Journal of Business Ethics

, Volume 140, Issue 2, pp 353–367 | Cite as

Corporate Social Responsibility and Management Forecast Accuracy

Article

Abstract

This study examines the association between corporate social responsibility (CSR) and management forecast accuracy. Using data from 1995 to 2009, we find that firms provide more accurate earnings forecasts in the face of CSR activities. We also find that the positive association between CSR and management forecast accuracy is only present for the post-regulation period of 2001–2009, after the introduction of disclosure regulations intended to mitigate managers’ opportunistic behavior. These findings are consistent with the notion that managers strive to improve the quality of financial disclosure following superior CSR performance in the recent period.

Keywords

Corporate social responsibility Disclosure regulation Financial disclosure Management forecast accuracy 

Introduction

This study examines the association between corporate social responsibility (CSR) and the quality of management forecasts. Specifically, we focus on management earnings forecast errors to test whether CSR-oriented firms put more or less emphasis on the accuracy of financial disclosure primarily intended for shareholders rather than other stakeholders (e.g., communities, employees, customers, suppliers). Following prior studies (e.g., Kim et al. 2012; Siegel and Vitaliano 2007), we consider CSR firms as those that appear to demonstrate corporate social actions to meet ethical expectations of various stakeholders, based on criteria established by Kinder, Lydenberg, and Domini’s (KLD’s) Research & Analytics (2006). Specifically, Siegel and Vitaliano (2007) define CSR as managerial activity to advance a social good beyond what is required by law on behalf of stakeholders in society. Although prior research has examined the association between CSR and financial performance (Orlitzky et al. 2003; Shropshire and Hillman 2007), little is known about whether managers provide high-quality disclosure about financial performance when they engage in CSR activities. This research question is particularly important from an ethical perspective because financial disclosure is considered as a form of social responsibility (Gelb and Strawser 2001; Jo and Kim 2008).

Management earnings forecasts are voluntary disclosures about future earnings expectations. Managers have a large amount of discretion over the accuracy of their earnings forecasts (Hirst et al. 2008). If managers want to build a reputation for forthcoming disclosure, they will provide more accurate and less biased information in their voluntary disclosure (King et al. 1990). In their survey-based study, Graham et al. (2005, p. 54) report that “92.1 % of the survey respondents believe that developing a reputation for transparent reporting is the key factor motivating voluntary disclosures.” In contrast, managers might issue less accurate and more optimistic forecasts when they have incentives to disclose good earnings news for their own interests (Noe 1999; Rogers and Stocken 2005).

In line with managers’ disclosure incentives for corporate reputation versus personal gains, prior literature provides competing predictions for the association between CSR and financial reporting quality. The transparent disclosure hypothesis based on stakeholder theory (Freeman 1984) predicts that when managers engage more in CSR-related activities, they provide more accurate information in their voluntary forecasts to maintain overall corporate reputation. In contrast, the opportunistic disclosure hypothesis based on agency theory (Jensen and Meckling 1976) predicts a negative association between CSR and management forecast accuracy under the assumption that managers are rather opportunistic in financial disclosure, issuing more optimistic forecasts to portray a rosy picture of their firms’ CSR performance. Our empirical results offer strong support to the transparent disclosure hypothesis, suggesting that CSR-oriented firms adopt a forthcoming disclosure policy that mitigates information asymmetry between managers and outside stakeholders.1

To proxy for firm-specific CSR, we use the aggregate mean score of corporate social ratings, classified as strengths, in KLD data for the areas of community relations, employee relations, diversity issues, product issues, and environmental issues, consistent with prior studies (e.g., Kim et al. 2012).2 This score represents a firm’s overall strength in its relationships with primary stakeholders other than capital suppliers (Hillman and Keim 2001), and it is considered a well-established indicator of corporate social actions toward key stakeholders, such as employees, customers, and communities (Mattingly and Berman 2006).

Using data from 1995 to 2009, we estimate a firm-fixed effects regression to control for unobservable firm characteristics to the extent that they are stable over time (Garcia-Castro et al. 2010; Gormley and Matsa 2014). We find that CSR activities are, on average, positively associated with management forecast accuracy after controlling for uncertainty and known determinants identified from prior research. This finding is consistent with the transparent disclosure hypothesis that managers want to provide a more accurate view of firm performance in response to their CSR investments in order to enhance overall corporate reputation.

We also find that the positive association between CSR and management forecast accuracy is significantly strong in the post-Regulation Fair Disclosure (Reg FD) period; in contrast, we find no association between CSR and forecast accuracy in the pre-Reg FD period. In October 2000, the Securities and Exchange Commission enacted the Reg FD to provide equal access to management’s material information for a wide range of stakeholders by prohibiting selective disclosure to a few favored interested groups. In July 2002, the Sarbanes–Oxley Act (SOX) also came into effect to promote the interests of stakeholders as well as shareholders subsequent to a multitude of management wrongdoings. Both regulatory actions are designed to discipline managers’ opportunistic behavior in disclosure practices (Heflin et al. 2012) and CSR-related activities (Roush et al. 2012). Consistent with a potential shift in managerial behavior, we find that the positive relationship between CSR and management forecast accuracy is only limited to the time period (2001–2009) subsequent to the adoption of new regulations in the early 2000s.

This study is related to, but differs from, that of Dhaliwal et al. (2012), who find a positive association between analyst forecast accuracy and the issuance of stand-alone CSR reports. We examine management forecast accuracy, whereas they evaluate analyst forecast accuracy. Management forecast accuracy reveals information about whether insider managers provide more accurate earnings forecasts than outside stakeholders, including analysts. In contrast, analyst forecast accuracy concerns whether financial analysts as an outside information intermediary have more accurate assessments of future earnings presumably after they gather and process publicly available information, such as management forecasts issued by firm managers. In contrast with Dhaliwal et al. (2012), this study examines the accuracy of management earnings forecasts, which are considered the fundamental source of public information that analysts are likely to rely on when forming their opinions about the value of the firm (Williams 1996).

We also differ from Dhaliwal et al. (2012) by focusing on the importance of CSR strengths rather than the existence of CSR reports. Dhaliwal et al. (2012) use the publication of CSR reports to proxy for management disclosure of non-financial information, whereas our study employs a measure of CSR strengths to capture the level of firms’ corporate social performance. It is important to assess the strength of managers’ CSR actions rather than their exposition of CSR practices because the issuance of CSR reports in itself does not guarantee that a firm engages in strong CSR activities for stakeholders’ interests.3 In addition, Dhaliwal et al. (2011) find that firms issuing CSR reports experience a significant decrease in transaction costs only when they have superior CSR performance. This finding suggests that managers can develop corporate reputation and increase firm value only when they demonstrate the strength of CSR activities to stakeholders, not necessarily when they publish CSR reports to describe both socially responsible and irresponsible actions.

This study contributes to two broad streams of literature. First, it contributes to the accounting literature by examining the role of CSR in the quality of voluntary disclosure about future earnings. To the best of our knowledge, this is the first study to document a positive association between management forecast accuracy and CSR activities, consistent with the notion that firms with superior CSR performance also try to improve the quality of their earnings forecasts in the subsequent year.

Second, this study contributes to the CSR and business ethics literature by providing initial evidence of the significant improvement in managers’ ethical behavior regarding financial disclosure in conjunction with CSR performance since the passage of new regulations in the early 2000s. This empirical finding is in line with the argument that firm managers perceive high-quality disclosure as ethical and socially responsible behavior (Gelb and Strawser 2001; Kim et al. 2012). In addition, our study should be of interest to regulators, who tend to encourage managers to disclose more accurate information (Choi et al. 2010). In general, our finding is consistent with the regulations achieving their intended objectives of improving disclosure accuracy and promoting managers’ social responsibility toward society.

We begin by discussing related literature and developing the hypotheses. Then, we describe the data, present the research design, and explain the empirical findings. After discussing the implications of this study, we provide concluding remarks.

Motivation from Prior Research

Shropshire and Hillman (2007) note that much of the prior literature focuses on an association between CSR and financial performance. Focusing on earnings management, a few studies (e.g., Kim et al. 2012) have examined whether managers influence financial reporting quality in response to CSR activities.4 The current study differs from earnings management studies that examine managers’ potential manipulation of reported earnings in two ways. First, we focus on the quality of management disclosure about expected earnings. Unlike reported earnings, management forecasts reveal ex ante information about managers’ expectations of the value of their CSR activities. In particular, the accuracy of management earnings forecasts is useful to evaluate whether managers have strategic motives to build disclosure reputation (Hirst et al. 1999; Hutton and Stocken 2009) or to mislead external parties about financial performance (Rogers and Stocken 2005).

Second, management earnings forecasts are voluntary and unaudited in nature, whereas publicly listed firms must provide detailed and audited information about the actual realization of earnings. Thus, managers can exert more discretion over the information content of their voluntary disclosure about expected earnings performance. For example, managers do not need to provide detailed information to justify their earnings expectations (Baginski et al. 2004). Similarly, managers have a large amount of decision-making power over the quality of earnings forecasts, such as forecast accuracy (Hirst et al. 2008). Therefore, with more managerial discretion, management forecast accuracy provides a more powerful test of distinguishing between transparent disclosure and opportunistic disclosure practices.5

We further articulate the importance of distinguishing between forecast accuracy and financial performance for stronger inferences from empirical tests. Because the accuracy of management forecasts can be easily verified through actual realized earnings (Cheung et al. 2010), forecast accuracy allows us to assess whether managers have ex ante underlying motives to be transparent or opportunistic about the value of CSR investment. In other words, if managers engage in opportunistic expectations management by issuing optimistically biased earnings forecasts, ex post verification of expected earnings with actual earnings helps confirm the opportunistic nature of managerial behavior ex ante.

A different conclusion might be reached with respect to actual earnings performance. Some managers might engage in impression management to appear socially responsible without real CSR investment. If they are opportunistic in CSR, they are also likely to engage in earnings management to report higher financial performance ex post. For example, Enron was highly praised as a socially responsible company and won multiple CSR awards before its fraudulent business activities led to its collapse in 2001, negatively affecting the interests of many stakeholder groups, including its own employees ex post (Chandler and Werther 2013). In hindsight, Enron engaged in strategic impression management for CSR, and the company also manipulated its earnings to appear profitable and to mislead outside perceptions of the value of the firm. The Enron case shows that if we simply focus on realized earnings, we might come to an erroneous conclusion that a positive association between CSR and financial performance arises from managerial intention to increase the interests of both shareholders and stakeholders when, instead, managerial opportunism drives the positive empirical association.

Hypothesis Development

Traditional stakeholder theory posits that managers have ethical obligations to satisfy the interests of various stakeholder groups (e.g., communities, employees, customers, suppliers) beyond capital providers’ interests (Donaldson and Preston 1995; Freeman 1984; Garriga and Melé 2004). In particular, Garriga and Melé (2004) point out that stakeholder theory is normative in its origin since it is grounded in managers’ ethical concerns about their impact on society as well as their fiduciary duties to protect various stakeholders’ interests.

Jones (1995) incorporates information economics and further develops stakeholder theory under the assumption of information asymmetry between managers and stakeholders. He proposes that establishing a reputation for acting honestly and ethically is important because such a reputation can mitigate information asymmetry-related costs, which in turn can increase the total firm value. Consistent with this argument, Cho et al. (2013) document empirical evidence of a negative association between CSR engagement and information asymmetry in the equity market. In sum, the theory predicts that managers known for engaging in CSR-oriented activities for various stakeholders will experience greater market success while managers whose actions undermine a reputation for honesty and integrity will experience lesser market success.

Corporate reputation links stakeholder theory with a firm’s voluntary disclosure practices (Bhojraj et al. 2012). King et al. (1990) suggest that managers have economic incentives to maintain a reputation for high-quality disclosure by providing accurate and unbiased earnings forecasts. They further argue that such a reputation reduces transaction costs associated with information asymmetry in the market. Similarly, Miller and Bahnson (2002, p. 169) note that “establishing a pattern and reputation for being forthcoming, honest, and timely will boost stock prices and cut capital costs, simply because these actions will reduce uncertainty and risk for capital market participants.”

If managers engage in CSR activities to enhance corporate reputation for honest and ethical behavior on behalf of stakeholders’ interests, they are also likely to develop a reputation for transparent and accurate earnings forecasts. To the extent that both disclosure and CSR activities reflect managers’ reputation-building efforts to enhance the value of the firm, we expect a positive association between CSR and management forecast accuracy. We state the transparent disclosure hypothesis as follows:

Hypothesis 1a

(transparent disclosure hypothesis) Ceteris paribus, there is a positive association between CSR and management forecast accuracy.

The underlying assumption in agency theory literature is that managers are opportunistic by nature, pursuing their own interests in the absence of disciplining mechanisms. On the basis of this assumption, Friedman (1970) and Jensen (2001) argue that managers might invest in value-decreasing projects in the name of CSR to advance their own personal agendas. Specifically, Jensen (2001, p. 14) states that, “stakeholder theory plays into the hands of managers by allowing them to pursue their own interests at the expense of the firm’s financial claimants and society at large.” Consistent with this view, prior studies argue that managers engage in CSR-related activities to conceal the impact of their misdemeanors by imparting a “socially good” image on the firm (Hemingway and Maclagan 2004) or to insure a firm against a potential loss of reputation in the case of adverse events, such as product recalls (Minor and Morgan 2012).

Focusing on the potential manipulation of reported earnings, prior research provides some empirical support for the argument that CSR activities are negatively associated with financial reporting quality. Petrovits (2006) finds that managers strategically adjust the amount of charitable contributions to make reported earnings appear more persistent over time. Salewski and Zülch (2012) document evidence consistent with managers engaging in more earnings management when they increase investments in social actions designed to benefit stakeholders.

To the extent that managers’ opportunistic motives dominate both disclosure and stakeholder policies, we expect a negative association between CSR and the quality of earnings forecasts. Specifically, if managers invest in CSR-related activities in an opportunistic manner, they are also likely to be opportunistic in financial disclosure about future earnings performance. If so, they might issue less accurate and more biased earnings forecasts to conceal the opportunistic use of firm resources for potentially value-irrelevant projects in the name of CSR. We state the opportunistic disclosure hypothesis as follows:

Hypothesis 1b

(opportunistic disclosure hypothesis) Ceteris paribus, there is a negative association between CSR and management forecast accuracy.

In the early 2000s, U.S. policy makers enacted new regulations (i.e., Reg FD and SOX) to deter managers’ opportunistic behavior in financial reporting and corporate governance practices. Specifically, Reg FD prohibits private communications with selected investors and analysts, mitigating information asymmetry. Thus, a broader group of stakeholders has equal access to material information from management. In a similar vein, the stated objective of SOX is to improve “the accuracy and reliability of corporate disclosures” (Hamilton and Trautman 2002, p. 87). In addition, as a response to a series of high-profile accounting scandals during 2000–2001, SOX aimed to prevent financial fraud and accounting irregularities by enhancing managers’ corporate responsibility. Since the introduction of Reg FD and SOX in October 2000 and July 2002, respectively, two streams of literature have emerged to examine the effectiveness of the regulations.

First, in general, the Reg FD literature (e.g., Heflin et al. 2012) provides empirical evidence consistent with an improvement in the information environment in the post-Reg FD period. Heflin et al. (2012) find that management earnings forecasts have become less optimistic and more accurate, leading to increased credibility. Herrmann et al. (2008) document that analyst forecast optimism has also decreased following the implementation of Reg FD, perhaps due to a reduction in management forecast optimism. Canace et al. (2010) also find that Reg FD is effective in reducing managers’ tendency to influence market expectations with biased earnings forecasts. These studies suggest that the overall quality of management forecasts has improved subsequent to the enactment of Reg FD.

Second, the SOX literature examines whether SOX mitigates agency problems associated with managers’ potential mismanagement of firm resources. Li et al. (2008) find that managers have become less likely to manipulate reported earnings subsequent to SOX, leading to a potential increase in financial reporting quality. They interpret this finding consistent with a reduction in agency costs, to the extent that earnings management is associated with the presence of agency problems. Hochberg et al. (2009) document that when firms were subject to agency problems before SOX, managers were more likely to lobby against strict implementation of SOX. Leuz et al. (2008) find some support for the argument that SOX affects firms’ going-private decisions because of its negative effect on agency costs. The overall findings are consistent with the notion that SOX mitigates managers’ opportunistic behavior.

Because the adoption timing of Reg FD is close to that of SOX (in the early 2000s), it is inherently difficult to distinguish the effects of the two regulations. Thus, we assume that both Reg FD and SOX complement each other in mitigating managers’ opportunistic behavior. To the extent that the newly adopted regulations achieve their stated goals of improving both disclosure quality and managers’ corporate responsibility, we expect a significant improvement in the association between CSR and management forecast accuracy. Therefore, we propose the following hypothesis:

Hypothesis 2

Ceteris paribus, the association between CSR and management forecast accuracy is stronger in the post-Reg FD period than in the pre-Reg FD period.

Sample Selection and Measurement Timeline

Sample Selection

Table 1 outlines the sample selection procedures. We obtain corporate social ratings from the KLD Stats database.6 This database provides CSR scores for over 3,000 firms on an annual basis. Specifically, KLD database includes both strength ratings and concern ratings for five categories of community relations, employee relations, diversity issues, product issues, and environmental issues in accordance with criteria established by KLD Research & Analytics (2006). We focus on these five categories since they directly relate to the welfare of various stakeholders that can affect or be affected by the firm’s actions, consistent with prior research (e.g., Hillman and Keim, 2001; Kim et al. 2012).7 In each subcategory, KLD provides a number of positive and negative indicators for each strength and concern activity. In Table 2, we show specific subcategory items under the five strength and concern ratings used to construct our CSR and CSIR (Corporate Social Irresponsibility) measures. After evaluating KLD data, prior studies have concluded that KLD is a widely accepted and extensively used measure of corporate social performance relevant to influencing the interests of primary stakeholder groups in society (e.g., Chatterji et al. 2009; Mattingly and Berman 2006).
Table 1

Sample selection

Total firm-year observations in the KLD ratings data from 1995 to 2009

 

26,894

Less

 Insufficient observations when merging KLD with the first annual point and range forecasts for EPS in the First Call CIG (Company Issued Guidance) data

(18,104)

 

 Insufficient observations for actual EPS in the First Call Actual data

(358)

 

 Insufficient observations for accounting data in COMPUSTAT

(572)

 

 Insufficient observations for analysts’ forecast data available in the First Call Summary data set

(1,715)

 

 Insufficient observations for CRSP stock return data in obtaining stock return volatility (STDRET)

(21)

 

 Insufficient observations for quarterly earnings data in calculating earnings volatility (EARNVOL)

(17)

 

 Missing observations in obtaining other explanatory variables

(529)

 

Sample for testing hypotheses

 

5,578

Table 2

KLD subcategory items for five strength and concern ratings

Strength

Individual items

Community relations

Charitable giving, innovative giving, non-US charitable giving, support for housing, support for education, volunteer programs, and other strengths

Employee relations

Union relationship, cash profit sharing, employee involvement, retirement benefits strengths, health and safety strengths, and other strengths

Diversity issues

CEO diversity, promotion, board of directors, work/life benefits, women and minority contracting, employment of the disabled, gay and lesbian policies, and other strengths

Product issues

Quality, R&D innovation, benefits to the economically disadvantaged, and other strengths

Environmental issues

products and services, pollution prevention, recycling, clean energy, and other strengths

Concern

Individual items

Community relations

Investment controversies, negative economic impact, tax disputes, and other concerns

Employee relations

Union relationship concerns, health and safety concerns, workforce reduction, retirement benefits concerns, and other concerns

Diversity issues

Controversies, non-representation of women, and other concerns

Product issues

Product safety, marketing/contracting concerns, antitrust, and other concerns

Environmental issues

Hazardous wastes, regulatory problems, ozone-depleting chemicals, substantial emissions, agricultural chemicals, climate change, and other concerns

We merge the KLD data with the First Call CIG (Company Issued Guidance) database, which contains firm managers’ annual earnings forecasts and analyst forecast information during the fiscal years from 1995 to 2009. We begin the sample period in 1995 because of the lack of management forecast observations before 1995 in the First Call CIG database (Anilowski et al. 2007). In addition, prior to 1995, KLD data do not report the CUSIP (Committee on Uniform Security Identification Procedures) variable used to merge KLD with management forecast data. We restrict the sample to annual management forecasts of EPS denoted in U.S. dollars and eliminate earnings forecasts with missing CUSIPs, duplicate observations, and forecasts that are not comparable because of a merger or accounting change. We exclude open-ended and qualitative forecasts, focusing on point and range forecasts because we can ex post identify forecast accuracy relative to realized EPS for the quantitative forecasts. We obtain stock return data from CRSP and necessary accounting information from quarterly and annual COMPUSTAT files, requiring each firm-year observation to have sufficient data to test the hypotheses. The final sample contains 5,578 firm-year observations consisting of 1,588 unique firms.

Measurement Timeline

We measure firm-specific CSR in year t − 1 and use the first management forecasts for EPS in year t following the end of year t − 1. We assume that managers consider the impact of CSR activities on future earnings when they issue earnings forecasts for year t. Consistent with this assumption, many prior studies (e.g., Waddock and Graves 1997) have focused on the subsequent year’s earnings performance when examining the effect of CSR on future financial performance.

Research Design

Test of Hypotheses 1a and 1b

To test Hypotheses 1a and 1b, we examine whether management forecast accuracy is positively or negatively associated with CSR activities after controlling for uncertainty and other known determinants identified in prior research. Using the sample period of 1995–2009, we estimate regression model (1) with firm-fixed effects to exploit within-firm variation and to control for unobserved heterogeneity associated with time-invariant firm attributes (Garcia-Castro et al. 2010; Gormley and Matsa 2014).

Regression model (1):
$$ACCURACY = \beta _{0} + \beta _{1} CSR + \beta _{2} CSIR + \beta _{3} EARNVOL + \beta _{4} STDRET + \beta _{5} HORIZON + \beta _{6} DISP + \beta _{7} | {MF} | + \beta _{8} ANALYST + \beta _{9} SIZE + \beta _{{10}} RET + \beta _{{11}} ROA + \beta_{{12}} LOSS + \beta _{{13}} BADNEWS + \sum {\beta_{i} } {\text{Year fixed effects}} + \sum {\beta _{j} } {\text{Firm fixed effects}} + \varepsilon.$$
In estimating model (1), the main focus is the coefficient (β1) of the CSR variable that captures the aggregate strength of firm-specific CSR activities. A positive (negative) coefficient would indicate that managers increase (decrease) earnings forecast accuracy subsequent to CSR activities, consistent with Hypothesis 1a (1b).

Measurement of Main Variables

The dependent variable ACCURACY in regression model (1) captures the relative difference between managers’ earnings-per-share (EPS) forecasts for year t and the actual realization of EPS. This variable reflects whether managers issue more or less accurate forecasts subsequent to CSR activities in year t − 1. Specifically, ACCURACY is the negative of the absolute value of (Management Forecast minus Actual EPS) multiplied by 100, deflated by lagged assets per share. We define Management Forecast as either a point estimate or the mid-point of the range estimates of the firm’s annual EPS in year t. Actual EPS is the realized EPS for year t. A higher value in ACCURACY indicates a more accurate forecast relative to the actual realization of earnings. We use lagged assets per share as a deflator for this variable, consistent with prior research (e.g., Feng et al. 2009; Li et al. 2012).

To proxy for firm-specific CSR activities, we construct the CSR variable, which is the equally weighted sum of KLD’s positive screens, classified as strength ratings, in year t − 1 for five categories of the KLD strength ratings data (i.e., community relations, employee relations, diversity issues, product issues, and environmental issues), consistent with prior research (e.g., Hillman and Keim 2001; Ioannou and Serafeim 2014; Kim et al. 2012). Specifically, using the total strength score within each category for a firm, we construct the aggregate value of the mean scores across those five ratings criteria. The CSR variable represents managers’ overall CSR activities to enhance various relationships with primary stakeholder groups (communities, employees, customers, and suppliers) other than capital suppliers. These CSR activities can affect a firm’s financial performance (Orlitzky et al. 2003), depending on whether managers’ underlying motivation is to build corporate reputation or to conceal their opportunistic behavior (Cennamo et al. 2009).

Measurement of Control Variables

We construct CSIR (corporate social irresponsibility) as a control variable for a firm’s social actions that have negative implications on stakeholder interests. Specially, CSIR is the equally weighted sum of KLD’s negative screens, classified as concern ratings, in year t – 1 for five categories of community relations, employee relations, diversity issues, product issues, and environmental issues. Recent studies (e.g., Mishra and Modi 2013) highlight the importance of distinguishing between CSR strength and CSR concern because they are theoretically distinct constructs. Unlike CSR strength, CSR concern captures firms’ irresponsible social actions, which are not associated with managers’ strategic motives to enhance corporate reputation or to camouflage their opportunistic behavior (Strike et al. 2006).

We also include a variety of known determinants to control for managerial uncertainty, external demand for managers’ private information, and financial performance at the time of management forecasts. Control variables for uncertainty are earnings volatility (EARNVOL), stock return volatility (STDRET), forecast horizon (HORIZON), analysts’ forecast dispersion (DISP), and firm-specific earnings shocks (|MF|) embedded in management forecasts. We include analyst following (ANALYST) and firm size (SIZE) to capture market demand for managers’ private information. To control for financial performance, we include abnormal stock returns (RET), return on assets in the previous year (ROA), and indicator variables for loss firms (LOSS) and firms with bad earnings news (BADNEWS). Finally, we include year- and firm-fixed effects. Specific definitions of the control variables are stated as follows.

EARNVOL is the standard deviation of quarterly earnings over at least six of the eight quarters before the management forecast date, scaled by the absolute value of the mean EPS. Ajinkya et al. (2005) find a positive association between earnings volatility and firms’ tendencies to provide less accurate forecasts. We use earnings volatility as a proxy for managers’ difficulty in forecasting future earnings due to uncertain operating environments. We expect management forecast accuracy to be negatively related to earnings volatility.

STDRET is the standard deviation of daily raw returns during 180 days before the management forecast date. We require at least 100 trading days to calculate this variable. Baginski and Hassell (1997) and Chen et al. (2011) use stock return volatility to proxy for uncertain business environments. We expect management forecast accuracy to be lower when business environments become more uncertain.

HORIZON is the log value of the number of calendar days between management earnings forecast and the end of the forecasting year. When issuing earnings forecasts earlier in the year, managers are likely to face more uncertainty and provide less accurate forecasts in earnings (Ajinkya et al. 2005). Thus, we expect a decrease in forecast accuracy in response to a longer forecast horizon.

DISP is the standard deviation of security analysts’ earnings forecasts before the management forecast date, scaled by the absolute value of the mean earnings forecast. Ajinkya et al. (2005) and Chen et al. (2011) suggest that greater inter-analyst disagreement in earnings forecasts is associated with more volatile business environments, possibly leading to less accurate management forecasts. Thus, we expect a negative association between analyst’s forecast dispersion and management forecast accuracy.

|MF| is the absolute value of Management Forecast, defined as either a point estimate or the mid-point of the range estimates of the firm’s annual EPS in year t. We use this variable to capture firm-specific earnings shocks. In general, managers are likely to have more difficulty in generating more accurate earnings forecasts when they face greater earnings shocks (Baginski et al. 2002), leading to a decrease in management forecast accuracy.

ANALYST is the log value of 1 plus the number of analysts following the firm before the management forecast date. Bhushan (1989) and Baginski and Hassell (1997) use analyst following to proxy for the market demand for managers’ private information. Consistent with this idea, Lang and Lundholm (1996) find a positive association between analyst following and the market’s perception of the informativeness of management disclosure. In contrast, with respect to management earnings forecasts, Karamanou and Vafeas (2005) find a negative relationship between analyst following and forecast accuracy, and Ajinkya et al. (2005) find no significant association between analyst following and forecast accuracy, based on a different sample period. Thus, we do not predict a sign for the relationship between ANALYST and management forecast accuracy.

SIZE is the log value of total assets at the beginning of the forecasting year. Bhushan (1989) posits that firm size reflects operational complexity, suggesting a negative association between firm size and earnings forecast accuracy. In contrast, Choi et al. (2010) argue that managers in larger firms improve disclosure quality with more firm resources in hand, issuing more accurate earnings forecasts. Thus, we do not predict a sign for this variable.

RET is the buy-and-hold abnormal stock returns during the 180 days before the management forecast date. We require at least 100 trading days to calculate this variable. Lang and Lundholm (1993) find that stock performance is positively associated with analysts’ ratings of disclosure quality, whereas McNichols (1989) finds that past stock performance is negatively correlated with management forecast accuracy. Thus, we do not predict a sign for this variable.

ROA is the ROA (income before extraordinary items deflated by total assets) at the beginning of the forecasting year. Gong et al. (2011) find that the prior year’s ROA is negatively associated with management forecast accuracy in the current year. In contrast, Baik et al. (2011) find that past three-year earnings performance after industry adjustment is positively related to management forecast accuracy. Thus, we do not predict a sign for this variable.

LOSS equals 1 if the firm reports a loss at the beginning of the forecasting year and 0 otherwise. Yang (2012) predicts and finds that loss firms have less accurate earnings forecasts because managers have more difficulty in forecasting earnings for loss firms than for profit firms.

BADNEWS equals 1 if management’s forecasted EPS is less than the mean analysts’ forecasted EPS and 0 otherwise. Karamanou and Vafeas (2005) construct this variable to capture management’s greater exposure to litigation risk. They find that managers issue more accurate forecasts in the presence of bad earnings news relative to pre-existing analysts’ expectations. Thus, we predict a positive association between BADNEWS and management forecast accuracy.

Test of Hypothesis 2

Hypothesis 2 states that in the post-Reg FD period, managers are more likely to increase the accuracy of earnings forecasts in line with their CSR activities. To test this hypothesis, we construct POST, an indicator variable that equals 1 for firm-year observations in the post-Reg FD period and 0 otherwise. Then, we estimate the model (2) and examine whether the association between forecast accuracy and CSR activities is significantly stronger in the post-Reg FD period of 2001–2009 than in the previous period. Control variables are defined as in the previous section.

Regression model (2):
$$\begin{aligned} ACCURACY = \gamma _{0} + \gamma _{1} CSR + \gamma _{2} POST + \gamma _{3} CSR \times POST{\text{ }} + {\text{Control variables in model (1)}} \\ + \sum {\gamma _{i} } {\text{ Year fixed effects }}+ \sum {\gamma _{j} } {\text{ Firm fixed effects }} + {\text{ }}\nu\\ \end{aligned},$$
where POST equals 1 if the fiscal year is after 2000 in the post-Reg FD period and 0 otherwise.

In estimating model (2), the primary variable of interest is the interaction term (\(\gamma_{3}\)) between CSR and POST. This variable reflects an inter-temporal change in the overall association between CSR and forecast accuracy. Hypothesis 2 predicts a positive coefficient on CSR × POST, which indicates a stronger association between CSR and management forecast accuracy in the post-Reg FD period. Such a result would be consistent with the argument that managers increase the quality of voluntary disclosure about CSR activities subsequent to the introduction of new regulations (i.e., Reg FD and SOX) in the early 2000s.

Empirical Results

Descriptive Statistics and Correlations

Table 3 presents variable definitions in the empirical models. Table 4 shows descriptive statistics for these variables. In Table 4, we report that the mean (median) value in ACCURACY is −1.789 (−0.808) for the entire sample. Unreported statistics for signed forecast errors suggest that the majority of sample observations contain optimistically biased forecasts (i.e., forecasted EPS greater than realized EPS). This is consistent with prior studies’ findings that long-horizon annual management forecasts tend to be more optimistic (e.g., Bergman and Roychowdhury 2008; Choi and Ziebart 2004). In particular, Bergman and Roychowdhury (2008) suggest that managers provide more optimistic long-horizon forecasts to maintain optimistic earnings valuations among market participants. This is consistent with the idea that some managers might sacrifice forecast accuracy for potential benefits associated with optimistically biased forecasts (Rogers and Stocken 2005).
Table 3

Variable definitions

Variable

Calculation

ACCURACY

The negative of the absolute value of (Management Forecast minus Actual EPS) multiplied by 100, deflated by lagged assets per share. Management Forecast is defined as either a point estimate or the mid-point of the range estimates of the firm’s annual EPS in year t

CSR

The equally weighted sum of KLD positive screens in year t − 1, classified as strengths, in the areas of community relations, employee relations, diversity issues, product issues, and environmental issues

CSIR

The equally weighted sum of KLD negative screens in year t − 1, classified as concerns, in the areas of community relations, employee relations, diversity issues, product issues, and environmental issues

EARNVOL

The standard deviation of quarterly EPS over at least six of the eight quarters before the management forecast date, scaled by the absolute value of the mean EPS

STDRET

The standard deviation of daily raw returns during 180 days before the management forecast date, requiring at least 100 trading days

HORIZON

The log value of the number of calendar days between management earnings forecast and the end of the forecasting year

DISP

The standard deviation of security analysts’ earnings forecasts before the management forecast date, scaled by the absolute value of the mean earnings forecast

|MF|

The absolute value of Management Forecast, defined as either a point estimate or the mid-point of the range estimates of the firm’s annual EPS in year t

ANALYST

The log value of 1 plus the number of analysts following the firm immediately before the management forecast date

SIZE

The log value of total assets in year t − 1

RET

The buy-and-hold abnormal stock returns during 180 days before the management forecast date, requiring at least 100 trading days

ROA

Income before extraordinary items scaled by total assets in year t − 1

LOSS

1 if the firm reports a loss in year t − 1, and 0 otherwise

BADNEWS

1 if management’s forecasted EPS is less than mean analysts’ forecasted EPS, and 0 otherwise

POST

1 if the fiscal year is after 2000 in the post-Reg FD period, and 0 otherwise

Table 4

Descriptive statistics (n = 5,578)

Variable

Mean

SD

Q1

Median

Q3

ACCURACY

−1.789

2.498

−2.319

−0.808

−0.249

CSR

0.339

0.465

0.000

0.200

0.400

CSIR

0.324

0.364

0.000

0.200

0.400

EARNVOL

1.364

3.181

0.252

0.457

1.013

STDRET

0.024

0.012

0.015

0.021

0.029

HORIZON

285.998

79.778

260

321

334

DISP

0.068

0.113

0.019

0.035

0.067

|MF|

2.109

1.386

1.150

1.850

2.725

ANALYST

10.704

6.1

6

9

15

SIZE

7.829

1.679

6.586

7.718

8.975

RET

0.019

0.215

−0.115

0.003

0.134

ROA

0.056

0.070

0.026

0.054

0.090

LOSS

0.087

0.282

0

0

0

BADNEWS

0.478

0.500

0

0

1

POST

0.903

0.295

1

1

1

We winsorize all continuous variables at the 1st and 99th percentiles. We use the untransformed variables for both HORIZON and ANALYST for ease of interpretation. The logged values are included in the following regression analyses

In Table 4, the mean value of CSR is 0.339, and the mean value of CSIR is 0.324. The average firm appears to engage in both positive and negative types of CSR activities. The mean value of POST (i.e., an indicator variable for the post-Reg FD period) is 0.903. Approximately 90 % of observations stem from the post-regulation period of 2001–2009, while the remaining 10 % come from 1995–2000.8

Descriptive statistics for several control variables are worth noting. The mean (median) value in unlogged HORIZON is 285.998 (321), reflecting a relatively long time between the forecast date and the end of the forecasting year. The mean (median) value of unlogged ANALYST is 10.704 (9), suggesting that many analysts follow the sample firms. Control variables for firm performance suggest that sample firms are, on average, profitable. Average values in ROA and LOSS are 0.056 and 0.087, respectively. Approximately 8.7 % of the sample observations experience losses before the issuance of management forecasts. The mean value in the BADNEWS indicator variable is 0.478. This suggests that in slightly less than half the observations, the value of management forecasts is lower than the mean value of pre-existing analysts’ expectations of future EPS.

Table 5 provides a correlation matrix among the main variables. ACCURACY is positively correlated with CSR, consistent with a potentially positive impact of CSR on forecast accuracy. ACCURACY is also positively correlated with POST. This positive correlation might reflect prior studies’ findings that management forecast accuracy improved in the post-Reg FD period (Heflin et al., 2012). ACCURACY is negatively related to three proxies for uncertainty (i.e., STDRET, HORIZON, and DISP), consistent with prior studies (e.g., Baginski and Hassell 1997). A positive correlation between ACCURACY and BADNEWS is consistent with the argument that managers tend to provide more accurate forecasts to mitigate litigation risk when providing bad earnings news relative to market expectations (Karamanou and Vafeas 2005). Subsequent multivariate analysis is necessary for testing our hypotheses because it is important to control for correlated variables, such as uncertainty and firm size, in examining the association between CSR and management forecast accuracy.
Table 5

Correlation matrix

 

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(1) ACCURACY

1

              

(2) CSR

0.07

1

             

(3) CSIR

0.14

0.38

1

            

(4) EARNVOL

−0.01

0.03

0.01

1

           

(5) STDRET

0.16

0.14

0.10

0.14

1

          

(6) HORIZON

0.05

0.03

0.07

−0.02

−0.02

1

         

(7) DISP

0.11

0.10

−0.01

0.16

0.32

0.02

1

        

(8) |MF|

0.01

0.24

0.23

0.16

0.27

0.08

0.31

1

       

(9) ANALYST

0.02

0.35

0.22

0.07

0.08

0.14

0.11

0.24

1

      

(10) SIZE

0.26

0.54

0.48

0.05

0.29

0.08

0.18

0.51

0.51

1

     

(11) RET

0.05

−0.01

−0.02

−0.01

0.06

0.01

0.06

0.05

−0.03

−0.03

1

    

(12) ROA

0.15

0.06

0.05

0.23

0.17

0.04

0.36

0.18

0.13

0.06

0.03

1

   

(13) LOSS

0.04

0.06

−0.02

0.36

0.27

−0.02

0.42

0.24

0.07

0.12

−0.01

0.62

1

  

(14) BADNEWS

0.16

0.01

0.07

0.04

0.06

0.04

0.09

0.16

0.00

−0.01

0.14

0.07

0.06

1

 

(15) POST

0.14

0.11

0.06

0.01

0.04

0.25

0.09

0.06

0.06

0.15

−0.02

0.06

0.05

0.10

1

Amounts in bold are significant at the 0.05 level

Regression Results for Hypotheses

Table 6 presents the results from estimating the firm-fixed effects regression model (1). With firm-fixed effects, we control for unobserved time-invariant heterogeneity across firms to help mitigate correlated omitted variable bias.9 All reported tests are two-tailed.
Table 6

Regression of management forecast accuracy on CSR, POST (i.e., an indicator variable for the post-Reg FD period), their interaction term, and control variables

 

Dependent variable = ACCURACY

(1)

(2)

Estimate

SE

Estimate

SE

CSR

0.397**

(0.157)

−0.351

(0.287)

POST

  

2.707***

(0.377)

CSR × POST

  

0.752***

(0.242)

CSIR

0.0185

(0.170)

−0.00170

(0.170)

EARNVOL

−0.0277***

(0.0103)

−0.0277***

(0.0103)

STDRET

−9.471**

(4.367)

−9.240**

(4.363)

HORIZON

−0.569***

(0.0705)

−0.558***

(0.0705)

DISP

−2.223***

(0.378)

−2.186***

(0.378)

|MF|

−0.600***

(0.0421)

−0.603***

(0.0421)

ANALYST

−0.206

(0.128)

−0.207

(0.128)

SIZE

1.719***

(0.117)

1.711***

(0.116)

RET

0.0104

(0.147)

0.000216

(0.146)

ROA

0.0845

(0.831)

0.146

(0.830)

LOSS

0.263

(0.165)

0.277*

(0.164)

BADNEWS

0.159**

(0.0644)

0.155**

(0.0643)

Intercept

−12.91***

(0.923)

−12.68***

(0.917)

Year effects

Yes

 

Yes

 

Firm-fixed effects

Yes

 

Yes

 

N

5,578

 

5,578

 

Adj. R2

0.484

 

0.485

 

See Table 3 for variable definitions

p < 0.10; ** p < 0.05; *** p < 0.01 (two-tailed)

As column (1) of Table 6 shows, management earnings forecast accuracy is positively associated with CSR at a p-value less than 0.05 (β = 0.397), after controlling for uncertainty and known determinants. This finding provides support for Hypothesis 1a (transparent disclosure hypothesis), suggesting that when managers are more involved with CSR activities, they issue more accurate earnings forecasts in the subsequent year. In contrast with CSR, we do not find an association between CSIR and management forecast accuracy. This is consistent with the argument that when managers engage in socially irresponsible activities, their motives do not stem from either reputation building or strategic impression management (Strike et al. 2006).

As column (2) of Table 6 shows, the interaction term between CSR and POST is significantly positive at a p-value less than 0.01. This finding provides support for Hypothesis 2, indicating that in the post-Reg FD period, managers’ increased forecast accuracy is associated with their CSR activities. This is consistent with the argument that newly mandated regulations in the early 2000s decreased managers’ opportunistic behavior in their disclosure and investment activities (e.g., Cheng et al. 2013; Heflin et al. 2012).

Since KLD data expand its coverage over time, it is possible that newly added firms during the post-regulation period influence the inter-temporal association between CSR and management forecast accuracy. To mitigate this concern, we examine the subset of firms existing in both the pre- and post-regulation period. In unreported analysis, we continue to find a significantly positive interaction term between CSR and POST, similar to the result based on the full sample.10 This finding provides strong evidence consistent with the argument that in the face of CSR engagement, firms already existing in the pre-regulation period appear to improve the quality of management forecasts subsequent to the new disclosure regulation in the early 2000s (e.g., Heflin et al. 2012).

For the control variables, columns (1) and (2) of Table 6 show that ACCURACY is negatively associated with several proxies for managerial uncertainty (earnings volatility, stock return volatility, forecast horizon, analysts’ forecast dispersion, and firm-specific earnings shocks reflected in the absolute value of management forecasts). In addition, ACCURACY is positively related to firm size. This finding lends support to Choi et al.’s (2010) argument that larger firms tend to provide high-quality forecasts to maintain disclosure reputation. Last, we find that firms with bad earnings news (i.e., BADNEWS) provide more accurate forecasts potentially to avoid exposure to legal liability (Karamanou and Vafeas 2005).

Discussion

This study has two main objectives. The first is to document an empirical link between CSR and the quality (i.e., accuracy) of management earnings forecasts. Our analysis provides strong evidence of a positive association between CSR and management forecast accuracy. Prior research has not focused on managers’ financial disclosure regarding future earnings performance associated with CSR. This is surprising given an extensive number of studies examining (and providing mixed evidence on) the relationship between CSR and financial performance (Orlitzky et al. 2003). The interpretation of these extant studies boils down to the underlying argument regarding whether managers are inherently ethical or opportunistic in their CSR activities. Focusing on financial disclosure, the current research contributes to this debate by providing empirical evidence consistent with ethical financial reporting in support of stakeholder theory. Specifically, we document a positive association between the prior year’s CSR performance and the quality of the current year’s management forecasts.11

The second objective in this study is to investigate a potential shift in managers’ forecasting behavior related to CSR in the recent time period. We find that a strong positive association between CSR and management forecast accuracy is only present after new disclosure regulations came into effect in the early 2000s. Recently, a new stream of literature has emerged to examine the effectiveness of accounting regulations in disciplining managerial behavior (Koch et al. 2013). Related to this literature, our findings suggest that managers have enhanced the quality of earnings forecasts associated with their CSR activities in the recent period following regulatory actions. To the extent that a shift in CSR-related financial disclosure reflects the impact of regulations on managerial behavior, it appears that managers have become more conscientious of their corporate actions involving CSR and earnings forecasts over time.

This study argues that managers regard financial disclosure as a form of ethical and socially responsible behavior (Gelb and Strawser 2001). Financial disclosure has an important influence on managers’ ethical behavior because the accuracy of information disclosure is considered the first fundamental step for ethical communication in the business ethics literature (Holley 1998; Ruppel and Harrington 2000; Whitener et al. 1998). Because firm managers have private information that is not available to external parties, they can choose the precision of information disclosure to outside stakeholders (Choi et al. 2010). Unethical managers will disclose less accurate information to mislead investors about the underlying value of the firm, as demonstrated by the myriad corporate scandals in the late 1990s and early 2000s (Rockness and Rockness 2005). Against this backdrop, our study offers first evidence that socially responsible managers behave in a more ethical manner with the provision of more accurate earnings forecasts after the introduction of disclosure regulations. Because this evidence is consistent with a positive effect of recent regulations, it also has a strong policy implication that disclosure regulations likely play an important role in managers’ ethical financial reporting as well as their investment behavior.

Overall, our study is consistent with the extant view of business ethics in financial reporting from prior research. Frecka (2008) states that unethical behavior is likely to be more severe when a high level of information asymmetry exists between managers and outsiders. Because firm managers commit to high-quality CSR and forecasting policies to mitigate such information problems (Hirst et al. 2008; Jones 1995), they are less likely to engage in unethical financial reporting in the first place. Consistent with this idea, Jo and Kim (2008) find that firms with high-quality disclosure are less likely to engage in unethical behavior in the form of aggressive earnings manipulation because of an improvement in information environments. In addition, Choi and Pae (2011) find that a high level of ethical commitment has perpetuating effects on future financial reporting quality. This study extends this line of research on ethical financial reporting by focusing on the quality of voluntary disclosure rather than that of mandatory reporting examined in the prior studies.

Although this study provides some important implications on ethics in financial reporting, it is subject to at least two limitations. First, while KLD ratings are widely used and recognized as an appropriate database to construct CSR measure for a large number of firms (Chatterji et al. 2009), they are inherently qualitative in nature based on KLD researchers’ evaluations of strong or weak CSR activities. KLD data are also concentrated in relatively larger firms, which might result in a sample selection bias. To corroborate prior research findings based on KLD data, future research might use an alternative source of CSR measures, examining both small and large firms. In particular, an interesting avenue for future research might be to investigate the evolution of CSR and financial reporting practices as firm size grows over time.

Second, because of data limitation for both management forecasts and CSR, the sample is restricted to firms in the United States. This might limit the generalizability of our findings to the international setting because country-specific differences in financial reporting might play a significant role in CSR disclosure practices in general (Salewski and Zülch 2014). Future research might investigate the role of institutional differences in CSR-related financial disclosure under different reporting regimes, which will further enrich our understanding of the determinants affecting the relationship between CSR and ethical financial reporting.

Conclusion

While many prior studies examine a potential role of CSR in financial performance, they have not examined whether managers provide high-quality disclosure about firm performance when they engage in CSR activities. To our knowledge, the current research is the first to document empirical evidence of a positive association between CSR and management forecast accuracy in the post-Reg FD period. This finding is consistent with the stakeholder theory-based argument that managers’ consideration of various stakeholders’ interests in society has an important implication on their financial reporting practices.

Footnotes

  1. 1.

    Jones (1995) posits that managers have incentives to commit to ethical behavior, thus developing a reputation for acting honestly and ethically. This commitment enhances mutual trust and cooperation with various stakeholder groups, such as communities, employees, customers, and suppliers. Such a reputation for honest and ethical behavior helps mitigate transaction costs associated with information asymmetry (Cho et al. 2013; Kim et al. 2014). Consistent with stakeholder theory, recent survey evidence indicates that reputation building is the primary motivation behind CSR (Adam Friedman Associates 2012).

  2. 2.

    See also Berman et al. (1999), Hillman and Keim (2001), Coombs and Gilley (2005), Shropshire and Hillman (2007), Benson and Davidson (2010), and Benson et al. (2011).

  3. 3.

    A firm might simply explain its future plans to address current problems in its CSR report. For example, DeTienne and Lewis (2005) use the accusations made against Nike Corporation for its poor labor practices to show that the company began publishing a detailed CSR report in 2005 as part of its efforts to address public concerns. Nike initiated in-depth CSR disclosure to explain the current status of labor conditions and its commitment to increase factory workers’ interests against the backdrop of the accusations. This case illustrates that the issuance of CSR reports does not necessarily indicate that the company has strong CSR practices.

  4. 4.

    Using earnings management as a proxy for the quality of financial reporting, prior studies provide mixed evidence for the association between CSR and financial reporting quality, based on different sample periods and countries. Choi and Pae (2011) and Kim et al. (2012) find a negative relationship between earnings management and CSR-oriented activities, whereas Prior et al. (2008) and Salewski and Zülch (2014) provide some evidence for a positive relationship between earnings management and CSR activities.

  5. 5.

    Building on stakeholder theory, Cennamo et al. (2009) argue that managerial discretion is also crucial for understanding managers’ underlying intentions behind CSR engagement. In general, managers have great latitude of action in their CSR policies (Goll and Rasheed 2004; Wood 1991). When managers have a large degree of discretion in various corporate policies, such as disclosure and investment, their actions have direct implications on the value of the firm (Hambrick and Finkelstein 1987) and social value of corporate actions (Wood 1991). Thus, it is important to consider managers’ ex ante disclosure behavior regarding future earnings to gain a better understanding of whether they possess value-increasing reputation-building motives or value-decreasing opportunistic motives with respect to CSR activities.

  6. 6.

    Since 2010, KLD data is renamed and marketed as “MSCI ESG Historical Data and STATS.” We use the term KLD to make it consistent with prior research using this database.

  7. 7.

    KLD also provides corporate governance category as well as exclusionary category items (i.e., alcohol, gambling, military, firearms, nuclear power, and tobacco) which only pertain to concern ratings. This study does not use these additional categories because corporate governance is considered as a distinct construct from CSR (Kim et al. 2012), and exclusionary category items do not reflect managers’ discretionary choices in CSR (Kim et al. 2014).

  8. 8.

    Unreported analysis shows that results are qualitatively similar when the sample is limited to the 1995–2002 period to ensure roughly similar numbers of observations in the pre-Reg FD (n = 539) and post-Reg FD (n = 607) periods.

  9. 9.

    We conduct a Hausman (1978) test to check for the existence of endogeneity in our data. This test examines whether the unique error terms are correlated with the regressors in the model. The null hypothesis of the test is that unique errors are not correlated with the regressors. The Hausman test of our data strongly rejects the null hypothesis, with a Chi-square of 242.34 and a p-value less than 0.0001. This, in turn, indicates that the use of the firm-fixed effects model is more appropriate in our data than the random effects model. When the null of no endogeneity is rejected, the fixed effect estimator is still consistent from a statistical standpoint, but the random effect estimator is inconsistent. Therefore, Gormley and Matsa (2014) suggest that researchers use the consistent fixed effect estimators to obtain adequate inferences from regression analyses.

  10. 10.

    Herrmann et al. (2008) note that the analysis of a constant sample of firms existing in both pre- and post-regulation periods mitigates concerns that correlated omitted variables might affect empirical results. The disadvantage of using a constant sample is a potential loss of statistical power due to a smaller sample size after imposing survivorship. The sample size decreases from 5578 to 1599, but the coefficient on CSR × POST remains significant at the 0.01 level (β = 0.650), alleviating concerns about correlated omitted variables. When we further limit the sample to the 1995–2003 period, to have an approximately equal number of observations in the pre- and post-Reg FD periods, the sample size decreases to 837 (i.e., 426 in the pre-period and 411 in the post-period). Despite the significant reduction in sample size, the coefficient on CSR × POST continues to be significant at the 0.05 level (β = 0.586), confirming the robustness of our results.

  11. 11.

    We caution readers that the chronological ordering of events does not conclusively establish causality. It is only a necessary, but not sufficient, condition for a cause-and-effect relationship.

Notes

Acknowledgments

The author thanks an anonymous referee, Rong Huang, Marcus Burger, Sangwan Kim, and Seoyoung Lee for valuable comments. This paper also benefitted from the constructive feedback of audiences at the 2014 Management Accounting Section Midyear Meeting and the 18th IESE International Symposium on Ethics, Business, and Society. The author acknowledges the financial support provided by Desautels Faculty of Management and the Internal Social Sciences and Humanities Development Grant program at McGill University.

References

  1. Adam Friedman Associates. (2012). Corporate social responsibility: Who’s responsible? Summary available at http://www.prweb.com/releases/prweb2012/11/prweb10144917.htm.
  2. Ajinkya, B., Bhojraj, S., & Sengupta, P. (2005). The association between outside directors, institutional investors and the properties of management earnings forecasts. Journal of Accounting Research, 43, 343–375.CrossRefGoogle Scholar
  3. Anilowski, C., Feng, M., & Skinner, D. J. (2007). Does earnings guidance affect market returns? The nature and information content of aggregate earnings guidance. Journal of Accounting and Economics, 44, 36–63.CrossRefGoogle Scholar
  4. Baginski, S., & Hassell, J. (1997). Determinants of management forecast precision. The Accounting Review, 72, 303–312.Google Scholar
  5. Baginski, S., Hassell, J., & Kimbrough, M. (2002). The effect of legal environment on voluntary disclosure: Evidence from management forecasts issued in U.S. and Canadian markets. The Accounting Review, 77, 25–50.CrossRefGoogle Scholar
  6. Baginski, S., Hassell, J., & Kimbrough, M. (2004). Why do managers explain their earnings forecasts? Journal of Accounting Research, 42, 1–29.CrossRefGoogle Scholar
  7. Baik, B., Farber, D. B., & Lee, S. (2011). CEO ability and management earnings forecasts. Contemporary Accounting Research, 28, 1645–1668.CrossRefGoogle Scholar
  8. Benson, B. W., & Davidson, W. N, I. I. I. (2010). The relation between stakeholder management, firm value, and CEO compensation: A test of enlightened value maximization. Financial Management, 39, 929–964.CrossRefGoogle Scholar
  9. Benson, B. W., Davidson, W. N, I. I. I., Wang, H., & Worrell, D. L. (2011). Deviations from expected stakeholder management, firm value, and corporate governance. Financial Management, 40, 39–81.CrossRefGoogle Scholar
  10. Bergman, N., & Roychowdhury, S. (2008). Investor sentiment and corporate disclosure. Journal of Accounting Research, 46, 1057–1083.Google Scholar
  11. Berman, S. L., Wicks, A. C., Kotha, S., & Jones, T. M. (1999). Does stakeholder orientation matter? The relationship between stakeholder management models and firm financial performance. Academy of Management Journal, 42, 488–506.CrossRefGoogle Scholar
  12. Bhojraj, S., Libby, R., & Yang, H. (2012). Analyzing guidance at the firm level: The effect of reputation-building and learning-by-doing on guidance frequency and guidance properties. Working paper, Cornell University.Google Scholar
  13. Bhushan, R. (1989). Firm characteristics and analyst following. Journal of Accounting and Economics, 11, 255–274.CrossRefGoogle Scholar
  14. Canace, T., Caylor, M., Johnson, P., & Lopez, T. (2010). The effect of regulation fair disclosure on expectations management: International evidence. Journal of Accounting and Public Policy, 29, 403–423.CrossRefGoogle Scholar
  15. Cennamo, C., Berrone, P., & Gomez-Mejia, L. R. (2009). Does stakeholder management have a dark side? Journal of Business Ethics, 89, 491–507.CrossRefGoogle Scholar
  16. Chandler, D. B., & Werther, W. B, Jr. (2013). Strategic corporate social responsibility: Stakeholders, globalization, and sustainable value creation. Thousand Oaks, CA: Sage.Google Scholar
  17. Chatterji, A., Levine, D., & Toffel, T. (2009). How well do social ratings actually measure corporate social responsibility? Journal of Economics and Management Strategy, 18, 125–169.CrossRefGoogle Scholar
  18. Chen, S., Matsumoto, D., & Rajgopal, S. (2011). Is silence golden? An empirical analysis of firms that stop giving quarterly earnings guidance. Journal of Accounting and Economics, 51, 134–150.CrossRefGoogle Scholar
  19. Cheng, M., Dhaliwal, D., & Zhang, Y. (2013). Does investment efficiency improve after the disclosure of material weakness in internal control over financial reporting? Journal of Accounting and Economics, 56, 1–18.CrossRefGoogle Scholar
  20. Cheung, Y. L., Jiang, P., & Tan, W. (2010). A transparency disclosure index measuring disclosures: Chinese listed companies. Journal of Accounting and Public Policy, 29, 259–280.CrossRefGoogle Scholar
  21. Cho, S. Y., Lee, C., & Pfeiffer, R. J, Jr. (2013). Corporate social responsibility performance and information asymmetry. Journal of Accounting and Public Policy, 32, 71–83.CrossRefGoogle Scholar
  22. Choi, J., Myers, L., Zang, Y., & Ziebart, D. (2010). The roles that forecast surprise and forecast error play in determining management forecast precision. Accounting Horizons, 24, 165–188.CrossRefGoogle Scholar
  23. Choi, T. H., & Pae, J. (2011). Business ethics and financial reporting quality: Evidence from Korea. Journal of Business Ethics, 103, 403–427.CrossRefGoogle Scholar
  24. Choi, J., & Ziebart, D. (2004). Management earnings forecasts and the market’s reaction to predicted bias in the forecast. Asia-Pacific Journal of Accounting and Economics, 11, 167–192.CrossRefGoogle Scholar
  25. Coombs, J. E., & Gilley, K. M. (2005). Stakeholder management as a predictor of CEO compensation: Main effects and interactions with financial performance. Strategic Management Journal, 26, 827–840.CrossRefGoogle Scholar
  26. DeTienne, K. B., & Lewis, L. W. (2005). The pragmatic and ethical barriers to corporate social responsibility disclosure: The Nike case. Journal of Business Ethics, 60, 359–376.CrossRefGoogle Scholar
  27. Dhaliwal, D. S., Li, O. Z., Tsang, A., & Yang, Y. G. (2011). Voluntary nonfinancial disclosure and the cost of equity capital: The initiation of corporate social responsibility reporting. The Accounting Review, 86, 59–100.CrossRefGoogle Scholar
  28. Dhaliwal, D. S., Radhakrishnan, S., Tsang, A., & Yang, Y. G. (2012). Nonfinancial disclosure and analyst forecast accuracy: International evidence on corporate social responsibility disclosure. The Accounting Review, 87, 723–759.CrossRefGoogle Scholar
  29. Donaldson, T., & Preston, L. E. (1995). The stakeholder theory of the corporation: Concepts, evidence, and implications. Academy of Management Review, 20, 65–91.Google Scholar
  30. Feng, M., Li, C., & McVay, S. (2009). Internal control and management guidance. Journal of Accounting and Economics, 48, 190–209.CrossRefGoogle Scholar
  31. Frecka, T. J. (2008). Ethical issues in financial reporting: Is intentional structuring of lease contracts to avoid capitalization unethical? Journal of Business Ethics, 80, 45–59.CrossRefGoogle Scholar
  32. Freeman, R. E. (1984). Strategic management: A stakeholder approach. Boston: Pitman.Google Scholar
  33. Friedman, M. (1970). The social responsibility of business is to increase its profits. New York Times, 13 September, 122–126.Google Scholar
  34. Garcia-Castro, R., Ariño, M. A., & Canela, M. A. (2010). Does social performance really lead to financial performance? Accounting for endogeneity. Journal of Business Ethics, 92, 107–126.CrossRefGoogle Scholar
  35. Garriga, E., & Melé, D. (2004). Corporate social responsibility theories: Mapping the territory. Journal of Business Ethics, 53, 51–71.CrossRefGoogle Scholar
  36. Gelb, D. S., & Strawser, J. A. (2001). Corporate social responsibility and financial disclosures: An alternative explanation for increased disclosure. Journal of Business Ethics, 33, 1–13.CrossRefGoogle Scholar
  37. Goll, I., & Rasheed, A. A. (2004). The moderating effect of environmental munificence and dynamism on the relationship between discretionary social responsibility and firm performance. Journal of Business Ethics, 49, 41–54.CrossRefGoogle Scholar
  38. Gong, G., Li, L. Y., & Wang, J. J. (2011). Serial correlation in management earnings forecast errors. Journal of Accounting Research, 49, 677–720.CrossRefGoogle Scholar
  39. Gormley, T. A., & Matsa, D. A. (2014). Common errors: How to (and not to) control for unobserved heterogeneity. Review of Financial Studies, 27, 617–661.CrossRefGoogle Scholar
  40. Graham, J., Harvey, C., & Rajgopal, S. (2005). The economic implications of corporate financial reporting. Journal of Accounting and Economics, 40, 3–73.CrossRefGoogle Scholar
  41. Hambrick, D. C., & Finkelstein, S. (1987). Managerial discretion: A bridge between polar views on organizations. Research in Organizational Behavior, 9, 369–406.Google Scholar
  42. Hamilton, J., & Trautmann, T. (2002). Sarbanes-Oxley Act of 2002: Law and explanation. Chicago: CCH Incorporated.Google Scholar
  43. Hausman, J. A. (1978). Specification tests in econometrics. Econometrica, 46, 1251–1271.CrossRefGoogle Scholar
  44. Heflin, F., Kross, W., & Suk, I. (2012). The effect of regulation FD on the properties of management earnings forecasts. Journal of Accounting and Public Policy, 31, 161–184.CrossRefGoogle Scholar
  45. Hemingway, C., & Maclagan, P. (2004). Managers’ personal values as drivers of corporate social responsibility. Journal of Business Ethics, 50, 33–44.CrossRefGoogle Scholar
  46. Herrmann, D. R., Hope, O. K., & Thomas, W. B. (2008). International diversification and forecast optimism: The effects of Reg FD. Accounting Horizons, 22, 179–197.CrossRefGoogle Scholar
  47. Hillman, A. J., & Keim, G. D. (2001). Shareholder value, stakeholder management, and social issues: What’s the bottom line? Strategic Management Journal, 22, 125–139.CrossRefGoogle Scholar
  48. Hirst, D. E., Koonce, L., & Miller, J. (1999). The joint effect of management’s prior forecast accuracy and the form of its financial forecasts on investor judgment. Journal of Accounting Research, 37, 101–124.CrossRefGoogle Scholar
  49. Hirst, D. E., Koonce, L., & Venkataraman, S. (2008). Management earnings forecasts: A review and framework. Accounting Horizons, 22, 315–338.CrossRefGoogle Scholar
  50. Hochberg, Y. V., Sapienza, P., & Vissing-Jørgensen, A. (2009). A lobbying approach to evaluating the Sarbanes-Oxley Act of 2002. Journal of Accounting Research, 47, 519–583.CrossRefGoogle Scholar
  51. Holley, D. M. (1998). Information disclosure in sales. Journal of Business Ethics, 17, 631–641.Google Scholar
  52. Hutton, A., & Stocken, P. (2009). Prior forecasting accuracy and investor reaction to management earnings forecasts. Working paper, Boston College.Google Scholar
  53. Ioannou, I., & Serafeim, G. (2014). The impact of corporate social responsibility on investment recommendations: Analysts’ perceptions and shifting institutional logics. Strategic Management Journal, 35, 1–23.CrossRefGoogle Scholar
  54. Jensen, M. C. (2001). Value maximization, stakeholder theory, and the corporate objective function. Journal of Applied Corporate Finance, 14, 8–21.CrossRefGoogle Scholar
  55. Jensen, M., & Meckling, W. (1976). Theory of the firm: Managerial behavior, agency cost and capital structure. Journal of Financial Economics, 3, 305–360.CrossRefGoogle Scholar
  56. Jo, H., & Kim, Y. (2008). Ethics and disclosure: A study of the financial performance of firms in the seasoned equity offerings market. Journal of Business Ethics, 80, 855–878.CrossRefGoogle Scholar
  57. Jones, T. (1995). Instrumental stakeholder theory: A synthesis of ethics and economics. Academy of Management Review, 20, 404–437.Google Scholar
  58. Karamanou, I., & Vafeas, N. (2005). The association between corporate boards, audit committees, and management earnings forecasts: An empirical analysis. Journal of Accounting Research, 43, 453–486.CrossRefGoogle Scholar
  59. Kim, Y., Li, H., & Li, S. (2014). Corporate social responsibility and stock price crash risk. Journal of Banking & Finance, 43, 1–13.CrossRefGoogle Scholar
  60. Kim, Y., Park, M. S., & Wier, B. (2012). Is earnings quality associated with corporate social responsibility? The Accounting Review, 87, 761–796.CrossRefGoogle Scholar
  61. Kinder, Lydenberg, and Domini (KLD) Research & Analytics. (2006). Getting started with KLD stats and KLD’s ratings definitions. Boston: KLD Research & Analytics.Google Scholar
  62. King, R., Pownall, G., & Waymire, G. (1990). Expectation adjustment via timely management forecasts: Review, synthesis, and suggestions for future research. Journal of Accounting Literature, 9, 113–144.Google Scholar
  63. Koch, A. S., Lefanowicz, C. E., & Robinson, J. R. (2013). Regulation FD: A review and synthesis of the academic literature. Accounting Horizons, 27, 619–646.CrossRefGoogle Scholar
  64. Lang, M., & Lundholm, R. (1993). Cross-sectional determinants of analyst ratings of corporate disclosures. Journal of Accounting Research, 31, 246–271.CrossRefGoogle Scholar
  65. Lang, M., & Lundholm, R. (1996). Corporate disclosure policy and analyst behavior. The Accounting Review, 71, 467–492.Google Scholar
  66. Leuz, C., Triantis, A., & Wang, T. Y. (2008). Why do firms go dark? Causes and economic consequences of voluntary SEC deregistrations. Journal of Accounting and Economics, 45, 181–208.CrossRefGoogle Scholar
  67. Li, C., Peters, G. F., Richardson, V. J., & Weidenmier Watson, M. (2012). The consequences of information technology control weaknesses on management information systems: The case of Sarbanes–Oxley internal control reports. MIS Quarterly, 36, 179–203.Google Scholar
  68. Li, H., Pincus, M., & Rego, S. O. (2008). Market reaction to events surrounding the Sarbanes–Oxley Act of 2002 and earnings management. Journal of Law and Economics, 51, 111–134.CrossRefGoogle Scholar
  69. Mattingly, J. E., & Berman, S. (2006). Measurement of corporate social action: Discovering taxonomy in the Kinder Lydenburg Domini ratings data. Business and Society, 45, 1–27.CrossRefGoogle Scholar
  70. McNichols, M. (1989). Evidence of information asymmetries from management earnings forecasts and stock returns. The Accounting Review, 64, 1–27.Google Scholar
  71. Miller, P., & Bahnson, P. (2002). Quality financial reporting. New York: McGraw-Hill.Google Scholar
  72. Minor, D., & Morgan, J. (2012). CSR as reputation Insurance: Primum non nocere. California Management Review, 53, 40–59.CrossRefGoogle Scholar
  73. Mishra, S., & Modi, S. B. (2013). Positive and negative corporate social responsibility, financial leverage, and idiosyncratic risk. Journal of Business Ethics, 117, 431–448.CrossRefGoogle Scholar
  74. Noe, C. F. (1999). Voluntary disclosures and insider transactions. Journal of Accounting and Economics, 27, 305–326.CrossRefGoogle Scholar
  75. Orlitzky, M., Schmidt, F. L., & Rynes, S. L. (2003). Corporate social and financial performance: A meta-analysis. Organization Studies, 24, 403–441.CrossRefGoogle Scholar
  76. Petrovits, C. (2006). Corporate-sponsored foundations and earnings management. Journal of Accounting and Economics, 41, 335–361.CrossRefGoogle Scholar
  77. Prior, D., Surroca, J., & Tribo, J. (2008). Are socially responsible managers really ethical? Exploring the relationship between earnings management and corporate social responsibility. Corporate Governance, 16, 160–177.CrossRefGoogle Scholar
  78. Rockness, H., & Rockness, J. (2005). Legislated ethics: From Enron to Sarbanes–Oxley, the impact on corporate America. Journal of Business Ethics, 57, 31–54.CrossRefGoogle Scholar
  79. Rogers, J., & Stocken, P. (2005). Credibility of management forecasts. The Accounting Review, 80, 1233–1260.CrossRefGoogle Scholar
  80. Roush, P. B., Mahoney, L. S., & Thorne, L. (2012). The effects of public pressure on CSR behavior in a capital market experiencing excessive moral debt. Accounting and the Public Interest, 12, 87–105.CrossRefGoogle Scholar
  81. Ruppel, C. P., & Harrington, S. J. (2000). The relationship of communication, ethical work climate, and trust to commitment and innovation. Journal of Business Ethics, 25, 313–328.CrossRefGoogle Scholar
  82. Salewski, M., & Zülch, H. (2014). The association between corporate social responsibility (CSR) and earnings quality—Evidence from European blue chips. Working paper, HHL Leipzig Graduate School of Management.Google Scholar
  83. Shropshire, C., & Hillman, K. (2007). A longitudinal study of significant change in stakeholder management. Business and Society, 46, 63–87.CrossRefGoogle Scholar
  84. Siegel, D. S., & Vitaliano, D. F. (2007). An empirical analysis of the strategic use of corporate social responsibility. Journal of Economics & Management Strategy, 16, 773–792.CrossRefGoogle Scholar
  85. Strike, V. M., Gao, J., & Bansal, P. (2006). Being good while being bad: Social responsibility and the international diversification of US firms. Journal of International Business Studies, 37, 850–862.CrossRefGoogle Scholar
  86. Waddock, S. A., & Graves, S. B. (1997). The corporate social performance–financial performance link. Strategic Management Journal, 18, 303–319.CrossRefGoogle Scholar
  87. Whitener, E. M., Brodt, S. E., Korsgaard, M. A., & Werner, J. M. (1998). Managers as initiators of trust: An exchange relationship framework for understanding managerial trustworthy behavior. Academy of Management Review, 23, 513–530.Google Scholar
  88. Williams, P. A. (1996). The relation between a prior earnings forecast by management and analyst response to a current management forecast. The Accounting Review, 71, 103–115.Google Scholar
  89. Wood, D. J. (1991). Corporate social performance revisited. Academy of Management Review, 16, 691–718.Google Scholar
  90. Yang, H. I. (2012). Capital market consequences of managers’ voluntary disclosure styles. Journal of Accounting and Economics, 53, 167–184.CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media Dordrecht 2015

Authors and Affiliations

  1. 1.Desautels Faculty of ManagementMcGill UniversityMontrealCanada

Personalised recommendations