Do Good Investment Managers Give Away Great Ideas? It’s Entirely Rational According to New Study
In the competitive, performance-driven world of professional investment management, why would any rational manager share a potentially market-beating investment idea with rival fund managers? And are the ideas they do share any good?
Several months ago, I wrote about SumZero, a relatively new private online community where buy-side managers “share actionable ideas with one another and grow their professional networks.” Cross-sectional data on buy-side stock recommendations is relatively difficult to come by, but a trio of researchers was granted access to SumZero‘s data goldmine: a representative sample of the roughly 4,000 peer-rated investment ideas shared on the site, which they found had earned an average market-adjusted cumulative abnormal return of 4.03% in the 45 days after being posted. Short recommendations, similarly, generated immediate and significant declines in price. Based on these findings, the authors wrote, “the evidence suggests buy-side recommendations have investment value.” So it seems that rational investment managers do share good ideas.
At least one reader was nonplussed by these findings, though, and wrote the following critique in a comment:
This has been going on for decades [with] good old fashioned phone calls and ICQ messages. SumZero sounds like yet another “downstream” idea sharing platform that is mainly designed as another way for funds to “talk their book.” Of course stocks go up, particularly smaller ones, when they promote them. The real question is about stocks and ideas that go up over months and years, not 45 days. Everyone wants a short cut. Success requires doing your own work and having your own investment process. Collaboration is a great idea among buy-side firms but not thousands of them.
These observations are spot on, up to a point, if a separate and recently updated academic study is any guide. In that paper, “Do Fund Managers Identify and Share Profitable Ideas?,” a trio of b-school professors — including two authors of the SumZero study, Drexel University’s Wesley Gray and Rice University’s Steven Crawford — turn their exploration of investment idea sharing to another online community for money managers that boasts a rarer vintage: Value Investors Club (VIC), which was launched in 2000 by Joel Greenblatt and John Petry, cofounders of New York hedge fund Gotham Capital, and caps membership at just 250 analysts. To maintain their membership status, analysts must submit at least two “high quality” investment ideas each year and rate at least 20 other security recommendations posted to the site.
One fortuitous by-product of VIC’s decade-long track record is a rich data set that enables the researchers to study the performance and trading volume impact of private idea sharing over longer time periods — from one month to three years. The study also investigates three key points raised in the comment above: whether the impact of sharing is more pronounced for smaller-cap stocks versus larger ones; what drives “rational” analysts to share any stock recommendations with competitors (hint: it’s about more than just “talking your book”); and whether collaboration is more valuable among smaller groups of analysts and not, as the commenter argues, among “thousands of them.”
First, the headline results: Based on an analysis of 3,175 recommendations posted to VIC between 1 January 2000 and 31 December 2011, stock picks on the site delivered average estimated annualized alpha of between 6% and 24%, depending on the methodology applied and the time period studied. With that much alpha being generated you might expect arbitrageurs to bid up VIC stock picks — and quickly. In fact, the longer the time period studied, the lower the magnitude of the incremental return, which, the authors say, “supports the notion that the stock market incorporates new information identified by VIC members into stock prices over time.”
Still, as the commenter above suggested, returns for long recommendations posted to VIC are generally lower when the researchers restrict their sample to large firms (in fact, the returns are not significant over many time horizons). The authors also report that there is virtually no alpha associated with recommendations of stocks with greater than $2 billion in market cap over any time period, and thus conclude that “VIC members’ stock-picking skill may be limited to smaller companies in which the market is presumably less efficient.”
So why would this elite group of 250 investment analysts share any market-beating stock picks at all? The authors do find evidence supporting the “talking your book” hypothesis: Abnormal trading volume is significantly higher in the month after a long or short recommendation is posted to the site — although, as with performance, the effect diminishes over time. These results suggest “that VIC recommendation events drive a significant amount of trading in recommended securities.” And not without reason: VIC members are predominately long-biased, fundamentals-based hedge fund managers each with assets under management of between $50 million and $250 million, suggesting that collectively the community of 250 analysts runs between $12.5 billion and $62.5 billion in capital.
Even so, “talking your book” is only part of the motivation for idea sharing. In a 2008 article in American Economic Review (PDF) cited by the paper’s authors, Harvard economist Jeremy Stein outlined a theory of information exchange which posited that rational competitors would share investment ideas if doing so provided access to feedback that ultimately improves the idea. Stein also hypothesized that the most valuable ideas remain localized within a smaller group.
The researchers tested both of these ideas by analyzing commenting patterns on VIC, which permits member analysts to mark their comments as either “private,” in which case they are accessible by active VIC members only in perpetuity, or “public,” in which case the comments become available to nonmembers along with the original investment thesis after a 45-day lock-up period. The professors find that long and short recommendations receive 11.35 and 13.13 comments on average, respectively, by roughly five different VIC analysts. Many of the comments are provided by the original author of the idea, suggesting “there is a conversational, give-and-take nature” between the recommendation’s author and VIC members, a finding that supports Stein’s collaboration theory.
VIC’s differentiating factor is its strict membership cap and the elite pedigree of those investment managers who are granted access, so the significant alpha-generating potential of the ideas that are posted there is perhaps the best evidence that the most valuable ideas are — just as Stein also predicted — more likely to be shared in smaller groups. But the authors take this analysis one step further by noting that VIC analysts are more likely to mark comments on the highest-rated investment ideas as “private.” This suggests “a positive relationship between the perceived quality of an idea (as proxied by the idea’s rating) and how widely information is shared in the market (as proxied by the percentage of private comments.)
When it comes to sharing high-quality investment ideas, it seems, the size of the community may well be the enemy of the good.