At first glance, the question seems straightforward — it's an aspect of our cognition that predisposes us to make errors. However, with the significant expansion of the list of cognitive biases with the success of behavioural economics, one might question what these biases truly are and what they mean about human cognition. Regular readers of LessWrong will already be familiar with the idea that these cognitive 'mistakes' are often edge cases stemming from cognitive processes that are adaptive and efficient in helping us make suitable decisions in real-life situations. Gigerenzer popularised this perspective with his notion of "fast and frugal heuristics", and nowadays cognitive scientists are adopting this viewpoint when explaining cognitive biases. They posit that our cognition is "resource rational" and performs well given its finite resources, particularly time.

Nevertheless, the discovery of a long list of biases harming our ability to make good decisions is frequently used as a summary of the discipline of behavioural economics. The notion that a core outcome of the discipline is the representation of people as irrational was a crucial selling point in Ariely's book "Predictably Irrational".

I would argue that the concept of "biases" in this literature isn't as simple as it may appear. Unlike visual illusions, where biases can be evaluated as misperceptions compared to an objective reality, in the realm of decisions, biases are gauged against a benchmark of rational decisions. A notable point made by Gigerenzer and Goldstein in their 1996 article is that Kahneman and Tversky's school of thought critiqued the economic principles of rationality as descriptive principles, but maintained them as normative principles. They upheld the view that people should adhere to these principles.

This position, however, is debatable. Firstly, individuals frequently breach the economic "axioms" of rationality, even when economists painstakingly elucidate them. An axiom is supposed to be a self-evidently true principle. If individuals choose not to follow the economists' axioms after thorough explanations, it's unclear whether economists can assert they violate evident principles of rationality. Secondly, some principles are simply beyond human cognition's grasp. An example often presented as a simple principle is completeness (the notion that people should be capable of choosing between any set of goods at no cost). Combinatorics quickly shows that the number of such sets in a typical supermarket is astronomical, far exceeding the number of atoms in the universe. It's unsurprising that in many decision situations, people hesitate and are uncertain about their preferred option. They need time to make up their minds, a process entirely overlooked by the rational economic approach when completeness is assumed.

Thus, the accumulation of "biases" presented as key findings from behavioural economics largely stems from an unrealistic benchmark retained from traditional economic models, coupled with an underappreciation of the complexity of the problems humans face in real-life situations and the efficiency of our cognitive processes in making sound decisions most of the time. I believe this perspective is changing in the field, and there is renewed interest in functional/adaptive explanations of human cognition, a point also underscored by Benabou and Tirole in 2016. The recent scandals involving behavioural scientists may hasten this reassessment of how we should view human decision-making processes.

This post builds upon the Substack post linked, where I expand on these points.