The New York Times quotes Charles Hall, professor of systems ecology at SUNY-ESF, summing up what I find ultimately problematic with trying to use thermodynamics to understand economics:
"The basic issue is very fundamental: Why should economics be a social science, because it's about stuff?"Actually, it's not. And this is where this type of approach begins to go off course. Economics is the study of what people value and how it affects their behavior. What is most of interest to economists is the behavior, not the stuff. People do indeed value stuff, but one of the basic tenets of economics is that this value is subjective. Value is created when people trade things, but this has little, if any, relationship to how much entropy was created in that physical transfer. Actions that produce large degrees of physical entropy (explosions) can destroy value, while actions producing small levels of entropy can create large value (trading a diamond for a cup of water that saves a life). And this isn't even getting into the fact that people value a lot of intangible things too (time, poetry, love, music, etc.). How do you evaluate the entropy of those?
This "new" approach to economics is not unlike what early economists debated a couple hundred years ago when people were convinced that value was determined based on the amount of labor put into the creation of a good. However, this line of thinking was eventually replaced with notions of subjective value which depend upon marginal utility. (See the Diamond-Water Paradox.)
Hall and his colleagues are trying to use objective measurements of "energy return on investment" and entropy, but applying it in such a way that misses some of the most basic insights of economics. This might be more understandable if these insights weren't something found in every Econ 101 textbook.
Read the whole New York Times piece. I heard several presentations in Santa Fe that dealt with the same issues and had the same problems in their approach. If you image a group of economists developing the new field of 'physinomics' and attempting to explain Newton's law of universal gravitation in terms of decreasing transaction costs, you'll get an idea of what a lot of this sounds like in reverse.
While neoclassical economics may have it's share of problems, I don't think its inattention to thermodynamics is one of them. I'm afraid trying to apply physical notions such as entropy to economics may ultimately end up generating a lot of heat, but very little light.
(HT Tyler Cowen)