915
n
Liaisons and intrigues
Take investment banks. Investment banks often have a research
department that employs analysts who provide recommenda-
tions on whether we should buy or sell the shares of a particular
company. A recommendation to sell by such an analyst is
quite a pain (in all sorts of body parts) for a company because
their inuence can be substantial, in the sense that the rm’s
share price will likely fall as a result of this recommendation.
A recommendation to buy is obviously a much more cheerful
event.
Such investments banks, however, often also have a corporate
nance department which is trying to secure deals to advise
companies on things like debt, equity, or M&A transactions. The
potential conict of interest is clear: if a bank’s analyst gave a sell
recommendation for a (potential) client company, this is going
to be one valued client who is “not amused”!
Some companies are known to select only those investment
banks to represent them on their deals who have awarded them
positive stock recommendations in the past. In any case, if a
bank is chatting up a client on the left side but recommending
“sell their shares; now!” on the right side, this client might just
tell them to f-off.
So how do investment banks deal with such potential conicts of
interest? “Chinese walls!” they’ll so rmly declare that you will
almost believe it. “We have Chinese walls in our rms, so that the
corporate nance guys cannot pressure or even have lunch with
the security analyst dudes”. Yeah, right . . .
Professors Mathew Hayward and Warren Boeker – at the time
based at the London Business School – investigated exactly this
issue. They selected 70 companies from a variety of industries
(e.g., biotech, oil and gas, restaurants, telecom), collected a total
of 8,169 analyst ratings that had been issued for them, and
analyzed the investment banks that had been involved in their
equity, debt, and M&A deals. Then they statistically examined
whether analysts made more positive recommendations for those
rms who their banks were also serving as clients. The answer
was “heck, yes”; in 80 percent of the cases, analysts would rate