Analyst Earnings Forecast Revisions Research Articles

Insider Trading Behavior Following Analyst Earnings Forecast Revisions

Insider Trading Behavior Following Analyst Earnings Forecast Revisions

Cross‐firm return predictability and accounting quality

AbstractWe test the hypothesis that if poor accounting quality (AQ) is associated with poor investor understanding of firms’ revenue and cost structures, then poor AQ stocks likely respond more slowly than good AQ stocks to new non‐idiosyncratic information that affects both sets of firms. Consistent with this, results indicate that stock returns of good AQ firms significantly positively predict one‐month‐ahead stock returns to industry‐ and size‐matched poor AQ firms. In testing a delayed‐information‐processing mechanism behind the cross‐firm return predictability, we find that: (i) analyst earnings forecast revisions (FR) mimic the return patterns, as FR of good AQ firms significantly positively predict one‐month‐ahead FR of matched poor AQ firms; (ii) cross‐firm return predictability is concentrated in months with substantial news arrival, including months with Federal Open Market Committee (FOMC) rate announcements, but not in no‐news months; (iii) cross‐firm return predictability is stronger when the good AQ predictor firms have a richer information environment than poor AQ firms as proxied by analyst following, institutional ownership, and the presence of a Big 4 auditor. Collectively, the results uncover a new relation between accounting quality and stock return dynamics.

Detecting Earnings Information in Repurchase Announcements Using the Luck of the (Double) Irish

Previous studies disagree about whether open market repurchase announcements contain positive, negative, or no information about earnings expectations. In this comprehensive study, I use a novel measure of analyst earnings forecast revisions that corrects for undetected biases found in prior studies. While I find no evidence of average changes in earnings expectations across a 28-year sample of repurchase announcements, I find positive post-announcement forecast revisions in a subsample of firms likely to be using “Double Irish” offshore subsidiary structures to delay or avoid U.S. taxes on firm profits. Tests of this subsample utilize a quasi-natural experiment related to a 2007 IRS repurchase regulation which imposed additional repurchase-related costs on these firms. The results are robust to a variety of test specifications and proxies for earnings. This link between repurchase costs and positive earnings revisions is broadly consistent with signaling models of payouts and suggests that managers may announce especially costly repurchases when earnings expectations are underestimated by outsiders.

An Analyst by Any Other Last Name: Country Favorability and Market Reaction to Analyst Forecasts

This paper investigates whether subjective perception such as favorability influences market reactions to analyst earnings forecast revisions. Using the Gallup poll on Americans’ overall opinion about foreign countries, we find that analysts with last names associated with more favorable countries of origin elicit stronger market responses to their forecast revisions. The findings are more pronounced when firms are largely held by individual investors and when analysts have last names whose origins are easier to infer. Following the 9/11 terrorist attacks, analysts with Middle Eastern origins suffer a significant decline in market responses to their forecast revisions. Further, we find no systematic relation between forecast quality and favorability of analysts’ countries of origin, suggesting that our finding is unlikely to be driven by rational weighting based on forecast quality. We also find no evidence that analysts with more favorable countries of origin have better career outcomes. Taken together, our findings suggest that investors’ subjective feeling about an analyst can influence how they process the information in the analyst forecasts.

Accounting Quality and the Speed of Price Adjustment: New Evidence

We examine whether cross-firm return predictability is associated with accounting quality (AQ), and find that stock returns of good AQ firms significantly positively predict one-month-ahead stock returns to industry- and size- matched poor AQ firms. In testing a delayed-information-processing mechanism behind the cross-firm return predictability, we find that: (i) Analyst earnings forecast revisions (FR) mimic the return patterns, as FR of good AQ firms significantly positively predict one-month-ahead FR of matched poor AQ firms; (ii) Cross-firm return predictability is concentrated in months with substantial news arrival, but not in no-news months; (iii) Cross-firm return predictability is stronger when the good AQ predictor firms have a richer information environment than poor AQ firms as proxied by analyst following and institutional ownership. Collectively the results uncover a new relation between accounting quality and stock return dynamics.

Discrepancy between I/B/E/S Actual EPS and Analysts’ Inferred EPS

Users of I/B/E/S data generally act as if I/B/E/S reported actual earnings represent the earnings analysts were forecasting when they issued their earnings estimates. For example, when assessing analyst forecast accuracy, users of I/B/E/S data compare analysts’ forecasts of EPS with I/B/E/S reported actual EPS. I/B/E/S states that it calculates actuals using a “majority rule” indicating that its actuals often do not represent the earnings that all individual analysts were forecasting. We introduce a method for measuring analyst inferred actuals, and we assess how often I/B/E/S actuals do not represent analyst inferred actuals. We find that I/B/E/S reported Q1 actual EPS differs from analyst inferred actual Q1 EPS by at least one penny 39 percent of the time during our sample period, 36.5 percent of the time when only one analyst follows the firm (and hence this consensus forecast is based on the “majority rule”), and 50 percent of the time during the last three years of our sample period. We document two adverse consequences of this phenomenon. First, studies failing to recognize that I/B/E/S EPS actuals often differ from analyst inferred actuals are likely to obtain less accurate analyst earnings forecasts, smaller analyst earnings forecast revisions conditional on earnings surprises, greater analyst forecast dispersion, and smaller market reaction to earnings surprises than do studies adjusting for these differences. Second, studies failing to recognize that I/B/E/S EPS actuals often differ from analyst inferred actuals may make erroneous inferences.

I/B/E/S Reported Actual EPS and Analysts' Inferred Actual EPS

ABSTRACTUsers of I/B/E/S data generally act as if I/B/E/S reported actual earnings represent the earnings analysts were forecasting when they issued their earnings estimates. For example, when assessing analyst forecast accuracy, users of I/B/E/S data compare analysts' forecasts of EPS with I/B/E/S reported actual EPS. I/B/E/S states that it calculates actuals using a “majority rule,” indicating that its actuals often do not represent the earnings that all individual analysts were forecasting. We introduce a method for measuring analyst inferred actuals, and we assess how often I/B/E/S actuals do not represent analyst inferred actuals. We find that I/B/E/S reported Q1 actual EPS differs from analyst inferred actual Q1 EPS by at least one penny 39 percent of the time during our sample period, 36.5 percent of the time when only one analyst follows the firm (hence, this consensus forecast is based on the “majority rule”), and 50 percent of the time during the last three years of our sample period. We document two adverse consequences of this phenomenon. First, studies failing to recognize that I/B/E/S EPS actuals often differ from analyst inferred actuals are likely to obtain less accurate analyst earnings forecasts, smaller analyst earnings forecast revisions conditional on earnings surprises, greater analyst forecast dispersion, and smaller market reaction to earnings surprises than do studies adjusting for these differences. Second, studies failing to recognize that I/B/E/S EPS actuals often differ from analyst inferred actuals may make erroneous inferences.

LEADER OF THE PACK: STOCK COVERAGE NETWORK POSITIONS AND INFORMATION LEADERSHIP AMONG SECURITY ANALYSTS.

The article studies how the positions of stocks within the coverage network of sell-side security analysts influence their ability to exert leadership in the provision of information on the stocks they follow. The importance of information leadership to the circulation of information within financial markets and in determining market value is explored. Information leadership is defined as the ability of an analyst's earnings forecast revisions to influence other analysts to revise their own. An argument for stocks' positions in the coverage network is also presented.

Do analysts leak information to preferred customers?

Our research question is whether financial analysts leak proprietary information to their preferred customers by warning them of future earnings forecast revisions. We explore this question by monitoring investors’ trading behavior during the weeks prior to analyst earnings forecast revisions using a unique stock transactions data set from Finland. We do not find evidence of large investors systematically being warned of future earnings forecast revisions. However, our results indicate that the very largest investors show trading behavior partly consistent with being informed about future earnings forecast revisions.

Analyst Recommendations, Mutual Fund Herding, and Overreaction in Stock Prices

This paper documents the tendency of mutual fund managers to follow analyst recommendation revisions when they trade stocks, and the impact of analyst revisioninduced mutual fund herds on stock prices. We find that mutual fund herds follow consensus revisions in analyst recommendations, controlling for common investment signals that affect both analyst revisions and mutual fund trading. Consensus upgrades result in herds of funds buying a stock, while consensus downgrades result in even bigger herds of funds selling. Our most important finding is that mutual fund herding impacts stock prices to a much larger degree during our sample period (1994 to 2003) than during prior-studied periods. Further, we find the first evidence that mutual funds appear to overreact when they herd in their trades - stocks heavily bought by herds tend to underperform their size, book-tomarket, and momentum cohorts during the following year, while stocks heavily sold outperform. These reversal patterns are even stronger when herds of mutual funds (especially funds with poor performance records) follow analyst recommendation revisions. An investment strategy that accounts for the direction of both analyst revisions and mutual fund herding generates a return (adjusted for size, book-to-market, and momentum) exceeding six percent during the following year. Our results remain robust when we condition fund herding on analyst earnings forecast revisions instead of recommendation revisions. Overall, our study finds that the interaction between sell-side analysts and mutual fund managers plays an important role in setting prices in equity markets.

Retail Investor Sentiment and Return Comovements

ABSTRACTUsing a database of more than 1.85 million retail investor transactions over 1991–1996, we show that these trades are systematically correlated—that is, individuals buy (or sell) stocks in concert. Moreover, consistent with noise trader models, we find that systematic retail trading explains return comovements for stocks with high retail concentration (i.e., small‐cap, value, lower institutional ownership, and lower‐priced stocks), especially if these stocks are also costly to arbitrage. Macroeconomic news and analyst earnings forecast revisions do not explain these results. Collectively, our findings support a role for investor sentiment in the formation of returns.

Analyst earnings forecast revisions and the persistence of antidumping relief

Panel corrected standard errors with instrumental variables and effects are invoked to assess the significance of earnings forecast revisions around critical dates in non-steel AD petitions filed in 1985–1987. These petitions were filed between two important US trade law revisions (1984 and 1988), and the period encompasses significant stock market advances and declines. Event studies have been invoked to assess the value of AD petitions. However, they do not estimate the temporal distribution of any abnormal returns. Because analysts make quarterly earnings forecast revisions over several horizons, we can assess the short and long run value of petitions. We find that AD petitions tend to depress earnings forecasts in the year of the petition. However, second year earnings forecasts tend to be revised upwards. There is no effect on five year (long term) earnings growth forecasts. Hence any benefits of protection do not persist. There is evidence that analysts anticipate the filing by revising forecasts in the three months in advance of the filing. We also find that AD petitions do not affect the accuracy of forecasts.

Retail Investor Sentiment and Return Comovements

Using a database of more than 1.85 million transactions made by retail investors over a six-year period (1991-96), we show that these trades are systematically correlated − i.e., individuals buy (or sell) stocks in concert. Moreover, as predicted by noise trader models, we find that systematic retail trading explains return comovements for stocks with high retail concentration (i.e., small-cap, value, lower institutional ownership, and lower-priced stocks), especially if these stocks are also costly to arbitrage. Macro-economic news and analyst earnings forecast revisions do not explain these results. Collectively, our findings support a role for investor sentiment in returns formation.

Analyst Earnings Forecast Revisions and the Pricing of Accruals

We investigate the relation between two market anomalies to provide insights into analysts’ role as information intermediaries. Prior research finds that accruals and analyst earnings forecast revisions predict future returns. We find that the accrual and forecast revision strategies generate hedge returns of 15.5% and 5.5% when implemented independently. Strikingly, a combined strategy that uses forecast revisions to refine the accrual strategy generates a hedge return of 28.5%. Firms with consistent accrual and forecast revision signals have less persistent accruals and earnings. We also find that accruals can be used to refine the forecast revision strategy—high accruals are associated with overoptimism in analyst forecasts. Our findings indicate that although forecast revisions reflect information about accrual and earnings persistence beyond that reflected in the level of current year accruals, investors do not fully incorporate this information into their valuation assessments.

Analyst Earnings Forecast Revisions and the Pricing of Accruals

We investigate the relation between two market anomalies to provide insights into analysts' role as information intermediaries. Prior research finds that accruals and analyst earnings forecast revisions predict future returns. We find that the accrual and forecast revision strategies generate returns of 15.5% and 5.5% when implemented independently. Strikingly, a combined strategy that uses forecast revisions to refine the accrual strategy generates a return of 28.5%. Firms with consistent accrual and forecast revision signals have less persistent accruals and earnings. We also find that accruals can be used to refine the forecast revision strategy - high accruals are associated with overoptimism in analyst forecasts. Our findings indicate that although forecast revisions reflect information about accrual and earnings persistence beyond that reflected in the level of current year accruals, investors do not fully incorporate this information into their valuation assessments.

Analysts' Reactions to Warnings of Negative Earnings Surprises

Both the popular press and research literature report increasing use of short-horizon earnings guidance (referred to hereafter as warnings), primarily before negative earnings surprises.1 While it is claimed that investors and analysts have little tolerance for disappointments, feel misled by the failure to warn, lose confidence in management, and vote with their feet by selling off their shares in companies which do not warn, Kasznik and Lev [1995] reported findings suggesting the opposite. Specifically, they found that, controlling for the magnitude of the negative earnings surprise, firms that warn tend to experience more negative stock price reactions, and more negative analyst earnings forecast revisions, than those providing no warning. In this paper, we investigate analysts' reactions to qualitative warnings of adverse earnings and attempt to reconcile analysts' more negative forecast revisions, as documented by Kasznik and Lev [1995], and the apparently conflicting anecdotal evidence that suggests more positive responses

Do Managers' Accrual Decisions Speak Louder Than Words?

This paper examines whether analyst earnings forecast revisions around management earnings forecasts are conditioned on managers' previous accrual decisions. We find that analysts rely less heavily on management earnings forecast information when it conflicts with prior signals about earnings performance contained in accruals. Moreover, we find that analyst earnings forecast errors decline by a smaller amount around management earnings forecasts when management earnings forecast information and prior accrual signals are inconsistent. These findings suggest that the influence of management earnings forecasts on analysts' earnings expectations is contingent upon managers being able to convince analysts as to the informativeness of these disclosures in light of prior signals about earnings performance contained in accruals.

Evidence on the Possible Underweighting of Earnings-Related Information

This study proposes and tests hypotheses related to three issues. (1) Do analysts underestimate the persistence (permanent component) of earnings forecast errors? (2) Do investors use analysts' earnings forecast revisions to reassess the persistence of previous forecast errors? (3) Do investors systematically underestimate the persistence of previous forecast errors signaled by forecast revisions? In their investigations of post-earnings-announcement drift, Rendleman, Jones, and Latan6 [1987], Bernard and Thomas [1989; 1990], and Freeman and Tse [1989] present evidence suggesting that investors systematically underestimate the persistence of earnings forecast errors. I use analysts' expectations (their forecasts) to test the hypothesis that market participants underestimate persistence (issue (1)). Freeman and Tse also show that investors reevaluate the persistence of earnings innovations in the light of subsequent earnings announcements. I investigate whether investors similarly reassess persistence in the light of subsequent analyst earnings forecast revisions (issue (2)). Finally, if investors both underestimate the persistence of forecast errors at earnings announcements and reevaluate persistence when forecasts are revised, then they may underestimate the persistence of earnings forecast errors signaled by forecast revisions. That is, investors may

Unsystematic Security Price Movements, Management Earnings Forecasts, and Revisions in Consensus Analyst Earnings Forecasts

Empirical evidence suggests that information about a firm's future earnings prospects is important to investors. Griffin [1976], Givoly and Lakonishok [1979; 1980], and Imhoff and Lobo [1984] reported statistically significant abnormal stock returns before, at, and after announcements of analyst earnings forecast revisions. In addition, Imhoff and Lobo documented positive correlation between the magnitudes of the forecast revisions and the abnormal returns. These findings are consistent with the hypothesis that investors find analyst forecast revisions informative. Another source of information about future earnings is projections from corporate officials. Several studies have found unsystematic stock price changes prior to and/or immediately surrounding the public release of management earnings forecasts (e.g., Foster [1973], Nichols and Tsay [1979], Patell [1976], Jaggi [1978], Penman [1980], and Waymire [1984]). The conclusion from this research is that management forecasts also convey information to investors.