THE dealers peer at their computer screens, frantically buying and selling
shares. Everybody knows the shares are overpriced but still they buy. The hope
is that somebody else will pay even more. This is the greater-fool theory at
work: the notion that there is always a greater fool out there willing to pay
more. As time grows short, the share price holds remarkably constant. In fact,
as the underlying value drops, the trading price actually rises. In this market,
there is no shortage of fools it seems.
But suddenly the buyers dry up. And when everybody realises that nobody wants
to pay these inflated prices, panic sets in. As the final moments of trading
approach, prices plunge. The bubble has burst.
A back-room scene from the recent global stock market plunge? Not quite. This
is an experimental economics lab at the California Institute of Technology in
Pasadena鈥攐ne of a number of kinds of places where economists can study the
economic behaviour of real people. The results of these experiments plus the
arrival of some clever and much-needed mathematics could help them design
better, more realistic economic models. Economics labs could become as important
for economists as particle accelerators are for high-energy physicists. If this
approach can be made to work on a larger scale, economics is set to become a
truly experimental science rather than one marooned in the 18th century.
Advertisement
Until about 20 years ago, economics was an armchair science: theoreticians
watched how the economy behaved, then tried to model its behaviour. 鈥淔or
decades, economists worked almost exclusively on very rational models of human
behaviour just because those models lend themselves to neat mathematical
analysis,鈥� says Caltech economist Colin Camerer.
In these theoretical worlds, people facing economic choices collect all
relevant 鈥渞ational鈥� information, analyse alternatives, and develop complex,
sophisticated strategies that they can adapt to unexpected twists of fate. They
constantly analyse potential returns and calculate probabilities to wring the
maximum value from their money鈥攁n idea known as expected utility theory.
Conventional economists also assume that people who routinely use their money
irrationally, lose it and are swept off the economic Monopoly board. And they
reckon that most people act sensibly enough most of the time so that the
consequences of foolish choices don鈥檛 seriously affect the economy.
Wrong, says Richard Thaler, an economist at the University of Chicago. In an
influential paper published in 1985, he and a colleague demonstrated
mathematically that erratic financial behaviour 鈥渋s neither fatal nor
immediately self-defeating鈥�. In fact, nonrational spenders have a continual
effect that cannot be ignored. That, Thaler believes, is a fundamental reason
why standard economics fails. 鈥淢ost models predict that there will be no such
thing as unemployment,鈥� he says. So events such as the Great Depression cannot
be explained by conventional theories which predict that a share price is a fair
reflection of its value. 鈥淪tandard theory doesn鈥檛 just miss by a smidgen, it鈥檚
often embarrassingly false,鈥� adds Thaler.
Camerer and Thaler are among a growing number of behavioural economists who
want to rewrite economic theory. By translating human foibles into algebra, they
plan to craft more reliable economic models and to explain persistent
disparities between economic theory and everyday life that have dogged economics
since Adam Smith.
According to Thaler, 鈥減eople鈥檚 actual behaviour differs from rational choice
in systematic ways鈥濃�攚ays that can be measured, quantified mathematically
and used to tweak standard economic equations. 鈥淎ny insight that can be
described verbally can be put into an equation and inserted into an economic
analysis,鈥� says George Loewenstein, an economist at Carnegie Mellon
University.
One prime example of the conflict between the expectations of conventional
economics and human behaviour is the way people bet at the races. Here, the
logical strategy is to bet on horses that are most likely be among the top
three. In time, this should return a steady, if modest, profit.
But often more than half of all bets are laid on long shots, horses that
statistically have the smallest chance of winning. This is a strategy that on
average gives the smallest possible return. 鈥淎bout twice as much money is bet on
them than can be justified by their mathematical chances of winning,鈥� says
Camerer. People simply give too much credence to the improbable.
Another human trait is to place too much value on what we have now
(see 鈥淐all it quits鈥� 快猫短视频, 23 May, p 40)
and too little on what we could have tomorrow. Last year, Camerer, Loewenstein, Thaler and
colleague Linda Babcock found just this effect at work among taxi drivers in New York City. A
taxi driver鈥檚 income varies from day to day depending on the number of fares he
or she carries. In theory, any drivers wanting to maximise their earnings would
work longer on good income days and quit early on bad days.
But Camerer鈥檚 group reasoned that real people would think differently. They
predicted that the drivers would set daily targets and work until they reached
their goal. Sure enough, they found that New York鈥檚 taxi drivers consistently
quit early on high-wage days but worked longer hours on days when they were
earning less. If the goal is to maximise earnings while minimising the hours
worked, this is a poor strategy. Camerer and his team worked out that the
drivers could increase their earnings by 8 per cent simply by working the same
number of hours every day and by 15 per cent if they worked longer on good days.
But humans, it seems, set their income goals not in terms of a lifelong plan,
nor year by year, but one day at a time.
Can鈥檛 bear losing
These behavioural economists think they know what is really going on. They
point out that the taxi drivers work longer and harder to avoid missing their
daily targets than they do to gain a good day鈥檚 unexpected bounty. Most humans,
it seems, will work harder to avoid losing than they will to make a gain. This
principle is called loss aversion and it is a basic tenet of behavioural
economics. Conventional economics cannot gracefully explain the drivers鈥�
irrational choice, says Camerer. But behavioural economists can by the simple
expedient of 鈥渁ssuming people open a mental account at the beginning of the day,
close it at the end, and hate closing an account in the red鈥�.
To explain this, Camerer believes people have some kind of emotional
reference point that they have difficulty moving from. In another much-repeated
study, people are asked to imagine that they live near a factory chimney that
spews pollutants into the air. When they are asked how much their community
would pay to remove the pollutants, 鈥渢hey鈥檙e willing to pay some reasonable
amount鈥攎aybe $10 per person per year鈥�, Camerer says.
Now ask the same people to imagine that they live in an area with clean air
and that someone is going to build a factory with a chimney nearby that will
foul it. How much money would they demand as payment for making their clean air
dirty? The same people typically set a price at least ten times as high as they
would be willing to pay to clean air that鈥檚 already polluted. Some say they
couldn鈥檛 be paid enough. Clearly the emotional reference points that people set
when they think about clean air are different from those they set when thinking
about dirty air.
Conventional economics also assumes that people not only take the long view
but that they also act on it slavishly. The economist鈥檚 term for this iron
discipline is exponential discounting. 鈥淚t means that in theory, people鈥檚
patience, their degree of willingness to wait for something, doesn鈥檛 fluctuate
with time,鈥� explains David Laibson, an associate professor of political economy
at Harvard who has studied time-discounting extensively. 鈥淚n expected utility
theory, if you give a worker the choice of having a 15-minute break one day or a
30-minute break on a later day, he鈥檒l always choose the 30-minute
break鈥攚hether his choice is between today and tomorrow or between 100 days
from now and 101.鈥� But real workers think differently.
鈥淧eople show great patience over the long term but great impatience in the
short run,鈥� notes Laibson. In reality, workers tend to choose a 15-minute break
today over a 30-minute break tomorrow. But they鈥檒l also choose a 30-minute break
101 days from now instead of a 15-minute break 100 days from now.
Similarly, people have a long-term preference to be active savers, but at any
given moment the desire for a big holiday or splurging on a fancy meal prevents
them from acting on that long-term plan. 鈥淭he drive for instant gratification
overwhelms the long-run intention,鈥� he says.
Not everyone believes in the potential of behavioural economics. Sceptics
shrug off the exercise. 鈥淏ehavioural economics is an attractive area because it
looks so promising. But I don鈥檛 think it is. There鈥檚 nothing you can do with it
other than look into these occasional paradoxes,鈥� says Merton Miller, an
emeritus professor of economics at the University of Chicago Graduate School of
Business.
Save and prosper
Not so, say the behaviourists. 鈥淗ow people trade off the present against the
future is a very important consideration in economics,鈥� Camerer explains. 鈥淚t
has implications for spending versus savings, for investment planning, and a
range of similar decisions.鈥� Laibson points out that adults in the US are saving
only about a third as much as they鈥檒l need to support themselves in retirement.
鈥淗ow do we design mechanisms that most effectively help savers?鈥� he asks.
His goal is to combine the discoveries of behavioural economics with
mathematical simulations to identify what consistently makes savers act on their
long-term desire to save. The results could help beleaguered governments develop
policies that will prompt people to save more on their own.
But progress towards such lofty goals is slow for two reasons. First,
behavioural economics is a young science. Its practitioners are still shaping
the building blocks they鈥檒l need to erect more comprehensive theories. Secondly,
says Thaler, 鈥淢acroeconomics is just hard. Even current, well-developed theories
haven鈥檛 had all that much success in explaining fundamental things such as
unemployment.鈥� Behavioural economists are betting that the odds are with them.
鈥淥ur goal is that 20 or 30 years from now, the term `behavioural economics鈥�
won鈥檛 be necessary because economics will be, by definition, behavioural in its
approach,鈥� says Thaler. 鈥淓conomists should allow for the possibility that people
are human.鈥� Now there鈥檚 a thought.