Extreme Earnings Research Articles

Attention-driven reaction to extreme earnings surprises

We investigate the relationship between investor attention and stock returns in the context of extreme earnings surprises. We propose a novel mechanism that describes this interaction: high attention to very positive and very negative earnings news results in faster incorporation of information into stock prices, an overreaction effect, and a subsequent partial reversal. We test this mechanism using post-announcement abnormal returns and measure investor attention using internet search volume. We confirm that abnormal attention to earnings announcements is positively related to post-announcement abnormal returns when earnings surprises are very positive and negatively related when earnings surprises are very negative. More importantly, we argue that investors exhibit attention-driven overreactions to these extreme earnings surprises since the initial effects of abnormal attention on abnormal returns are subsequently partially reversed.

Analysts' identity negotiations and politeness behaviour in earnings calls of US firms with extreme earnings changes

PurposeFinancial analysts' roles and incentives mean that they have conflicting identities to maintain towards investors and firm managers. The authors study how analysts adopt various politeness strategies in their questioning to establish socially desirable identities in the Q&A of publicly accessible earnings calls.Design/methodology/approachThe study is based on a sample of US firms with extreme earnings changes. 46 transcripts of end-of-year earnings calls were investigated with the help of linguistic discourse analysis, drawing on frameworks of face and linguistic politeness. For each transcript, the authors identified the structure of the face-threatening acts (FTAs) that arise when analysts ask probing questions and ascertained what specific politeness strategies, if any, are used by analysts to mitigate those FTAs. The authors examine how analysts perform identities through politeness in language and compare analysts' politeness behaviour and identity construction in the increasing earnings sub-sample with the decreasing earnings sub-sample.FindingsAnalysts negotiate different identities according to specific social contexts, promoting their identity as (1) competent professionals when firms report problematic performance by asking questions in a confrontational manner with few politeness strategies and (2) dependents of the firm by asking questions in a more polite manner when firms experience satisfactory performance. Analysts aim to present a socially desirable face in Q&A to influence managers' and investors' perceptions.Practical implicationsThe study raises awareness about linguistic politeness as a communication strategy in the Q&A in earnings calls. It thereby enables managers and analysts to use linguistic politeness consciously and strategically and to recognise such use by others.Originality/valueThis study complements existing literature on earnings conference calls as part of external corporate communications by focusing on analysts' use of language when interacting with manages. To the best of our knowledge, this paper is the first to show that politeness underpins analysts' language use as a device for identity negotiations. This is important to understand because analysts' identities vis-a-vis managers and investors is closely related to the stability of the financial system.

Measuring human capital in South Africa across socioeconomic subgroups using a latent-variable approach

Human capital is a complex concept to measure given it is an unobserved latent construct. Education is a fundamental dimension of human capital and thus an education-based approach is the one most widely used. However, the international literature recommends a latent-variable approach to measuring human capital. This study thus aims to measure human capital in South Africa, a country experiencing extreme earnings and education inequalities, using a latent-variable approach and the National Income Dynamic Study (NIDS) dataset. The findings are that parental education is associated with the largest amount of variance in (latent) human capital, while the health indicator captures the least variance. Furthermore, the (latent) human capital variable provides a valuable measure to profile the distribution of human capital by socioeconomic subgroups.

Two Inconsistent Behavioral Biases

Our study explores two well-known anomalies that seem to be inconsistent with one another. Investors underreact to extreme quarterly earnings surprises resulting in returns drift that may persist for several quarters. This behavioral bias suggests that investors believe that extreme earnings surprises are mostly transitory during the earnings announcement period, but modify their beliefs later as new information arrives in subsequent quarters. In contrast, investors overreact to extreme performance of mutual funds; they pour more money into mutual funds that have recent outstanding performance and withdraw funds from mutual funds that have recently experienced extreme low performance. As these extreme patterns in performance tend to reverse in subsequent quarters, this behavior by investors suggests they erroneously consider most of the extreme performance to be permanent. We do not attempt to explain this inconsistency for retail investors but provide a potential explanation for this behavior among professional investors.

Exploiting Earnings Persistence to Better Measure Market Expectations in Detecting Post-Earnings Announcement Drift

ABSTRACT We construct an alternate proxy for earnings expectations that does not depend on analyst following and exploits differential earnings persistence across earnings levels. We examine the usefulness of our proxy in the context of post-earnings announcement drift (PEAD) where the random walk model is still widely used in the literature. We show that a refined trading strategy using our alternative proxy results in statistically significant and economically large increases in PEAD relative to the random walk model. Furthermore, using a subsample where analysts' forecasts are available, we show that the PEAD associated with our alternate proxy is incremental to the PEAD associated with analyst forecast errors. Our earnings expectation proxy can be used in other settings where extreme earnings or earnings surprise portfolios are the focus of the analysis and researchers want to include firms without analyst following.

Financial Reporting and Consumer Behavior

We show that financial reporting spurs consumer behavior. Using granular GPS data, we show that foot-traffic to firms’ commerce locations significantly increases in the days following their earnings announcements. Foot-traffic increases more for announcements with extreme earnings surprises, that correspond to firms’ fiscal year-ends, that occur outside of Fridays, and that elicit greater internet search volume, consistent with earnings announcements spurring consumer behavior by garnering attention. Consumer activity also rises with reductions in solvency risk among firms selling durable goods, consistent with consumers responding to information about firms’ longevity conveyed by their earnings. Using demographic information, we show financial reporting disproportionately affects foot-traffic in populations more likely to consume financial news. Collectively, these results suggest earnings announcements serve a marketing function by drawing attention to firms, and that a byproduct of the financial reporting process is that it shapes consumer behavior.

The monitoring role of the media: Evidence from earnings management

AbstractIn response to the recent debate on the media, this paper examines the effect of media coverage on firm earnings management. Even if prior studies (Dyck et al., 2010; Miller, 2006) have documented the media's role in detecting and deterring accounting fraud (or extreme earnings management), it is unclear ex ante whether the media amplifies or curbs less egregious earnings management. Our results show that media coverage is negatively associated with both accrual‐based and real earnings management, suggesting that the media serves as an external monitor that curbs managers’ opportunistic earnings management behaviors. Further analyses show that the effect of media coverage on earnings management is more pronounced when monitoring from auditors is weak and when the other information intermediaries are active. Overall, the findings suggest a monitoring role of the media in firm financial reporting practices.

The Speed of Earnings Anticipation: Evidence from Daily Analyst Forecasts

Prior to a firm’s earnings announcement, the capital market receives earnings-relevant information from a wide array of sources. While prior research isolates specific incentives that influence the timeliness of particular disclosures, our understanding of how these forces combine to influence the aggregate timeliness of earnings information to the market is limited. In this study, I use daily analyst forecasts to examine the arrival of earnings information to the market and assess how its timeliness varies with the direction and magnitude of the earnings news. Despite managers’ career-driven incentives to leak good news and withhold bad news, I document that the flow of earnings information to the market is significantly more timely for bad news than for good news. This asymmetry, however, is reduced for more extreme earnings news. That is, as the magnitude of the earnings news increases, good news becomes more timely whereas bad news becomes less timely. Collectively, my results are most consistent with a scenario where managers face a tradeoff between current stock price implications and future reporting reputation benefits.

Analystss Estimates of Cost of Equity Capital

We explore a large sample of analysts’ estimates of cost of equity capital (CoE) revealed in analysts’ reports to evaluate their determinants and ability to capture expected stock returns. We first document that CoE estimates are more likely to be provided by less experienced and less busy analysts and for harder-to-value firms. We also find that CoE estimates are significantly related to beta, size, book-to-market ratio, leverage and idiosyncratic volatility but not to profitability, investments or other return predictors. The CoE estimates also incrementally predict future stock returns, which possibly reflects analysts’ ability to garner information about expected returns through their direct interactions with investors. We also find that analysts increase their CoE estimates following extreme earnings surprises, indicating that companies with volatile earnings as perceived as more risky. Finally, based on a pair-wise comparison of CoE estimates with alternative expected return proxies (estimated from CAPM, Fama-French factor models or implied cost of capital models), we find that CoE estimates tend to be least noisy. We conclude that analysts’ CoE estimates, where available, are a useful proxy for expected stock returns.

(Presentation Slides) Do Individual Investors Cause Post-Earnings Announcement Drift? Direct Evidence From Personal Trades

This study tests whether naive trading by individual investors, or some class of individual investors, causes post-earnings announcement drift (PEAD). Inconsistent with the individual trading hypothesis, individual investor trading fails to subsume any of the power of extreme earnings surprises to predict future abnormal returns. Moreover, individuals are significant net buyers after both negative and positive extreme earnings surprises, consistent with an attention effect, but not with their trades causing PEAD. Finally, we find no indication that trading by individuals explains the concentration of drift at subsequent earnings announcement dates. The paper is available here: https://ssrn.com/abstract=1120495.

Testing the boundaries of the double auction: The effects of complete information and market power

We report boundary experiments testing the robustness of price convergence in double auction markets for non-durable goods in which there is extreme earnings inequality at the competitive equilibrium (CE). Following up on a conjecture by Smith (1976a), we test whether the well-known equilibrating power of the double auction institution is robust to the presence of complete information about traders’ values and costs and the presence of symmetric market power. We find that complete information is insufficient to impede convergence to CE prices; however, introducing market power consistently causes prices to deviate from the CE, whether or not subjects possess complete information. Our design highlights the value of boundary experiments in understanding how market institutions shape behavior, and our findings help delineate the limits of the double auction institution to generate competitive outcomes.

Testing the Boundaries of the Double Auction

We report boundary experiments testing the robustness of price convergence in double auction markets for non-durable goods in which there is extreme earnings inequality at the competitive equilibrium (CE). Following up on a conjecture by Smith (1980), we test whether the well-known equilibrating power of the double auction institution is robust to the presence of complete information about traders' values and costs and the presence of symmetric market power. We find that complete information is insufficient to impede convergence to CE prices; however, introducing market power consistently causes prices to deviate from the CE, whether or not subjects possess complete information. Our design highlights the value of boundary experiments in understanding how market institutions shape behavior, and our findings help delineate the limits of the double auction institution to generate competitive outcomes.

Under-or-overreaction: Market responses to announcements of earnings surprises

We test whether the well-documented market reaction to the announcements of earnings surprises is a manifestation of an investor underreaction or overreaction to extremely good or bad earnings news. Using the market reaction in the three-day period surrounding the announcements of extreme earnings surprises (i.e., SUE) in quarter Qt as a reference point, we show that firms reporting a high (low) SUE in subsequent quarter Qt+1 that confirms their initial quarter Qt SUE ranking in the same highest or lowest SUE quintiles generate an incremental price run that moves in the same direction as that of the initial SUE. However, the price impact of the confirming SUE signal is weaker than that of its initial SUE. Our findings are robust to the Fama-French three-factor daily regression extended by the momentum factor and a number of other robustness tests. Our result is not consistent with the prevalent view that investors underreact to earnings news. To the contrary, the evidence suggests an initial investor overreaction to extreme SUE signals.

PERSISTENCE, VALUE RELEVANCE, AND ACCRUALS QUALITY IN EXTREME EARNINGS AND CASH FLOW SITUATIONS

ABSTRACT Purpose: To analyze the impact of extreme earnings and cash flows on the persistence, value relevance, and accruals quality of Brazilian public firms. Originality/gap/relevance/implications: The present study contributes to the literature by analyzing extreme earnings and cash flows. Because they change the results of the period in which they occur in an unusual manner, this study suggests that studying their behaviors and impacts on the quality of the information disclosed by companies tends to contribute to optimal decision making in the capital market. Key methodological aspects: The analyzed sample comprised non-financial public companies traded on the BM&FBovespa stock exchange from 2005 to 2014, with estimates made using quantile regression. Summary of key results: The main results indicate that cash flows are more persistent than earnings because the accruals component of the latter makes them less persistent. Extreme values of both earnings and cash flows affect the persistence of these variables, with a strong and negative effect on cash flows. Finally, extreme values of accruals negatively affect accruals quality because, compared to moderate quantiles, they significantly alter the standard deviation of extreme quantiles. Key considerations/conclusions: The presence of extreme values in this study's focus variables negatively influenced market information quality, reducing its predictive power and, consequently, its relevance. Furthermore, controlling for extreme values, it was possible to observe the incremental information content of these two metrics.

The Monitoring Role of Media: Evidence from Earnings Management

In response to the recent debate on the media, this paper examines the effect of media coverage on firm earnings management. Even if prior studies (Miller 2006; Dyck et al. 2010) have documented the media’s role in detecting and deterring accounting fraud (or extreme earnings management), it is unclear ex-ante whether the media amplifies or curbs less egregious earnings management. Our results show that media coverage is negatively associated with both accrual-based and real earnings management, suggesting that the media serves as an external monitor that curbs managers’ opportunistic earnings management behaviors. Further analyses show that the effect of media coverage on earnings management is more pronounced when monitoring from auditors is weak and when the other information intermediaries are active. Overall, the findings suggest a monitoring role of the media in firm financial reporting practices.

Board sub-committees and earnings quality

The purpose of this study is to examine whether an overlap between compensation and audit committee memberships in firms affects earnings quality. We use a sample of 625 firm-year observations comprising data collected from Australian publicly listed firms between 2004 and 2010. Four measures of earnings quality are utilised to test our hypothesis. We find that firms with compensation and audit committee overlap exhibit higher earnings quality. Such sub-committee overlap is also more effective when overlapped members are independent and financial experts. Additional findings also indicate that female overlapped members are effective in reducing extreme earnings management. Findings have clear practical and regulatory implications in terms of the efficacy of sub-committee overlap.

The incremental value relevance of cash flows and earnings affected by their extremity

Purpose In contrast to earlier studies, the most recent studies on the incremental value relevance of earnings and cash flows from operations find that both earnings and cash flows have incremental value relevance beyond each other. An interesting question that follows is whether these findings hold after controlling the extremity of earnings and cash flows. This study, therefore, aims to examine the incremental value relevance of earnings and cash flows in the following four cases: moderate earnings and moderate cash flows, moderate earnings and extreme cash flows, extreme earnings and moderate cash flows and extreme earnings and extreme cash flows. Design/methodology/approach To evaluate the incremental value relevance (information content) of earnings and cash flows for each of the four cases mentioned above, we examine the statistical significance of the slope coefficients for regression of returns on both unexpected earnings and unexpected cash flows from operations. Findings The results show that (i) both moderate and extreme earnings have incremental value relevance beyond both moderate and extreme cash flows, (ii) moderate cash flows have incremental value relevance beyond both moderate and extreme earnings and (iii) extreme cash flows lack incremental value relevance beyond moderate earnings; however, they (extreme cash flows) have incremental value relevance beyond extreme earnings. These results suggest that earnings and cash flows have incremental value relevance. However, only in cases when cash flows are extreme and earnings are moderate, cash flows do not possess incremental value relevance. In further analysis, we find that the value relevance for cash flows and earnings decreases when they are extreme and transitory. Moreover, the value relevance for cash flows increases when they are moderate (not extreme) and the other competing measure (earnings) is transitory and extreme. Practical implications The results support the idea that earnings and cash flows from operations complement each other in explaining variation in returns. However, when cash flows are extreme and less informative, investors rely more on earnings in firm valuation, especially when earnings are moderate. Because earnings are unlikely to persist to be permanent across the years, these results can be interpreted as indicating that cash flows and earnings information are used jointly by investors. Originality/value In contrast to previous studies, we control for the extremity of earnings and cash flows when evaluating the incremental value relevance of earnings and cash flows from operations.

An alternative model for the analysis of detecting electronic industries earnings management using stepwise regression, random forest, and decision tree

This study attempts to diagnose the detecting electronic industries' earnings management by integrating suitable soft computing methods. Accounting earnings information is a very crucial element for corporate stakeholders to determine their stock prices and evaluate their supervision and management authority's performance, while it is also essential information for measuring corporate value. Hence, whether an enterprise can faithfully express its true economic meaning over its financial statements and how the management handles its earnings have turned out to be a popular issue widely discussed by researchers. Detecting public companies' earnings management is an important and challenging issue that has served as the impetus in many academic studies over the last few decades. Data mining technique and machine learning methods have also been commonly applied by accounting and financial personnel to other fields of studies. The study used the stepwise regression and random forest techniques to screen the variables in the first place, followed by adopting three kinds of decision trees including Chi-squared automatic interaction detector, classification and regression trees and C5.0 to establish a model and find out if the tested enterprise had extreme earnings manipulation. The results show that the proposed hybrid approach (RF+C5.0) has the optimal classification rate (the accuracy rate is 91.24 %) and the lowest occurrence of Type I error and Type II error. Also, as discovered from the rules set of the final additional testing, an enterprise's operating cash flow, times interest earned ratio and previous period's discretionary accruals play a decisive role in affecting its extreme earnings management.

The relative information content of cash flows and earnings affected by their extremity

Purpose – Many studies examine the relative information content of earnings and cash flows from operations. Most studies find that earnings have higher information content than cash flows. An interesting question that follows is whether these findings hold after controlling the extremity of earnings and cash flows. The purpose of this paper is to examine the relative information content of earnings and cash flows in the following four different cases: first, moderate earnings vs moderate cash flows, second, extreme earnings vs moderate cash flows, third, moderate earnings vs extreme cash flows, and fourth, extreme earnings vs extreme cash flows. Design/methodology/approach – To assess the relative information content of earnings and cash flows for each of the four cases mentioned above, the authors compare the explanatory power for regression of returns on unexpected earnings relative to regression of returns on unexpected cash flows. Therefore, the author compares the adjusted R2 of the model with earnings variables and the model with cash flows variables using Vuong's test, that examines the statistical significance of the difference between adjusted R2s of the rival (non-nested) models, and interpret a statistically higher adjusted R2 as an indicator for higher relative information content. Findings – The results show that: first, when both earnings and cash flows are moderate, earnings are more highly associated with stock market price changes than cash flows, second, when both earnings and cash flows are extreme, earnings also have greater relative information content than cash flows, third, when the extremity differs between earnings and cash flows, the moderate variable is superior to the other extreme variable in explaining security returns. These results suggest that earnings are definitely more value relevant than cash flows. However, only in cases when cash flows from operations are moderate and earnings are extreme, cash flows possess higher information content than earnings. Practical implications – The explanatory power for stock returns will be higher for earnings or cash flows depending on which is more highly persistent. This result reverses the conventional finding of the superiority of earnings over cash flows in explaining security returns. Originality/value – In contrast to previous studies, the authors control for the extremity of earnings and cash flows when evaluating the relative information content of earnings and cash flows from operations.

It's All Overreaction: The Post Earnings Announcement Drift

We test whether the well-documented post-earnings-announcement drift is a manifestation of an investor underreaction or overreaction to extremely good or bad earnings news. Using the market reaction to extreme earnings surprises (i.e., SUE) in quarter Qt as a reference point, we show that firms reporting a SUE in subsequent quarter Qt 1 that confirms their initial quarter Qt SUE ranking in the same highest or lowest SUE quintiles generate an incremental price run that moves in the same direction as that of the initial SUE. However, the price impact of the confirming SUE signal is weaker than that of its initial SUE. Our findings are robust to the Fama-French three-factor daily regression extended by the momentum factor and a number of other robustness tests. Our finding is not consistent with the prevalent view that investors underreact to earnings news. To the contrary, the results suggest an initial investor overreaction to extreme SUE signals.