for the IPOs recommended by underwriters is −18.1 percent after two years,
compared with a mean excess return of +45 percent for the IPOs recom-
mended by nonunderwriters.
These results demonstrated that underwriter recommendations, on average,
underperform those of nonunderwriters. They reveal that the best indicator
for long-term performance of an IPO is not whether the underwriter rec-
ommends, but what the more unaffiliated sources say. Stocks recommended
by nonunderwriter analysts do well in the long run, with or without the un-
derwriter analyst’s blessing, and similarly stocks not recommended by
nonunderwriter analysts do poorly.
As we discussed earlier, this bias may have its roots in an investment
bank’s agency relationship with the IPO firm from which it receives sizable
underwriting fees. Alternatively, it may be a result of some cognitive behav-
ior of analysts. That is, it is possible that underwriter analysts genuinely be-
lieve that the firms they underwrite are better than the firms underwritten
by other investment banks and history (or research) is unlikely to change
their priors. This reasoning is a direct outcome of what Kahneman and
Lovallo (1993) label “the inside view.”
MARKET EFFICIENCY AND BIASES 407
!" #$% &''#( )
)
$%
* ") + , -+ + + ,
+
. . + . Figure 11.3. Cumulative mean buy-and-hold size-adjusted return for companies
conducting IPO in 1990–1991 conditional upon source of brokerage recommenda-
tions. Cumulative return begins at the IPO price. Source: Roni, Michaely, Kent L.
Womack. 1999. Conflict of Interest and the Credibility of Underwriter Analyst Rec-
ommendations. The Review of Financial StudiesSpecial 1999, 12 (4):653–86.