Gee, and I thought it was just that I didn't eat right or take the time to exercise. Who knew all this was going on in my head.

The "convex utility function" thingy made me laugh. It is a good euphemism -- instead of calling somebody fat, well, let's just say her utility function is pretty convex.

Eh, it's not all that weird thinking that different incentives are going to produce different behavior.

2: The "concave utility function" makes a pretty good euphemism as well.

To derive maximum value, one should make use of both the convex and concave utility functions.

ATM.

But... convex utility functions don't really exist. They'd imply increasing utility returns to the given attribute or product, which is extremely unlikely because it is very likely to cause an out-of-balance situation where people give up everything else to pursue the increasing returns. Sure, it might work for the few really pathological anorexics and bulemics, but it's a bad model for the majority of the population.

The idea of the attractiveness of thinness as a mostly relative comparisons sounds good. If everyone is mostly trying to be thinner than those around them, and thinness is closely correlated with class, wealth, age, certain occupations and localities, etc., then you'll end up with some people with all the factors pointing toward thinness having to work very hard to stay on top while others can slide a bit further into obesity as those below them on the thinness ladder get fatter.

it's not all that weird thinking that different incentives are going to produce different behavior

Sure, according to Econ 101...

They'd imply increasing utility returns to the given attribute or product, which is extremely unlikely because it is very likely to cause an out-of-balance situation where people give up everything else to pursue the increasing returns.

An example given elsewhere in the paper is elite athletes -- the difference between being third and second at Wimbleton isn't as great as the difference between being second and winning. (I've seen this described elsewhere by someone noting that figure skating is the highest-stakes event in the Olympics -- the difference between being a champion and a runner-up is the difference between a career skating under your own name, or in a Goofy suit.)

This paper irritates me, and I haven't even read it. What is it trying to get at, exactly?

Yes, a convex utility function could theoretically explain the "diet in the face of societal gluttony" phenomenon. But is it a good explantion? That the reason some people are getting skinnier and skinnier is because they're "deliberately choosing" to "invest more" in slimness principally because the greater the weight gap existing between fat and thin, the happier the thin people are? Does that really "make sense when you think about it"? (To anyone other than an economist struggling to explain all human phenomena in terms of costs and benefits and rational choices?)

Also, even accepting the worldview of an economist, shouldn't this yield some testable hypotheses: that extreme thinness would be more common in areas where overall comparison population was relatively fat, for example? Lots of (rationally chosen!) anorexia in Mississippi and Alabama, much less in Manhattan. Lots of anorexia among the relatively plumper middle-aged demographic, relatively less among the generally fit and trim college crowd. I think these hypotheses are demonstrably false.

the greater the weight gap existing between fat and thin, the happier the thin people are?

Wait, but is it that the *magnitude* of the weight difference is what's important? I'd imagine that it's indirect - fatter people and more of them make thinness more valuable, and the result may be a tendency toward increasing thinness, but as a result of the effort to secure the fat/thin contrast as it becomes more valuable, rather than a "5% thinner = 5% better" kind of calculation.

That was not very artfully worded.

I think it makes sense in terms of people under social pressure to separate themselves from the masses; once unusually low weight is a marker for, say, being a famous pretty person (fashion industry powerful person, whatever), it seems perfectly reasonable that you'd get an arms-race effect among people jockeying for those positions, doesn't it?

I'm not sure, actually, that the concave/convex utility functions bit works -- what makes more sense to me is that it's all keeping up with the Joneses, and the question is who your reference population of Joneses is.

an arms-race effect

Let's leave Professor Althouse out of this.

And even if this paper is right on, it seems obvious that the dynamic described is vastly outweighed (ha!) by the familiar common cause of more fatness and higher status for thinness - availability of cheap, low-nutrient calories, and scarcity of leisure time.

7: Well, that's true, but it only ever works for very small portions of the population who are able to top out.

There are also increasing returns to education for really smart people, since a college degree adds much more to their earnings than the final high school degree and a professional/grad degree adds even more. But then you top out.

The way the convex/concave thing comes in is this: if you have a convex utility function, then while more of something is good, twice as much of it isn't twice as good. If you have a concave utility function, then twice as much of something is more than twice as good. So if your neighbor gets fatter, and the only thing you care about is how fat you are compared to your neighbor, then with a convex utility function you will tend to keep a fixed differential with your neighbor. If you have a concave utility function, then you have more "relative thinness", which makes you want even more relative thinness, since twice as much relative thinness is twice as good.

9- the excerpted analysis is all about the magnitude of the difference.

10- it seems to me that anything sensible you've said here has nothing to do with the article (at least based on the excerpt).

Eh, I cite Reynolds' Defense: I was just linking, not claiming that what I was linking to made any sense. (Well, other than claiming that it 'made sense when you think about it' but, um, I didn't type that bit. I was kidnapped! and miscreants retitled the post! Which was originally "I have absolutely no opinion!")

This thread reminds me of a couple of jokes I from time to time.

1. My plan to bring a nonstop flow of desserts to work to fatten everyone around me, making myself look relatively more attractive.

2. My plan to urge all good looking men to turn gay, and thus raise my relative standing in the remaining pool of hetero guys.

14: but that works! At least IME. I'm skinny anyway, but when my weight fluctuates naturally to its lowest points I find myself wanting to be even thinner because I'm kind of enjoying the sheer unusualness of my slimness. I like fitting in to the smallest sizes in the shop. Um, it gives me an irrational and completely abhorrent sense of superiority. In other words getting stupidly thin triggers a kind of latent-anorexic tendency in me. Luckily I both lack willpower and love my boobs. I must read this paper now. Dammit.

(having read the abstract it sounds like they aren't proposing that the majority of the population have this convex utility function. So I think it works.)

My plan to urge all good looking men to turn gay, and thus raise my relative standing in the remaining pool of hetero guys.

Hilarious. With a sense of humour like yours, no wonder your ranking needs a lift.

Good progress, by the way; not that you asked.

Why on earth use a dead president handle for that comment?

Because of his ongoing personal story, under that name because of the subject matter. See last sentence of 19.
A project worthy of the Hallel, in this week, particularly Ps. 113.

See last sentence of 19.

Ah. I should have thought of that. Seems so weird to me to have that be something to hide in this day and age.