Wednesday, May 1, 2013

Energy Errors

Charles Mann has an interesting article on unconventional oil and gas in the Atlantic. Lots of good stories and  he cites Morris Adelman so it is a worthwhile article. But two glaring errors induce this post.

First, is this concept of EROEI. Let's quote Mann:
Economists sometimes describe a fuel in terms of its energy return on energy invested (EROEI), a measure of how much energy must be used up to acquire, process, and deliver the fuel in a useful form. OPEC oil, for example, is typically estimated to have an EROEI of 12 to 18, which means that 12 to 18 barrels of oil are produced at the wellhead for every barrel of oil consumed during their production. In this calculation, tar sands look awful: they have an EROEI of 4 to 7.
I am fully aware that some analysts use this concept, but I cannot believe that any economist would. The whole point of economics is to compare the value of inputs and the value of outputs. Not their physical unit measures. If you use lots of low-cost energy to create high value energy that is a good thing even if the quantity of the low-cost energy is huge. In this example, Mann talks about barrels of oil for barrels of oil. Sometimes the measure is in BTU's or some other physical unit, but the same error arises. We use coal to produce electricity because electricity is a more valuable type of energy -- try powering your ipad directly with coal.  Later in the article Mann talks about Econ 101, but if he took it he would not make this error.

Second, Mann goes on and on about the resource curse. The idea that resource abundance causes deteriorating economic performance has been much studied. But as Mike Alexeev and Robert Conrad clearly demonstrated it is an elusive curse. Alexeev and Conrad's paper should have ended these discussions, but they endure. This despite the fact that the empirical work that purports to find a resource curse is flawed for many reasons. But the most important is the failure to consider what the wealth of the country would be without the resources. That is correct, this is actually ignored.

To find the resource curse cross-country regressions are employed to explain growth performance. You always need some measure of initial GDP since growth slows down as you get richer. But most resource abundant countries discovered their oil before 1960 or 1970, the typical years used for initial GDP. So resource abundance led to higher GDP but its effect in the regression is explained by initial income not natural resource wealth (no matter how mismeasured -- I leave that for another time).

This fundamental error pervades the analyses. You typically find that a country like Russia has very high corruption for its level of GDP and this is explained by its resource abundance. But if Russia had less resources it would have much lower GDP! The reason Russia is off the regression line is not that its oil makes it more corrupt but that its oil makes it wealthier than its institutions and other fundamentals would suggest. Take away the oil does not make it richer!

No comments: