equity. This thinking leads to a cross-sectional prediction, namely that the
investment of equity-dependent firms should be more sensitive to gyrations
in stock price than the investment of non-equity dependent firms.
Other than this equity-dependence mechanism, there are other channels
through which investor sentiment might distort investment. Consider the
case where investors are excessively optimistic about a firm’s prospects.
Even if a manager is in principle interested in maximizing true value, he
faces the danger that if he refuses to undertake projects investors perceive
as profitable, they may depress stock prices, exposing him to the risk of a
takeover, or more simply, try to have him fired.^36
Even if the manager is rational, this does not mean he will choose to
maximize the firm’s true value. The agency literature has argued that some
managers may maximize other objectives—the size of their firm, say—as a
way of enhancing their prestige. This suggests another channel for invest-
ment distortion: managers might use investor exuberance as a cover for
doing negative NPV “empire building” projects.
Finally, investor sentiment can also affect investment if managers put
some weight on investors’ opinions, perhaps because they think investors
know something they do not. Managers may then mistake excessive opti-
mism for well-founded optimism and get drawn into making negative NPV
investments.
An important goal of empirical research, then, is to try to understand
whether sentiment does affect investment, and if so, through which chan-
nel. Early studies produced little evidence of investment distortion. In ag-
gregate data, Blanchard, Rhee, and Summers (1993) find that movements
in price apparently unrelated to movements in fundamentals have only
weak forecasting power for future investment: the effects are marginally
statistically significant and weak in economic terms. To pick out two partic-
ular historical episodes: the rise in stock prices through the 1920s did not
lead to a commensurate rise in investment, nor did the crash of 1987 slow
investment down appreciably. Morck, Shleifer, and Vishny (1990) reach
similar conclusions using firm level data, as do Baker and Wurgler (2002a):
in their work on capital structure, they show that not only do firms with
higher B/M ratios in their past have more equity in their capital structure
today, but also that the equity funds raised are typically used to increase
cash balances and notto finance new investment.
More recently though, Polk and Sapienza (2001) report stronger evi-
dence of investment distortion. They identify overvalued firms as firms with
58 BARBERIS AND THALER
(^36) Shleifer and Vishny (2003) argue that in a situation such as this, where the manager feels
forced to undertake some kind of investment, the best investment of all may be an acquisition
of a less overvalued firm, in other words, one more likely to retain its value in the long run.
This observation leads to a parsimonious theory of takeover waves, which predicts, among
other things, an increase in stock-financed acquisitions at times of high dispersion in valua-
tions.