New York, NY – In a recent article published in the March 2007 issue of CFA Institute Conference Proceedings Quarterly, Harvard Graduate School of Business Professor Paul M. Healy conducted multiple academic studies and came to a startling conclusion that the equity research produced by sell-side firms, including broker-dealers and investment banks, was superior to the research produced by buy-side firms. The following article reviews the conclusions of these studies and addresses our concerns about the validity of these conclusions.
As we have mentioned numerous times in the past, a large number of sell-side analysts have migrated to the buy-side for a variety of reasons including increased regulations on the sell-side and decreased analyst compensation. Simultaneously, buy-side firms have been spending more on their internal research capabilities. Given these trends, Professor Healy and his team assumed that buy-side analysts would outperform their sell-side brethren. However, the results of his study did not support this belief.
To assess these issues, Professor Healy conducted a number of studies to determine which group (sell-side versus buy-side) was more optimistic and more “accurate”. Unfortunately, Healy found that he could not readily obtain sufficient buy-side data to conduct these studies in the manner he would have liked.
Ultimately, one significant buy-side firm provided Healy and his team with data for the seven year period 1997 to 2004. According to Healy, this buy-side firm has “a good reputation, a value-oriented investing style, and a bias toward employing career analysts.”
Healy tested for the following metrics for both the buy- and sell-side analysts:
- EPS forecast optimism relative to the consensus for the same firm, time period, and forecast horizon;
- EPS forecast accuracy relative to the accuracy of the consensus forecast for the same firm, time period, and forecast horizon;
- BUY / SELL / HOLD recommendation optimism; and
- The stock performance of recommendations after controlling for overall market performance and for a variety of risk factors.
Based on these criteria, Healy and his team discovered that “buy-side earnings forecasts were far more optimistic relative to the consensus than those of the sell side.” Healy, however did not stop here, as he and his team tested to determine which group – the sell-side or buy-side had more accurate earnings estimates. Here again, Healy found that sell-side analysts had much more accurate earnings estimates than the buy-side.
BUY / SELL / HOLD RECOMMENDATIONS
Healy also evaluated the optimism of buy-side and sell-side recommendations, as well as the performance of these two groups over the sample period. Interestingly, the group discovered that the recommendations of the buy-side firm tracked were less optimistic than the recommendations of the group of sell-side analysts evaluated – a finding inconsistent with the study on earnings estimates.
Sell-side analysts made “Strong Buy” recommendations 59.1% of the time, compared to buy-side analysts who made “Strong Buy” recommendations only 44.1% of the time. In addition, buy-side analysts were willing to make “Under-Perform” or “Sell” recommendations much more often than sell-side firms (13.6% versus 5.6%).
According to Healy’s study, both sell-side and buy-side analysts beat the S&P 500 over this time period. On average, buy-side analysts outperformed the S&P 500 index by 3.5% per annum, whereas sell-side analysts outperformed the index by about 6.5% per year. Healy also noted that the superior performance of sell-side analysts was even stronger using risk-adjusted returns. Average annual risk-adjusted returns for the sell side were about 7.5% in comparison to buy-side analysts, who generated risk adjusted returns of just more than 1% per year.
SOME POSSIBLE EXPLANATIONS
Healy went on to try and explain why this buy-side firm underperformed analysts at various sell-side firms and came up with a number of reasons, including more transparency and competition on the sell-side, less rigorous performance measurement on the buy-side, more access to helpful information (management access) on the sell-side, as well as a number of other factors. Ultimately, Healy concluded that his conclusions seemed to be validated by the data – that sell-side analysts were better at generating excess returns than their buy-side brethren.
It is clear to the team at Integrity Research Associates that the study undertaken by Professor Healy and his colleagues is an extremely important one that requires additional follow-up. Unfortunately, it is also evident to us that Healy’s study was flawed in many ways making his conclusions less than convincing. A few of these problems include:
Self Fulfilling Prophecy of Sell-Side Recommendations: First, it must be noted that sell-side analysts have one very large advantage over buy-side analysts – and that is the marketing impact of selling retail investors on their recommendations. Analysts at value shops, on the other hand, typically try to find companies whose “fair value” is greater than the “market value” of these stocks, and wait until the market adjusts to this under valuation. These analysts cannot easily influence other investors to support their recommendations.
Study Period was an Anomaly: It is also not surprising to us that sell-side firms outperformed this buy-side firm over much of the time frame studied. After all, most sell-side firms were aggressively touting technology and internet stocks (even though these stocks could not be supported by earnings or cash flow) from 1998 to 2000. Investors, believing their brokers’ ratings, snapped up these shares, supporting the BUY recommendations.
Buy-side analysts for a value oriented fund, were probably less encouraged by the outlook as they saw very few companies that were good “value plays”. The period 2001 to 2004, however, was not as good for growth oriented sell-side firms as the markets plunged and then recovered. Value investors (like this buy-side firm) probably made up some ground during this period.
Limited Data Set: It is also clear that this study was based on a limited and dangerously flawed data set. Not only was this data for an extremely short period (this 7 year period covered only one real cycle), but it also included data from only one buy-side firm. In addition, this buy-side firm is an acknowledged “value investor” – a firm that is likely to under perform the general markets in bull markets (a dominant part of this time period).
Underlying Data Problems: Probably one of the most significant problems with this study is not associated with the work done by Professor Healy and his team, but rather with the reliability of the data he used to ascertain the quality of “sell-side” recommendations. A few months ago, an article published in Barron’s revealed that three professors (Alexander Ljungqvist of NYU, Christopher J. Malloy of the London Business School, and Felicia C. Marston of the University of Virginia) found almost 55,000 changes that had been made in the I/B/E/S database of stock-analyst recommendations maintained by Thomson Financial. These alternations included eliminating analyst names, eliminating Strong Buy recommendations, and adding Sell recommendations from the period 1993 to 2002. Consequently, the recommendations currently shown in the I/B/E/S database reveal annual gains that are 15% to 42% better than the originally recorded recommendations.
As a result, while we find Professor Healy’s work to be commendable, we must conclude that the “jury is still out” on who produces better equity research – the sell-side or the buy-side. However, Professor Healy does pose a very serious question for the buy-side, and that is, “Can the buy-side justify their research spend, either on their internal research capabilities or external research?” Given the countless conversations we have had with buy-side directors of research, we must admit that very few firms have built the internal systems or the processes to justify how they are spending, or how much they are spending, on investment research.