Readers of The Economist may have been surprised to read in its 26 July 2014 “Free exchange” section on page 63, or in its online version, the “clear” conclusion that solar and wind power are “the most expensive way of reducing greenhouse-gas emissions,” while “nuclear plants…are cheaper,” so governments are foolish to boost renewables and mothball nuclear.
In each of the past three years, the world has invested more than a quarter-trillion dollars to add over 80 billion watts of renewables (excluding big hydro dams). That growth is accelerating: solar power is scaling faster than cellphones. Big European utilities lost €0.5 trillion in market cap, as an Economist cover story featured, not because renewables couldn’t compete, but because they competed all too well, wiping out old power plants’ profits. The same is happening to some well-running U.S. nuclear plants, now facing closure as uneconomic just to operate.
That full-page article highlights a May working by Charles R. Frank, Jr. (economics Ph.D. 1963), a nonresident fellow at the nonpartisan and notably debate-friendly Brookings Institution. His is in international development and finance. I daresay most experts on the economics of technologies and climate change had never heard of him—but they have now. As soon as The Economist featured his paper, their inboxes and Twitter feeds lit up with incredulity: could his conclusions possibly be true?
They’re not (and yes, I’ve written The Economist a letter saying so). My detailed critique at www.rmi.org/frank_rebuttal explains why, and cites two other reviews and a podcast. But for anyone who knows the subject, Dr. Frank’s conclusions don’t even pass the giggle test. He finds that new wind and solar power are the least, and new nuclear power and combined-cycle gas generation are the most, cost-effective ways to displace coal-fired power—just the opposite of what you’d expect from observing market prices and choices.
So are Dr. Frank’s odd findings artifacts of errors in his methodology, his data, or both? Both, but there are so many mistakes that just nine data points can carry the whole load. My colleague Titiaan Palazzi reconstructed Dr. Frank’s spreadsheets, reproduced his results, then simply updated the nine most egregiously outdated figures to those in the latest official historical statistics (not forward-looking projections) from the U.S. Energy Information Administration, Department of Energy, Nuclear Energy Institute, and similarly authoritative sources.
Presto! The conclusions flipped. Instead of gas combined-cycle and nuclear plants’ offering the greatest net benefit from displacing coal plants, followed by hydro, wind, and last of all solar, the ranks reversed. The new, correct, story: first hydro (on his purely economic assumptions), then wind, solar, gas, and last of all nuclear—still omitting efficiency, which beats them all.
Beneath Dr. Frank’s wrong answer, however, lurks a useful question. He adopts the distinguished economist Prof. Paul Joskow’s 2011 valid thesis that the way power-sector investments are chosen—lowest long-run economic cost—is incomplete, because different technologies generate power at different times, creating different amounts of value. Of course value as well as cost should be considered. But interestingly, this case suggests that if we use correct and up-to-date cost and performance data, the cost- and value-based calculations yield the same priorities, whether judged from the perspective of financial investment or climate-protection effectiveness. That is, adjusting for different resources’ time of generation, though theoretically nice, doesn’t change the result; cost-benefit analysis gives the same answer as a simple cost comparison. The resulting best-buys-first sequence would also gain even more value if other hidden costs, risks, and benefits were counted too.
Making a splash—intentional or not—with a flawed analysis that doesn’t survive more careful scrutiny is nothing new. My esteemed Stanford colleague Dr. Jon G. Koomey cowrote a 2002 Annual Review of Energy and the Environment paper (here) called “Sorry, Wrong Number: The Use and Misuse of Numerical Facts in Analysis and Media Reporting of Energy Issues.” Its abstract says: “Students of public policy sometimes envision an idealized policy process where competent data collection and incisive analysis on both sides of a debate lead to reasoned judgments and sound decisions. Unfortunately, numbers that prove decisive in policy debates are not always carefully developed, credibly documented, or correct. This paper presents four widely cited examples of numbers in the energy field that are either misleading or wrong. It explores the origin of those numbers, how they missed the mark, and how they have been misused by both analysts and the media. In addition, it describes and uses a three-stage analytic process for evaluating such statistics that involves defining terms and boundaries, assessing underlying data, and critically analyzing arguments.” It’s a bracing read, with a nice summary and update.
The diligent Dr. Frank has collected not just one wrong number but a flotilla, together driving a false conclusion that gained a prominent platform in The Economist. The analytic lesson: rapidly changing data quickly pass their sell-by date.
It’s too early to guess whether prompt refutations will prevent the distressing phenomenon Dr. Koomey describes, whereby media and advocates fond of a false thesis (or who don’t know any better) keep repeating it long after it’s been decisively debunked. Time will tell. But your ability to stay well-informed and to exercise your critical faculties can help build sound public discourse. If you hear a claim that sounds nutty, maybe it is. If it is, say so. As biologist Prof. E.O. Wilson wrote, “Sometimes a concept is baffling not because it is profound but because it’s wrong.”
Amory B. Lovins, Cofounder and Chief Scientist, Rocky Mountain Institute, 2317 Snowmass Creek Road, Snowmass CO 81654. Tel. (303) 245-1003, firstname.lastname@example.org, www.rmi.org