At 03:46 PM 8/17/01 -0400, you wrote:
>All well-taken. The political problem, as I see it, is that
>critics of these people have no counter-science, theory, or
>evidence. They are reduced to emotionalism. The best they
>can do is ask people like me to find errors in the other
>side's arguments. But all I can do is find errors given
>the operating premises of the other side, which is like
>spotting them a pair of touchdowns.
but what about the doubts by NC types about the NC economists' own theory?
See below, for example. In addition, there's a growing body of experimental
evidence indicating that we should reject the “economists’ canonical
assumption that individuals are entirely self-interested: in addition to
their own material pay-offs, many experimental subjects appear to care
about fairness and reciprocity, are willing to change the distribution of
material outcomes at personal cost, and are willing to reward those who do
not, even when these actions are costly to the individual” (Henrich,
Joseph, Robert Boyd, Samuel Bowles, Colin Camerer, Ernst Fehr, Herbert
Gintis, and Richard McElreath. 2001. In Search of Homo Economicus:
Behavioral Experiments in 15 Small-Scale Societies. _American Economic
Review_. 91(2) May: 73-78.) If the standard economic view of humanity is
avoided, perhaps the standard version of cost-benefit analysis will be
replaced...
August 11, 2001
FINANCE & ECONOMICS
Averse to reality
A GREAT deal of economic theory turns on how people cope with risk-one of
the least escapable facts of economic life. The model that most economists
rely on when they need to take account of risk in their pure or applied
research is expected-utility theory. The trouble is, this theory has
implications so absurd that it cannot be true.
People are commonly observed to be "risk averse" in everyday life-that is,
they reject better-than-fair gambles. Suppose you were offered a 50-50 bet
that paid you $11 if you won and cost you $10 if you lost. Given the odds
and the pay-offs, the expected return for accepting this bet is 50 cents
(50% of $11 less 50% of $10); and since this is a positive number, the
gamble looks attractive. The fact that many people do turn down such bets
does not trouble expected-utility theory. It has an explanation:
diminishing marginal utility. As your wealth rises, each extra dollar is
worth less to you than the previous one. Because the utility of extra
wealth declines, it is not necessarily illogical to attach a lower
subjective value to the upside of the gamble (50% of $11) than to the
downside (50% of $10). All seems well: the facts and the theory sit
comfortably together.
Unfortunately, if you think about it, they do not, as Matthew Rabin of the
University of California, Berkeley, and Richard Thaler of the University of
Chicago point out in a recent article*. Consider the bet described in the
previous paragraph, and imagine some unremarkably risk-averse person who
turns it down. Now ask yourself this: knowing nothing else about the
person, and assuming expected-utility theory to be true, how big a prize
would you need to offer in a 50-50 bet to persuade him to risk losing $100?
Knowing he turned down the $11 prize, you might guess it would have to be
more than $110. Would $220 be enough? The expected pay-off of that bet
would be $60 (50% of $220 less 50% of $100). Looks good-yet our putative
risk-avoider would still turn it down. Things get worse. What about $2,000?
He would turn that down as well. All right, $20,000. No, still too risky.
Very well, $2m; wait, what the heck, $2 billion. Still no. Given only what
you know about this risk-avoiding person, plus the truth of
expected-utility theory, you are forced to conclude that he will reject any
50-50 bet costing $100, regardless of the prize.
Risk-aversion of this degree is literally insane-yet rational, according to
the theory. What is going on? To understand what the theory is doing, as Mr
Rabin and Mr Thaler explain, you need to follow along with some arithmetic.
Suppose that the person's initial wealth is W. Then rejecting the original
lose-$10 gain-$11 bet implies that on average he values each of the dollars
between W and (W + 11) by at most ten-elevenths of the average value he
puts on dollars between W and (W - 10). This implies that the value he puts
on the W + 11th dollar is at most ten-elevenths of the value he puts on the
W - 10th dollar. In effect, then, our subject's marginal utility of wealth
falls as his wealth rises, and rises as his wealth falls, at a rate of
around 10% for every change of $21. This phenomenally powerful multiplier
so inflates the value he attaches to a loss of $100, and so deflates the
value he attaches to any gains, that no gain can be big enough to make the
bet seem attractive.
The absurdity, as Mr Rabin and Mr Thaler emphasise, is not a trick
reflecting particular assumptions, but is wired into the standard theory.
"Expected-utility theory says risk attitudes derive solely from changes in
marginal utility associated with fluctuations in lifetime wealth. Hence,
the theory says that people will not be averse to risks involving monetary
gains and losses that do not alter lifetime wealth enough to affect
significantly the marginal utility one derives from that lifetime wealth."
The theory, in other words, implies that people should be risk-neutral
towards gambles involving small stakes-but they aren't.
Odds and ends
The question is, how to make sense of the fact that people will reject
small-stake gambles and yet accept, as they are wont to, moderate-stake
gambles provided the terms are good. The authors call for an approach based
on two ideas (both mentioned on earlier occasions in this space): loss
aversion and mental accounts.
Loss aversion is the idea that people feel the pain of a loss more acutely
than the pleasure of a gain of equal size: changes in wealth, and their
direction, are what count, regardless of levels. This directly explains why
people turn down even very small gambles with positive expected gains.
Mental accounting plays a complementary role. It is the idea that people
judge financial risks in isolation, rather than alongside overall wealth
and other risks. Small, better-than-fair gambles may look irresistible in
relation to total wealth, because any losses will be negligible in that
context; judged in isolation, especially given loss aversion, such gambles
are much easier to turn down.
The authors argue that such decision isolation is pervasive, and explains
many otherwise perplexing features of economic life-from the
"equity-premium puzzle" to the otherwise contradictory facts that (a)
lotteries are popular and (b) people are willing to pay outlandish prices
to insure themselves against easily affordable losses (as with, for
instance, optional extended warranties on consumer durables). The evidence
against the expected-utility approach seems overwhelming, and the broad
shape of an alternative, thanks to the earlier work of Mr Thaler and
others, is reasonably clear. The greatest puzzle, perhaps, is that the old
theory has not yet been discarded.
*"Anomalies: Risk Aversion" By Matthew Rabin and Richard Thaler. Journal of
Economic Perspectives, Volume 15, Number 1.
Copyright © 2001 The Economist
Jim Devine [EMAIL PROTECTED] & http://bellarmine.lmu.edu/~jdevine
