The weakness of the mean reversion evidence also means that people should be careful in timing the market.
"A lot of people got out of the stock market in 1997
because of a belief in mean reversion," he said, "and they missed out on several more good years."
The Elusive Equity Premium
Barberis's analysis also takes into account the fact that the equity premium is difficult to estimate exactly. The equity premium is the average amount by which stocks beat bonds; if the equity premium is expected to be high, then
a person should invest more in stocks.
"Historically, the equity premium has been high, about 6 percent or so per year," said Barberis. "But even 60 years of data doesn't let us estimate it exactly. It might be less than 4 percent or as high as 8 percent."
The inherent uncertainty about the equity premium
is itself a form of risk, often known as estimation risk.
"Estimation risk is another reason why young people should not be too aggressive in their equity investments."
Barberis's results hold true for investors regardless of
their choices in portfolio allocation. His research tested three different portfolio strategies:
--Buy and Hold
In this case, an investor with a long horizon chooses
an allocation at the beginning of the first year and does
not touch the portfolio again.
--Myopic Rebalancing
In this case, the investor chooses some arbitrary rebalancing interval, say one year for the long-horizon investor.
He or she then chooses an allocation at the beginning of the first year, knowing that he or she will reset the port-
folio to that same allocation at the start of each year. This
is a myopic strategy because the investor does not use
any new information he or she may have learned during the course of a year.
--Optimal Rebalancing
This is the most sophisticated strategy. Assume again
that the rebalancing interval is one year. In this case, the
investor chooses his or her allocation today, knowing
that at the start of every year he or she will reoptimize the portfolio using new information.
Overall, Barberis's research favors caution. Sensible
portfolio allocation decisions require that an individual proceed very carefully. A long-horizon investor who
ignores estimation risk and the weakness of the mean
reversion evidence may overallocate to stocks by
a sizeable amount.--Soo Ji Min
Nicholas Barberis is an associate professor of finance.
"Investing for the Long Run When Returns Are Predictable" was published in the February 2000 issue of the Journal of
Finance. The paper also received the TIAA-CREF Institute's 2001 Paul A. Samuelson Award for Outstanding Scholarly Writing on Lifelong Financial Security.
