Beyond Reason: How we really make decisions

By Edward Foster

When our decisions are irrational, neuroeconomist Al do Rustichini can help us understand why

Al do Rustichini
Students have trouble with the economist's standard assumption that people act rationally. They know, and so do we, that it is not literally true.

Despite these reservations, most economists think that in most circumstances the assumption can form the basis for good predictions. But people don't always do what seems rational. For behavior that deviates systematically from what economists think of as rational, the new field of neuroeconomics is yielding some explanations.

Aldo Rustichini, professor of economics at Minnesota, has been a leader in this emerging field, which uses dramatic new tools such as functional Magnetic Resonance Images (fMRIs) and Positron Emission Tomography (PET) scans to track brain activity while relying as well as on older tools for measuring phenomena such as eye movements, heartbeats, and galvanic skin response.

The core of decision theory resides in the brain

We might call Daniel Bernoulli's 1738 resolution of "the St. Petersburg paradox" the beginning of decision theory. The paradox arises from a lottery based on a series of coin tosses, with the payoff made when the first head comes up: $2 if on the first toss, $4 on the second, $8 on the third, and so on. The "paradox" is that despite the infinite "expected value" (theoretical average payoff) of a ticket to that lottery, the probability of a large payoff is remote, and no one would be willing to pay very much. Bernoulli resolved the paradox by proposing that the wealthier the lottery player becomes, the lower the value of extra winnings. His theory that lottery choices are based on maximizing subjective expected utility (SEU), subject to diminishing marginal utility of wealth, became a staple of decision theory and of economics.

Neuroeconomics has developed firm evidence that Bernoulli was right, and that the decisionmaking process summarized by SEU is hardwired into the brain -- and not just the human brain, but also the brains of other mammals. Moreover, even for humans the choices do not come through painstaking logical thought (which occurs in the prefrontal cortex) but rather from the parietal cortex (which integrates sensory information, is independent of language, and is shared by all mammals). Assessments are quick, and approximate. The result is the same for strategic decisions analyzed by game theory.

People do not always follow the theory

Economists and psychologists have found systematic deviations from the SEU theory. Here are three famous examples.

The Allais paradox:
Maurice Allais (Econometrica 21, 503 [1953] confronted subjects with two (hypothetical!) choices between pairs of lotteries. Each column in the table below describes prizes in a lottery, with each cell listing a possible prize and the probability of receiving that prize (M stands for million):

EXPERIMENT 1
EXPERIMETENT 2
Lottery 1A
Lottery 1B
Lottery 2A
Lottery 2B
$1M 89%
$1M 89%
$0 89%
$0 89%
$1M 1%
$0 1%
$1M 1%
$0 1%
$1M 10%
$5M 10%
$1M 10%
$5M 10%

With 89 percent probability, lotteries 1A and 1B will give the same $1M outcome, and lotteries 2A and 2B will give the same zero outcome. Each pair of lotteries differs only in the second and third rows, and the differences between A and B are identical in experiments 1 and 2. SEU says that, if you prefer an 11 percent chance of $1M to a 10 percent chance of $5M in Experiment 1, you should make the same choice in Experiment 2. In fact, most subjects choose 1A and 2B. But Allais did not present the choices as above; instead, he showed them:

EXPERIMENT 1
EXPERIMETENT 2
Lottery 1A
Lottery 1B
Lottery 2A
Lottery 2B
$1M 100%
$1M 89%
$0 89%
$0 90%
$0 1%
$1M 11%
--- 1%
$5M 10%
Lottery 1A is a sum of $1M to be received with certainty.

Rustichini and his colleagues (J. Dickhaut et al., Proc. Natl. Acad. Sci. U.S.A. 100, 3536 [2003]) have shown that when one of the alternatives is to be received with certainty, mental processing of the choice occurs in a different part of the brain than when both alternatives are risky: parallel processors are at work, with different kinds of problems routed to different processors.

The Ellsberg paradox:
Daniel Ellsberg (Q. J. Econ. 75, 643 [1961]) also confronted subjects with two pairs of lotteries, based on balls in an urn; the experiment illuminates the difference between risk (events with known probabilities) and ambiguity (events with unknown probabilities). The urn is known to hold 30 blue balls, and 60 balls that are each red or yellow, with the number of each unspecified. Then two experiments are run. The first offers the choice of betting that a ball drawn at random from the urn will be blue (probability of winning: 30/90) or red (probability of winning: unknown, call it x/90 where x is some number between 0 and 60). After making that choice but before drawing a ball from the urn, the subject is offered a better deal: the choice of betting that a ball drawn at random from the urn will not be blue (probability of winning: 60/90) or that it will not be red (probability of winning: 1 - x/90). The table shows the two sets of choices and the winners' payoffs.

EXPERIMENT 1
EXPERIMETENT 2
Bet on Blue
Bet on Red
Bet Against Blue
Bet Against Red
$100 30/90
$100 x/90
$100 60/90
$100 1-x/90

In experiment 1, most players choose blue, which implies according to SEU theory that they think x is less than 30. But in experiment 2 most choose to bet against blue, which implies that they think that x is greater than 30. To make sense out of these paired choices, economists have proposed that people do not like ambiguity; instead of forming opinions -- based on no information -- that x is above or below 30, they simply prefer to avoid placing a bet on the outcome about which they have no information.

Neurological studies suggest that this hypothesis is correct, and that parallel processors are again at work; the experiments with healthy subjects are confirmed by observing that subjects with brain lesions in the area where it is thought that these ambiguous choices are processed do not show the aversion to ambiguity that healthy subjects exhibit. When a choice is ambiguous, it appears that most people might consider the worst outcome that might occur and choose so as to maximize that worst outcome (what has been called a maximin strategy).

The ultimatum game:
The first of two players is given $10 and told to propose a division of the money between the two. The second player can either accept or reject the proposal; if she rejects it, neither gets anything. Game theory says that the first player should offer the least possible amount to the second player, who should accept it (since it is better than nothing). However, real life does not go like that. Even when the game is played anonymously on a computer, those first players who are foolish enough to offer only $2 or even $3 to the second are likely to end up with nothing; anticipating that, most first players offer more, with many offering an even split. Both introspection and brain scans suggest that there are again parallel processors at work, in this case with an emotional reaction to unfair treatment overriding the cool rational response of the game theorist.

Neuroeconomics illuminates the failures of standard theory

In each of these examples, observation of brain activity shows that the hardwired circuits of standard decision theory are not always in control. The brain contains alternative decisionmaking processes and an executive to decide which process to use for each task.

This should not surprise us. We have probably all experienced a sudden fright, with the hairs on our arms standing up and adrenalin pumping reflexively, all before we could consciously identify the cause of the fright -- another case of parallel processors in different parts of the brain, with one responding much faster than the other.

Elucidating the different mental paths followed in different kinds of decisions and pursuing the question of why the brain has developed as it has helps us understand when and why standard economic theory will not apply. But Rustichini's ambitious agenda for neuroeconomics goes further. To explain that agenda requires a brief excursion to the work of Adam Smith.

Adam Smith's research agenda

While Smith is known chiefly for The Wealth of Nations (WON, 1776), he considered an earlier, now neglected, work equally important: The Theory of Moral Sentiments (TMS, 1759). He intended the two books, together with a third on jurisprudence that he never completed, to be a "unified inquiry into the nature of human society" (Rustichini, Games and Econ. Behavior, 52, 205 [2005]). In Rustichini's summary, the question of TMS is: "How can human societies, based on people who are fundamentally selfish, function properly?" And Smith's answer is that the success of society rests on sympathy. "The desire for approval of . . . fellow human beings restricts men to moral behavior. In turn, this desire for approval is grounded on the innate affect of sympathy" (Ibid.). And sympathy arises, Smith argues through introspection, from imagining ourselves in the other's situation.

Aldo Rustichini 's research agenda

There is a sound neurological basis for Smith's introspective conclusion. Mirror neurons fire not only when one takes an action but also when one observes another taking the same action; neuroscience research suggests that we mimic in our brain the neural activity that causes our own action in order to understand the other. Imagining ourselves in the other's situation does more than make us sympathetic: it helps us anticipate the other's actions and respond to them.

Rustichini wants to extend the explorations of neuroeconomics to complete Smith's ambitious research agenda, by providing a unified theory of human behavior built on theories of economists and psychologists and tested by the tools of neurobiologists. We would perhaps then understand how we each address Rabbi Hillel's famous questions: "If I am not for myself, then who will be for me? And if I am only for myself, what am I? And if not now, when?"

Categories

Pages

Powered by Movable Type 4.31-en

About this Entry

This page contains a single entry by cla published on June 23, 2008 6:00 PM.

Paving the Way was the previous entry in this blog.

A Capital Investment is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.