Saturday, February 24, 2007

With us or against us

Playing soccer today, I found myself asking, where is the self-interest? Economists assume humans are, for the most part, rational and self-interested. And when we're not, it's called altruism, or sympathy, or commitment, and it's modeled as an exception to the rule of individual utility maximization.

As Larry Iannaccone quotes David Friedman, this is "false, but useful."

But on the soccer field, it's not even useful. You can't predict how things will unfold by assuming each player is maximizing his shots on goal, his saves, his assists, etc. You can say he is maximizing his utility, but what the hell is that? His utility is almost completely determined by the team's objective, which is to win. And this is of course at odds with the opposing team's objective. By thinking of group interest, you can predict, to some extent, which way he'll push the ball. He'll pass to his own team mates, he'll shoot on the opponent's goal, etc. He won't take the ball and do cool tricks with it, generally, unless it serves the objective of winning.

Now, not to fret, my fellow libertarians and economists, I still think in most situations we are largely self-interested. For instance, within the firm, we are largely motivated by our wages. And within the nation, we are largely motivated by our self-interest, be it money or whatever. But not entirely. For instance, voting is a mystery when viewed as self-interested behavior. Where's the payoff? But it makes sense when viewed as satisfying the group interest, where the relevant group may be one's nation, political party, neighbors, friends, or office mates. Within the family, group-interest and self-interest are probably about equally important, and also quite intertwined.

The bottom line is I think we should, in many cases, model behavior as a combination of self-interest and group-interest. And, to make it mathematically tractable, assume the two are separable, so it's a nice linear combination.

Many questions arise:
1) How do we weight these interests for various groups? Where do we get the data to make that determination?
2) What is the optimal size of the group? Is it whatever it is? How did we arrive at 11 for the size of a soccer team? And is the EU appropiate as a political or even economic unit?
3) Are some groups detrimental (I'm thinking of nations)? Is that because of the size? Or because of the tendency of politicians to screw with the economy? Cliques can be very limiting, and often innovation comes from the outside, a'la Granovetter's Strength of Weak Ties. On the other hand, forming associations helps to alleviate public goods problems, and facilitates networking, and allows Smithean specialization and the division of labor.
4) To what degree are various groups substitutes for each other, e.g. the State for religion? And to what degree are they substitutes for individual interests? How do groups shape individual interests, and vice-verca? This week I saw Jack Citrin present this paper, which indicates that ethnicity, i.e. identifying primarily with one's ethnic group, and nationalism are to a large degree substitutes, with nationalism largely replacing ethnicity over time among immigrants.

Here are a few of the more common groups I can think of:
1) Geographic - a) Nation, b) Region, c) State, d) City, e) Neighborhood
2) Economic - a) Currency union, b) Trade linked, and free-trade zones
3) Politics - a) Party, b) Issues
4) Language and Cultural heritage
5) Physical attributes - a) Ethnicity, b) Color, c) Gender, d) Other features
6) Intellect - a) College degree, b) Conversational skills, c) Comedic skills, d) Analytical skills, e) Social skills, f) Cultural interests - theater, art, music
7) Income - a) Job status, b) Other signs of wealth
8) Company/Firm
9) Sports team - a) Participant, b) Spectator
10) Religion - a) Denomination, b) Church


Someday Scientist said...

Altruistic behaviour is, in fact, often accompanied by underlying self-interest.

In animals, when altruistic behaviour occurs, it is often a form of kin-selection: they put themselves at risk for the sake of passing on their gene pool - through relatives. this normally occurs in social animals that live in groups, such as rodents.

there is something called "reciprocal altruism" in which one anmal does something for the other, and the stability of the group allows for the other to reciprocate at some point in their life. this normally only happens in very stable groups in which all those included in the population are certain that the other will have ample oppurtunity to return the favour, such as in humans.

so, not even altruism is truly altruistic. there's always an element of "how can this work out for me?" even if it's buried deep within the subconcious.

aes said...

Optimal group size is an interesting social question, for despite every institution's instinct to "grow," there is a bureaucratic-like point of diminishing productivity which eventually forces the group to subdivide (like we were talking about with the ASA community group). On a soccer field for example, you could comfortably add more players if you expanded the field, but I am sure the added distance would reduce communication among the team as a whole. On top of that, the increased ratio of players to the the ball would reduce individual participation, causing a player to question her contribution to the team. Plus, it's no fun standing around the field when the ball is stuck on the other end for most of a half. So it seems the optimal group size would be whatever allows optimal participation and "utilization" of individuals in the accomplishment of a given objective (or group interest).

Will McBride said...

Someday Scientist, I understand reciprocity, and agree it's behind much of what we call altruism. But this is pick-up soccer, often completely anonymous, where no one exchanges names and different people show up every week. I play the same way regardless, e.g. without any expectation of collecting on my reputation for being a fair and skillful player. Yes, it can be reduced to self-interest, in that the game is most fun when everyone plays this way, i.e. as a team. But for modeling purposes, it is more direct, and maybe simpler, to treat it as group-interest. Otherwise, you're just jamming whatever makes you happy into the utility function, which is dangerously close to a tautology.

Of course, the problem is that we don't know what makes groups tick. We don't have a law of demand for groups. Sociologists seem to know something, if you can sort through the mumbo-jumbo. I think Granovetter is quite insightful. And psychologists, like Judith Harris.

Will McBride said...

aes, you play soccer, so you know there's a fair amount of flexibility in the optimal size. A pick-up game of 7 on 7 works out pretty well, as long as you shrink the field accordingly. Even 2 on 2 is kind of fun, but more predictable than 7 on 7, which is more predictable than 11 on 11. Predictability is probably a bad thing here. So predictability is minimized, subject to the crowding problem and field size problem you mention.

Maybe that's why we engage in massively wasteful and arcane social democracies. It adds unpredictability, the spice of life. So don't ask an economist where inflation is heading, or who will be elected. Don't ask anybody.