A recent article in The Economist describes a blog post by Charles Frank of the Brookings Institute in which he questions the methods that have been used to compare renewable energy sources with more traditional sources like coal, gas and nuclear.
Drawing on the work of Paul Joskow of MIT, Frank claims that the generally accepted levelized cost models, which essentially divide the total lifetime system cost by the total amount of electricity produced, do not adequately discount the value of renewables sources like solar and wind based on their intermittent nature. Joskow’s reasoning is that since these intermittent sources vary their output at different times of the day and the year, that should be reflected in their value, since the demand for, and the price of electricity also varies throughout the day, at least in the commercial market.
So, given that wind, for example, produces electricity mostly at night, when the power is less valuable, that should be reflected in the value of a wind investment. Solar, on the other hand produces mostly at mid-day, when the power is most valuable, so it may be getting short-changed by the levelized cost approach.
Frank started with Joskow’s premise, then went on to perform a detailed analysis of various energy sources, based on avoided emissions and avoided costs, which revealed, he says, that contrary to popular belief, solar and wind are the least cost-effective way of producing low carbon electricity, followed by hydro, nuclear, and finally at the top of the list is combined cycle gas turbine power. Written from the perspective of building new electric generation capacity, Frank concludes, “Assuming that reductions in carbon dioxide emissions are valued at $50 per metric ton and the price of natural gas is not much greater than $16 per million Btu, the net benefits of new nuclear, hydro, and natural gas combined cycle plants far outweigh the net benefits of new wind or solar plants.”
The problem with an analysis like this is, given the rapidity with which renewable energy costs are dropping, trying to compare them with traditional sources is akin to trying to catch a falling knife. Frank’s data was obsolete by the time the ink dried on the page.
In addition, the analysis is highly sensitive to the eventual market price for carbon, which could swing the results dramatically. Also not considered is the impact of energy storage which could easily neutralize the liabilities that form the basis for Frank’s thesis.
But Amory Lovins, of the Rocky Mountain Institute, has already demonstrated that the perceived need for storage has been overstated. Smart grid and smart grid operators will be able to dynamically reconcile supply with demand far more effectively than originally thought, through what Lovins calls choreography. Yes, wind and sun are intermittent, but they are also fairly predictable. The tools available for and acceptance of demand management (can you wait a few seconds for a spike to pass before your air conditioner kicks in?) are also gaining in importance.
Given the broad coverage this story received, Lovins felt the need to weigh in here with a written rebuttal to correct what he felt were significant errors that led to erroneous conclusions.
Lovins points out that correcting nine of Frank’s incorrect assumptions reverses the conclusion to fall in line with the order that the marketplace has consistently chosen as the best investment for the money: hydro, wind, solar, gas, and nuclear. Says Lovins, Frank assumed the both wind and solar were twice as costly and half as productive as they actually are today. He also assumed gas productivity to be twice what it is and ignored the impacts of both methane leakage (a very serious greenhouse gas contributor) and price volatility. Indeed the climate impact of methane leakage (not to mention the environmental impacts of fracking) is severe enough to have caused some experts to question the value of switching away from coal.
Rounding out the list, are out of date assumptions regarding the construction cost (low by half) and operating cost (low by 80%) of nuclear plants. Frank’s assertion leaned heavily on the premise that a good deal of new generating capacity is needed (which it isn’t), and he did so without taking energy efficiency opportunities into account. While Frank has brought a fresh analytical slant to this important problem, when applied with the most recent data it produces the same conclusion as the prevailing approach. In other words, although the variation of the value of electricity produced at different times had not been taken into account by prior analyses, the impact of doing so was not sufficient to change the overall conclusion.