How do we know what causes human behavior?

The way to improve economics

Posted Mar 10, 2011

Humans are complicated. Most animals are excellent strategists, they figure out how to get the most food or the best mate very rapidly and act to reach these goals. Humans sometimes make decisions poorly because we "over-think" choices, worrying about future consequences and who might know about them. We often (rightly) modify our behavior because of the action of our enlarged prefrontal cortices, the uniquely human part of the brain, that lets us imagine possible futures based on present actions.

This "over-thinking" plays havoc with social scientists who seek to predict behavior by asking people to report the reasons for their actions. The brain seldom accurately gives an account of this information, calling into question much survey data. Economists know this so they focus on observation and then try to identify people's motivations presuming they are rational and self-interested. These two assumptions are very much in question in many settings. As a result, the information we all want from scientists--the true cause of some human predilection--is difficult to obtain. Yet scientists of all stripes inevitably use the "C" word in their writing and public statements.

While the identification of causation is a dilemma scientists have struggled with for several centuries, a new approach from my lab offers behavioral scientists a way to solve this problem. As long as, that is, they don't mind rolling up their sleeves.

The approach I am advocating (along with my colleague Dr. Moana Vercoe) starts with the insights from the father of the scientific method, the British natural philosopher Francis Bacon (1561-1626). Bacon advocated the inductive method in which experiments are run with carefully varying conditions so that the general case reveals itself. Bacon's idea was to go from the specific to the general by ruling out incorrect inferences. This is the standard approach today in experimental sciences.

Since the early twentieth century, economics has followed the model of theoretical physics and has worked on a deductive basis. That is, economists typically sit in their offices and use logic (sometimes looking at data) to build mathematical models of the general case. From the general case, specific empirically testable implications are derived and sometimes tested. The fatal conceit in this endeavor is the belief that the economy is a system that, like classical mechanics, can be fully described by a set of power laws.

This is wrong. The economy is the aggregation of individual decisions made by often poorly informed, cognitively compromised individuals who are trying to do the best they can in their lives. Summing up of these decisions leads to a constantly evolving adaptive and self-organizing set of dynamics that is poorly described by immutable laws.

So how can we figure out what causes what? This is the holy grail of science--understanding and then controlling one's environment. Experimental economics, pioneered by Nobel laureate Vernon Smith, has given economics an important boost in the Baconian direction. Neuroeconomics has gone further by providing direct data on the brain mechanisms producing behavior. Yet, the vast majority of neuroeconomics studies are based on correlational data, for example, through the use of brain imaging technologies.

My lab focuses on how the brain causes behavior by using drugs to turn on o r off parts of the brain. For example, we showed that the brain chemical oxytocin causes generosity by infusing synthetic oxytocin into the human brain and had people make choices with money. Similarly, we showed that testosterone makes men selfish by administering this hormone to men and then asking them to make economic decisions.

Human beings are more than a bag of chemicals, but chemicals are the software that runs our lives. By changing the software code pharmacologically, I have found that most behaviors are conditional--that is, they depend on a larger set of conditions than social scientists have so far considered. For example, we have shown that some decisions in women depend the stage of their menstrual cycles. Human beings are adaptive creatures, and chemicals are the way we rapidly we adapt to new situations.

Let's call this the "brains on drugs" research method. So why don't more economists use this approach? It is messy, expensive and requires going beyond traditional training in economics. Francis Bacon provided a reason why new approaches are not adopted by scientists that he called idola theatri. This "idolatry" is following the academic dogma no matter what.

This dogma hasn't worked so well recently. That drugs can improve economics is revolutionary. Viva la revolucion!