Sunday, May 16, 2021

How We Choose What We Do

 From Inference Review, Vol. 6, Issue 1:

The reviewer's mini-bio at Inference:

Jean-Paul Fitoussi is Professor Emeritus of Economics at the Institut d’études politiques de Paris and Professor at Luiss Guido Carli University. He is also a member of the Center for Capitalism and Society at Columbia University. In January 2021, Fitoussi was elected Vice President of the Fondation nationale des sciences politiques. His last books are Measuring What Counts: The Global Movement for Well-Being, with Joseph Stiglitz and Martine Durand, and Comme on nous parle: L’emprise de la novlangue sur nos sociétés


Willful: How We Choose What We Do
by Richard Robb
Yale University Press, 256 pp., $18.00.

How do we choose? None of us knows. What is known precisely is not complete, and what is known completely is not precise. Richard Robb’s research lies on the boundary between economics and philosophy; it is a ragged boundary and so a tough discipline. In thinking about choice, Robb commits to a thesis that is both economic and philosophical: people choose by calculating what is optimal or by acts that escape calculation. In looking to optimization, Robb is in good company. Kenneth Arrow, Gérard Debreu, Paul Samuelson, Amartya Sen, and Herbert Simon have all employed and criticized the canonical answer: consumers maximize their utility under budgetary constraints; entrepreneurs, their profits under technological constraints.

Arrow argued that making a rational choice through calculation is more than complex—it is often impossible. When trying to determine the outcome of an action, the real challenge is not so much factoring in one’s own rationality, but that of others. No computer can solve an optimization problem on these terms. If we do not have any idea about the consequences, how then do we make choices?1

George Akerlof proposed adding an objective to the optimization problem: people do not simply want to economize or maximize their profits, they “want to be ‘rich and famous’—the and-famous part of the expression not being redundant.”2 The desire to be famous sometimes leads people to make irrational choices. Clearly, another approach is needed.3

Simon, in turn, proposed that humans exhibit bounded rationality: an agent looks merely for a satisfactory solution because he does not have the cognitive means to achieve optimization. This form of rationality allows greater freedom to act since the agent does not have to seek an unattainable optimum. In this sense, bounded rationality is an accommodation between the theory of rational choice and that of free action.

Robb wishes to maintain some distance from the theory of rational choice. As both a scientist and an engineer, he is, by nature, a pragmatist. This leaves him no room for dogmatism. Why, he asks, throw out an instrument such as rational choice theory when it could still be utilized, and when there is no comparable replacement available? And this all the more so when its shortcomings are known: rational choice theory presents a mechanical vision of the world, reducing every life to a mere choice.

Robb acknowledges a larger class of actions within economics that seem stubbornly retrograde to rational decision theory. “Early on in graduate school,” he writes,

my classmates and I stumbled on behavioral economics, which was then emerging as an alternative to rational choice orthodoxy. Cognitive biases were documented in all sorts of lab experiments. In one famous experiment, subjects were indifferent between receiving $10 immediately and receiving $21 in one year. They were also indifferent between paying $10 immediately and paying $15 in one year. Since a rational person ought to be willing to trade small amounts of cash now for cash in one year at a single discount rate, whether paying or receiving, this discrepancy was interpreted as evidence of “gain-loss asymmetry”—meaning that people need more compensation to delay gains than they are willing to pay to delay losses.4

However curious or compelling the examples of behavioral economics, in the end, Robb is persuaded, they can be mostly folded within the ambit of optimization:

Behavioral economics assumes that people understand their preferences, but that defects in their mental apparatus impair decision-making. At least one hundred and fifty behavioral biases have been identified, mostly through laboratory experiments, from the “ambiguity effect” (ruling out options when we can’t assign probabilities to possible outcomes) to the “zero-risk bias” (spending unwarranted amounts to reduce small risks to zero while ignoring bigger ones). Presumably, once people are made aware of their biases, they will try to correct them, choose more wisely, and become better off. Until then, the field seeks to build more accurate models of behavior.5

In this respect, irrational behavior is no more a violation of decision theoretic principles than a perturbed planetary orbit is a violation of Newtonian mechanics.

There yet remain actions that are well defined but neither rational nor frankly irrational. Some actions, Robb argues, are undertaken for their own sake “without regard to whether [they are] better than some alternative.”6 No calculations are involved. Actions of this kind “cannot be ranked against, or traded for, other actions.”7 They are neither rational nor irrational. An act of this sort, Robb writes, is undertaken for-itself. These actions make sense, Robb argues, “only if we accept that an activity can matter beyond its ostensible purpose.”8