“A false sense of precision”

“Whenever you hear the term ‘Darwinian’ from anyone other than historians of science, assume the crash position; it’s going to get real ugly.”

The quote is from a blogger known as Mike the Mad Biologist. The title of the post is When Economists Misunderstand Biology, an entry he wrote in response to economist Russ Roberts’ piece called What is economics good for?. In Roberts’ opening paragraph, he refers to his previous argument that macroeconomics is “deeply flawed and not a science”. He goes on to describe that economist Friedrich Hayek (the original anti-Keynesian) felt that to label economics a science gave “a false sense of precision and understanding.”

The comparisons to Darwinism are a small part of an interesting post. Most of the post is a discussion by Roberts of economics and economic theory and philosophy, coming down to a statement that absolutely resonated with my existing prejudices about modeling economic systems: “… the system is simply too complex to model with too many unobservable variables to allow systematic understanding.”

As I implied in my previous post, one of the unobservable variables is the human mind, which—observably–does not drive behavior in a way to maximize utility. In the last 30-40 years psychologists, neuroscientists, and even behavioral and neuro-economists have been struggling to understand what makes us tick. I would argue that each brain is as complex a system as a planet-wide economy.

For one thing, it not only has internal circuitry that is somewhat changeable, it has inputs, both from the environment and from the rest of the body. Some inputs are consciously recognized and many are unconscious, and they can skew our thinking. In the last post I talked about unconscious bias, but there is also the HALT rule from 12-step programs: Don’t make any decision when you’re hungry, angry, lonely, or tired. My central question in the previous post was this: How does information–or the perceived value of any given piece of information–affect decision making in general, and economic decision making in specific? The valence we put on information is not only affected by cognitive biases, but also by physiological state.

Model that, economists.

Mike the Mad Biologist had a point when he called out Roberts for picking Darwinian evolution as the parallel or simile for economics. Roberts says, “Darwinism, like much of economics, exploits tautological reasoning.” He’s confusing a layman’s understanding of Darwin’s theories with the modern work of evolutionary biologists, and assumes it’s all about the fossil record. His point appears to be that biologists know what they can’t know, and that it really is just a “useful way of organizing our thinking.” There is a difference between observational approaches, such as those used astrophysics and paleontology*, and experimental approaches, such as those used in particle physics or molecular evolution.

Roberts says, “We do not expect a biologist to forecast how many squirrels will be alive in ten years if we increase the number of trees in the United States by 20%. A biologist would laugh at you.” We would laugh at you because the data from which you ask us to extrapolate are incomplete. Squirrels do not live, or die, by trees alone. Will increased cover increase the coyote population and thus predation? Are the trees near residential areas with bird feeders? Will the presence of trees increase bird populations and thus the likelihood that people will put out bird feeders, providing food for squirrels?

In any science, measurement or measureability is key. One of the goals of a theoretical model is to test it. The data from behavioral economics and neuroeconomics should be turning standard economic theory and modeling on its head, but if it is, that hasn’t trickled into the world of finance and business. They still believe that the “quants” can predict the future, because there is an abiding unconscious bias toward the ‘scientism’ of economics that Robert’s argues against in his post. (And I recommend the post, despite this argument about his simile.)

Next time I really will write about split-brain patients, neuroimaging, and creating unconscious biases on purpose.

*There is some interesting experimental work in paleontology.