Buy side analysts demonstrate less accuracy in their buy/sell/hold recommendations than sell side analysts, according to a study conducted by professors at Harvard Business School. The study tracked the recommendations and earnings forecasts from analysts at a large investment manager and compared them to those of sell side analysts. In its conclusion, the authors question why the buy side is investing heavily in hiring internal analytic staff if their performance recommendations are less accurate than the sell side. The answer may be that buy/sell/hold recommendations may not be the primary value-add from buy side analysts.
In their study, “Do Buy-Side Analysts Out-Perform the Sell-Side?”, published in March 2007, four Harvard Business professors and a professor from UNC-Chapel Hill, examine the performance of buy-side recommendations and earnings forecasts. They conclude that buy-side analysts are worse stock pickers:
“The buy-side firm analysts’ stock recommendations are less optimistic than their sell-side counterparts, consistent with buy-side analysts facing few conflicts of interest. However, the buy-side analysts make markedly more optimistic earnings forecasts than analysts on the sell-side. Findings on the buy-side analysts’ performance are remarkably consistent: …returns for their Strong Buy/Buy recommendations are lower than those of their sell-side counterparts.”
There are a few caveats to the study, however. First, the buy side data came from one large Boston-based firm (Putnam?) in which the analysts typically remain analysts and very few become portfolio managers. Second, the period studied was from 1997 to 2004, and the study results showed that the sell-side analytic advantage disappeared after Reg FD was put in place in 2000. Another explanation for the decline in relative sell-side performance would be the bear market, when the more optimistic sell-side recommendations would under-perform.
The study examines a variety of potential explanations for buy side under-performance: lower quality analysts, differences in incentives, ability to retain high quality analysts and willingness to fire low quality analysts, and whether the sell side had an information advantage prior to Reg FD. The study puts the blame on the retention of low quality analysts in the firm studied:
“We conclude that roughly one-third of the buy-side analysts’ poor earnings
forecast performance is attributable to the investment firm’s higher retention rate for low
quality analysts.”
While the study showed that the relative out-performance of the sell side declined after 2000, the authors were reluctant to attribute the effect to Reg FD without more data.
The question we have is whether stock recommendations and earnings forecasts are the right metrics. We repeatedly hear from the buy side that stock recommendations are not the most important factor in evaluating sell side research. As one example, in our review of small cap research providers, we found that the bulge firms generally had better performance for stock recommendations, yet the buy side strongly preferred regional brokers and alternative research providers for their access to small cap management, knowledge of the small cap space, and more experienced small cap analysts. While stock recommendations and earnings forecasts are convenient for measurement and analysis, they may not capture the true value-added of investment research. These are harder to quantify: investment insights, industry knowledge, and access to company management.