Centre for Economic Policy Research

2.5
Evidence of increasing conflicts in the 1990s
While individual cases of conflict of interest have recently figured prominently,
some empirical studies find support for their presence, even years before the stock
market boom. To identify how a potential conflict of interest was exploited it is
necessary to examine how the actual information was transmitted by analysts and
used by investors. 

---

22 Conflicts of Interest in the Financial Services Industry
The idea that investors could be ‘fooled’ seemed unlikely to many researchers
in academic finance, given the strong prior that analysts’ reports and recommen-
dations contain little new information because the market are governed by 
rational expectations. Beginning with Cowles (1933), it was long believed that 
recommendations of equity analysts did not influence the market, producing
abnormal returns. In a seminal article, however, Grossman and Stiglitz (1980)
argued against a naïve informational efficiency in markets. They pointed out that
market prices would not reflect all available information, otherwise there would
be no return on the millions of dollars spent every year by investment banks on
research. Accumulated recent evidence based on very detailed data bears out this
insight and reveals that analysts’ reports do move the markets. One prominent
example (Womack, 1996) examined 150,000 analysts’ comments for the period
1989-91 and found that buy recommendations produced a 3% price increase and
sell recommendations a 4.7% drop in a three-day event window. In addition to
this notable asymmetric response, which indicated more news came from the less
frequent sell recommendations, there was also considerable drift in prices in 
subsequent months, suggesting that full adjustment was not immediate. 
Furthermore, given that they can influence the market, what is even more 
striking is analysts’ tendency to be overoptimistic. Studying IPOs during an 
earlier, relatively quiet period, 1975-87, Rajan and Servaes (1997) found a 
strongly optimistic bias in analysts’ behaviour. The more underpriced an IPO, the
larger following of analysts it attracted. Analysts then systematically over-
estimated the earnings of these companies, with their longer-term forecasts being
more (excessively) optimistic. Rajan and Servaes also found that more firms 
complete IPOs when analysts are especially optimistic about growth prospects,
consistent with their finding that more firms conduct IPOs when seasoned firms
in their industries are trading at historically high multiples.
If analysts do influence the market and their information is biased, is the new
information they provide aimed at exploiting a conflict of interest? The answer to
this question should be found in the differential behaviour of analysts at under-
writing and non-underwriting banks. If conflicts of interest are minor and the
informational advantages gained by combining underwriting and brokerage are
dominant, then there are implications for the reception of analysts’ information.
Owing to their key position, lead underwriters’ analysts’ reports should carry extra
weight and their predictions should be unbiased and more accurate than those of
other equity analysts. Consequently, the market should react more to their
announcements than to reports of other analysts. Their recommendations should
have more predictive power of future prices and give investors better investment
results. If conflicts of interest dominate informational advantages, however, lead
underwriter analysts will issue recommendations that are biased toward being
overly optimistic. Underwriter analysts will also issue relatively more positive 
recommendations for firms that trade poorly in the IPO aftermarket. In a 
rational market, participants should then discount underwriter analysts’ 
recommendations relative to non-underwriter analysts.
13
Examining data for seasoned equities between 1983 and 1988, Dugar and
Nathan (1995) find that while underwriters’ analysts are optimistically biased,
their earnings forecasts are just as accurate on average as those of non-underwriter
analysts. They uncover some limited evidence that investors rely relatively less on
underwriters’ analysts since market reaction around the report dates of 
non-underwriter analysts was greater than the reaction for underwriter analysts,
although the difference was not statistically significant. This finding offers some
support for the contention that investors are not ‘fooled’ by the optimism of
underwriter analysts.

---