Monday, December 9, 2013

Just Because You're Paranoid Doesn't Mean They're Not Out To Get You: Revealed Preference Edition

Giacinto Gimignani, An Angel and a Devil Fighting for the Soul of a Child, via Wikimedia Commons

Maybe you've had an inchoate sense that dark forces are aligning against you. Maybe you're scared. Maybe you're too bored to think about it for more than five minutes. Well you're in luck, because to preferred clients of TKIN like yourself, we are proud to offer our premium service: we think it through so you don't have to.

This week, how advertising and cost-benefit analysis bond in unholy matrimony, spawning The Policy Methodology From Hell.

As we all know, if you're wondering how to make policy decisions, one popular answer is "cost-benefit analysis." And as we all know, if you're wondering what to count as a cost and what to count as a benefit, one popular answer is an "economic" one. What a person prefers is what is a benefit to them; giving up what is preferred is a cost; where for a preference set to be rational just means that it obeys formal consistency axioms like transitivity.

Maybe less well-known is that if you're wondering how to tell what preferences a person has, one popular answer is that the preferences to count are "revealed preferences." If you chose x over y, you manifested a preference for x; if you paid good money for that future-item-in-a-landfill tchotchke, you showed that you revealed a tchotchke-related preference.

It's obviously not news that there's something shady about the use of revealed preferences to reason about what is good and bad. What people choose is shaped by what options they have. And in classic "sour grapes" fashion, people who have no access to a given option sometimes come not to prefer it -- their preferences are "adaptive." Also, people might choose on the basis of false information. No one thinks the person who mistakenly ingests pesticide in food has some kind of latent preference for death.

But one another interesting fact about revealed preferences doesn't come up as often, overshadowed as it is by all these other things. That has to do with weakness of will. It's evidently a problem, because if you thought choosing y was for the best, but you caved to pressure/cravings/madness for x, then your choosing x plausibly does not plausibly reflect a way that x is good for you, and so counting your revealed preference as a preference would be a mistake.

It's easy to get lost in the philosophical thickets of weakness of will, meandering around like a drunk person, trying to figure out what could possibly determine the difference between "I chose it even though I thought it best not to" and "I chose it because I changed my mind" and other important distinctions. But perhaps we can say this: there is sufficient agreement on the concept to allow for a robust research program in the social sciences. Roy Baumeister and his colleagues have conducted study after study measuring the effects various things have on weakness of will.

One conclusion they came to will surprise no one who has lived a human existence, and that is that the will is a thing that can be worn down.

Ask people to exercise their will -- eating radishes instead of chocolate, keeping a straight face when something is funny, etc. etc. -- and they'll become less and less able to effectively exercise their will. They call this "ego depletion." The "active self" is a "limited resource."

Among other weird things, this means that they more you have to resist temptations, the more you deplete your ego. The more you deplete your ego, the less you're able to resist temptations. So the more environmental factors there are requiring resistance -- the more bakeries you have to walk past, the more click-bait you have to ignore, the more advertised consumer items you have to not buy, the more likely you'll exhibit the weakness of will.

And this is where it gets interesting. Because when you put "the preference taken into account in cost-benefit reasoning fail to track your good when there's weakness of will" together with "the more you have to resist, the less you can resist," what you get is that the massive forces always present in a consumer society, urging you to BUY EAT DRINK ENTERTAIN YOURSELF YOU DESERVE IT, are actually making your revealed preferences less likely to be for your actual preferences.

Let's do an example. In one of their Freakonomics books, Levitt and Dubner said that fighting climate change through change in behavior was pointless because "It's not that we don't know how to stop polluting the atmosphere. We don't want to stop, or aren't willing to pay the price."

Their evidence for the implicit claim about preferences is our behavior: our behavior reveals preferences for convenience and fun over environmental goals. Plausibly, our actual preferences are sometimes for clean air and a world inhabitable by our children even when immediate forces make it hard to resist the temptations of climate damaging behavior. So basing judgments about our good based on our revealed preferences is a planning FAIL.

What's crazy about this whole thing is that the more other stupid temptations we have to resist -- the more we're subjected to BUY EAT DRINK ENTERTAIN YOURSELF YOU DESERVE IT -- the more likely we'll fail to act in accordance with our judgments, and thus the more likely that cost-benefit analysis will track our preferences for stupid consumer goods instead of, you know, clean air and water.

Remember the old picture of the angel on one shoulder and the devil on the other? This is like the devil has two kinds of minions: those shouting at you to listen and obey, and those who watch you obey and conclude: Well, we know what will make that person better off, nudge nudge wink wink.

You see why I called it an unholy alliance, eh?

1 comment:

SEO Moz said...
This comment has been removed by a blog administrator.