A recent paper by Nathan Swen, of the McCombs School of Business, University of Texas, compares the “timing of information acquisition” by hedge funds and mutual funds, focusing especially on the release of broker-dealer analyst reports.
I won’t follow Swen’s argument, which you can study for yourself here. I will instead just use his interest as a way of catalyzing my own thoughts on an underlying subject. Specifically: do broker/dealer analysts move markets and, if so, how much? And precisely how?
Did the price of XYZ stock really fall yesterday because an analysts for, say, Morgan Stanley issued a downgrade? Consider that after the market has closed on Tuesday, the wire services ran a story headlined, “Morgan Stanley downgrades XYZ.” And the price dropped on Wednesday morning. It is common enough to hear the story then cited as the cause of the drop.
Is a Brand Name Moving the Market
But one line of thought would indicate that the words “Morgan Stanley downgrades XYZ” can’t tell Mr. Market anything he doesn’t already know. So on some abstract level those words ‘shouldn’t’ move the market. The report itself is likely a compilation of publicly available data, after all. Joe Smith has worked his way through a lot of SEC filings and research specific to XYZ’s market or markets and has added the benefit of his own biases or inferences. Everything there except Smith’s biases, and inferences that may be affected by them, is already known to the market.
If Joe Smith himself isn’t moving the market, perhaps it is his attachment to the words “Morgan Stanley” that is doing so. But is that rational?
Suppose we are confronted with solid statistical evidence that the words “Morgan Stanley downgrades” on a wire service do generally precede a fall in price, and that “Morgan Stanley upgrades” generally precedes a rise in price. It still does not follow that Joe Smith, with or without the help of the brand name, is moving the market. For what this establishes is a combination of correlation with temporal priority, which when combined are known as Granger causation. Granger causation is still far from ordinary common-sense, I-never-heard-of-David-Hume causation.
What we mean when we speak ordinary unphilosophical English and say that A caused B is that B happened because of A: B was plucked out of the realm of the merely possible and brought into the world of the actual by A. Correlation (even if temporary priority is added) surely doesn’t get us to that.
After all, since we’re assuming in our thought experiment that Morgan Stanley’s analyst worked from publicly available information, we can further hypothesize that a lot of traders and in-house buy-side managers reached the same conclusion at the same time that Joe Smith did. This in itself might be enough to explain why “Morgan Stanley downgrades” regularly precedes a stock price fall.
The Causal Nexus
Yet if it is simply a matter of everybody reaching the same conclusion at more-or-less the same time from the same data, then the stock price fall would have happened even if Joe Smith had had nothing to say, or even if he had through some analytical failure reached the contrary conclusion.
Now I should bring Swem (who has worked at Longbow Capital Partners and Goldman Sachs) back into our discussion, at least in gratitude for having kicked it off. Swem finds as a statistical matter that hedge funds anticipate B/D analyst upgrades in the coming quarter, and that they defy such upgrades in the following quarter.
That finding might still leave us in some uncertainty as to the causal nexus. On the one hand it might indicate that the hedge funds saw the same data as did Smith, but saw it first or at least acted upon it before he made his analysis public. But on the other it might also suggest that the hedge funds take Joe Smith’s more public upgrading as a signal that they should sell, because now the retail investors will be doing the buying, and it’s time to get off that train. Or on the third hand we might suppose, as Swem also says, that the B/D analysts get their information from the hedge funds, who strategically disclose it to them precisely for this reason – to sell into its publication.