Most buy-side analysts would reflexively say “yes.” But proving so with hard numbers is tricky for a simple reason: Buy-side firms hire analysts to support their own portfolio managers and rarely, if ever, make their stock picks accessible to the investing public.* After all, even the all-knowing Bloomberg terminal can’t provide you with historical equity recommendations from an extensive cross-section of analysts at top hedge funds (notwithstanding what’s disclosed in 13D filings, which cover large shareholders only).
That’s why large sample studies of buy-side calls have been virtually nonexistent. Until now.
As I have written elsewhere, a growing variety of investment blogs, websites, and online communities are creating new avenues for investment analysis to emerge — and for professional investors, even those at rival firms, to share research and both comment on, and rate, each others’ ideas. As one site in particular, SumZero, gains in popularity, it is also providing a living laboratory in which to study the value of buy-side investment recommendations.
SumZero is a private online community built exclusively for buy-siders to “share actionable ideas with one another and grow their professional networks.” (Hence the name, which is the opposite of “zero sum” and reflects the site’s ethos that even rival managers are better off sharing information.) Launched in 2008, the site now boasts more than 6,200 members, including analysts and portfolio managers at many leading funds — from the Baupost Group and Och-Ziff Capital Management Group to SAC Capital Advisors, Bridgewater, and Farallon Capital Management, to name just a few. SumZero members use Facebook-like features to network privately with each other. The cornerstone of the website, though, is a unique database of nearly 4,000 peer-rated investment calls by member analysts.
A new research paper has tapped into this novel assemblage of buy-side stock picks in order to answer the question posed by its working title: “Do Buy-Side Recommendations Have Investment Value?” Authored by Wesley Gray of Drexel University, Bryan Johnson of Creighton University, and Rice University’s Steven Crawford and Richard A. Price III, the paper analyzes a subset of SumZero’s ideas database, which is composed of 2,135 long and short calls on U.S. equity securities submitted to the website by 1,112 analysts from 910 different funds.
The findings: the long picks earned an average market-adjusted cumulative abnormal return of 4.03% in the 45 days after the investment thesis was posted to SumZero. Short picks demonstrated an average cumulative abnormal short-term return of nearly −1% (measured in the first and second days after issuance of the recommendation). Prices of the short calls, on average, continued to decline through day nine, the authors reported, at which point the average cumulative abnormal return was roughly −2% (see chart below).
Mean Cumulative Abnormal Returns of the Average Stock Pick Posted to SumZero*
*From 45 trading days prior to the recommendation through 45 trading days after the recommendation.
Of course, it’s reasonable to wonder whether this outperformance is merely a temporary momentum effect, or whether it holds over longer investment horizons. Alas that question is beyond the scope of the study and difficult to assess, according to the authors. For starters, the data simply aren’t there yet: SumZero has only been around since 2008, and the majority of stock picks were submitted in 2010. In addition, over longer time periods, “any extra returns could be related to a risk factor we have not controlled for,” co-author Wesley Gray said in an interview. (This critique is frequently leveled at studies involving longer time horizons, noted Gray. For more context on this issue, see “Market Efficiency, Long-Term Returns, and Behavioral Finance” by Eugene Fama.)
Despite these limitations, the authors were able to show that buy-side analysts influence asset prices by bringing new information to the marketplace.
They also uncovered some interesting nuances. With regard to long picks, for example, the authors found that small-cap value stock calls tended to experience the most positive returns immediately after being posted to SumZero. Conversely, the bigger the market-cap of the stock, the lower the abnormal return. And on the short side, the study found evidence that analysts from the smallest firms provided the most informative recommendations, judging by the price action of their picks in the first few days following the recommendation. One explanation, according to the authors, is that boutique hedge funds might employ higher-quality analysts.
So if members of SumZero are enjoying positive price effects from their investment calls, who is losing out? Put another way, who is selling at the wrong time, thus creating — on the long side, at least — the undervalued opportunities that SumZero’s buy-side analyst community is profitably exploiting?
The answer, the authors found, is big institutional investors. By gathering data from the Thomson Reuters Institutional (13F) Holdings database and examining how total institutional ownership changes before and after the release of a buy-side call on SumZero, the authors demonstrate that institutions overall “are net sellers of stocks recommended long by buy-side analysts on SumZero.com.” The authors therefore find evidence of “wealth transfers” flowing from the broader institutional marketplace to the firms represented on SumZero.
One interesting avenue for future inquiry, the authors note, would be to identify which types of institutional investors are “losing” to the “smart money” (their words) on SumZero. In fact, co-authors Gray and Crawford have already begun the research, and the initial findings, according to Gray, “are not encouraging for many ‘branded’ institutional investors.”
Investment committees and pension fund boards may need to prepare for some challenging conversations ahead.
*Conversely, there is a voluminous literature analyzing sell-side analyst recommendations. This June 2008 literature review provides a helpful overview.