On the Street, it's all relative. There are times when a fund is as good as short a given name even if, in practice, it holds no short position in that name whatsoever. In those instances, the fund will want whatever shares it is "short" to underperform.
Take, for instance, Apple (AAPL) and Capital Group's American Funds AMCAP Fund. The AMCAP Fund held 360,000 Apple shares as of Dec. 31, 2012. One would be hard-pressed to say AMCAP was "short Apple." Yet, relatively speaking, AMCAP really was short Apple in the sense that it was underweight Apple. AMCAP held just 0.75% of the fund in Apple shares, whereas Apple has a S&P weighting of 3.11% as we can see from the SPDR S&P500 ETF (SPY) holdings. The weighting is even larger if we take into account such indexes as the Nasdaq 100.
So although AMCAP would not make money the way a short seller would if Apple plunged, it would still gain vs. a benchmark if Apple underperformed. Such a fund thus would have a stake in seeing a stock like Apple underperform even while holding Apple shares.
If Apple underperforms by 40% while a fund holds just 0.75% of it, instead of the 3.1% weight, the fund will gain (3.1% - 0.75%) x 40% vs. the index, or 9.4% on that position alone. This type of positioning -- in this case, aggressively underweight Apple equity -- can sometimes explain the behavior of "tactical upgrades/downgrades" by eager-to-please sell-side brokers. It is well-known that upgrades and downgrades have market impact, at least for a few days. So someone, let's say, downgrading Apple near the end of the quarter or upgrading a position where the fund company is hugely overweight can have significant impact.
Another Example: Pacific Crest's Actions Going Into Quarter-End
Yesterday, Pacific Crest pulled a Netflix (NFLX) upgrade out of its hat. Today it suddenly saw this coming weakness in Apple iPad sales next quarter. Are these coincidences? They might be. But it sure is impressive that American Funds' AMCAP fund was both severely underweight Apple and severely overweight Netflix (at 1.1% of the fund's assets -- a larger weight than Apple).
While the impact is minor on the whole, somewhere around 2.5% x 3% or just 7.5 bps (0.075%), having a broker simultaneously upgrade something you're deeply overweight in one day, and then downgrade something you're deeply underweight in the next day is still very nice.
Being underweight a stock is sometimes almost the same as being short the stock. A fund that underweights Apple will want to see Apple underperform, and will be happy if brokers turn negative on it. On their side, the brokers will see an advantage in conforming to their clients' desires, and thus might see their own opinions on what is worth how much influenced by those same desires.
Seen in this light, the Netflix upgrade as well as the Apple downgrade by Pacific Crest were surely very aligned with American Funds' AMCAP positioning.