Today I want to begin a review of two different accounts of human preference. One is George Stigler and Gary Becker’s seminal ‘De Gustibus Non Est Disputandum’. The other will be Bryan Caplan’s ‘Stigler-Becker versus Meyers-Briggs: why preference-based explanations are scientifically meaningful and empirically important’.
As setup last time, Stigler and Becker are seeking to extend the extend the use of neoclassical marginalist analysis beyond traditional markets:
[W]e are proposing the hypothesis that widespread and/or persistent human behavior can be explained by a generalized calculus of utility-maximizing behavior…
They focus on four areas: addiction, custom and tradition, advertising, and fashions and fads. All of these analyses are framed with ‘the new theory of consumer choice’, which is a rather clever extension of utility-maximizing theory. Normally, when modelling consumer utility, consumers choose a bundle of goods from the marketplace based on whatever gives them the most utility, and then consumption takes place immediately.
But when you buy a sack of potatoes, you don’t eat them as soon as you finish with checkout. The end product you will consume is not yet finished; the potatoes have to be transformed into a delicious massaman curry through further use of labor time and combination with other goods.
So instead of instantaneous consumption, consumers or household use goods from the market to create ‘commodities’, which are then used to produce utility. Simply by adding another step to the consumption process, a whole new range of behavior can be modeled.
The next extension to consider is the possibility of adding ‘capital’ to this production process. Capital is anything that aids in production and is not labor; it is not relegated to physical machinery. Practice is a form of capital building; you can work a little bit faster next time, get a little higher quality. You could buy a really nice pan, or take a cooking course. This notion of building ‘consumption capital’ is the key to the argument, as applied to a range of human behaviors.
One of the main categories of behavior Stigler and Becker analyze is addiction. Normally behavior like addiction is seen as a change in preference: you didn’t have a taste for cigarettes, then you were exposed to them however many times, and now you have a preference for them. But using the capital consumption framework, both good and bad addictions can be explained holding tastes constant. The examples considered are music (good) and heroin (bad).
Imagine consuming music is not merely the act of listening in the moment; surely that yields some utility, but as with other arts, experience and education can be combined with the listening experience to obtain even greater utility. For example, my appreciation of Beethoven’s symphonies was greater enhanced by learning about and listening to his contemporaries, the dominant structure of music at the time, and how he pushed against the formalism of the day.
So too it is with the Beatles; when I listen to the Fab Four, I’m not merely appreciating the production, the chord changes, and the songcraft; I’m also thinking about the cultural moment, the turn in the post-war British and American culture, the pressures of fame, interpersonal group dynamics, human frailty, and so on. All of this comes from having spent years increasing my ‘consumption capital’ of music. (For the other fans in the audience, someone has created an augmented Anthology podcast that I really enjoyed).
And when one has built up such consumption capital, the return the additional music listening becomes even greater; a music ‘addiction’ has formed:
On this interpretation, the (relative) consumption of music appreciation rises with exposure not because tastes shift in favor of music, but because its shadow price falls as skill and experience in the appreciation of music are acquired with exposure.
With a harmful addiction, the math is the same with the signs flipped. Consuming heroin destroys consumption capital in this model, so more and more of the drug is needed subsequently to obtain the same level of utility. Again, no change in tastes is necessary to explain the junkie’s behavior.
Of course, we know that in the real world, people *do* have differing susceptibilities to addiction, some of which seems to simply be genetic. But this shouldn’t be taken to mean the Stigler-Becker model is wrong; it may often be more useful to model addiction with the consumption capital concept than to develop a model based on brain chemistry.