Editor’s Note: This post is part of a discussion on the net benefits of low and no-carbon electricity technologies. Discussants include Charles Frank, a nonresident senior fellow in Global Economy and Development at Brookings and Amory Lovins, cofounder and chief scientist of the Rocky Mountain Institute.
Regular visitors to Planet Policy will be aware that Mr. Amory Lovins of the Rocky Mountain Institute has produced a rebuttal to my recent Brookings working paper, The Net Benefits of Low and No-Carbon Electricity Technologies. In his critique, Mr. Lovins makes some valid points for which I am thankful, but he often exaggerates the effect of his criticisms, many of which are based on methodological errors or draw on unreliable data. My analysis, after taking account of Mr. Lovins’s helpful suggestions, is based on what I believe is the best available existing information on each of the five technologies addressed in the Brookings working paper. Over time, technological developments may change the data and conclusions, but I prefer leaving predictions about future technology to others more qualified.
The valid points Mr. Lovins raises have virtually no effect on the overall conclusions of the paper: the ranking of the five technologies remains the same, except that nuclear drops from second to third. Wind continues to rank number four and solar ranks number five by a large margin. Mr. Lovins has many concerns, more than I can address within the confines of a single blog. So what follows represents the first of three articles, each tackling various issues. I begin here with a discussion of capacity factor and capital cost assumptions, which are most critical in driving the results.
Capacity Factors for Wind and Solar. Mr. Lovins wrote that I should use more recently published U.S. Energy Information Administration (EIA) capacity factors rather than my earlier estimates. He is correct. I have redone the analysis (and corrected a minor error in one of the formulas that I used) using a weighted average of the more recent EIA data—31.2 percent for wind and 19.9 percent for solar with the following results:
Wind Solar Hydro Nuclear Gas CC
Net Benefits per MW-Year $32,591 ($161,941) $156,830 $261,262 $476,636
The ranking of the different technologies remains the same, and the order of magnitude is not substantially different than in my original paper.
Mr. Lovins also maintained that capacity factors reported by the U.S. Department of Energy (DOE) 2012 Wind Technologies Market Report, updated to 2013 on August 19, 2014, and Utility-Scale Solar 2012, are even better than the recent capacity factors reported by the EIA. However, the DOE capacity factors are not appreciably different than those of the EIA. For example, the average DOE wind capacity factor is 32.1 percent between 2006 and 2013 while the weighted average EIA capacity factor between 2009 and 2013 is 31.2 percent. Furthermore, the DOE data are subject to sampling error, while the EIA data are based on a survey of all wind and solar projects, not just a sample.
Capacity Factor for Gas Combined Cycle. Mr. Lovins asserts that I should have used the historic actual average capacity factor for gas combined cycle plants. It is inappropriate to use a historical average for old natural gas plants to compare a new, highly efficient gas plant that would rank above older gas plants in the merit order and therefore run far more often than the average for old gas plants.
Furthermore, a new gas plant would rank above coal plants if CO2 emissions were priced at $50 per ton and as a result would be used much more often than any coal plant. Thus, with the right price signals, and without high penetration of overly subsidized wind and solar production, there is no reason to believe that a new combined cycle plant would not operate at its maximum capacity factor of approximately 92 percent. Wind and solar plants, on the other hand, can never increase their capacity factors beyond the recent levels of approximately 30 percent and 20 percent, respectively, because of the limitations imposed by the availability of wind and sun.
Capital Costs. Mr. Lovins claims that I have assumed “wind and solar plants that cost about twice as much to build as they do.” Mr. Lovins suggests that the 2012 capital cost estimates of wind and solar published by the DOE are superior to the 2012 capital cost estimates of the EIA, but gives no rationale for such a conclusion.
Wind Capital Costs. The DOE calculated the weighted average capital cost wind projects completed in 2012 to be $1,940 per kilowatt. I use the EIA 2012 estimate of overnight capital cost of $2,213 per kilowatt, about 14 percent, not 100 percent, higher. The weighted average cost for wind projects in 2013 was $1,630 per kilowatt, lower than in 2012, but based on a very small sample. The wind capacity installed in the U.S. in 2013 was 92 percent lower than in 2012 because of the uncertainty about renewal of production tax credits. Only the least costly projects were likely to have been built in 2013.
The DOE data comes from a very wide range of sources of varying quality and use inconsistent definitions of cost. The DOE cautions analysts to use their data mainly to infer capital cost trends from year to year and not to place much significance in the absolute magnitude of its estimates.
Solar Capital Costs. The DOE report Utility-Scale Solar 2012, published September 2013, calculated the weighted average capital cost of 113 utility-scale solar projects completed in 2012 to be $3.90 per watt or $3,900 per kilowatt (pages 3 and 4). In my paper, I use the EIA 2012 estimate of $3,873 per kilowatt, slightly less than the DOE estimate, not 100 percent higher as Mr. Lovins has maintained.
In his final analysis, Mr. Lovins reduces the capital cost for solar to $2,000 per KW, citing a study by Lazard dated August 2013. However, the Lazard study gives no rationale or any sources to support its use of $2,000 per kilowatt as the capital cost for solar electricity.
Conclusion. The main reasons for the low ranking of wind and solar technologies are high capital costs and low capacity factors. Nothing in Mr. Lovins critique changes that result.