|
 |
| Research Paper 16 | |
 | | Prospect Theory and Asset Prices
Winner
of the 2000 FAME Research Prize
Authors
Nicholas
BARBERIS - University of Chicago
Ming HUANG - Stanford University
Tano
SANTOS - University of Chicago
Date
September
2000
This paper has now been published and is no longer available as
a part of our Research Paper Series. The reference to this paper is:
Barberis,
N., Huang, M., Santos, T. (2001): "Prospect Theory and Asset Prices". The Quarterly Journal
of Economics, Vol. CXVI, Issue 1.
Abstract
We
study asset prices in an economy where investors derive direct utility not only from consumption but
also from fluctuations in the value of their financial wealth. They are loss averse over these fluctuations
and the degree of loss aversion depends on their prior investment performance. We find that our framework
can help explain the high mean, excess volatility and predictability of stock returns, as well as their
low correlation with consumption growth. The design of our model is influenced by prospect theory and
by experimental evidence on how prior outcomes affect risky choice.
Executive
Summary
One of the most important topics studied by financial economists
is the behavior of aggregate stock markets. Over the past two decades, a number of puzzling features
of stock market movements have been identified. This paper is an attempt to resolve these puzzles.
There
are three main puzzles associated with aggregate stock market behavior: (i) the equity premium puzzle;
(ii) the volatility puzzle; and (iii) the predictability puzzle. We describe each one in turn.
The
equity premium is the average return on the overall stock market minus the return on
riskless
government bonds. The puzzle is that in most countries, the historical equity premium has been much
higher than our economic models would predict.
The volatility puzzle
is that stock market levels appear to move around too much. For example, ratios of price to earnings
in the U.S. stock market have often been very high.
The standard rationalization
of this is that investors must be expecting high cashflow and earnings in the future, and are
therefore happy to pay high prices today. However, historical data shows that high levels of price-earnings
ratios are not, on average, followed by higher earnings. In this sense, it is a puzzle why prices were
so high to begin with. This is the volatility puzzle.
Historical data
also shows that the price-earnings ratio can predict future returns on the stock market. High levels
of the price-earnings ratio have generally led to lower subsequent returns, and low levels of the ratio
to higher returns. This evidence is known as the predictability puzzle.
Our
paper tries to make sense of these findings. Most traditional models assume that investors only receive
utility from consumption. We depart from this framework by arguing that investors also receive direct
utility from another source, namely changes in the value of their financial wealth. This second type
of utility need have nothing to do with consumption. For example, if you suffer a big loss in the stock
market, you may experience a sense of regret at the decision to invest in the first place; you may interpret
the loss as a sign that you are a second-rate investor, dealing your ego a painful blow; and you may
feel humiliation in front of friends and family when words leaks out. Whatever the reason, our model
assumes that when thinking about how much to invest in the stock market, people take this additional
source of utility into account.
This second source of utility has two
features which are motivated by research in the psychology literature. First, investors are loss averse,
which means that they are more sensitive to losses in financial wealth than to gains. Second, how loss
averse they are may change over time depending on their previous investment results. If they have recently
made a lot of money in the stock market, they may be less nervous, or less loss averse, because any
loss they incur will be cushioned by their prior gains. However, if they have recently been burnt by
painful losses in the stock market, they may be more nervous about any additional setbacks, in other
words, more loss averse.
In this paper, we show that this framework may
be helpful in resolving the three puzzles outlined earlier. First, we find that our model predicts large
equity premia, in line with those observed in the data. The reason is that the investors in our model
are loss averse: they are much more sensitive to losses than to gains, and therefore they are uncomfortable
with the frequent fluctuations of the stock market and demand a large average premium to compensate
them for this risk.
To understand how we resolve the volatility puzzle,
suppose that the stock market receives some good news about earnings. This will push the stock market
up, generating substantial gains for investors. Now that they have gains, investors will be less loss
averse, because these gains will cushion any subsequent losses. Since they are less risk averse than
before, they are prepared to pay even more for stocks, and push stock market prices even higher. Therefore,
a changing degree of loss aversion may explain why prices appear to move more than is justified by news
about earnings.
The resolution of the predictability puzzle is similar. After a good
piece of earnings news, the stock market goes up, generating gains for investors, who become less loss
averse and push the stock market even further up. Since their prior gains make them feel more comfortable,
investors demand a lower average return as compensation for staying in the stock market. Therefore,
high prices are on average followed by lower returns, in line with the findings of predictability in
the data.
Remarkably, very few models have been proposed that can address
all three of these puzzles. Other promising models rely on other behavioral ideas such as overreaction
to explain phenomena like the volatility puzzle. More testing of these models in the next few years
will give us more insight into which factors are truly important for understanding aggregate stock market
behavior.
| |
|
|
|
