Recently the BBC’s Horizon programme presented a friendly introduction to the world of behavioral economics. I am a fan of Horizon. In the past it was a programme that dealt with some serious scientific questions. They also provided a critical view of some mathematical finance

However, this programme presented an uncritical view of behavioral economics and painted Daniel Kahnemann as the saviour of the human race. I exaggerate, but I was pretty annoyed with this programme because it provided false confidence in the work of these scientists.

So, what’s so wrong with behavioral economics? It begins with a complete disregard for how economics actually works. I have written elsewhere that economics is a descriptive. It takes your observed behavior as optimal for you and tries to rationalize why it is optimal. This was discussed in my post here . As a result, the starting point for traditional economic analyses is that we are observing equilibrium behavior, in other words after the forces of evolution have made you adjust your behavior. Our mathematical models are used to provide a more analytical framework for understanding these phenomena. I liken this to the work that mathematical biologists do when analyzing animal behavior, where animals are treated as fitness maximizers. No one believes that animals actively do this, however, we do believe they behave as if they did this. Why should human beings be any different in that regard?

Behavioural economists believe that people make mistakes. This is a deep as their ideas get. When you observe a phenomenon, the starting point for their analyses is to assume that something is going wrong and that it is their job to determine what is causing these errors. We all make mistakes. We make arithmetic mistakes. However, we do not re-write the rules of mathematics to adapt to these errors. Behavioural economists on the other hand want to do this.

The Horizon programme begins with the example of New York taxi driver. This is taken from a paper by Camerer, Babcock, Loewenstein and Thaler (of Nudge fame) which can be found here The idea of the paper is simple. If you look at trip data of taxi drivers you see the following: on busy days they clock off too early and on slow days they clock of too late. In other words, they should be working more on the busy days because they can make more money on those days and transfer that wealth to times when business is slow. The explanation for this is that taxi drivers use a rule of thumb that tells them to hit a certain amount of earnings in a day. They have preferences with specific reference points. This all seems pretty good evidence for sub-optimal behavior and gives the behavioral economist his/her raison d’être.

How would a non-behavioural economist deal analyse this situation? They would start by first saying that taxi drivers were doing what was best for them and that their behavior revealed something about the decision problem they were facing. Given that a taxi driver is optimizing in some dimension, we need to understand more about the actual job they are doing. We cannot assume we fully understand the trade-offs each driver is taking into account by just looking at trip data. We need more data to understand what they are doing. Simple data has to be added to get a more complete picture:

• weather
• traffic
• day of the week
• location
• cumulative hours of the shift worked

These seem like variables that a taxi driver would care about as well. Well, why are we not analyzing the decision problem with this data as well? The original paper did not do this, Henry Farber did this. He showed that if these factors are included in the data analysis there is no reason to believe that taxi drivers are behaving sub-optimally. The original paper just did not fully understand the decision problem of a taxi driver and presented a toy version of that job in order to make a point.

So, it has been shown that the taxi driver example is not factually correct. But why is this worrying? The problem is that behavioral economists want to use these “insights” to shape public policy and to legislate taking into account these behavioral quirks. If we are all basically stupid and prone to errors, it is up to the behavioural economists to decide how things should be. There are many high-profile examples of this already happening: the behavioural insights team in the UK government, Cass Sunstein is already in the Obama government. These would fundamentally alter our freedoms to engage in certain activities, for example writing specific types of legal contracts. These are incredible infringements on our freedoms driven by bad science.