agents to behave more rationally.^2 In many important economic choices
(e.g., career choice, marriage, saving for retirement), if one agent makes a
poor choice (picks the wrong career or spouse, saves too little) no profit
opportunity is created. You may think that my wife will soon realize what
a mistake she has made in marrying me, but (as far as I know) there is no
way for you to sell my marriage prospects short, and even if you could, it
might not alter the behavior of me or my unlucky wife. Missing markets
prevent arbitrage, thus allowing irrationality to persist.
For years, many financial economists believed that in the financial world,
the existence of well-functioning markets (including opportunities to sell
short) implies that irrational agents will not affect asset prices. If I (and other
confused investors) buy the wrong stock and drive up the price too high,
smart investors like you can sell that stock short (unlike my marriage). In a se-
ries of papers Brad DeLong et al. started to undercut this notion (see their
chapter in Volume I), but I think scope of “limits to arbitrage” was much bet-
ter understood after the paper by Andrei Shleifer and Robert Vishny (Chapter
2) appeared. Shleifer and Vishny show that arbitrageurs need long horizons to
be able to bet successfully on slow-moving market mispricing, and since real-
world arbitrageurs need to bet with other people’s money in order to have suf-
ficient capital to affect prices, they need their investors to have long horizons.
Shleifer and Vishny show that in the annoying circumstances in which prices
temporarily move even further away from rationality, arbitrageurs lose money
and, as a result, their investors may withdraw funds. The chapter contains
what now appears to be an uncanny prediction of events that soon followed,
especially the downfall of Long Term Capital Management (LTCM).
Chapters 3 and 4 document two particularly dramatic illustrations of
limits to arbitrage, namely violation of the law of one price. Kenneth Froot
and Emil Dabora discuss so-called twin stocks, such as Royal Dutch and
Shell, whose prices should be linked by a simple formula (since by charter
the earnings are divided according to a 60:40 ratio). Nevertheless, prices
have diverged from this true fundamental value by as much as 35 percent.
(Ironically, one of the trades that LTCM had on when they collapsed was a
bet on the Royal Dutch Shell spread to converge.) Owen Lamont and I dis-
cuss the equally bizarre case of Palm and 3Com in Chapter 4. In the midst
of the Internet bubble, 3Com announced a spin-off of their Palm division,
maker of spiffy handheld computers. 3Com held a wildly successful IPO
for Palm, in which a small portion of the shares were sold, the rest were to
be distributed in a few months to 3Com shareholders. The weird part was
that the market value of 3Com was, for several months, less than the value
of the Palm shares they owned. Several other similar negative valuations
were observed around the same time.
xii THALER
(^2) See for example, Arrow (1982), Akerlof and Yellen (1985), Russell and Thaler (1985),
and Haltiwanger and Waldman (1985).