Downloading . . . If the document you requested does not automatically load, please click here. STOCK PRICES AND THE BARRON'S 'RESEARCH REPORTS' COLUMN Ki C. Han and David Y. Suk Abstract We examine stock price reactions to securities recommendations by investment firms. We focus on the market reaction on the date an investment firm issues a research report and the date the report is subsequently covered in the Barron's 'Research Reports' column. The results show significant stock price effects on both the issuance date and the Barron's publication date, although investors can obtain all necessary information about recommendations on the issuance date, which is the first public announcement. Thus, the results suggest the market reaction to the coverage in Barron's is separate from that to the investment firms' recommendations. The media coverage is responsible for the wider dissemination of information to the investing public, partly because the cost of acquiring information from research reports before the Barron's publication may not be trivial. The empirical results also show that the initial market reaction to stocks with positive recommendations is reversed within 5 trading days, supporting the price pressure hypothesis. INTRODUCTION Investment firms issue research reports, which include recommendations, on many firms. Investors may react to these recommendations on the date the reports are issued. However, these reports are also covered subsequently in the Barron's 'Research Reports' column. If investors react to the Barron's publication of these recommendations, it may be that the Barron's column has information content separate from the research reports issued previously. If the stock market is efficient in incorporating new information into stock prices, a secondary dissemination of the information should not affect stock prices. However, recent studies document significantly abnormal stock returns upon the publication of analysts' recommendations, which are considered secondary since they are given to the analysts' clients before publication. For example, Davies and Canes (1978) find significant abnormal returns on the publication date of analysts' recommendations in the Wall Street Journal's (WSJ) 'Heard on the Street' column. Liu, Smith, and Syed (1990) examine WSJ's 'Heard on the Street' column from 1982 to 1985 and report significant announcement-date returns for recommendations. Syed, Liu, and Smith (1989) find that where information about the forthcoming WSJ's 'Heard on the Street' column is leaked before its publication, the column still causes a significant market reaction on its publication date. More recently, Barber and Loeffler (1993) analyze the effect of second-hand information on the behavior of stock prices using WSJ's 'Dartboard' column from 1988 to 1990 and report abnormal returns for stocks recommended by investment analysts. They attribute these abnormal returns to both naive buying pressure and the information content of the analysts' recommendations. Beneish (1991), however, argues that the publication of analysts' information constitutes a primary dissemination. He states that, to establish their reputation, analysts have an incentive to reveal information through the media before revealing it to their clients. In this study we examine stock price reactions to investment firms' recommendations on the date investment firms issue research reports and on the date the reports receive subsequent coverage in the Barron's 'Research Reports' column. We examine three stock price effects associated with the recommendations. First, we look for a stock price effect caused by information leakages before the investment firms issue the reports. These leakages could occur through the activities of the investment firms and market professionals, such as arbitragers and shark watchers, who closely monitor individual stocks. Second, we look for a stock price effect between the report issuance date and the Barron's publication date. Most of the reports issued by investment firms can be purchased from Investext, a division of Thomson Financial Services. Accordingly, a market response might be expected on the report date and/or the following day. Third, we look for a stock price effect on the Barron's publication date because of media coverage. Since a gap separates the report issuance date and the Barron's publication date (15 days in our sample), Barron's should be considered a source of second-hand information. In an efficient market we expect to see relatively little stock price reaction to the Barron's coverage. We find a significant stock price effect on both the report issuance date and the Barron's publication date. The well-documented media coverage effect may explain these findings. Lee, Ready, and Seguin (1994) find a relation between media coverage on New York Stock Exchange trading halts and market activity. Dopuch, Holthausen, and Leftwich (1986) find a significantly negative stock price reaction to media disclosures of qualified audit opinions, even when the media coverage is not the first public disclosure of the event. Similarly, Stice (1991) reports a market reaction to WSJ earnings announcements, but not to the preceding filings of 10-K and 10-Q reports with the Securities and Exchange Commission (SEC). He also finds no evidence of a stock price reaction for the 8-day period between the SEC filing date and the WSJ publication date. According to these studies, a price response may be observed at the time of publication if the publication constitutes a wider dissemination of the news. Another implication from these studies is that it may be costly to acquire information from 10-K or 10-Q filings. Many investors are not sophisticated in collecting data about recommendations before the news coverage partly because information search is a costly process. As a result, information on securities recommendations by investment firms is disseminated relatively slowly, until the media provide coverage. SAMPLE The initial sample is collected from the Barron's 'Research Reports' column from January 1, 1991, through December 31, 1991. To be included in the final sample, firms that are the subject of research reports must have sufficient trading data on the Center for Research in Security Prices (CRSP) tape during the estimation period. If TABLE 1 Descriptive Statistics For 601 Recommendations In The Barron's Research Reports Column From January 1, 1991, To December 31, 1991 Note: (1) Firm size is the market value of equity measured 25 days before the report issuance date. (2) Scholes-Williams beta is measured over 150 days, beginning 175 days before and ending 26 days before the report issuance date. (3) Share price is measured 25 days before the report issuance date. (4) If the interval between the report issuance date and the Barron's publication date is less than 5 trading days, we delete the recommendation from the sample. the difference between the report issuance date and the Barron's publication date is less than 5 trading days, the sample firm is deleted. In Table 1, our final sample consists of 601 recommendations, which include 521 positive recommendations, 43 neutral recommendations, and 37 negative recommendations. Table 2 provides a breakdown of the type of recommendations. Most of research reports are positive recommendations. Only 13% of the reports contain neutral or negative recommendations. TABLE 2 Types Of Investment Firms' Securities Recommendations Published In The Barron's Research Reports Column From January 1, 1991, To December 31, 1991 A sample research report follows: Barron's, June 3, 1991: "Sizzler International (SZ - NYSE) by A.G. Edwards (16 1/8, May 13): We recommend purchase by aggressive investors looking for a company that combines growth and strong financial characteristics." The gap between the report issuance date and the Barron's publication date for Sizzler International is 14 trading days (the report was issued on May 13 and published on June 3), which is close to the mean interval (14.9 days) for the entire sample. Other descriptive statistics for the sample are provided in Table 1. Average firm size is about $3 billion, suggesting that investment firms tend to research large firms. Market beta is 1.30, which is higher than average. Mean share price is about $30, which is high since the sample firms are relatively large firms. STOCK PRICE RESPONSE We use the Scholes and Williams (1977) procedure, which corrects for the bias in the market parameters, to examine stock price reactions around the report issuance date and the Barron's publication date. Abnormal return is measured as the difference between raw return and expected return. The expected return is estimated over 150 days, from day -175 through day -26 relative to the report issuance date. TABLE 3 Daily Average Abnormal Returns (%) Around Investment Firms' Research Reports Date (RD), And The Barron's Announcement Date (BD) For 601 Securities Recommendations From January 1, 1991, To December 31, 1991 In Table 3, average abnormal returns before the report issuance date (RD) for stocks with positive recommendations are statistically significant at the 1% level for RD-3 (0.33%, t = 2.70) and significant at the 5% level for RD-2 (0.23%, t = 1.99). These results suggest that before the positive recommendations were officially issued by the investment firms, speculation and predictions by other security analysts and expert observers may have affected returns. Alternatively, insiders may have leaked information about the forthcoming recommendations (see Pound and Zeckhauser (1990)). However, stock price reactions for stocks with neutral and negative recommendations during this period are only marginally significant. As expected, strong stock price effects are observed on the report issuance date and the following day for the positive and negative recommendations. For the positive recommendations, the abnormal returns are 0.31% (t = 2.64) on RD and 0.37% (t = 3.20) on RD+1. For the negative recommendations, the abnormal returns are -0.86% (t = -2.23) on RD and -0.71% (t = -2.15) on RD+1, showing the strongest negative market reactions during the entire event period. The results also suggest an asymmetrical stock price reaction may exist between positive recommendations and negative recommendations during the event period. In fact, the market reaction to the negative recommendations is about two times stronger than the reaction to the positive recommendations. Given that most of recommendations are positive, the market may be expected to react stronger to the negative recommendations than to the positive recommendations because investment firms may be reluctant to report negative news unless they are confident about the research reports they investigated. This argument suggests the information content of negative news is more credible than that of positive news. For the neutral recommendations, we do not find any significant abnormal returns during this period. During the interim period, after the report issuance date and before the Barron's publication date, (RD+2 through BD-1), no significant stock price effect is, in general, observed. Although some significant abnormal returns are detected for the neutral group (BD-5 and BD-3), these returns are canceled out. On the Barron's publication date, the highest positive abnormal return is observed for the positive recommendations (0.54% with t = 4.25). The second-strongest daily average abnormal return is observed for the negative group (-0.71% with t = -2.32). The results indicate the publication of the 'Research Reports' column significantly affects stock prices. No abnormal returns are observed for the neutral group. For the positive recommendations, note that the market reaction on the Barron's publication date is stronger than that on the report issuance date. We compare the results in Table 3 with those of previous studies that examine stock price reactions to the WSJ's recommendations. For example, Liu, Smith, and Syed (1990) examine the WSJ's 'Heard on the Street' column and report announcement-date returns of 1.54% for buy recommendations and -1.99% for sell recommendations. The abnormal returns for our sample are smaller than those reported in previous studies because the investment firms have already made the research reports available to their clients before the publication of the Barron's column. Our empirical findings that the market reacts to the Barron's column even though the information is second-hand raise the issue of market efficiency. According to the semistrong form of the efficient market hypothesis, all publicly known information is quickly impounded in security prices so that no investor can use it to earn excess returns. The magnitudes of abnormal returns from the report issuance date to the Barron's publication date are approximately 1% for positive recommendations and 2% for negative recommendations. These abnormal returns are not significantly large enough for investors to earn returns after transactions costs. These results are consistent with Davies and Canes's (1978) study on the effect of the 'Heard on the Street' column on stock prices.1 The initial price response to the positive recommendations on the Barron's publication date is almost completely reversed within 5 trading days. From day 2 through day 4 following the publication date, the stocks with positive recommendations experience a significantly negative abnormal return of -0.65 percent. Therefore, the results suggest the stock price reaction to the recommendation announcement is due to naive buying pressure.2 However, we do not detect such a reversal for the negative recommendations. Our results support the findings by Barber and Loeffler (1993), who investigate WSJ's 'Dartboard' column and find that the positive abnormal returns realized upon the announcement date are partially reversed within 25 trading days. The results in Table 3 tend to support the media coverage hypothesis documented by Stice (1991) in the accounting literature and by Lee, Ready, and Seguin (1994) in the finance literature. The findings also support the price pressure hypothesis documented in Davies and Canes (1978), Liu, Smith, and Syed (1990), and Barber and Loeffler (1993). CONCLUSIONS We examine the stock price reactions to investment firms' research reports on their issuance date and on the subsequent Barron's publication date. The results show significant stock price effects at both the report issuance date and the Barron's publication date, although investors can obtain all necessary information on the report issuance date, which is the first public announcement. Thus, the results suggest that the Barron's publication causes a price reaction separate from the reaction to the report issuance. Our findings support the media coverage hypothesis recently documented in the accounting and finance literature. It suggests that information about securities recommendations by investment firms is not fully reflected in prices until the subsequent Barron's coverage. The empirical results also show that the initial price response for the stocks of positive recommendations is reversed within 5 trading days, supporting the price pressure hypothesis. ENDNOTES 1. We thank an anonymous referee for suggesting this line of inquiry. 2. We again thank an anonymous referee for making this observation. REFERENCES [1] Barber B. and D. Loeffler, "The Dartboard Column: Second-Hand Information and Price Pressure," Journal of Financial and Quantitative Analysis 28, June 1993, pp. 273-284. [2] Beneish, M., "Stock Prices and the Dissemination of Analysts' Recommendations," Journal of Business 64, 1991, pp. 393-416. [3] Davies, P.L. and J. Canes, "Stock Prices and the Publication of Second-hand Information," The Journal of Business 51, 1978, pp. 43-56. [4] Dopuch, N., R. Holthausen, and R. Leftwich, "Abnormal Stock Returns Associated with Disclosures of 'Subject To' Qualified Audit Opinions," Journal of Accounting and Economics 8, 1986, pp. 93-117. [5] Lee, C, M. Ready, and P. Seguin, "Volume, Volatility, and New York Stock Exchange Trading Halts," Journal of Finance 49, March 1994, pp. 183-214. [6] Liu, P., S. Smith, and A. Syed, "Stock Price Reactions to the Wall Street Journal's Securities Recommendations," Journal of Financial and Quantitative Analysis 25, September 1990, pp. 399-410. [7] Pound, J. and R. Zeckhauser, "Clearly Heard on the Street: The Effect of Takeover Rumors on Stock Prices," The Journal of Business 63, 1990, pp. 292-308. [8] Scholes, M. and J. Williams, "Estimating Betas from Nonsynchronous Data," Journal of Financial Economics 5, December 1977, pp. 309-327. [9] Stice, E., "The Market Reaction to 10-K and 10-Q Filings and to Subsequent The Wall Street Journal Earnings Announcements," The Accounting Review 66, January 1991, pp. 42-55. [10] Syed, A.A., P. Liu, and S.D. Smith, "The Exploitation of Inside Information at the Wall Street Journal: A Test of Strong Form Efficiency," The Financial Review 24, 1989, pp. 567-579.