When Rules of Thumb Fail

Faulty Exits Tuesday's column discussed a research paper, "Selling Fast and Buying Slow: Heuristics and Trading Performance of Institutional Investors," that shows managers to be more adept at buying U.S. stocks than at selling them. Per the authors' analysis, the managers' new purchases outperformed their existing holdings (before expenses were considered) over the ensuing 12 months.

(The paper, which is in its draft stage, is authored by Klakow Akepanidtaworn, Rick Di Mascio, Alex Imas, and Lawrence Schmidt. Its title riffs on this book, from Nobel Laureate Daniel Kahneman.)

Not so with divestitures. Sales that occurred on days without earnings announcements—those being a different story—fizzled. To phrase the matter another way, the stocks that the managers rejected would have been among their stronger performers. Portfolios constructed from their aggregate sales decisions, on days in which there were no earnings announcements, would have fared well.

Isn’t it ironic?

The Mutual Fund Evidence Morningstar's research team, it must be confessed, is skeptical. Morningstar cannot duplicate the authors' inquiries, which use daily trade information. Those inputs come from a specialized institutional database; they cannot be found in public filings. However, when Morningstar's researchers have studied mutual fund flows, they have generally found the opposite effect.

That is, funds that receive high cash flows typically disappoint. If portfolio managers are indeed adept at buying stocks, then heavy inflows should have benefited them, by spurring them to make purchases. That has not been so. Whenever Morningstar has examined the subject, most recently in Maciej Kowara's December 2018 Morningstar FundInvestor article (sorry, no link available), the company's researchers have found that U.S. equity funds that are flooded with cash subsequently perform worse, not better.

Kowara observed the reverse, as well. In his study, U.S. equity funds that underwent high redemptions enjoyed relatively good future returns. Again, one would have expected otherwise given the outcome of the Selling Fast paper. Mutual funds forced to jettison stocks, so as to meet redemptions, should have suffered relatively poor returns, because on days without earnings announcements, the Selling Fast authors found that such trades worked poorly.

Possible Explanations It's not difficult to reconcile the two sets of findings. For example, it may be that mutual fund flows are a signal, not a cause. Mutual fund investors time their transactions badly, both by buying hot funds that invest in overbought sectors, and by selling cold funds that are primed for a rebound. By this hypothesis, mutual fund managers didn't create the results; their shareholders did.

Also, the broad group of mutual fund managers is not as skillful as the group of institutional investors surveyed for the Selling Fast paper. The institutional managers who volunteered their portfolios to be analyzed by the Selling Fast authors are self-selected. Their portfolios are more concentrated than most, they trade less frequently, and they have better overall performance.

That disparity, presumably, does not affect the primary tenet, that portfolio managers approach their buy decisions differently than they do their sells, but it would make mutual fund managers likelier to lag because of cash infusions. Receiving a large amount of money presents a challenge, of putting the money to work quickly, so the cash does not sit. Mutual fund managers, overall, do not have enough excess skill with their buy decisions to make such rapid trades.

Quick Decisions? The central puzzle is why investment managers seem to be better buyers than sellers. Although cloaked in behavioral-finance language, their hypothesis is as the paper's title suggests: Managers think long and carefully about what new investments they will purchase, but not so long, nor so carefully, about what they will sell. Their sell orders, in contrast, are likely to come from "heuristics"—rules of thumb, as it were.

As guesses go, these strike me as sound. Portfolio managers are by nature studious. Most come from the right schools, with the right grades. They are trained to come to their exams well-versed. That is generally how they approach their buy decisions. They purchase stocks not on whims, but instead because they think they have learned enough about a company to have an advantage over other investors. Which, if the Selling Fast report is to be believed, holds true.

They do not seem to prepare as thoroughly for their sell decisions.

Anecdotally, I can attest to that through my past career as a mutual fund analyst, when I conducted several thousand portfolio-manager interviews. (That figure is true.) Typically, managers would provide long, thoughtful answers when asked about a new position, but would respond with a glib “I took some profits” or “I didn’t like the direction the company was heading” when asked about their sells.

Less anecdotally, the Selling Fast paper finds that when unloading a stock, investment managers are influenced by its past returns. They are significantly more likely to move issues that are either their very best or very worst performers. There is no such pattern with their buys. Apparently, managers who would never purchase a stock because of its price return are quite happy to consider technical analysis when selling.

That also matches with my personal experience.

The Next Topic While I am convinced by Selling Fast's central argument, that professional stock managers tend to sell fast and buy slow, I am wary of its final supposition. As the authors acknowledge, it is strange that the nonannouncement sales fare poorly. If managers lack skill when divesting their holdings, then one would expect a null result—as with, for example, tossing darts. But the managers perform worse than neutral. Their sell decisions actively hurt them.

This implies that the discarded stocks possess characteristics that make them strong future performers. The authors think that may occur because, in addition to watching their stocks’ returns, managers are likely to trade stocks that are closest to mind—the securities that they think about most frequently. These are likely to have been their best investment ideas. Thus, managers inadvertently damage their funds by giving up too quickly on their best-researched holdings.

Wouldn’t that be a wonderful second irony? Aesthetically, I hope that their hypothesis proves to be correct. But they have work ahead of them to demonstrate that is so.

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.