Tuesday, February 18, 2020

Instrumental and epistemic rationality - worth thinking deeper about decision behavior and intelligence



Rationality has been defined as either being instrumental and epistemic. See the work of Keith Stanovich, the cognitive scientist focused on the psychology of reasoning and others for more details. Given this construct, it is relevant to think about the different forms of rationality to help explain why investors make good and bad decisions. Investor may act rational in that they are trying to meet their goals. The problem is their belief system may be wrong.

Rationality is not the same as intelligence, so someone who is intelligent may not always act rational. Similarly, if there is a range of intelligence, we should also expect a range of rationality.  It is not just a yes-no issue.

Instrumental rationality, according to Stanovich, means behaving in the world so that you get exactly what you most want, given the resources (physical and mental) available to you. You are and act like an optimizer for your goals. 


However, there is another form of rationality. Epistemic rationality is concerned with how well beliefs map onto the actual structure of the world. 


This is often referred to as theoretical or evidential rationality. This is where there is more potential for irrationality. You are wrong or irrational because your beliefs do not match with reality. There is rationality in your implementation, but it can be based on wrong beliefs or belief updating. If you are epistemic rational, you will be skeptical of unfounded beliefs and change beliefs as new evidence is presented.

This combination is the rationality that we often talk about as economists in an expected utility framework. It is process driven can be described through the axioms of choice. If we follow those axioms, then we are behaving rationally. If we act to fulfill our goals, we are rational regardless of the view of others. Given the axioms of choice, it relatively easy to measure whether someone is acting rational or committing errors, just check for violations of consistency. However, these behavioral tests do not measure intelligence, and do not tell us much about rationality of the individual


Rationality can be consistent with wrong beliefs. There can be optimization of beliefs that are wrong or have the wrong likelihoods. Someone can follow the axioms of choice but make bad decisions on beliefs. This closer breakdown of rationality does not dismiss biases and errors in decision-making. Rather, these differences in rationality require us to look deeper on why mistakes are made. 

There can be degrees of rationality like degrees of intelligence. You are on both an intelligence and rational spectrum. The rational person blends the reflective mind of beliefs and goals with algorithmic mind that processes information. To truly understand the behavioral finance revolution requires not just an understanding of decision errors but the forms of rationality and why they may change. Picking the right portfolio manager is looking at the blend of rationality and intelligence. 

No comments: