مقایسه عملکرد توصیه ای سهام بانک های سرمایه گذاری و شرکت های تحقیقاتی مستقل

From January 1996 through June 2003, the average daily abnormal return to independent research firm buy recommendations exceeds that of investment bank buy recommendations by 3.1 basis points (almost 8 percentage points annualized). Investment bank buy recommendation underperformance is more pronounced following the NASDAQ market peak (March 10, 2000) and strikingly so for buy recommendations on firms that recently conducted equity offerings. In contrast, investment bank hold and sell recommendations outperform those of independent research firms by 1.8 basis points daily (View the MathML source412 percentage points annualized). These results suggest reluctance by investment banks to downgrade stocks whose prospects dimmed during the bear market of the early 2000s, as claimed in the SEC's Global Research Analyst Settlement.

On April 28, 2003, the Securities and Exchange Commission (SEC) announced a historic agreement with ten of the largest investment banks.1 This agreement, known as the Global Research Analyst Settlement, was the culmination of extensive investigations by Congress, New York Attorney General Elliot Spitzer, the SEC, and other regulators into potential conflicts of interest among security analysts employed by investment banking firms. Alleging numerous incidents of analysts compromising the integrity of their research in order to generate investment banking business, the agreement requires the ten firms to pay $875 million in penalties and disgorgement of profits, $80 million for investor education, and $432.5 million to fund independent research. In addition to these payments, the investment banks must separate their investment banking and research departments and add a number of specific disclosures to their research reports. They must also provide independent securities research to their retail clients, in order to “…ensure that individual investors get access to objective investment advice…”. 2 Motivated by this last requirement, and the arguably implicit assumption that the recommendations of independent research firms are superior to those issued by investment banks, this study compares the overall performance of the stock recommendations of the two types of firms. We also test the implicit (if not explicit) allegation by the SEC in its individual complaints against the investment banks that analyst conflicts of interest resulted in a reluctance to downgrade buy-rated stocks during the bear market of the early 2000s. If banking analysts were issuing buy recommendations during this period when holds or sells were deserved, then we should observe their buy recommendations underperforming those of independent research firms over this time. Moreover, with potential analyst conflicts of interest arguably strongest for covered firms with recent investment banking activity, underperformance should be greatest for the buy recommendations issued on those firms. Conversely, with banking analysts reluctant to issue holds and sells during the bear market, such recommendations, when issued, are likely to reflect quite unfavorable news; consequently, they should earn more negative returns than those of analysts at independent research firms, who presumably do not have a similar reluctance to issue hold and sell recommendations. Since the SEC's allegation focuses on the bear market, we make no relative performance predictions for the bull market. Our analysis utilizes the First Call database, which contains almost 335,000 recommendations issued on more than 11,000 companies by 409 securities firms. We partition these recommendations into those issued by investment banks and those provided by independent research firms (defined here as either pure research firms or firms with research and brokerage activities, but without investment banking business). Each of these two samples is further subdivided into buy recommendations (including upgrades to buy or strong buy, and initiations, resumptions, or reiterations with a buy or strong buy rating), and hold and sell recommendations (including downgrades to hold, sell, or strong sell as well as initiations, resumptions, or reiterations with a hold, sell, or strong sell rating). We then compute daily buy-and-hold abnormal returns to the buy and hold/sell recommendation portfolios, controlling for market risk, size, book-to-market, and price momentum effects. For our entire sample period of 1996 through mid-2003, the independent research firms’ buy recommendations outperform those of the investment banks, earning a daily abnormal return that is a significant and economically large 3.1 basis points higher (almost 8 percentage points annualized) than that of the buy recommendations of the investment banks. In contrast, the hold/sell recommendations of the investment banks outperform those of the independent research firms, generating a negative average daily abnormal return that is a significant and economically large 1.8 basis points greater in magnitude (View the MathML source412 percentage points annually) than that of the hold/sell recommendations of the independent research firms. We next partition our time frame into the period prior to March 10, 2000, the date of the NASDAQ market peak (sometimes referred to hereafter as the bull market) and the period commencing on that date (sometimes referred to hereafter as the bear market).3 We find that during the bull market the average daily abnormal return to investment banks’ buy recommendations exceeds that of the independent research firms, but only by a statistically insignificant 0.4 basis points. In contrast, during the bear market investment banks’ buy recommendations underperform, on average, by a significant and quite large 6.9 basis points per day, or more than 17 percentage points annualized. More strikingly, during this latter period the investment bank buy recommendations that are outstanding subsequent to equity offerings significantly underperform the corresponding recommendations of independent research firms by 8.7 basis points, or almost 22 percentage points yearly. For analysts’ hold/sell recommendations, we find that investment bank outperformance is concentrated almost entirely in the bear market, where investment bank hold/sell recommendations outperform those of the independent research firms by a significant 3.5 basis points per day. Furthermore, those investment bank hold/sell recommendations outstanding subsequent to equity offerings significantly outperform those of the independent research firms during this period by an even greater 8.8 basis points (22 percentage points annually). Taken together, our results suggest that a significant portion of the underperformance (outperformance) of investment banks’ buy (hold/sell) recommendations is due to a reluctance on the part of banking analysts to downgrade stocks whose prospects dimmed during the bear market, as claimed in the SEC's Global Research Analyst Settlement. Since the SEC chose to sanction only ten investment banks, it is natural to ask whether there is any difference in the performance of their recommendations relative to those of the non-sanctioned banks. We address this question by partitioning the investment banks in our sample into three groups: the ten sanctioned banks, non-sanctioned investment banks that, like the sanctioned ones, were lead or joint-lead underwriters in at least one equity offering during our sample period (this group is referred to hereafter simply as lead underwriters), and the non-sanctioned investment banks that were syndicate members of at least one offering during our sample period, but were never a lead or joint-lead underwriter (referred to hereafter as syndicate members). This partition is similar to that used by Cowen et al. (2003). We find that the buy recommendations of all three investment banking categories underperform those of the independent research firms during our sample period, by a daily average that ranges from 2.2 basis points (for the syndicate members) to 3.5 basis points (for the sanctioned banks). This is true, in particular, for the subset of recommendations outstanding both during the bear market and subsequent to equity offerings, where the average daily underperformance ranges from 5.9 basis points (for syndicate members) to 9.2 basis points (for lead underwriters). This uniform underperformance suggests that differentiating between the sanctioned and non-sanctioned banks, in terms of the requirement that independent research be provided to clients, may not be justified. Comparing performance across investment banking categories for our whole sample period, we find that syndicate members’ buy recommendations earn significantly higher returns than those of the sanctioned banks, as well as higher (albeit not statistically significantly higher) returns than those of the lead underwriters. The outperformance is especially pronounced during the bear market, where the average daily abnormal return of the syndicate members’ buy recommendations exceeds those of the sanctioned banks and lead underwriters by a significant 2.7 and 2.5 basis points, respectively. The syndicate members’ outperformance relative to the other investment banks is consistent with the overall superiority of independent research firms’ buy recommendations, since a number of syndicate members are essentially independent research firms. At such firms, investment banking activity consists solely of distributing shares allocated to them by lead underwriters; they do not actively seek lead underwriter roles themselves. Perhaps the best-known example is Sanford Bernstein, which has been described as “…one of the more independent research houses – it only has a small syndicate business…” (Financial Times, September 8, 2003, p. 26). Our paper contributes to the extant literature in several ways. First, by directly comparing the recommendation performance of investment banks and independent research firms, it provides empirical support for analyst conflict-of-interest claims behind the Global Research Analyst Settlement. Second, it employs a more expansive definition of an investment bank to include all banks that participated in at least one equity offering during our sample period (rather than just including the lead banks from each covered firm's offerings), thereby recognizing that all investment banking analysts face potential conflicts of interest. Third, it documents similarities between the recommendation performance of the sanctioned banks and other investment banks, thus calling into question the appropriateness of requiring only the sanctioned banks to provide independent research to their clients. Finally, it finds sharp differences between the performance of analyst recommendations during the bull and bear markets, highlighting the importance of including the recent bear market period in any analysis of analyst recommendations. Recently, a few papers have independently examined issues similar to those we explore here.4Clarke et al. (2004) find that the long-term average abnormal returns for upgrades issued by analysts at investment banks, independent research firms, and brokerages are insignificantly different from each other; the same is true for downgrades. In drawing their conclusions, Clarke et al. do not separately analyze abnormal returns for the bull and bear market periods or for stocks with investment banking business around the time of recommendation issuance; this likely accounts for their not finding any significant recommendation return differences. Also working against their finding significant differences, Clarke et al. use a 250-day holding period for all recommended stocks, whether or not the recommendations were dropped or changed during this time. Cliff (2004) compares the average abnormal return to the recommendations of analysts at independent research firms to that of analysts at lead investment banks. Like us, he finds that independent research firms’ buy recommendations significantly outperform those of the lead investment banks, while their sell recommendations underperform. He does not, however, provide comparisons of abnormal returns for the bull market and bear market periods or for firms with and without recent securities offerings. He also does not examine the performance of the recommendations issued by analysts at non-lead investment banks. Lin et al. (2005) look at the time pattern of analysts’ recommendation revisions. They find that analysts employed by lead investment banks are significantly slower than other analysts to revise downward their buy and hold recommendations. The evidence is not as strong, though, with respect to downgrades from strong buy. While the results of our analysis suggest that some of the research issued by investment banking analysts was biased, our findings must be approached cautiously, given that they hold for a relatively narrow window, coinciding with a period of time that has been the subject of intense media and regulatory attention. Consequently, we cannot rule out the possibility that our findings are the product of a media-driven hindsight bias, rather than indicative of the existence of biased research on a significant scale. It must also be emphasized that our results pertain to our sample of securities firms and recommendations, on average. As such, they cannot be used to conclude, or bolster the contention, that research on any particular stock by any given investment bank was biased. This is an important point to recognize in the current legal environment, where a large number of claims have been brought by investors against investment banks alleging specific instances of biased research. The plan of this paper is as follows. In Section 2 we describe the recommendation sample used in the study. This is followed by a discussion of our research design in Section 3. Section 4 presents descriptive statistics for our sample, while Section 5 compares the recommendation returns of investment banks and independent research firms. Robustness tests are presented in Section 6. In Section 7 we examine whether the recommendation returns of the ten sanctioned banks differ from those of the non-sanctioned banks. A summary and conclusions appear in Section 8.

Motivated by the requirement that ten of the largest investment banks begin providing independent securities research to their clients, this study has compared the performance of recommendations issued by analysts at investment banks with those prepared by analysts at independent research firms (securities firms without investment banking business). Over the February 1996–June 2003 time period we find that the buy recommendations of independent research firms outperform those of investment banks by an average of 3.1 basis points per day. Investment bank hold/sell recommendations, in contrast, outperform those of the independent research firms by 1.8 basis points daily, on average. The outperformance of independent research firms’ buy recommendations is concen- trated in the bear market period, where they generate an average daily abnormal return that is 6.9 basis points greater than that of the investment banks’ buy recommendations. Moreover, during this period, independent research firm buy recommendations that are outstanding subsequent to equity offerings outperform those of the investment banks by a quite large 8.7 basis points. These results, taken as a whole, are consistent with allegations in the Global Research Analyst Settlement that at least some investment banking analysts were reluctant to downgrade stocks whose prospects weakened during the bear market. We go on to separately analyze the performance of the recommendations of the ten banks sanctioned in the Global Research Analyst Settlement , those of non-sanctioned lead underwriters, and those of non-lead syndicate members. Overall, we find that the buy recommendations of each investment banking category significantly underperform those of the independent research firms. This uniform underperformance calls into question the appropriateness of the SEC’s requirement that only the ten sanctioned banks provide independent research to their clients. We further find some evidence that the buy recommendations of the syndicate members outperform those of the sanctioned banks and lead underwriters. This is consistent with the notion that some syndicate members are, in essence, hybrids of investment banking and independent research firms. It is important to keep in mind that our results apply to a relatively narrow window, coinciding with a period of intense media and regulatory scrutiny into potential analyst conflicts of interest. This opens up the possibility that our findings reflect hindsight bias, rather than evidence of biased research. It is also important to recognize that our results apply to our sample taken as a whole, and do not imply that the research of all investment banking analysts, or of any particular analyst, was biased during this period.