Revealed Preference

Monday, February 15, 2016
By dreeves

Paul Samuelson

The doctrine of revealed preference — that you can infer someone’s utility function based wholly on what they choose to do — has an illustrious history. John Locke said “the actions of men are the best interpreters of their thoughts.” And Ludwig von Mises said “the scale of values or wants manifests itself only in the reality of action.” [1] Strong words. Revealed preference remains the orthodox economics position. (Needless to say behavioral economics begs to differ.)

Tyler Cowen of Marginal Revolution applies the doctrine to procrastination and time management thusly:

All people are equally good at time management, but some people are more willing than others to admit that they are doing what they want to do, while others maintain the illusion they wish they were doing something else.

Will Wilkinson tears that apart in detail. In particular he points out the distinction between “doing what you’re motivated to do” and “acting in your own self-interest”.

Here’s a more concise reductio proof: People use commitment devices; therefore by revealed preference their preferences don’t match their actions. QED.

In other words, my use of a commitment device proves (by the orthodox economists’ own criterion) that I really do wish to be doing something else.

For any real economists in the audience I should point out that I’m being a little glib here. I do want to pile on with Will Wilkinson against Tyler Cowen but I’m not really trying to debunk Nobel prize winning economist Paul Samuelson who pioneered revealed preference in consumer theory. In fact, when Beeminder is sufficiently mainstream then the doctrine becomes perfectly reasonable: Whatever you’re doing, with the help of Beeminder [2], must indeed be what you most want to be doing!



[1] Thanks to Josh Jordan for both pointing me at those quotes (in the comments of a previous Beeminder blog post) and for first articulating the proof by contradiction.

[2] And friends.

Tags: , , , , , ,

  • Micro(chip)managed

    Viewed too narrowly, Cowen’s view leaves out a whole host of psychological motivations that are invisible to the observer (and often to the agent). He could be read as saying “When you watch 6 hours of Netflix instead of doing your research work, it’s because *that* (where “that” is, specifically, to watch Netflix) is what you *want* to do.” even at times when, in reality, the ‘thing you want to do’ is to avoid the anxiety of starting on a project you fear, having nothing at all to do with the tv binge. Actions need to be construed to include the approach/avoidance of mental states as well, whether or not that approach/avoidance is intentional, or his view would seem to leave out a rather large segment of motivations, IMO.

    The narrow view can be rejected as not complete enough an account. The view that includes avoidance/approach behaviours might seem a little trivial (it’s not that you’re bad at managing time, it’s that part of what you *want* is to be avoiding doing things that …make you nervous, remind you of unpleasant things, require a large investment for an unlikely reward, etc. etc. etc.), but it might still be worth remembering that it’s usually not *time* that we’re managing in these cases. Assuming there is a desire holding some otherwise undesirable behaviour in place might help suss it out and deal with it more directly.

    Or, and this is the way I prefer, you can overrule it all with Beeminder, making the desire to avoid paying your pledge trump everything!

    (PS – Of course, Cohen also leaves out all of the people for whom it *is* time they’re having trouble managing, like those with issues with executive functioning.)