|
|
More Factoids October 04 Oct F-Too September 04 Sept F-Too August 04 July 04 July F-Too More Factoids Q4-03 Index Q3-03 Index Q2-03 Index Q1-03 Index Q4-02 Index Q3-02 Index Q2-02 Index Q1-02 Index Q4-01 Index Q3-01 Index Q2-01 Index Q1-01 Index Q4-00 Index Q3-00 Index Q1-00 Index Links by Topic Biz Links Business News Economic Reports Stock Exchanges Searches |
After years of such miracles, investors finally seem to be wising up to the fact that an extra penny of profit is not only meaningless but may also be evidence of earnings management and, therefore, bad news. After all, the practice can hide what's genuinely going on in a company's books. A study by Thomson Financial examined how many of the 30 companies in the Dow Jones industrial average missed, met or beat analysts' consensus earnings estimates during each quarter over the last five years. It also looked at how the companies' shares responded to the results. Over the period, on average, 46.1% - met consensus estimates or beat them by a penny. Pulling off such a feat in an uncertain world smacks of earnings management. "It is not possible for this percentage of reporting companies to hit the bull's-eye," said Bill Fleckenstein, principal at Fleckenstein Capital. "Business is too complicated; there are too many moving parts." The precision has a purpose, of course: to keep stock prices aloft. According to Thomson's five-year analysis, companies whose results came in below analysts' estimates lost 1.08% of their value, on average, the day of the announcement. The loss averaged 1.59% over five days. Executives have lots of levers to pull to make their numbers. Lowering the company's tax rate is a favorite, as is recognizing revenues before they actually come in or monkeying with reserves set aside to cover future liabilities. If all else fails and a company faces the nightmare of an earnings miss, its spinmeisters can always begin a whispering campaign to persuade Wall Street analysts to trim their estimates, making them more attainable. Their stock might drift downward as a result, but the damage is not usually as horrific as it is when earnings miss the target unexpectedly. So it is not surprising that the strategy has become so widespread and that fewer companies in the Thomson study are coming in below their target these days. For the first three quarters of 2004, 10.9% missed their expected results, down from 11.7% in 2003 and 25% in 2002. At the heart of earnings management is - what else? - executive compensation. The greater the percentage of pay an executive receives in stock, the bigger the incentive to produce results that propel share prices. A study published last year by Qiang Cheng, assistant professor of accounting at the University of British Columbia's Sauder School of Business, and Terry Warfield, assistant professor of accounting at the University of Wisconsin in Madison, found that managers with high equity incentives are more likely to report earnings that meet or just beat analysts' forecasts than are managers who have low equity incentives. The study also found that managers with high equity incentives are less likely to report large positive earnings surprises, perhaps choosing instead to set aside the extra money for a rainy day. This practice is known as earnings smoothing. Unfortunately for executives on the beat-the-number treadmill, an extra penny isn't what it used to be. In 1998, Dow components that beat their numbers by a cent saw their stocks rise 0.78% the day of the announcement, according to Thomson. This year, the increase has averaged 0.15%. Returns over five days were flat in 1998, while stocks of companies beating estimates by a cent have lost an average 1.42% this year. This may be why the percentage of companies meeting or beating the estimate by a penny has also declined recently. This year, 35% of companies in the Dow average are in this category, down from 60% in 1998. Genuine gains now go only to companies that exceed estimates by more than a cent. This year, shares of these companies have risen 0.64% on average on the day of the announcement. Over five days, there was an average 1.91% gain. Meanwhile, the number of Dow component companies reporting results that are more than a penny above expectations has climbed significantly in the last few years, rising to 54.3% today from 27.2% in 2002. This means one of two things: executives have figured out that a penny doesn't jump-start a stock anymore and are speeding up the treadmill, or they are playing the earnings smoothing game less and less. Let's hope it's the latter. Companies should refuse to participate in the earnings management charade and investors should stop demanding it. "The marketplace wants quarterly measurement, but it has to be put into perspective," said Michael Thompson, director of research at Thomson. "What's more important is long-term viability of a company, and investors should be more respectful of that." Yes, investors who reward companies for the myth of making their numbers are a big part of the problem. But so are those in the media who report breathlessly when companies beat estimates by a penny or two. One company that successfully shuns the earnings game is the Progressive Corporation. Rather than whisper to analysts what they can expect from upcoming results, Progressive publishes monthly, in-depth reports on its Web site. The reports, which it began in May 2001, include an income statement, a balance sheet and segment information. Also included are changes in loss reserve estimates, a number that is closely watched by investors and which contributes to wide swings in earnings. "We never make promises to hit an earnings number," said Tom King, Progressive's treasurer. "We don't have predictable earnings, but it doesn't matter. We have been very clear on our financial reporting policies, and as a consequence we have not attracted those investors for whom meeting earnings numbers is important." How have Progressive's shareholders fared? Interestingly, the stock has been less volatile than the Standard & Poor's 500-stock index since the monthly updates began. It's not clear who got the earnings game going: executives or investors. But it's past time for it to stop. As the Progressive example shows, those companies that continue the charade do it by choice.
Out late last week, Alpha magazine's second-annual ranking of hedge funds' favorite research analysts provides a peek at how the rules are changing for Wall Street stock scribes. Stock-research analysts no longer are rewarded for touting a firm's investment-banking deals. That longtime practice was prohibited last year after 10 securities firms agreed to pay $1.4 billion to settle regulators' accusations that their analysts wrote overly rosy stock reports to help win underwriting assignments. More firms followed this year with similar settlements. Today, the best hope of clinching a seven-figure paycheck, recruiters say, is to cater to the most-lucrative trading clients. Increasingly, those clients are hedge funds, relatively lightly regulated investment vehicles for wealthy individuals and institutions that often trade rapidly in and out of stocks and, in some cases, sell stocks "short," essentially betting on their decline. Michael Flood, a Wall Street headhunter in New York with Westwood Partners, says analysts who can bring in significant trading revenue from hedge funds and other powerful investors are scarce, but can get $1 million to $4 million pay packages, if the hedge fund doesn't hire them first. "For the most part, they view analysts as a waste of time," he says. Wall Street research as a whole tends to be perennially optimistic. While there are more "sell" ratings nowadays, research-tracking firm Investars.com says the number of "sells" still totals just 13% of the 10 biggest firms' recommendations. Analysts with a skeptical streak stand out. Terry Darling, who covers oil-services and equipment for Goldman Sachs Group Inc., has come in first two years in a row among hedge funds for his ability to identify not just the best companies in his industry but the four or five worst, those surveyed said. "What a hedge fund looks for is much more of a versatile individual who thinks like a trader instead of just writing reports," says Michael Karp, co-founder of Options Group, an executive-search firm. Hedge funds generate more than $10 billion in revenue to Wall Street securities firms annually, or about 7% of these firms' total revenue, according to estimates from investment-advisory firm Hennessee Group LLC and Richard Strauss, a brokerage-industry analyst at Deutsche Bank AG. Most of it is tied to stock trading. While mutual funds and pension funds collectively hold more assets, hedge funds tend to trade more actively. Alpha magazine was launched last year by Institutional Investor magazine to cover the hedge-fund world. II, as the parent publication is called, already publishes a widely watched, sometimes controversial, analyst ranking that surveys a broad array of professional money managers, from mutual funds to pension funds. In past years, regulators, including New York Attorney General Eliot Spitzer, have criticized the II poll as a "popularity contest." Critics have accused Wall Street analysts of trumpeting stocks that big clients hold, or even tipping off favored investors when their ratings were about to change, as part of their efforts to win votes. II declined to comment. II uses those poll results to produce the Alpha rankings, retabulating just the hedge-fund votes to show how the fast-money crowd alone views big-firm research departments. Lehman Brothers Holdings Inc., which has snagged the top spot in the overall II poll two years in a row, also came in first this year with hedge funds. But other firms and analysts show striking differences. Goldman Sachs, which placed eighth in the overall poll, moved up to third place with hedge funds. Goldman is a leading provider of stock services known as prime brokerage to hedge funds, and has a specialist sales force that helps analysts set up meetings for hedge-fund traders with companies they cover. David Tenney, co-chief operating officer of global research for Goldman, says he encourages Mr. Darling and all the analysts to suggest trades. "When analysts have an idea for a stock that they believe will rise, we also encourage them to think about stocks that they believe might decline in value." Morgan Stanley moved up to second place from fourth. A Morgan Stanley spokeswoman says the firm wouldn't speculate about why it did better with hedge funds. The firm also is a leader in prime brokerage to hedge funds. In July 2003, Morgan Stanley announced it would no longer use analysts' II rankings to gauge performance. It also stopped providing II with photographs of its analysts. Two independent-research boutiques, Buckingham Research Group and Fulcrum Global Partners, have analysts that made the list of hedge-fund favorites, but didn't make the overall rankings. Manny Korman, director of research at Buckingham, says his 17 analysts don't focus on short-term trading ideas, but instead gear research for six-to-12-month performance. He says hedge funds appreciate that, too, as long as the analysts "do a good job with providing them investor intelligence." This page is NOT a final product, but in the first stages of being formed. So far I have taken research previously posted on three factoids pages and posted them here. The planned format: a list of the top 20 things investors should know about analysts' forecasts. Examples: (1) Analysts' opinions are relevant and are reflected in stock prices. (2) Consensus analysts' recommendations are too optimistic; (3) Being optimistic has historically helped analysts obtain inside information from the firms they cover; (4) Analysts underract to news of earnings surprises; (5) Analsyts' forecast errors are positively related to the prior earnings change. Think of this as a kind of analyst anti-inertia - for stocks with large earnings gains, it is often assumed that the large gains will not continue; (6) Analysts underreact to negative information but overreact to positive information; (7) But analysts' underreaction to past earnings news decreases with their experience; (8) Analysts tend to recommend glamour stocks, those with high P/E ratios, high past sales growth, and strong price momentum - even though value stocks tend to outperform growth stocks. I want a second list of the top 20 things investors should know about the markets reaction to analysts' findings - and changes to their recommendations and earnings forecasts. Examples: (1) There is a significant price impact associated with new buy and sell recommendations. Womack found that the initial reaction (three days) is a 3% rise for buy recommendations and a 4.7% drop for sells. He also found incremental positive returns of 2.4% for the month following buys and a -9.1% drift over the six months following a sell. (2) But other studies have found that recommendation changes have only a temporary effect on a stock's price and quickly reverse; (3) Still, negative changes are different in most studies - and the effect can linger for months. There is even a theory that negative changes should have more informational value to the market than positive changes. "Sells" are less frequent and more visible. And negative changes also come at a 'cost' to the analyst or his/her firm [in potential investment business and access to information]. If the costs of issuing a sell recommendation are greater, then the analyst's expected return for issuing these should also be greater. (4) Negative momentum arises because bad news disseminates slowly. Negative post-recommendation stock price drift is also consistent with investor "loss aversion". (5) Stocks favored by analysts out-perform stocks disfavored by analysts over most of the tested time periods. For the period of 1986-1996, a portfolio comprised of the most highly recommended stocks generated an average annual market-adjusted return of 3.97% while a portfolio of the least favorably recommended ones yielded an average annual market-adjusted return of -9.06%, a difference of over 13 percentage points. (6) But in 2000, the stocks least favorably recommended by analysts earned an annualized market-adjusted return of 48.66% while the stocks most highly recommended fell 31.20%, a return difference of almost 80 percentage points. Remeber what happened to 'glamor' stocks in 2000? (7) Given widespread unwillingness to sell short, more trades will result from a 'buy' than from a 'sell' recommendation; (8) Investors do not fully adjust for the tendency of affiliated analysts to issue overly optimistic recommendations for the client firms. I also want an alphabtically order list of research. So that's the original plan. Right now the information [and it's good information] is randomly posted here. Give me a month or so - say until Feb/March 2005 - to have this page looking right. I am still in the stage of gathering data. A. Dugar, and S. Nathan in 'The Effects of Investment Banking Relationships on Financial Analysts' Earning Forecasts and Investment Recommendations' [1995] and H. Lin and M. McNichols in 'Underwriting Relationships, Analysts' Earnings Forecasts and Investment Recommendations' [1998] find that analysts appear to favorably bias their recommendations for firms that have underwriting relationships with their brokerage firms. Even if a frim does not presently have a relationship with the covered firm, there is still an opimistic bias. Bradshaw, Richardson, and Sloan (2003), Michaely and Womack (1999), Jegadeesh, Kim, Krische, and Lee (2004) and Cowen, Groysberg, and Healy (2003) document that the prospect of getting investment banking deal induces analysts to express more optimistic views. J. Easterwood and S. Nutt in 'Inefficiency in Analysts' Earnings Forecasts: Systematic Misreaction or Systematic Optimism?' [1999] conclude that analysts' earnings forecasts are systematically overoptimistic. M. McNichols and P. O'Brien in 'Self-Selection and Analyst Coverage' [1997] find that consensus recommendations are biased because optimistic analysts are more likely to provide recommendations than are pessimistic analysts. Michaely and Womack (1999) show that investors do not fully adjust for the tendency of affiliated analysts to issue overly optimistic recommendations for the client firms. Agrawal and Chen (2004), however, point out that biases due to conflict of interest affect only recommendations and not earnings forecasts. S. Krische and C. Lee in 'Analyzing the Analysts: Are Stock Recommendations Informative?' [2001] and S. Stickel in 'Analysts Incentives and the Financial Characteristics of Wall Street Darlings' [2000] find that analysts tend to recommend glamour stocks, which have high market-to-book ratios, high price-to-earnings ratios, high past sales growth, and strong price momentum, although these stocks tend to underperform non-glamour stocks in the time period they examined. Krische and Lee further show that in and of itself, analysts' stock-picking patterns tend to reduce the effectiveness of their picks, because they fail to exploit systematic factors that lead to higher future returns. K. Womack in 'Do Brokerage Analysts' Recommendations have Investment Value?' [1996] and Stickel [1995] find that there is a significant price impact associated with new buy and sell recommendations. S. Stickel in 'The Anatomy of the Performance of Buy and Sell Recommendations' [1995] finds that several recommendation effects on price, such as the effects of recommendations made by large brokerage firms and higher-ranked analysts, are temporary and quickly reverse. B. Barber, R. Lehavy, M. McNichols, and B. Trueman in 'Can Investors Profit from the Prophets? Security Analyst Recommendations and Stock Returns' [2001] and Krische and Lee [2001] find that stocks favored by analysts out-perform stocks disfavored by analysts. Vinesh Jha, David Lichblau, and Haim Mozes in 'The Usefulness of Analysts' Recommendations' find that rather than providing new, private information, analysts appear to use their information-processing skills to corroborate or reject other publicly available information. It is not that analysts have information signals that other investors do not have; they may simply better understand the implications of these information signals.
First, think about this puzzle. Disagreement ought to make shares riskier, producing higher returns for those who can tolerate the ups and downs. Instead, the opposite is true. Investors tend to do better if they bet on safer stocks where there's broad agreement about the future. Lower risk and higher reward - not the way the world ought to work. To understand how economists measure these differences, consider Dell and Sears. The consensus (average) estimate is that the Texas computer maker will earn $1.53 per share next year. The range is narrow, with a high estimate of $1.57 and a low estimate of $1.49. At Sears, meanwhile, the consensus is $3.37. But the range is broader, with a high of $3.85 and a low of $2.95. There's clearly more agreement about Dell, where estimates cluster near the consensus. The statistical measure of this clustering is called standard deviation, which you can calculate by plugging all the estimates into a spreadsheet. This is how economists measure dispersion of opinion. For Dell, the magic number is 0.02. For Sears, it's 0.28. From an investor's point of view, low is good (lots of agreement); high is bad (lots of disagreement). So if you're investing purely on the basis of dispersion, buy Dell and sell Sears. Now switch from math to theory. The best explanation of the dispersion effect has to do with short selling. In 1977 an economist named Edward Miller pointed out that restrictions on selling short - everything from regulations requiring an uptick in price before each short sale to the cultural bias against bearishness - could distort markets. When pessimists are pushed to the sidelines, Miller reasoned, optimists call the shots. Under this logic, things are most likely to get out of whack when there are highly diverging views. The bulls buy, the bears sit on their hands, and prices are bid too high. Over time investors correct this distortion. As that happens, stocks that were overpriced because of differences of opinion do badly. Then there are the subtle pressures that also play a role. Anna Scherbina, who teaches at Harvard Business School, is the doyenne of dispersion. She has just written her third paper on the subject and believes that disagreement can also distort earnings estimates. Her work indicates that bearish analysts tend to drop coverage of companies rather than publish estimates that are far below the average. As their mothers taught them, better to say nothing at all than to be negative. So differences of opinion make for an overheated consensus, which pushes up share prices even more. Other economists focus on what causes analysts to disagree. The idea here is that companies with something to hide deliberately get fuzzy when they talk about their prospects. This leads to disagreement about estimates. Companies that really are doing well, on the other hand, communicate in a clear, no-nonsense way. So analysts generally agree about their prospects. Tong Yao at the University of Arizona points out that when managers provide clearer guidance about future earnings (as measured by high levels of analyst agreement), performance is better - in terms of earnings growth and share price. According to his research, these gains may persist for at least three years. Yao argues that "good" companies have an incentive to provide unbiased guidance, while "bad" companies have an incentive to shade the truth. Anna Scherbina, Harvard Business School, June 2004 Investors fail to fully account for optimistic bias associated with analyst disagreement. This bias arises for two reasons. First, analysts issue more optimistic forecasts when earnings are uncertain. Second, analysts with sufficiently low earnings expectations who choose to keep quiet introduce an optimistic bias in the mean reported forecast that is increasing in the underlying disagreement. Indicators of the missing negative opinions predict earnings surprises and stock returns. By selling stocks with high analyst disagreement institutions exert correcting pressure on prices. It has been empirically documented that analysts' opinions are reflected in stock prices. This is done in A. Abdel-khalik & Ajinkya [Returns to informational advantages: The case of analysts' forecast revisions (1982)], W. Forbes & L. Skerratt [Analysts' forecast revisions and stock price movements (1992)], D. Givoly & J. Lakonishok [The information content of financial analysts' earnings forecasts (1979)]. C. Gleason & C. Lee [Characteristics of price informative analyst forecasts (2000)], N. Gonedes, N. Dopuch, & S. Penman [Disclosure rules, information production, and capital market equilibrium: The case of forecast disclosure rules (1976)], E. Hawkins, S. Chamberlin, & W. Daniel [Earnings expectations and security prices (1984)], E. Imhoff & G. Lobo [Information content of analysts’ composite forecast revisions (1990)], and S. Stickel [Predicting individual analyst earnings forecasts Stickel (1990)]. To empirically assess whether the marginal investor adjusts for this bias I construct predictors of the earnings surprises based on this theory and investigate whether they also forecast stock returns. I estimate the bias that arises due to self-selection in coverage based on the decrease in analyst following of a firm over the past three months. I further assume that the reported forecast distribution is a truncated normal and the absent analysts would have issued forecasts in the missing left tail of the distribution. I then compute the standard deviation of thus defined “true” forecast distribution and assume that it represents the underlying level of uncertainty. Interacting the decrease in coverage with dispersion in the outstanding forecasts captures the degree to how pessimistic the withheld opinions could have been. The estimate of the bias is a significant negative predictor of the future earnings surprise, with a coefficient close to one, implying that it correctly captures the magnitude of the bias. That the bias is also negatively related to abnormal returns around earnings announcement days suggests that the marginal investor does not adjust earnings forecasts for self-selection. Among the stocks in the highest quintile of forecast dispersion, those that have experienced a decrease in analyst following over the past three months earned, on average, a 4.8% lower risk-adjusted annual return than the stocks with a nondecreased analyst following. Right-skewness of the forecast distribution indicates the absence of negative opinions among reported forecasts. Stocks in the highest skew-based quintile have outperformed the stocks in the lowest quintile by, on average, 2.76% per year on the risk-adjusted basis. The strong relation between forecast dispersion and future earnings forecast revisions accounts for the fact that earnings momentum fully explains the profitability of the trading strategies based on forecast dispersion and estimated bias due to self-selection in analyst coverage. Price momentum explains about half of the profit. Since the low future stock returns are at least partly caused by analysts’ unwillingness to report bad news, this result lends support to the hypothesis of Harrison Hong, Terence Lim, & Jeremy Stein [Bad news travels slowly: size, analyst coverage, and the profitability of momentum strategies (2000)] that negative momentum arises because bad news disseminates slowly. The empirical findings described above suggest that the marginal investor does not fully adjust for the forecast biases due to self-selection in coverage and earnings uncertainty. Diether, Malloy, and Scherbina (2002) document that dispersion is negatively related to future returns, although the authors offer a slightly different explanation for this phenomenon. They interpret analyst disagreement as indicative of disagreement among investors and invoke the E. Miller [Risk, uncertainty, and divergence of opinion (1977)] argument that in the presence of short-sale constraints stock prices will reflect the view of the more optimistic investors. When the disagreement is resolved, prices converge down to the fundamentals, earning low returns. If the predictive power of dispersion on future returns is caused by bounded rationality, sophisticated investors should sell high-dispersion stocks. Among users of analysts’ forecasts, institutional investors, in particular, mutual funds that engage in independent research, are better positioned to understand inherent biases. But due to short-selling restrictions mutual funds are able to sell only shares they already own. Moreover, they are able to correct mispricing only if they sell the stock in quantities sufficient to affect prices. Sadka and Scherbina (2004) argue that high-dispersion stocks continue to be mispriced because arbitrage is costly. They show that high-dispersion stocks have significantly higher trading costs than otherwise similar stocks. Ronnie Sadka, University of Washington Business School & Anna Scherbina, Harvard Business School, June 18, 2004 Examining mispricing of stocks with high levels of analyst disagreement about future earnings reveals a close link between mispricing and liquidity. Previous research finds these stocks often to be overpriced, but prices to correct down within a fiscal year as uncertainty about earnings is resolved. We conjecture that one reason mispricing has persisted is that these stocks have higher trading costs than otherwise similar stocks, possibly because some investors are better informed than the market maker about how to aggregate analysts’ opinions. As analyst disagreement increases so does the informational disadvantage of the marker maker, and trading costs rise. In the cross-section, less liquid stocks are, on average, more severely mispriced. Moreover, increases in aggregate market liquidity accelerate convergence of prices to fundamentals. As a result, returns of initially overpriced stocks are negatively correlated with the time series of innovations in aggregate market liquidity. Wei Xiong [Convergence Trading with Wealth Effects: An Amplification Mechanism in Financial Markets (2001)], and D. Gromb & Vayanos, [Equilibrium and welfare in markets with financially constrained arbitrageurs (2002)] show that in imperfect capital markets a further price divergence in the assets involved in the convergence trade might trigger additional demand for capital and thus force arbitrageurs to abandon potentially profitable positions. Dilip Abreu & Markus Brunnermeier [Synchronization Risk and Delayed Arbitrage (2002)] and Abreu & Brunnermeier [Bubbles and Crashes (2003)] establish that in a world in which two similar assets might differ in price indefinitely, arbitrageurs will not only forgo a convergence trade, but instead establish a long position in the overpriced asset anticipating a further price run-up. Markus Brunnermeier & Stefan Nagel [Hedge Funds and the Technology Bubble (2004)] provide empirical evidence for this having occurred during the 'tech bubble', when hedge funds held long positions in technology stocks they considered to be overpriced. Evidence presented in this paper of the relation between mispricing and liquidity augments a growing body of empirical literature on costly arbitrage. David Lesmond, Michael Schill, & Chunsheng Zhou, [The illusory nature of momentum profits (2004)], Robert Korajczyk & Ronnie Sadka [Are momentum profits robust to trading costs? (2004)], and Zhiwu Chen, Werner Stanzl, & Masahiro Watanabe [Price impact costs and the limit of arbitrage (2002)] who study the profitability of momentum trading strategies after accounting for transaction costs find that the momentum effect, as documented in the literature, could be largely eliminated by a small capital investment. Ronnie Sadka [The seasonality of momentum: analysis of tradability (2001)] reaches a similar conclusion about the January effect. Mark Mitchell, Todd Pulvino, & Erik Stafford [Limited Arbitrage in Equity Markets (2002)] and Malcolm Baker & Serkan Savasoglu [Limited Arbitrage in Mergers and Acquisitions (2002)] find that taking into account arbitrage costs greatly reduces potential profits in merger arbitrage. X. Gabaix, A. Krishnamurthy, & O. Vigneron [Limits of Arbitrage: Theory and Evidence from the Mortgage Backed Securities Market (2004)] document a relationship between mispricing and arbitrage costs in the mortgage-backed securities market, and Jeffrey Pontiff [Costly Arbitrage: Evidenc from Closed-End Funds (1996)] presents evidence that the mispricing of closed-end funds is closely related to the cost of arbitrage. Analyst disagreement about future earnings leads to a more optimistic than usual bias in the mean outstanding forecast. This relationship might be explained by analysts' incentives. Terrence Lim [Rationality and Analysts' Forecast Bias (2001)] hypothesizes that when earnings are highly uncertain, analysts are willing to add a higher optimistic bias to their estimates in exchange for inside information from management about a firm's future earnings. Anna Scherbina [Analyst Disagreement, Forecast Bias and Stock Returns (2004)] conjectures that analysts, who derive monetary benefits from issuing optimistic forecasts, add a higher bias to their private estimates knowing that they will be penalized less for being wrong when earnings are highly uncertain. Moreover, if analysts with extremely negative views choose not to reveal them the mean of the reported forecast distribution will be upwardly biased, more so the more negative the withheld opinions. This is likely to be the case when analyst disagreement is high overall. Because the marginal investor fails to fully account for the correlation between analyst disagreement and forecast bias, high-dispersion stocks are likely to be overvalued and to under-perform otherwise similar stocks in the future. [Diether, Malloy, and Scherbina (2002)]. Karl Diether, Christopher Malloy & Anna Scherbina [Differences of opinion and the cross-section of stock returns (2002)] find that mispricing is corrected in, on average, six months. Stocks with high dispersion tend to be smaller, possibly because smaller stocks are more opaque. After controlling for size, stocks with high dispersion tend to have higher analyst coverage, possibly because there is more demand for expert opinion when it is difficult to interpret available information. High-dispersion stocks tend to be value stocks that have done poorly in the past and have higher systematic risk (see Stephen Ciccone [Does Analyst Optimism about Future Earnings Distort Stock Prices? (2003)]. Jeffery Abarbanell [Univ of North Carolina] & Reuven Lehavy [Univ of Michigan] January 2003 We demonstrate the role of three empirical properties of cross-sectional distributions of analysts' forecast errors in generating evidence pertinent to three important and heretofore separately analyzed phenomena studied in the analyst earnings forecast literature: purported bias (intentional or unintentional) in analysts' earnings forecasts, forecaster over/underreaction to information in prior realizations of economic variables, and positive serial correlation in analysts' forecast errors. Four decades of research have produced an array of empirical evidence and a set of behavioral and incentive-based theories that address two fundamental questions: Are analysts’ forecasts biased? and Do analysts underreact or overreact to information in prior realizations of economic variables? This empirical literature has long offered conflicting conclusions and is not converging to a definitive answer to either question. We extend our analysis beyond a synthesis and summary of the findings in the literature by identifying the role of two relatively small asymmetries in the cross-sectional distributions of analysts' forecast errors in generating conflicting statistical evidence. We note that the majority of conclusions concerning analyst forecast rationality in the literature are directly or indirectly drawn from analyses of these distributions. The first asymmetry is a larger number and a greater magnitude of observations that fall in the extreme negative relative to the extreme positive tail of the forecast error distributions (hereafter, the tail asymmetry). The second asymmetry is a higher incidence of small positive relative to small negative forecast errors in cross-sectional distributions (hereafter, the middle asymmetry). The individual and combined impact of these asymmetries on statistical tests leads to three important observations. First, differences in the manner in which researchers implicitly or explicitly weight observations that fall into these asymmetries contribute to inconsistent conclusions concerning analyst bias and inefficiency. We present statistical evidence that demonstrates how the two asymmetries in forecast error distributions can indicate analyst optimism, pessimism, or unbiasedness. We also show how observations that comprise the asymmetries can contribute to, as well as obscure, a finding of apparent analyst inefficiency with respect to prior news variables, including prior returns, prior earnings changes, and prior forecast errors. For example, our empirical evidence explains why prior research that relies on parametric statistics always finds evidence of optimistic bias as well as apparent analyst underreaction to prior bad news for all alternative variables chosen to represent prior news. It also explains why evidence of apparent misreaction to good news is not robust across parametric statistics or across prior news variables, and why the degree of misreaction to prior bad news is always greater than the degree of misreaction to prior good news, regardless of the statistical approach adopted or the prior information variable examined. The evidence also highlights the importance of future research into the question of whether reported earnings are, in fact, the correct benchmark for assessing analyst bias and inefficiency. This is because common motivations for manipulating earnings can give rise to the appearance of analyst forecast errors of exactly the type that comprise the two asymmetries if unbiased and efficient forecasts are benchmarked against manipulated earnings. It is possible that some evidence previously deemed to reflect the impact of analysts’ incentives and cognitive tendencies on forecasts is, after all, attributable to the fact that analysts do not have the motivation or ability to completely anticipate earnings management by firms in their forecasts. The Zacks earnings forecast database contains approximately 180,000 consensus quarterly forecasts for the period 1985–1998. Lack of available price data reduced the sample size to 123,822 quarterly forecast errors. The data requirements for estimating quarterly accruals further reduced the sample on which our tabled results are based to 33,548 observations. One of the most widely held beliefs among accounting and finance academics is that incentives and/or cognitive biases induce analysts to produce generally optimistic forecasts (Brown [1993] and Kothari [2001]). Summary statistics associated with forecast error distributions raise doubts about this conclusion. One distinctive feature of the distribution is that the left tail (ex-post bad news) is longer and fatter than the right tail, i.e., far more extreme forecast errors of greater absolute magnitude are observed in the ex-post “optimistic” tail of the distribution than in the “pessimistic” tail. We refer to this characteristic of the distribution as the tail asymmetry. We find that 13% of the observations fall below a negative forecast error of -0.5, while only 7% fall above a positive error of an equal magnitude. Closer visual inspection of the data reveals a second feature of the distribution - a higher frequency of small positive forecast errors versus small negative errors.The incidence of small positive relative to small negative errors increases as forecast errors become smaller in absolute magnitude. We refer to this property of the distribution as the middle asymmetry. The prior studies that found that analysts have incentives to produce forecasts that fall slightly short of reported earnings: Degeorge, Patel, and Zeckhauser [1999], Matsumoto [2002], Brown [2001], Burgstahler and Eames [2002], Bartov, Givoly, and Hayn [2002], Dechow, Richardson, and Tuna [2003], and Abarbanell and Lehavy [2003a and 2003b]). Prior Studies: Degeorge, F., J. Patel, and R. Zeckhauser, 1999, “Earnings Management to Exceed Thresholds,” Bartov, E., D. Givoly, and C. Hayn, 2000, “The Rewards to Meeting or Beating Earnings Expectations,” Burgstahler, D., and M. Eames, 2002, “Management of Earnings and Analyst Forecasts,” Dechow, P., S. Richardson, and I. Tuna, 2003, “Why Are Earnings Kinky? An Examination of the Earnings Management Explanation,” Matsumoto, D., 2002, “Management's Incentives to Avoid Negative Earnings Surprises,” The statistics indicate that the tail asymmetry pulls the mean forecast error toward a negative value, supporting the case for analyst optimism. But the excess of small positive over small negative errors associated with the middle asymmetry is largely responsible for a significantly higher overall incidence of positive to negative forecast errors in the distribution, thus supporting the case for analyst pessimism. Abarbanell and Lehavy (2002) present evidence confirming the presence of the middle asymmetry. Finally, a zero median forecast error, which supports an inference of analyst unbiasedness, reflects the countervailing effects of the middle asymmetry and tail asymmetries. If a researcher relies on a given summary statistic to draw an inference about analyst bias, a relatively small percentage of observations in the distribution of forecast errors will be responsible for his or her conclusion. This is troublesome because extant hypotheses that predict analyst optimism or pessimism typically do not indicate how often the phenomenon will occur in the cross-section and often convey the impression that bias will be pervasive in the distribution (see studies suggesting that analysts are hard-wired or motivated to produce optimistic forecasts, e.g., Affleck, Graves, and Mendenhall [1990], Francis and Philbrick [1993], and Kim and Lustgarten [1998], or that selection biases lead to hubris in analysts’ earnings forecasts, e.g., McNichols and O’Brien [1997]). Abarbanell and Lehavy (2002) show that there is a very high correlation between observations found in the extreme negative tail of forecast error distributions and firms that report large negative special items, even when special items are excluded from the reported earnings benchmark used to calculate the forecast error. Brown (2001) finds that the mean and median forecasts in recent years indicate a shift away from analyst optimism and toward analyst pessimism. Brown, L., 2001, “A Temporal Analysis of Earnings Surprises: Profits versus Losses,” Number of Observations 33,548
Affleck-Graves, J., L. Davis, and R. Mendenhall, 1990, “Forecasts of Earnings per Share: Possible Sources of Analyst Superiority and Bias,” Francis, J., and D. Philbrick, 1993, "Analysts' Decisions as Products of a Multi-task Environment," Kim, C., and S. Lustgarten, 1998, “Broker-Analysts’ Trade-Boosting Incentive and Their Earnings Forecast Bias,” McNichols, M., and P. O’Brien, 1997, “Self-Selection and Analyst Coverage,” Would undergraduate business students, and MBA students make the same miscalculations as stock anaysts? That is the topic tackled below - and the finding is a qualified yes. Experimental Evidence of Reactions to Positive vs. Negative Information Douglas E. Stevens [Syracuse] & Arlington W. Williams [Indiana] August 2001 We find no evidence of systematic overreaction or optimism bias. Subject forecasts systematically underreact to both positive and negative information signals. Consistent with prior documented human decision bias, this underreaction is not reduced by increased experience with the forecasting series. Our results are consistent with two human decision biases found in prior studies of probabilistic judgments: the “anchoring and adjustment” heuristic (Tversky and Kahneman - "Judgment under uncertainty: Heuristics and biases" 1974) and the “gambler’s fallacy” (Kahneman and Tversky - "Subjective probability: A judgment of representativeness" 1972). gambler's fallacy - In physics this term is called “Maturity of Chances,” and can occur for example, if someone flips a coin 1,000 times. According to the law of averages, it is assumed that approximately 500 tosses will be heads and approximately 500 tosses will be tails. However let’s say that after 900 tosses, 600 were heads and only 300 were tails. Some people might say that tails are now “due,” so the remaining 100 tosses will be mostly tails. But previous rolls of the dice have no effect on future rolls - as it is with flips of a coin. anchoring and adjustment - We tend to base estimates and decisions on known ‘anchors’ or familiar positions, with an adjustment relative to this start point. We are better at relative thinking than absolute thinking. Example If asked whether the population of Turkey was greater or less than 30 million, you might give one or other answer. If then asked what you thought the actual population was, you would very likely guess somewhere around 30 million, because you have been anchored by the previous answer. Fried and Givoly (1982), O’Brien (1988), Francis and Philbrick (1993), Kang, O’Brien, and Sivaramakrishnan (1994), and Dreman and Berry (1995) provide evidence that analysts’ forecasts are overly optimistic. Studies finding evidence for systematic underreaction include Lys and Sohn (1990), Abarbanell (1991), Abarbanell and Bernard (1992), Ali, Klein, and Rosenfeld (1992), and Elliot, Philbrick, and Wiedman (1995). In contrast, DeBondt and Thaler (1990) find that changes in forecasted earnings per share are too extreme, consistent with systematic overreaction. Easterwood and Nutt (1999) regress the error in the forecast of annual earnings in the current period against the actual earnings change for the prior year. Consistent with Abarbanell and Bernard (1992), they find that forecast errors are positively related to the prior earnings change, suggesting that analysts underreact to prior earnings performance. Next, Easterwood and Nutt group firms into low, normal, and high values of prior year performance to test whether Abarbanell and Bernard’s (1992) results hold over all three partitions of earnings information. Easterwood and Nutt find that financial analysts underreact to negative information but overreact to positive information.Easterwood and Nutt’s evidence suggests that analysts are systematically optimisticconcerning the implications of new information. J Affleck-Graves, L. Davis, & R. Mendenhall in "Forecasts of earnings per share: Possible sources of analyst superiority and bias" [1990] compared the EPS forecasts of analysts on the Institutional Brokers Estimate System (IBES) to EPS forecasts generated by undergraduate business students and the Autoregressive Integrated Moving Average (ARIMA) time-series model. Student subjects were given 52 consecutive quarters of actual EPS data, in columnar and graphical form, and were asked to provide a forecast of the 53rd quarter. Each subject provided forecasts for 10 firms out of a total sample of 700 firms. For each firm, the 52 quarters of actual EPS data were also used to derive a forecast from the ARIMA time-series model. Affleck-Graves et al. found that the forecasts of both analysts and students exhibited significant positive bias (optimism) while forecasts generated by the time-series model did not. When errors above 100% were removed from the sample, analyst forecasts were found to be significantly more positively biased than student forecasts. Affleck-Graves et al. concluded that the optimism bias found in previous studies of analysts’ forecasts is not only attributable to economic incentives, but also to judgmental heuristics typical of all human decision-makers. There are a number of potential weaknesses in their study. First, the time-series data typically exhibited an upward trend due to inflation, so the optimism found in student forecasts is also consistent with subjects overestimating the trend component in the data. Second, the subjects were not paid for the accuracy of their forecasts, so there is some question as to the effort and attention the subjects devoted to the forecasting task. Finally, there was no feedback given to subjects, which reduced the opportunity for learning in the experiment. M Calegari & N Fargher in "Evidence that prices do not fully reflect the implications of current earnings for future earnings" [1997] examined the ability of student subjects to forecast the time series properties of EPS and the relation between their forecasts and market prices using an experimental market. In ten experimental sessions, undergraduate business students were presented with 15 years of quarterly EPS data. Calegari and Fargher found that while subjects’ forecasts were positively correlated with the innovation in the prior quarter earnings, forecasts were not significantly related to innovations in the three quarters immediately preceding the prior quarter.the underweighting exhibited in the forecasts and market prices was not reduced by experience with the forecasting and trading task. The underweighting result also persisted across positive and negative earnings innovations. Helbock Gunther & Martin Walker, 2000 This study presents evidence that forecast revisions respond asymmetrically to good and bad news. Analyst overreaction is greater in bad than good news years.
The study divided Wall Street into different camps based on the proportion of buy or sell ratings that research departments maintained on stocks they covered, classifying them as "optimistic" or "pessimistic" firms. They found that stock upgrades to buy issued by analysts at firms with the smallest percentage of buy recommendations significantly outperformed those of brokers with the greatest percentage of buys, by an average of 50 basis points per month. On the flip side, downgrades to hold or sell coming from brokers issuing the most buy recommendations significantly outperformed those of brokers issuing the fewest, by an average of 46 basis points per month. "We were trying to answer whether knowing those distributions could have been valuable for investors," said Lehavy. Whether a firm's overall proportion of buys to sells tended toward the optimistic or pessimistic side wasn't publicly available on a firm-by-firm basis until Sept. 9, 2002, when new regulations required firms to disclose their ratings distributions on every research report. Once the new regulations took effect, the differences between various brokerage firms' buy-sell distributions narrowed, and there was no longer an advantage to be had in trying to follow recommendations from analysts that are contrary to their firms' bent, according to Lehavy. In the authors' paper on the study, which is being reviewed by an accounting journal, they said they believe the new rules curbed some firms' tendencies toward issuing overly optimistic research. Prior Research from the above Authors: Consensus Analyst Recommendations and Stock Returns Brad Barber, Reuven Lehavy, Maureen McNichols, Brett Trueman - August 1998 - In this paper we document that an investment strategy based on the consensus [average] recommendations of security analysts earns positive returns. For the period 1986-1996, a portfolio of stocks most highly recommended by analysts eaned an annualized geometric mean return of 18.8%, while a portfolio of stocks least favorably recommended eaned only 5.78%. [In comparison, an investment in a value-wieghted market index earned an annulaized geometric mean return of 14.5%] These results are robust to partitions by time period and overall market direction, and most pronounced for small and medium-sized firms. the abnormal returns also persist when we allow a lapse of up to 15 days before acting on the investment recommendations. Using the Zacks database for the period 1985-1996, which includes over 360,000 recommendation from 269 brokerage houses and 4,340 analysts, we tracked in calendar time the investment performance of porfolios of firms grouped accoriding to their consussus analyst recommendations. These returns are gross of transaction costs and the market impact of trading. Investing in highly recommended securites requires an active trading strategy with turnover rates at times in excess of 400% annually. Academic theory and Wall Street practice are clearly at odds regarding this issue. On the one hand, the semi-strong for of market efficiency posits that investors cannot trade profitably on the basis of publicly available information, such as analyst recommendations. On the other hand, research departments of brokerage houses spend billions of dollars annually on security analysis, presumably because these firms believe it can generate superior returns for their clients. Our results provide surprisingly strong evidence that Wall Street may be right. Prior Studies on this Subject: Amoung the papers which previously examined the investment performance of security analysts' stock recommendations are Barber and Loeffler ["The Dartboard Column: Second-hand Information and Price Pressure" 1993], C. Bidwell ["How Good is Institutional Brokerage Research?" 1997], R. Diefenbach [also titled "How Good is Institutional Brokerage Research?" 1972], Groth, Lewellen, Schlarbaum, and Lease ["An Analysis of Brokerage House Securities Recommendations" 1979], Stickel [1995], and Womack [1996]. Copeland and Mayers ["The Value Line Enigma" 1982] studied the investment performance of the Value Line Investment Survey while Desai and Jain ["An Analysis of the Recommendations of the Superstar Money Managers" 1995] analyzed the return from following Barron's annual roundtable recommendations. Our study is most closely related to S. Stickel ["The Anatomy of the Performance of Buy and Sell Recommendations" 1995] and K. Womack ["Do Brokerage Analysts' Recommendations Have Investment Value?" 1996]. Using the Zacks database, Stickel studies the price impact of 16,957 changes in analyst recommendations over the 1988-1991 period and finds that recommendation changes from sell to buy [buy to sell] were accompanied by positive [negative] returns at the time of the announcement. Further, he documents that most of the price adjustment occureed during the first 30 days after the recommendation change. Using the First Call database, Womack analyzes the impact of 1,537 changes in analyst recommendation to or from strong buy or strong sell, for the top 14 US brokerage reasearch departments druing the 1989-1991 period. He finds significantly positive [negative] returns for the buy [sell] recommendation at the time of the announcement. He also documents a post-recommendation stock price drift lasting up to one month for buys and six months for sells. Brett Trueman, Spring 1994 The use of analyst forecasts as proxies for investors' earnings expectations is commonplace in empirical research. An implicit assumption behind their use is that they reflect analysts' private information in an unbiased manner. As demonstrated here, this assumption is not necessarily valid. There is shown to be a tendency for analysts to release forecasts closer to prior earnings expectations than is appropriate, given their information. Further, analysts exhibit herding behavior, whereby they release forecasts similar to those previously announced by other analysts, even when this is not justified by their information. "Analysts' Reputation Concerns and Underreaction to Public News: Theory and Evidence" by Sonya Lim, November 2002 Analysts underreact to past earning information: Richard Mendenhall ["Evidence on the Possible Underweighting of Earnings-Related Information" 1991], Jeffrey Abarbanell and Victor Benard ["Tests of Analysts' Overreaction/Underreaction to Earning Information as an Explanation for Anomalous Stock Price Behavior" 1992], Ahiq Ali, April Klein, and James Rosenfeld ["Analysts' Use of Information about Permanent and Transitory Earnings Components in Forecasting Annual EPS" 1992], John Jacob and Thomas Lys ["Determinants and Implications of the Serial-Correlation in Analysts' Earnings Forecast Errors" 2000], Jana Raedy and Philip Shane ["Horizon-Dependant Conservatism in Financial Analysts' Earnings Forecasts" 2000], Philip Shane and Peter Brous ["Investor and Value Line Analysts Underreaction" 2000] Analysts underract to past stock prices: April Klein ["A Direct Test of the Cognitive Bias Theory of Share Price Reversals" 1990], Lys and Sohn ["The Association Between Revisions of Financial Analysts' Earnings Forecast and Security-Price Changes" 1990], Jeffrey Abarbanell ["Do Analysts' Earnings Forecast Incorporate Information in Prior Stock Price Changes?" 1991] Analyst forecast error is predictable from prior revisions: John Elliott, Donna Philbrick, and Christine Wiedman ["Evidence from Archival Data on the Relation Between Security Analysts' Forecast Errors and Prior Forecast Revisions" 1995], Eli Amir and Yoav Ganzach ["Overreaction and Underreaction in Analysts' Forecasts" 1998], John Easterwood and Stacey Nutt ["Inefficiency in Analysts' Earnings Forecast: Systematic Misreaction or Systematic Optimism?" 1999] The stock market underestimates the implications of previous period earnings for future earnings: Bernard and Thomas ["Evidence that stock prices do not fully reflect the implications of current earnings for future earnings" 1990], Freeman and Tse ["The multi-period information content of earnings announcements: Rational delayed reactions to earnings news" 1989] Using an intrinsic value measure based on analysts' short- and long-term earning forecasts, Jing Liu ["Post Earnings Announcement Drift and Analysts Forecasts" 1999] finds that analysts underreact to earnings news more than the stock market. John Jacob and Thomas Lys ["Determinants and Implications of the Serial-Correlation in analysts' Earnings Forecast Errors" 2000] find that analysts exhibit similar patterns of serial-correlations in forecast errors across all the companies they follow, and that analysts following the same company also show similar patterns of serial-correlation. Their finds suggest that there are analyst- and company-specific factors underlying analysts' underreaction. Michael Mikhail, Beverly Walther, and Richard Willis ["The effect of esperience on security analysts underreaction and post earnings announcement drift" 2001] show that analysts' underreaction to past earnings news decreases with their experience, possibly becasue cognitive biases that lead to undrreactions are mitigated with experience. Inefficient processing of information due to psychological biases such as conservatism (Edwards [1968]) and overconfidence (Dale Griffin and Amos Tvershy ["The weighing of evidence and the determinants of overconfidence" 1992]) can result in underreaction. Geoffrey Friesen and Paul Weller ["Quantifying cognitive biases in analyst earnings forecasts" 2002] find that analysts are overconfident about the precision of their own information. Trueman [1990] shows that an analyst may be reluctant to revise a previously issued forecast upon receipt of new information. Forecast revision implies that the analyst's original information was inaccurate, and thus investors will lower their assessment of the analyst's ability to collect information in a tiimely manner. Analysts may issue optimistic forecast to gain access to inside information from management or to win investment banking business (Das, Levine and Sivaramakrishman [1998], Dugar and Nathan ["The effects of investment banking relationship on financial analysts' earnings forecasts and investment recommendations" 1995], Francis and Philbrick [1993], Harrison Hong and Jeffrey Kubik ["Analyzing the Analysts: Career Concerns and Biased Earnings Forecasts" 2001], Terrence Lim ["Rationality and Analysts' Forecast Bias" 2001]) Canice Prendergast and Lars Stole ["Impetuous Younsters and Jaded Old-Times: Acquiring a reputation for Learning" 1996] examine how individuals change their behavior on receipt of new information when they wish to acquire a reputation as fast learners, where the learning ability is reflected in the precision of their private signal. In the context of a manager making investment decision on projects over time based on his private information, they show that the manager first exaggerates his private information but later becomes too conservative. Brett Trueman ["On the Incentives for Security Analysts to Revise their Earnings Forecasts" 1990] and Prendergast and Stole [1996] show that reputational concerns create incentive for agents to underreact to information. Related Articles: Academic Studies on Analysts' Recommendations July 2003 Factoids - This is not a direct link - you must cursor down 12-14 times to see this article. If this kind of article is of interest, you may also want to view Investors Prefer Stocks in the News in the December 2003 Factoids - and once again you will need to cursor down, this time seven times should do it. unlisted author[s] Mergers and acquisitions (M&A), especially tender offers, generally provide abnormally positive returns to target firms, which creates an incentive for financial analysts to identify them ex ante. Numerous practitioner and academic articles suggest that merger targets are predictable [J. Hasbrouck, "The characteristics of takeover targets: Q and other measures", (1985), and M. Song and R. Walkling, "The impact of managerial ownership on acquisition attempts and target shareholder wealth", (1993)], so therefore we might expect analysts to incorporate this predictability component into their recommendations. In our sample, the average market adjusted return for target firms during the –1 day to +1 day period surrounding the announcement is 31.5 percent. We know of no other corporate event that generates such large returns during such a short period. In our 1998-2001 sample, we find analysts are incapable of identifying takeover targets through their recommendations nor can they distinguish between wealth creating and destroying tender offers. At the same time, however, we find no evidence of conflicts of interest - analyst ratings, the immediate market reactions around the recommendation, and the long-run performance of affiliated versus unaffiliated recommendations are virtually identical. There are several academic studies that show that following the advice of analysts would have proven profitable - those studies include: Barber, Lehavy, McNichols and Trueman {"Reassessing the returns to analysts’ stock recommendations" 2001 - for the period of 1986-1996, a portfolio comprised of the most highly recommended stocks, for example, generated an average annual market-adjusted return of 3.97 percent while a portfolio of the least favorably recommended ones yielded an average annual market-adjusted return of -9.06 percent, a difference of over 13 percentage points.], Jagadeesh, Kim, Krishe and Lee ["Analyzing the analysts: When do recommendations add value?" 2003] and Womack ["Do brokerage analysts’ recommendations have investment value" 1996]. However, a recent study by Barber, Lehavy, McNichols, and Trueman (2003) documents that adhering to the recommendations of analysts in 2000 and 2001 would have been “disastrous.” Specifically, they found that that the stocks least favored by analysts outperformed those most preferred by over 20 percentage points. Branson, Guffey, and Pagach (1998) and Irvine (2002) find short-term positive reactions to analyst recommendations while Barber, Lehavy, McNichols, and Trueman (2001), Jagadeesh, Kim, Krische, and Lee (2003) and Womack (1996) present evidence of long-run value stemming from analyst picks. A recent paper by Clarke, Ferris, Jayaraman, and Lee ["Peering into a cloudy crystal ball: Analyst recommendations preceding bankruptcy" 2003] investigates analyst behavior preceding bankruptcies. They find that analyst recommendations remain optimistic even as the firm approaches bankruptcy. In addition, they find no difference in ratings strength between affiliated and unaffiliated analysts. In the area of equity offerings, Bradshaw, Richardson, and Sloan ["Pump and dump: An empirical analysis of the relation between corporate financing activities and sell-side analyst research" 2003], Dugar and Nathan ["The effect of investment banking relationships on financial analysts’ earnings forecasts and investment recommendations" 1995] and Michaely and Womack ["Conflict of interest and the credibility of underwriter analyst recommendations" 1999] find that affiliated analyst recommendations are more favorable than unaffiliated analyst recommendations, which they interpret as consistent with the conflict of interest story. Bradley, Jordan, and Ritter (2003) find no such evidence when examining IPOs. Lin and McNichols (1988) find that short-term earnings forecasts are the same for affiliated and unaffiliated analysts, but longer-term forecasts by affiliated analysts tend to be more optimistic. Analysts tend to cover firms that they feel have good prospects (Lin and McNichols ["Underwriting relationships, analysts' earnings forecasts and investment recommendations" 1998]) Brad Barber, Reuven Lehavy, Maureen McNichols and Brett Trueman May 2001 After a string of years in which security analysts’ top stock picks significantly outperformed their pans, the year 2000 was a disaster. During that year the stocks least favorably recommended by analysts earned an annualized market-adjusted return of 48.66 percent while the stocks most highly recommended fell 31.20 percent, a return difference of almost 80 percentage points. Other papers examining the investment performance of security analysts’ stock recommendations are Barber and Loeffler (1993), Bidwell ["“How Good is Institutional Brokerage Research?" 1977], Diefenbach ["“How Good is Institutional Brokerage Research?" 1972], Dimson and Marsh ["“An Analysis of Brokers’ and Analysts’ Unpublished Forecasts of UK Stock Returns"1984], Groth, Lewellen, Schlarbaum, and Lease ["An Analysis of Brokerage House Securities Recommendations" 1979] Stickel ["The Anatomy of the Performance of Buy and Sell Recommendations" 1995], and Womack {"Do Brokerage Analysts’ Recommendations Have Investment Value?" 1996]. Paul Ryan and Richard J. Taffler Version: June 15, 2001 This paper shows that company share prices are significantly influenced by analysts recommendation changes, not only at the time of the recommendation change but also in subsequent months. The price reaction to new sell recommendations is greater than the price reaction to new buy recommendations and exhibits considerable post-recommendation drift which is consistent with initial underreaction to bad news. Ryan and Taggler directly investigate the existence of a post-recommendation announcement drift in the UK environment. They results demonstrate that analysts recommendation changes communicate valuable information to the market. Stock prices are significantly influenced, not only at the time of the recommendation change, but also in subsequent months. The price reaction to new sell recommendations is greater than the price reaction to new buy recommendations. However, these immediate price reactions appear to be incomplete showing considerable post-recommendation drift, particularly in the case of new sell recommendations. The price reaction to new sell recommendations is greater than the price reaction to new buy recommendations. We argue that this is associated with the potential costs of disseminating rather than gathering information per se. As new sell recommendations are less frequent and more visible an incorrect judgement on a sell recommendation is likely to be more costly to reputation than an incorrect buy recommendation when other analysts are likely to be making similar recommendations. Thus, if the costs of issuing a sell recommendation are greater, then the analyst's expected return for issuing these should also be greater. Post-recommendation drift is consistent with loss aversion (Shefrin, 2000) in the case of new sell recommendations. Specifically, downward drift in price continues for at least a six month period subsequent to the recommendation change. We find that the magnitude of the abnormal returns generated by new buy and sell recommendations is influenced by firm size. Smaller firms are especially pronounced in the case of new sell recommendations, in circumstances where a contemporaneous same-sign earnings forecast revision accompanies the recommendation change, and for those recommendation changes that skip a rank in the case of new buys. We also find that brokerage house reputation impacts the market for buy recommendations only, and that it takes the market some time to distinguish between those new buy recommendations that have superior investment value and those that do not. Narasimhan Jegadeesh, Joonghyuk Kim, Susan D. Krische and Charles C. Lee We find that analysts prefer high momentum stocks and growth stocks. On further analysis, we find that analyst recommendations are positively correlated with Momentum indicators but negatively correlated with Contrarian indicators. The stocks that receive more favorable recommendations typically have more positive price momentum, higher trading volume (turnover), higher past and projected growth, more positive accounting accruals, and more aggressive capital expenditures. We find that the level of the consensus analyst recommendation does not contain incremental information for the general population of stocks when it is used in conjunction with other predictive signals. For the subset of firms with favorable Momentum and Contrarian signals, we find that firms favored by analysts tend to outperform firms that are less favored. However, for the subset with less favorable quantitative signals, the stocks that analysts recommended most favorably actually underperform the stocks that they recommend less favorably. Perhaps, for this subset of firms, favorable analyst recommendations actually help delay the eventual convergence of price to the underlying fundamentals. The explanatory power of the change in the consensus analyst recommendation is more robust than that of the level of the recommendation. Our results suggest that if analysts' goal is to generate recommendations with greater predictive power for returns, they should more favorably recommend firms with lower trading volume, higher EP ratios, lower LTG and SG measures, more negative (income decreasing) accruals, and lower capital expenditures. When analyst recommendations conflict with a combined investment signal (the QScore), the QScore dominates. However, within individual QScore categories, analyst recommendations can be incrementally useful in returns prediction. Xia Chen October 2002 Do investors benefit from stock recommendationsl? Womack (1996) finds that the abnormal return in the 3-day window around recommendations is significantly positive for favorable and significantly negative for unfavorable recommendation changes. This evidence suggests that investors can better allocate their assets and earn abnormal profits based on stock recommendations. Barber, Lehavy, McNichols, and Trueman.["Can investors profit from the prophets? Security analyst recommendations and stock returns 2001] explicitly test this conjecture and, surprisingly, find that investors cannot profit from trading on stock recommendations after controlling for transaction costs. Results from intraday analyses confirm that institutional investors adjust their holdings in response to stock recommendations. There are more buyer-initiated than seller-initiated large trades (proxy for institutional tradings) around favorable recommendations, and more seller-initiated than buyer-initiated large trades around unfavorable recommendations. Using quarterly institutional holding data, we find that quarterly changes in institutional holdings are positively correlated with prevailing consensus stock recommendations.Average quarterly change in institutional ownership is 0.68% (-0.72%) for firms with favorable (unfavorable) consensus recommendations. This study contributes to the literature in several ways. First, we provide empirical evidence that institutional investors allocate their assets based upon stock recommendations, and benefit from doing so. Such evidence substantiates the demand for recommendations by institutional investors. This paper differs from Barber et al. (2001) by focusing on institutional investors and draws different conclusions for two reasons. First, Barber et al. assume that investors do not have timely access to stock recommendations. As a result, they exclude from their return calculations the abnormal returns around the release of recommendations, a substantial portion of the total abnormal returns associated with stock recommendations. However, institutional investors can capture these returns, at least partially, due to their timely access to recommendations. Second, Barber et al. deduct round trip transaction costs in their return calculations. The underlying premise is that in order to buy stocks with favorable recommendations, investors must sell other stocks to generate cash, and similarly, in order to sell stocks with unfavorable recommendations, investors must buy other stocks so as to use up the generated cash. In contrast, institutional investors likely have sufficient resources to utilize profitable opportunities without having to incur round trip transaction costs. David Dreman & Eric Lufkin, Forbes 10/21/1996 Dreman and Lufkin studied 78,695 analysts' consensus forecasts from 1973 to 1993. The odds are staggering against the investor who relies on fine-tuned earnings estimates. They estimate there is only a 1 in 170 chance that the analysts' consensus forecast will be within 5% for any 4 consecutive quarters. David Dreman and Michael A. Berry 1995 Dreman and Berry reviewed earlier research and studied 66,100 consensus estimates of Wall Street analysts from 1974 to 1991. There findings included the following. * "We demonstrate that consensus forecasts, revised as recently as two weeks prior to the end of the quarter for which the earnings forecasts were made, deviate significantly and consistently from actual earnings." * "Standardized errors are large uniformly across industries, indicating that even on a volatility-adjusted basis, analysts err indiscriminately across industries." * "...on average, large earnings surprises are the rule rather than the exception." * "A final conclusion of this study is that in spite of our earlier findings, analysts, money managers, and investors appear to ignore the industry's poor forecasting record, although it questions the viability of many important stock valuation methods." Kent L. Womack, March 1996 [summary from http://www.investorhome.com/analysts.htm] Womack documented earlier work that found little or no evidence that analysts recommendations produce abnormal returns. Using data from 1989 to 1991, and based upon a number of assumptions, he analyzed analysts recommendations (changes from extreme ratings - analysts use a variety of ranking systems) and found that the initial reaction (three days) is a 3% rise for buy recommendations and a 4.7% drop for sells. He also found incremental positive returns of 2.4% for the month following buys and a -9.1% drift over the six months following a sell (initial gains were "not mean-reverting"). Lawrence D. Brown 1996 [summary from http://www.investorhome.com/analysts.htm] Brown argued contrary to the conclusions of Dreman and Berry (see above) that (1) analysts forecasts are not "too large", (2) analysts' earnings forecasts are more accurate than naive models, (3) errors have not been increasing with time, (4) in the period following the study by Dreman and Berry, analysts have been too pessimistic (not optimistic), (5) and that the investment community places too little reliance on analysts' earnings forecasts. Mark Bagnoli and Messod Daniel Beneish Bagnoli and Beneish compared First Call analyst forecasts from 1/95 to 5/97 to 1,000 unofficial [or 'whisper'] forecasts of of quarterly earnings per share. Their analysis yields the following results. First, we find that whispers are, on average, more accurate than First Call forecasts and are better proxies for market expectations of earnings than are First Call forecasts, consistent with the claim in the professional press that whispers are increasingly becoming the true market expectation of earnings. Second, we show that trading strategies based on the relationship between whisper and First Call forecasts earn abnormal returns. Our results, when considered collectively, suggest that whispers contain information not contained in First Call analyst forecasts and that they appear to be widely enough disseminated so that at least part of this information is incorporated in stock prices prior to the earnings release. Jeffery Abarbanell & Reuven Lehavy 'Explaining Over/Underreaction in Analysts' Earnings Forecasts' 2003 Ronnie Sadka & Anna Scherbina 'Analyst Disagreement, Mispricing and Liquidity' 2004 Anna Scherbina 'Analyst Disagreement, Forecast Bias and Stock Returns' Note: stoped at [before] 'Inefficiency in Earnings Forecasts:' in doing the current list Jeffery Abarbanell & Reuven Lehavy 'Explaining Over/Underreaction in Analysts' Earnings Forecasts' 2003 A. Abdel-khalik & Ajinkya 'Returns to informational advantages: The case of analysts' forecast revisions' 1982 Dilip Abreu & Markus Brunnermeier 'Synchronization Risk and Delayed Arbitrage' 2002 Dilip Abreu & Markus Brunnermeier 'Bubbles and Crashes' 2003 Malcolm Baker & Serkan Savasoglu 'Limited Arbitrage in Mergers and Acquisitions' 2002 B. Barber, R. Lehavy, M. McNichols, & B. Trueman in 'Can Investors Profit from the Prophets?' 2001 Bradshaw, Richardson, and Sloan 'Pump and Dump: An Empirical Analysis of the Relation Between Corporate Financing Activities and Sell-side Analyst Research ' 2003 Markus Brunnermeier & Stefan Nagel 'Hedge Funds and the Technology Bubble' 2004 Zhiwu Chen, Werner Stanzl, & Masahiro Watanabe 'Price impact costs and the limit of arbitrage' 2002 Stephen Ciccone 'Does Analyst Optimism about Future Earnings Distort Stock Prices?' 2003 Cowen, Groysberg, and Healy 2003 Karl Diether, Christopher Malloy & Anna Scherbina 'Differences of opinion and the cross-section of stock returns' 2002 A. Dugar, and S. Nathan 'The Effects of Investment Banking Relationships on Financial Analysts' Earning Forecasts and Investment Recommendations' [1995] J. Easterwood & S. Nutt in 'Inefficiency in Analysts' Earnings Forecasts: Systematic Misreaction or Systematic Optimism?' 1999 W. Forbes & L. Skerratt 'Analysts' forecast revisions and stock price movements' 1992 X. Gabaix, A. Krishnamurthy, & O. Vigneron 'Limits of Arbitrage: Theory and Evidence from the Mortgage Backed Securities Market' 2004 D. Givoly & J. Lakonishok 'The information content of financial analysts' earnings forecasts' 1979 C. Gleason & C. Lee 'Characteristics of price informative analyst forecasts' 2000 N. Gonedes, N. Dopuch, & S. Penman 'Disclosure rules, information production, and capital market equilibrium' 1976 D. Gromb & Vayanos 'Equilibrium and welfare in markets with financially constrained arbitrageurs' 2002 E. Hawkins, S. Chamberlin, & W. Daniel 'Earnings expectations and security prices' 1984 Harrison Hong, Terence Lim, & Jeremy Stein 'Bad news travels slowly: size, analyst coverage, and the profitability of momentum strategies' 2000 E. Imhoff & G. Lobo 'Information content of analysts’ composite forecast revisions' 1990 Jegadeesh, Kim, Krische, and Lee 2004 Vinesh Jha, David Lichblau, & Haim Mozes in 'The Usefulness of Analysts' Recommendations' Krische and Lee 2001 Robert Korajczyk & Ronnie Sadka 'Are momentum profits robust to trading costs?' 2004 S. Krische & C. Lee in 'Analyzing the Analysts: Are Stock Recommendations Informative?' 2001 David Lesmond, Michael Schill, & Chunsheng Zhou 'The illusory nature of momentum profits' 2004 Terrence Lim 'Rationality and Analysts' Forecast Bias' 2001 H. Lin and M. McNichols 'Underwriting Relationships, Analysts' Earnings Forecasts and Investment Recommendations' 1998 M. McNichols & P. O'Brien in 'Self-Selection and Analyst Coverage' 1997 Michaely and Womack 1999 E. Miller 'Risk, uncertainty, and divergence of opinion' 1977 Mark Mitchell, Todd Pulvino, & Erik Stafford 'Limited Arbitrage in Equity Markets' 2002 Jeffrey Pontiff 'Costly Arbitrage: Evidenc from Closed-End Funds' 1996 Ronnie Sadka 'The seasonality of momentum: analysis of tradability' 2001 Ronnie Sadka & Anna Scherbina 'Analyst Disagreement, Mispricing and Liquidity' 2004 Anna Scherbina 'Analyst Disagreement, Forecast Bias and Stock Returns' 2004 S. Stickel 'Predicting individual analyst earnings forecasts' 1990 S. Stickel 'The Anatomy of the Performance of Buy and Sell Recommendations' 1995 S. Stickel 'Analysts Incentives and the Financial Characteristics of Wall Street Darlings' 2000 K. Womack 'Do Brokerage Analysts' Recommendations have Investment Value?' 1996 Wei Xiong 'Convergence Trading with Wealth Effects: An Amplification Mechanism in Financial Markets' 2001 Home Page Current Factoid Top Sites
|