Re-Thinking Fast and Slow
Abstract
Daniel Kahneman’s book Thinking, Fast and Slow (2011) has had a worldwide impact. The book’s insights are profound and have changed the thinking of both decision scientists and general audiences about how choices are made. Kahneman, however, claims that standard utility theory cannot explain these insights because it 1) lacks “reference points” from which gains and losses can be measured, 2) does not predict loss aversion, and 3) assumes preferences are stable (amid supposed counter evidence). These alleged failures of utility theory are what led Kahneman and Tversky (1979, 1991) to develop prospect theory. This brief article shows that a close reading of Thinking, Fast and Slow reveals fundamental oversights in these criticisms. Not only does loss aversion arise naturally within utility theory for rational economic agents with stable preferences, but the very measurements of gains and losses rely directly upon reference points. Rather than overturning the insights of prospect theory, proper use of utility theory and its indifference curve representations reveals these behavioral insights and places them within the sturdier, longer-established framework of neoclassical microeconomic theory.
Report on “Re-Thinking Fast and Slow”
I think this short paper makes an important contribution, and I recommend publication.
The only revision I suggest to the author is to reframe the argument in section 2 in terms of marginal rate of substitution (MRS) rather than in terms of WTP and WTA.
In fact, in the standard economic theory the author “defends”, indifference-curve analysis is completely ordinal in nature. This means, among other things, that the no meaning can/should be attached to the “distance” between two indifference curves. The use of WTP and WTA in section 2 suggests instead, involuntarily but misleadingly, that we can quantify the “distance” between two indifference curves by WTP or WTA, which is not the case.
If the argument is reframed in terms of the MRS between money and leisure (as the author already does in section 3) this misleading suggestion would disappear.
I appreciate the recommendation for publication and the suggested change of emphasis of MRS over WTP/WTA. My preference is to keep WTP and WTA as those are the terms used in the psych literature (e.g., the analysis of Thaler, Kahneman and Tversky, etc. invoke WTA/WTP) on this topic. The literature on endowment effects, framing effects, preference reversals, etc. use WTA/WTP much more than MRS (at least, in my reading of the lit). I agree with the author that the term “distance” could connote a cardinal measure. I use the term only once and only as pertains to the physical distance between the indifference curves in the plot. It appears unambiguous to me and I recommend leaving it as is. If others express concern that it will lead readers to misinterpret the results, I will revise it for publication.
Thanks again for your thoughtful comments.
John