Analyzing the Analysts
By James B. Stewart
June 5, 2001
WE NOW KNOW the truth about the performance of Wall Street analysts, and it's worse than I would have
dared to think.
Investars.com launched a new investment Web site last week with a fascinating analysis of Wall Street research
recommendations. Hypothetically buying on every upgrade and selling on each downgrade since January 1997 to the present,
Investars compiled returns for 19 firms covering the entire universe of stocks followed by research analysts, in itself a staggering
exercise in data processing.
Over the same period (through last week), the S&P 500 gained about 75%. The best of the Wall Street firms — Credit Suisse
First Boston — managed a return of just 7.63%, a breathtaking level of underperformance that should have every client who
follows such research flocking to an index fund.
How could this be? How could so much highly paid "talent" be so wrong?
Investars hypothesizes that it's because of the conflicts of interest that arise when the Wall Street firms act as investment bankers
to the companies their analysts cover. The resulting bullish bias is a problem, no doubt, but I'm not sure it explains this level of
underperformance. A bullish bias should be a boon in a period when the S&P rises 75%. I suspect the cause is the herd instinct.
This isn't the fear of being wrong — on the contrary — but the fear of being compared unfavorably to other analysts. If you
make a mistake, even a serious one, your career as an analyst will suffer only if your peers at rival firms disagreed with you and
made the right call.
That's one reason, I suspect, that the overall returns don't differ that much among firms. Only four earned positive results, and
the 10th-ranked firm, Bear Stearns, produced a return of -3.8%, a little over 10 percentage points behind Credit Suisse. The
herd instinct, unfortunately, leads analysts to issue upgrades when stocks are at market highs, and downgrade them when they
are at or near market lows. That, of course, is a recipe for a disaster of the magnitude Investars documents.
The real value of Investars' amazing database is ferreting out the one or two insightful analysts. Readers of this column may
recall that two months ago, I mentioned a case of dueling semiconductor analysts at Morgan Stanley and Salomon Smith Barney.
When that column appeared on April 17, the Salomon analyst had just upgraded the sector and Morgan Stanley's had
downgraded it. Each had a significant effect on the market the day they issued their recommendations. With this in mind, I
visited the Investars site to see what I could learn.
In research overall, Salomon Smith Barney is ranked third, slightly ahead of Morgan Stanley, which is fifth. In semiconductors,
Salomon is also on top, ranking third to Morgan Stanley's eighth. But these rankings change with individual companies. For
example Morgan Stanley ranked fourth among firms covering chip maker Xilinx (XLNX), a better track record than Salomon,
ranked ninth. Morgan Stanley downgraded Xilinx on April 16 with the stock at $38.99. Salomon Smith Barney upgraded it on
April 11 at a price of $40.55. But the top- ranked firm on Xilinx is UBS PaineWebber, which affirmed a Buy rating in December
2000 and has maintained it since. On Tuesday, nearly two months later, Xilinx was trading at $46. So far, Salomon Smith
Barney and UBS PaineWebber are thrashing Morgan Stanley.
Last week Morgan Stanley downgraded the fiber-optic sector, which also battered the prices of those stocks. With the
semiconductor exercise in mind, I checked the firm's record on Nortel Networks (NT), which Morgan Stanley downgraded on
May 30 with the price at $13.35. The result was startling: Morgan Stanley ranks 31 out of 33 firms that cover the stock, with a
total return of -75.59% since 1997. I checked Investars' "tradeplotter" chart, which showed every Morgan Stanley
recommendation on a graph of Nortel's price since 1997. In May 2000, it ranked Nortel as Outperform, a ranking it reiterated in
August with the stock near $80 a share. That turned out to be close to the all-time high. Morgan Stanley was still bullish last
October, and even as recently as March 7. It threw in the towel last week. On Tuesday, Nortel was at $13.66 — slightly above
the price where Morgan Stanley downgraded it.
The top firm on Nortel, Prudential, which has a 40.2% return on the stock, upgraded it in September 1998 and has maintained
its Accumulate ranking since. That doesn't inspire all that much confidence either, given Nortel's decline over the past year, but
still, why would anyone heed Morgan Stanley's recommendation on this stock?
This column recommended Nortel recently with the stock at $16.76, a recommendation I reiterate now that the stock has
dropped further. In other words, I'd do the opposite of Morgan Stanley, a firm I single out only because its recommendations in
two major sectors have had major market impact recently. Similar examples abound for every major firm. At the very least, I
urge investors to be skeptical of Wall Street research recommendations and the market moves they generate. If you doubt me,
check the evidence on Investars.
yahoo.smartmoney.com. |
