Explaining Things
From time to time I come across a pattern of observed behavior that appears inconsistent with my economist’s model of rational behavior. I look for a way of making sense of it that, contrary to the initial appearance, is consistent with, ideally implied by, my model. If I can’t, I try to find a way to make the behavior consistent with the model by adding an assumption neither implied by nor inconsistent with the model.
If neither of those works I may have to accept the fact that the model is wrong, that I am observing behavior that cannot be explained as rational. At that point I can either give up on understanding the behavior or switch to an alternative model such as evolutionary psychology, the subject of my previous post.
Some examples:
The Lottery-Insurance Puzzle
Many people buy insurance even though they know that, on average, they will pay in more than they will get back, since insurance companies have costs of operation that they have to cover. By buying insurance they demonstrate that they prefer a more certain future to a less certain. Many people buy lottery tickets. By doing so they demonstrate that they prefer a less certain outcome to a more certain.
Some people do both.
The standard economic explanation of why we buy insurance is declining marginal utility of income, the value to you of an additional dollar going down as the number of dollars you have goes up. Buying insurance trades dollars in the future where your house doesn’t burn down and you don’t need them very much for dollars in the future where it does and you do. A lottery ticket does the opposite, trades dollars at your current income for dollars in the hypothetical future where you win the lottery, have much more money, and so should value an additional dollar much less.
The simplest explanation for someone buying both insurance and lottery tickets is that he has a utility function for which marginal utility first decreases and then, above some level, increases.1 That is logically possible but if you are free to make any assumption about utility functions you like in order to explain observed behavior economics becomes a theory that predicts nothing, since any behavior can be explained by suitable assumptions.
One solution to that problem is to offer an explanation for the utility function that has testable implications — and test them.
Suppose you are poor, all of your friends and relatives are poor, and it is assumed in your social circles that if someone badly needs money any friend or relative who happens to have some should provide it. That norm provides a form of social insurance but makes it hard to get out of poverty, since working hard and spending little makes you more likely to be the person expected to pay for a friend’s doctor bill or car repair. Buying a lottery ticket trades dollars that you can’t count on keeping for yourself for a chance of enough dollars to get out of the slums, away from friends and relatives, start a new and better life.
That explanation has a testable implication, that lottery tickets should be bought mainly by people who are part of that sort of social network. A sufficiently competent and energetic researcher could test that prediction.
If it turned out to be false, we would need another explanation. One possibility is to drop the rationality assumption in favor of an alternative discussed in an earlier post.2 Because humans have limited resources of attention and intelligence they often use rules of thumb that are imperfectly rational. Probabilistic calculations are hard. Easier to classify all outcomes as certain, possible, or impossible. Fire insurance converts a bad outcome, having neither a house nor money to replace it with, from possible to impossible. Buying a lottery ticket converts a good outcome, being rich, from impossible to possible. Both look like good deals.
I prefer a third explanation for the puzzle. Buying a lottery ticket gives you a small chance of winning a lot of money. It also gives you a certainty of the pleasure of imagining what you would do with the money, a fantasy more vivid, more compelling, with a link to the real world, a way it could happen.
Consider the lotteries you have been offered for the price of a postage stamp by Publisher’s Clearinghouse and similar enterprises. The reward if you win is offered not in money but as a new Cadillac, a Caribbean vacation, an income of $20,000 a year for life. Possible prizes are lavishly illustrated with glossy photographs as a prop for your imagination.
It strikes me as the most plausible of the three explanations but I have not yet thought of a good test of the theory. Suggestions welcome.
Some More Examples
My post on the economics of vice and virtue attempted to explain two puzzles, the existence of apparently irrational crimes of passion and the existence of people who will not steal even if they are sure nobody is looking. It is tempting to explain the first as a failure of rationality, the second by assuming a utility function that values honesty. I instead show that apparently irrational actions follow from rational commitment strategies. An earlier post applied the same approach to explain rights, considered not as a legal or moral category but a description of human behavior. Also to explain what Hobbes got wrong, why a society without an all powerful ruler is not the Hobbesian war of each against all.
The post before this dealt with a collection of puzzles: predictable restaurant lines, price control to prevent a price rise but not to force a price drop, hyperbolic discounting, the endowment effect. To explain those puzzles I abandoned economic rationality in favor of evolutionary psychology.
Mea Culpa
Quite often there are alternative explanations for the same behavior. I have a post and a book chapter defending moral realism, the claim that moral judgements are not arbitrary tastes but the perception of a non-physical reality. That is an explanation of virtuous behavior that most people, at least most non-economists, are likely to find more plausible than virtue as a rational commitment strategy. Economists, even good economists, have offered explanations for predictable lines at restaurants that do not depend on a belief in just prices hardwired into us by evolution — why do I prefer one that is?
The answer is that my professional activity is in part scientific, in part aesthetic. I value explanations of the world for being true but that is not all I value them for.
My favorite example is Gary Becker’s model of altruism. Becker starts with the simplest way of fitting altruism into economics, the assumption that the beneficiary’s utility is an argument in the altruist’s utility function. He uses it to show that altruism can be consistent with evolution via the Rotten Kid Theorem, a proof that it is only in the rational interest of a rotten kid to kick his little sister if doing so is economically efficient.3
The model, while internally consistent, is inconsistent with observed behavior.4
But a thing of beauty is a joy forever.
This was the approach of a classic article by Friedman and Savage.
One can argue that, rather than dropping the rationality assumption, the approach expands it by including attention in the resource constraints faced by a decision maker.
Becker’s model of altruism and the Rotten Kid Theorem are described in Chapter 21 of my webbed Price Theory, under the subhead “The Economics of Altruism.”
A relevant data point is that most lottery winners end up blowing all their winnings within a couple of years and end up poorer than they were before.
I think a simpler explanation is that people like excitement and that scratching a lottery ticket creates some excitement. Thus lotteries are no more mysterious than rollercoasters or scary movies or sports games or whatever. Of course, no one would say that paying for a rollercoaster ride is irrational simply because there is a 100% chance of losing your money.