Corporate social responsibility and financial information quality

Yan, Shan Article Corporate social responsibility and financial information quality Journal of Business and Economic Studies (JBES) Provided in Cooperation with: Northeast Business and Economics Association (NBEA) Suggested Citation: Yan, Shan (2024) : Corporate social responsibility and financial information quality, Journal of Business and Economic Studies (JBES), ISSN 2576-3458, Northeast Business and Economics Association (NBEA), Port Jefferson, NY, Vol. 28, Iss. 1, pp. 99-111 This Version is available at: https://hdl.handle.net/10419/333858 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-sa/4.0/ Corporate Social Responsibility and Financial Information Quality 99 Corporate Social Responsibility and Financial Information Quality Shan Yan Department of Finance School of Business Administration Stetson University contact: [email protected] Abstract Corporate social responsibility has gained increasing popularity among managers and investors. Many firms commit significant resources and consider CSR as a part of their long-term strategic plan. Whether corporate social responsibility affects corporate financial information quality has been an important question. Using a big sample of corporations in the U.S., I study the relationship between corporate social responsibility (CSR) and corporate financial information quality. I find a significant and positive relationship between CSR and financial information quality. This relationship is robust to potential endogeneity concerns. These findings have important implications for financial disclosure, CSR, and corporate strategy. Keywords: Corporate Social Responsibility; Financial Reporting; Information Quality Introduction In classical corporate finance theory, firm managers are responsible for the value of the owners and the maximization of the shareholders’ wealth. Since the corporation is a joint complex for many stakeholders beyond shareholders and managers, in some cases, there are conflicts of interest between the shareholders and other stakeholders of the firm, such as debt holders, employees, customers, suppliers, the community, and different levels of government. As corporations play more important roles in business, community, and society, the overall responsibility of the corporations plays an important role in decision-making inside and outside the firm. In recent decades, Corporate social responsibility (CSR) has become an important part of US firms' strategy among many corporate operations. I can find many firms have increased their investment in CSR either voluntarily or because of pressure from shareholders. As in Christensen, Hail and Leuz (2021), many firms also actively publish annual CSR reports that provide detailed information about their CSR activities and achievements or devote large sections of their annual reports to a description of their CSR activities, such as ExxonMobil, IBM, Disney, etc. The essential idea of CSR is that businesses should not focus solely on maximizing profits but also consider their other impact, such as social and environmental impact, Carroll(1999). Companies that welcome CSR integrate social and environmental concerns into their business operations and stakeholder interactions, which could come from direct expectations from the shareholders. Some would argue that CSR will benefit the firm by enhancing its reputation and brand, mitigating the risk and negative impact, attracting more talent, engagement of employees and consumer preferences. Corporate Social Responsibility and Financial Information Quality 100 In this paper, I am interested in the relationship between CSR and corporate financial information quality. Corporate financial information refers to the characteristics and attributes of financial and non-financial information produced by a corporate accounting system. High-quality financial information is reliable, accurate, relevant, comparable, understandable, and consistent to the stakeholders of the company. Under the GAAP financial rules and accounting principles, managers have the discretion to report financial information. In recent research, we have evidence of different levels of accounting information quality, which could have a different impact on the stakeholders’ decisionmaking and overall welfare. The quality of corporate financial information plays a significant role in many areas of stakeholders’ decision-making. Xing and Yan (2019) find that financial information quality is significant and negatively related to systematic risk. There is also literature showing that financial information quality is associated with firm valuation, investor confidence, market efficiency, corporate governance, and many important aspects of corporations. Previous studies have examined the view of CSR on firm financial information quality. As in Kaya and Yazan (2019), the authors argue from the conceptual part that a better CSR corporation would potentially less manipulate the financial statements and earning management, leading to better financial information quality. In this study, I use a large sample of U.S. firms by constructing two measures of CSR and two measures of financial reporting information quality. In our main regression estimation, I find that there is a significant and positive relationship between CSR and financial information quality. As in many corporate finance studies, there is a concern about the potential endogeneity problem in CSR and financial information quality. I apply two-stage regressions and adopt the instrumental variables of religious ranks and blue states to address the endogeneity concerns. Our main results are robust to potential endogeneity concerns. Our study provides empirical evidence to support that higher corporate social responsibility is associated with better financial information quality. These findings have important implications for financial disclosure, CSR, and corporate strategy. The paper proceeds as follows. I first review the recent literature on CSR and financial information quality. Following, I discuss the sample and the methodology of the measures of CSR and the financial information quality. Then, I discuss the empirical results and robustness check, after which I conclude the paper. Literature review According to Carroll (1999), the concept of corporate social responsibility has existed in the business community, as early as 1930s and 1940s. Bowen (1953) gives an initial definition of the social responsibility of businessmen, “It refers to the obligations of businessmen to pursue those policies, to make those decisions, or to follow those lines of action which are desirable in terms of the objectives and values of our society”. Davis (1960) argued that social responsibility is a nebulous idea but should be seen in a managerial context. Johnson (1971) elaborated on the topic by presenting several different definitions or views of CSR and then proceeded to critique and analyze them. Carroll (1979) and William Frederick (2006) give several versions of the definition of CSR. In Carroll’s four-part definitional framework for CSR, “Corporate social responsibility encompasses the economic, legal, ethical, and discretionary (philanthropic) expectations that Corporate Social Responsibility and Financial Information Quality 101 society has of organizations at a given point in time” (Carroll 1979, 1991). I could summarize it as Carroll’s pyramid of corporate social responsibility. More recently, CSR has been connected and studied in many areas of business, such as marketing, business ethics, strategy, accounting, and finance. Cai, Pan and Statman (2016) and Liang and Renneboog (2017) show that country characteristics seem to be important in explaining firm’s CSR activities. Iliev and Roth (2023) studied the influence of boards on CSR. Borghesi, Houston and Naranjo (2014) argue that U.S. firms with women as corporate leaders have significantly higher CSR scores. McCarthy, Oliver and Song (2017) argue that a negative relationship exists between CSR and CEO confidence. Jian and Lee (2015) argue that there is a negative association between CSR and CEO pay. In contrast, Masulis and Reza (2015) find no evidence that CEOs are lower compensated for firms with large charitable contributions. Hong and Kacperczyk (2009) show that socially constrained institutions dislike socially irresponsible stocks. Benabou and Tirole (2010) argue that firms with better CSR could have different systematic risk exposures during a crisis. Albuquerque, Koshinen, and Zhang (2019) argue strong CSR firms could result in lower systematic risk due to product differentiation strategy. Jiraporn, Jiraporn, Boeprasert and Chang (2014) find that higher CSR activities are related to favorable bond ratings. Seltzer, Starks and Zhu (2022) argue a firm’s bond rating and yield spread are related to environmental and climate risk. Chava (2014) shows that the cost of capital is higher for firms with lower environmental profiles. Flammer (2021) finds no difference in yield spread between a firm’s green bonds and other bonds. Deng, Kang and Low (2013) argue that CSR improves firm value by the event of the merger announcement study. Tang and Zhang (2020) and Flammer (2021) find issuing green bonds is associated with positive stock market reactions. Information always plays a vital role in finance as as evidenced by Fama (1960) in his efficient market hypothesis. One important way to acquire public information is through corporate financial disclosures. Dechow and Dichev (2002) suggest a measure of the quality of working capital accruals and earnings. Rajgopal and Venkatachalam (2011) argue that information quality is associated with higher idiosyncratic return volatility and risk. Kim and Zhang (2013) suggest improving financial reporting transparency is important for firms and markets to reduce the tail risks. Xing and Yan (2019) argue that financial information quality is significantly and negatively related to systematic risk. Improving financial information quality could cause systematic risk to decrease. In this study, I am interested to study the connection between corporate social responsibility and financial information quality. With the increasing popularity of CSR in management and emphasis in firms’ strategy, how CSR interacts with the quality of financial information is an important question. Will better CSR lead to better information quality, hence influence many other aspects of firms’ decisions and valuation? It is an important application of CSR and financial information quality. Our contribution is to study whether higher CSR activity firms will report higher quality in financial reports or otherwise. Particularly, I take into account potential endogeneity problems and apply two-stage regressions to alleviate the potential endogeneity problem. Data and empirical methodology I start with all available observations jointly in CRSP and COMPUSTAT. To study the key variables in CSR measures and financial information quality measures, I need to construct the CSR Corporate Social Responsibility and Financial Information Quality 102 variables and financial information quality variables, in which I will discuss the methodology of key variable construction. Our final sample is a joint sample with non-missing key variables from the CSR measures, financial information quality measures, and other controls from CRSP and COMPUSTAT. Corporate social responsibility measure Following the recent literature, I would like to construct a sample for the CSR measures. As in Deng, Kang, and Low (2013), and prior CSR literature, I obtain the firm CSR measure from seven major dimensions: community, corporate governance, diversity, employee relations, environment, human rights, and product quality and safety. Each dimension is evaluated with positive, strength, or negative, concerns. Following the prior literature, I code positive evaluations as 1, and negative evaluations as -1. The simple score, CSR_ Direct, is constructed as a simple summation of all non-missing inputs, where I consider all dimensions without any adjustment. The sample has its limitations; the strength and weakness are not measured in the same dimensions across all sample years. The sample is not constructed uniformly across all years. There are variations of dimension across each year. An alternative way to adjust the CSR measure is to consider each dimension’s size and normalized by each dimension, the adjusted CSR score will overcome variation across each year and assign equal weights to each dimension. I follow the methodology as in Deng, Kang and Low (2013) to construct the weighted CSR score based on weighted dimensions for each year and each firm. To construct the adjusted measures, I normalize each dimension and sum the total normalized dimensions for each firm, which I control the variation across different dimensions and minimize the sample difference across each year. To compare and control for the construction, I also report the simple unweighted CSR score for comparison. Financial information quality measure For the firm’s financial information quality, I follow the literature, as in Rajgopal and Venkatachalam (2011), Dechow and Dichev (2002) and Xing and Yan (2019). I use the idiosyncratic volatility, which is related to the earnings quality and information disclosure. I first model the relationship between accruals and cash flows, 𝑇𝐶𝐴𝑖𝑡 = 𝛽0+ 𝛽1𝐶𝐹𝑂𝑖(𝑡−1)+ 𝛽2𝐶𝐹𝑂𝑖𝑡 + 𝛽3𝐶𝐹𝑂𝑖(𝑡+1)+ 𝑒𝑖𝑡 where i indexes firm and t for the time. TCA is total current accruals and CFO is cash flow from operations. There is a modified version of the above equation as in Francis et al (2005), 𝑇𝐶𝐴𝑖𝑡 = 𝛽0+ 𝛽1𝐶𝐹𝑂𝑖(𝑡−1)+ 𝛽2𝐶𝐹𝑂𝑖𝑡 + 𝛽3𝐶𝐹𝑂𝑖(𝑡+1)+ 𝛽4∆𝑅𝐸𝑉𝑖𝑡 + 𝛽5𝑃𝑃𝐸𝑖𝑡 + 𝑒𝑖𝑡 where i indexes firm and t for the time. ∆𝑅𝐸𝑉𝑖𝑡 is the change in revenues and PPE is the gross value of property, plant and equipment. I follow the approach in Xing and Yan (2019), to estimate the above equation in each of the 49 industry groups, where there are a least 20 firms in a year. I define the |DD| as the standard deviation of a firm’s residuals over 5 years as our measure for financial information quality. The higher the |DD|, the lower the information quality is. Corporate Social Responsibility and Financial Information Quality 103 Also as in Xing and Yan (2019), Jones (1991), I have an alternative measure of financial information quality. 𝑇𝐴𝑖𝑡 = 𝛽0+ 𝛽1(∆𝑅𝐸𝑉𝑖𝑡 − ∆𝐴𝑅𝑖𝑡)+ 𝛽2𝑃𝑃𝐸𝑖𝑡𝑖(𝑡−1)+ 𝛽3𝑅𝑂𝐴𝑖𝑡 + 𝑒𝑖𝑡 Where TA is the total accruals, and ∆𝐴𝑅𝑖𝑡 is the change in accounting receivable. I define the absolute value of the residual from the above equation estimation as the abnormal accruals, |ABACC|. A high value of abnormal accrual means lower financial information quality. Control variables Following the literature and isolating the effects of our key variables in CSR and financial information quality, I control for variables related to the CSR and information quality. I define the variables similarly to the literature. Size is the natural logarithm of total assets. Market-to-book is the ratio of the market value of assets divided by the book value of assets. ROA is net income before extraordinary items divided by total assets. Leverage is the ratio of total liabilities to total assets. R&D is R&D expenditure scaled by total assets. Net capital expenditures are the difference between capital expenditures and sales of property, plant, and equipment divided by total assets. Business segments is the number of business segments reported in the Compustat Segment Database. Sales Herfindahl is the sum of the squared ratios of segment sales to total sales. Summary statistics Our sample consists of jointly available datasets from CRSP, COMPUSTAT, and CSR dataset. I require there are no missing key variables after constructing the measures of weighted and simple CSR, financial information quality measure of |DD| and |ABACC|. In table 1, I present the summary statistics of our merged sample. The average and median of DD measure is 0.15 and 0.06, which is in line with previous literature. The average and median of |ABACC| is 0.22 and 0.06. For the CSR measure, I present the mean and median of direct CSR measures as -0.32 and -1, which is in line with literature. I also present the mean and median of weighted CSR measures as -0.19 and -0.2. For the control variables, I report the mean, median and standard deviation of size, market to book, return on assets, leverage, research and development expenses, net capital expenses, number of segments and sales Herfindahl index. Table 1 Summary Statistics This table reports summary statistics for the general sample. Accounting information quality-DD is calculated from the modified version of the Dechow and Dichev (2002) model. Accounting information quality-ABACC is the absolute value of abnormal accruals based on the Jones (1991) model. CSR Direct is the raw score of the corporate strength less the concern. CSR Weighted is the weighted corporate score in each category. Size is the natural logarithm of total assets. Marketto-book is the ratio of the market value of assets divided by the book value of assets. ROA is net income before extraordinary items divided by total assets. Leverage is the ratio of total liabilities to total assets. R&D is R&D expenditure scaled by total assets. Net capital expenditures is the difference between capital expenditures and sales of property, plant, and equipment divided by Corporate Social Responsibility and Financial Information Quality 104 total assets. Business segments is the number of business segments reported in the Compustat Segment Database. Sales Herfindahl is the sum of the squared ratios of segment sales to total sales. variable N Mean Standard Deviation Median DD 18082 0.15 0.26 0.06 ABACC 20770 0.22 0.45 0.06 CSR Direct 20770 -0.32 2.41 -1.00 CSR Weighted 20770 -0.19 0.53 -0.20 Size 20770 7.09 1.65 6.93 Market to Book 20770 2.13 1.43 1.65 ROA 20770 0.02 0.15 0.04 Leverage 20770 0.21 0.20 0.18 R&D 20770 0.05 0.09 0.00 NET CAPX 20770 -0.46 0.35 -0.36 Num Segments 20770 1.29 0.91 1.00 Sales Herf 20770 0.95 0.15 1.00 In table 2, I present the correlation coefficient between the key variables. Our financial information quality measures, DD and ABACC have the correlation coefficient of 0.38, which is similar to literature. The coefficient between the CSR direct measure and CSR weighted measure is 0.93. Our sample is similar to the related literature. Table 2 Correlation table This table reports summary statistics for the general sample. Accounting information quality-DD is calculated from the modified version of the Dechow and Dichev (2002) model. Accounting information quality-ABACC is the absolute value of abnormal accruals based on the Jones (1991) model. CSR Direct is the raw score of the corporate strength less the concern. CSR Weighted is the weighted corporate score in each category. Size is the natural logarithm of total assets. Marketto-book is the ratio of the market value of assets divided by the book value of assets. ROA is net income before extraordinary items divided by total assets. Leverage is the ratio of total liabilities to total assets. R&D is R&D expenditures scaled by total assets. Net capital expenditures is the difference between capital expenditures and sales of property, plant, and equipment divided by total assets. Business segments is the number of business segments reported in the Compustat Segment Database. Sales Herfindahl is the sum of the squared ratios of segment sales to total sales. DD ABAC C CSR Direct CSR Weighte d Size Market to Book ROA Leverag e DD 1 ABACC 0.38 1 CSR Direct -0.02 0.02 1 CSR Weighted -0.03 0.02 0.93 1 Corporate Social Responsibility and Financial Information Quality 105 Size -0.18 -0.09 0.21 0.15 1 Market to Book 0.18 0.1 0.06 0.04 -0.25 1 ROA -0.15 -0.08 0.08 0.07 0.25 -0.02 1 Leverage -0.09 -0.04 -0.03 -0.02 0.29 -0.17 -0.13 1 Empirical results Regression of Information quality on CSR score In table 3, I estimate the relation between the financial information quality and the corporate social responsibility. The dependent variables are the financial information quality measures, DD and ABACC as previously defined. The main explanatory variables are the CSR direct and weighted measures. I also estimate a parsimonious model with main control variables. I have similar results in our extended model and parsimonious model. In our tables, I report the results in the extend model. In specification 1 and 3, I use the direct CSR measure as our key explanatory variable; in specification 2 and 4, I use the weighted CSR measure as our key explanatory variable. In all specifications, I control the size, market to book, ROA, R&D expense, Net CAPX expenses, number of segments, and sales Herfindahl index. In all scenarios, I find a significant negative relationship between CSR measure and financial information quality. Since the financial information quality measures are negatively related to the information quality, the higher the CSR measure is associated with the better financial information quality. The impact of CSR scores is also economically meaningful. For the weighted CSR score, the regression results suggest that one standard deviation in CSR measure increase ABACC information quality by 7.7%, increase DD information quality by 4.6%. Those regression results support the relationship that higher CSR scored company is associated with better financial information quality. Table 3 Regression of CSR and information quality In the table, I report the regression of the accounting quality measure on the CSR measures. The dependent variables are the DD and ABACC. Accounting information quality-DD is calculated from the modified version of the Dechow and Dichev (2002) model. Accounting information quality-ABACC is the absolute value of abnormal accruals based on the Jones (1991) model. CSR Direct is the raw score of the corporate strength less the concern. CSR Weighted is the weighted corporate score in each category. Size is the natural logarithm of total assets. Market-to-book is the ratio of the market value of assets divided by the book value of assets. ROA is net income before extraordinary items divided by total assets. Leverage is the ratio of total liabilities to total assets. R&D is R&D expenditure scaled by total assets. Net capital expenditures is the difference between capital expenditures and sales of property, plant, and equipment divided by total assets. Business segments is the number of business segments reported in the Compustat Segment Database. Sales Herfindahl is the sum of the squared ratios of segment sales to total sales. (1) (2) (3) (4) DD DD ABACC ABACC Corporate Social Responsibility and Financial Information Quality 106 CSR Direct -0.001* -0.007*** (0.001) (0.001) CSR Weighted -0.013*** -0.032*** (0.003) (0.006) Size -0.012*** -0.012*** 0.001 -0.000 (0.001) (0.001) (0.002) (0.002) Market to Book 0.014*** 0.014*** -0.007** -0.007** (0.002) (0.002) (0.003) (0.003) ROA -0.069*** -0.068*** -0.195*** -0.196*** (0.022) (0.022) (0.032) (0.032) R&D 0.547*** 0.549*** -0.177*** -0.183*** (0.043) (0.043) (0.062) (0.062) NET CAPX 0.088*** 0.087*** 0.036*** 0.036*** (0.006) (0.006) (0.010) (0.010) Num Segments 0.002 0.002 0.012 0.013 (0.006) (0.006) (0.010) (0.009) Sales Herf -0.054 -0.055 0.117** 0.116** (0.034) (0.034) (0.058) (0.058) Constant 0.273*** 0.270*** -0.142** -0.140** (0.040) (0.040) (0.069) (0.069) Observations 18,082 18,082 20,770 20,770 R-squared 0.105 0.106 0.007 0.007 Robust standard errors *** p<0.01, ** p<0.05, * <0.1 Two-stage regression Following the literature and controlling an extensive set of control variables, I could still suffer from the endogeneity bias from unobservable omitted variables. There could be factors related to CSR and improve the information quality at the same time or different direction. To address this potential endogeneity problem, I apply the 2SLS regression using an instrumental variable, blue state, as in Deng Kang and Low (2013). The idea of the 2SLS regression is to find an instrumental variable, which is correlated with the explanatory variable, rather than the dependent variable. I use the instrumental variable in the first stage regression and use the predicted regression results from the first stage regression in the second stage to solve the inconsistent and biased regression estimates. I choose blue state as our instrumental variable, which is a dummy variable that equals one if a firm has headquarter in a Democratic state and Zero otherwise. As in Rubin (2008), high CSR score firms tend to choose the blue states as their headquarters. However, I could not relate the