 billion, or roughly $10,000 per kilowatt. The online date has slipped by nearly two years.
- In November 2009, CSP Energy disclosed that costs of the planned expansion of the South Texas nuclear station had risen from $13 to $17 billion, or over $6,000 per kilowatt. The original estimate for the project was $5.4 billion. Responding to this escalation, CSP reduced its ownership share in the project from 50 percent to 7 percent.
- The first "new generation" nuclear unit to begin construction, Finland's Olkiluoto 3, was originally scheduled to begin operation in 2009, and was projected to cost 3 billion euros. However, by May of 2009 costs had escalated to 6 billion euros, and the online date had been pushed to the second half of 2013.
A step in the right direction. Responding to these and other data, the US Energy Information Administration (EIA) has in the last year revised its assumed costs of new nuclear power plants, placing more emphasis on data from actual projects. EIA revised upward its assumed "overnight" cost of new nuclear units by 37 percent, from $3,902 to $5,339 per kilowatt. ("Overnight" costs do not include interest during construction.) This is a move in the right direction, but cost forecasts should be revisited frequently to ensure that they reflect the latest data and trends.
Moreover, it is important to consider the effects of subsidies when costing out various technologies for a study designed to inform long-term energy policy. Consumers pay the full cost of any new resource, whether they pay this cost through electricity rates or through their tax dollars. Quantifying the value of a $5 billion or $10 billion taxpayer loan guarantee is difficult, but it should not be ignored in responsible energy policy-making.
We need to do careful and current research on all resource types -- not just nuclear. For example, the cost of new wind projects fell to very low levels in the mid 2000s -- with the cheapest projects at or below $1,700 per kilowatt, or about 5 cents per kilowatt-hour. But the cost of new wind began trending up again in the late 2000s and has remained well above the lows of the last decade. The most economical new wind projects today are coming in at 7 cents to 8 cents per kilowatt-hour.
Photovoltaics, on the other hand, have seen dramatic cost reductions since mid-2009. Again, focusing on real market data:
- In 2009 and 2010, pursuant to a feed-in tariff (a posted offer to purchase renewable energy), the Sacramento Municipal Utilities District received offers for 100 megawatts of photovoltaic capacity in the range of 14 cents per kilowatt-hour.
- In early 2011, Southern California Edison reported bids for more than 2,500 megawatts of photovoltaic capacity at or below the time-of-use-adjusted Market Price Referent (MPR) in California. The MPR is an avoided cost number used by utilities in California to evaluate bids from potential power projects. The on-peak MPR in 2010 was in the range of 13 cents per kilowatt-hour.
All of these photovoltaic prices include a federal subsidy, so actual project costs are likely to be 3 cents to 4 cents per kilowatt-hour higher. But clearly photovoltaic costs have reached a major inflection point, and it will be crucial to stay abreast of these costs over the next several years and to rely on current data.
Best available data. As policy makers struggle to craft a response to climate change, billions of dollars are at stake, as well as unprecedented environmental impacts. While it is not always possible to rely entirely on empirical cost data, it is more important than ever to use the best data possible -- and that means data rooted in the real world.