“Solving the choice puzzle: Financial and non-financial stakeholders preferences in corporate disclosures”

The paper delves into the relationship between accounting conservatism, valued by financial stakeholders, and corporate social performance (CSP), esteemed by non-financial stakeholders. This study assesses the potential impact of financial reporting practices, specifically accounting conservatism, on a firm’s CSP activities, which has significant implications for diverse stakeholders. Employing an accrual-based proxy for accounting conservatism and the social contribution value per share from the Shanghai Stock Exchange as a proxy for CSP, the study utilizes a sample of 25,490 year-company observations of A-share listed companies on China’s Shanghai and Shenzhen stock exchanges spanning from 2008 to 2019. Empirical findings indicate a negative correlation between accounting conservatism and CSP. The study suggests that higher levels of social performance are associated with reduced conservatism in financial reporting, indicating that firms prioritize CSP over the interests of financial stakeholders by adopting less conservative financial reporting policies. Aligned with agency theory, these results underscore that socially responsible firms are less inclined to employ accounting conservatism in reporting earnings. This study establishes a connection be-tween firms’ unconventional and less traditional activities, such as CSP, and conservative financial reporting, offering valuable insights for investors, analysts, and regulators.


INTRODUCTION
Conservatism is among the key characteristics of financial accounting which has long been considered a major indicator of quality in financial reports and has impacted accounting practices for centuries (Basu, 1997;Cheng & Kung, 2016;Watts, 2003a).Conservatism implies "the asymmetrical verification requirements for gains and losses: the greater the difference in the degree of verification required for gains versus losses, the greater the conservatism" (Watts, 2003a, p. 208).Conservatism is associated with the contracting role of accounting (Cheng & Kung, 2016) and is instrumental in lessening the information asymmetry among stakeholders and mitigating agency problem ( Guo et al., 2020), however, examined whether stakeholder orientation of firms determined through their CSR activity manifested in dissimilar levels of the accounting conservatism compared with firms where CSR is in its infancy.Is the level of accounting conservatism influenced by the Chinese firm's efforts to strengthen its bonds with stakeholders through corporate social responsibility (CSR) activity?It is the question this paper seeks to find an explanation to, and, thus, contributing to this bourgeoning strain of literature through expanding its borders into China.Thus, from the perspective of stakeholder theory, a positive association should be expected, whereas the agency perspective assumes a reverse association between a firm's enhanced commitment to stakeholders and accounting conservatism.Therefore, the interrelationship between corporate endeavor on stakeholder relations and accounting conservatism is an issue that belongs to the empirical realm and should be treated accordingly.The bulk of extant studies give credence to the argument that managers in CSR active firms have a propensity for discipline in the provision of high-quality earnings information, both related to conservatism and earnings management (Almahrog et Sokil et al., 2020).Thus, examining the relationship between CSR and accounting conservatism could lead to useful findings and guide companies in the assessment of how sincere are firms about their stakeholders' involvement in CSR activities.Despite its merits, this strain of research is currently poorly presented in peer-reviewed literature.Moreover, to the best of the authors' knowledge, this study is the first investigating this relation on Chinese data covering the last decade, the jurisdiction deemed to be the world's largest economy by GDP measured through PPP (purchasing power parity) as opposed to more conventional MER (market exchange rates) (Allison, 2020).

LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT
The extant literature put forward two competing perspectives claiming two opposed to each other outcomes.The relationship between accounting conservatism and corporate social performance, viewed from the perspectives of stakeholder theory, presupposes a positive association, because enhanced stakeholders' engagement is expected to lead to the adoption of stakeholder perspectives in the financial reporting realm as well.Alternatively, agency theory presupposes a negative association contending that CSR activities are used as a disguise for others repugnant for stakeholders' activities.Following the discussion, this paper employs those theories and arguments utilized by the extant literature to establish two competing hypotheses on the association between CSP and accounting conservatism.

'Responsible' view/stakeholder theory
Preceding studies indicate that accounting conservatism fosters the exercise of caution in the recognition of income and assets and thus, might decrease the risk that the firm's financial prospects are exaggerated.Beaver and Ryan (2000) show that the "bias" component (incorporating conservatism as part of it) is instrumental in understating the book value of equity compared to the market value of equity and leading to a lower book-to-market ratio.Comparably, Givoly and Hayn (2000) testify that the last decade of the 20 th century has witnessed the patterns with an increase in conservative financial reporting over time (p.287) and that conservative reports reduce the risk that a firm's accounting-based measure of performance considerably outstrips its cash flows from operations.
The next argument is related to the fact that a wide array of stakeholders can benefit from conservatism in financial reporting.Watts (2003aWatts ( , 2003b) ) demonstrates that accounting conservatism restrains opportunistic behavior and improves contracting efficiency within the firm.Considering that stakeholders bear substantial downside risk from overstated accounting information, "conservatism provides stakeholders with risk protection by reporting a verifiable lower bound of the firm's net assets and earnings" (Guo et al., 2020).
In line with this argument Zhang (2008) illustrates the contracting benefits of accounting conservatism in the debt contracting process by helping lenders to assess net assets and presage more timely signals of default risk.LaFond and Watts (2008) show that conservatism lessens the manager's incentives and ability to manipulate accounting numbers and so "reduces information asymmetry and the deadweight losses that information asym- Moreover, managers are believed to make a kind of cost/benefits analysis that best suits their interest: accounting conservatism or CSR (Anagnostopoulou et al., 2020;Guo et al., 2020).Indeed, both of these activities can be costly for managers because accounting conservatism limits the level of compensation, usually tied to the level of profitability of the firm, and CSR, in turn, also requires additional investments, which often are strategic and which have a time lag to take the effect, and this, in turn, does not coincide with the shortterm approach of managers (Anagnostopoulou et al., 2020).Therefore, "strong engagement in CSR may be expected to act as window-dressing in cases where the management does not wish to incur the costs entailed in accounting conservatism" (Anagnostopoulou et al., 2020).Thus, CSR is a calculated choice by the manager who selects a lesser from two evils (from his/her private point of view Considering the foregoing, to assess the firms' commitment to stakeholders by their deeds embodied in the relevant financial reporting practices this study sets to investigate the association between accounting conservatism, measured as the asymmetric timeliness of recognition of economic gains and losses and corporate social performance (CSP).
Therefore, in compliance with the discussion above, the paper develops this hypothesis: Hypothesis A: Corporate social performance is positively related to accounting conservatism.SCVPS considers a firms' responsibilities and values created for manifold stakeholders (Table 2).

Independent variable:
Accounting conservatism.This paper refers to the method of Ahmed and Duellman (2007) and Givoly and Hayn (2000) to reflect the accounting accruals in the following period.Accounting conservatism results in negative accruals, since the higher the negatives, the higher the level of conservative accounting in corporate financial reporting.Therefore, the accounting conservatism method is: , EBEXTit DEPit OCFit Accruals TA where CONACC -accounting conservatism based on the accrual-based measure of conservatism for firm i in year t; EBEXT -income before tax and extraordinary items; DEP -depreciation charge for the year; OCF -operating cash flow; and TA -total assets.
Grouping adjustment variables.This paper uses the nature of the actual controller of a company to measure the nature of its equity that is, set the nature of equity grouping dummy variable (STATE).If the actual control of the sample company is state-owned, the value of this variable is 1, otherwise, it is 0. This paper uses the Herfindahl-Hirschman Index as well to measure corporate equity concentration.The mean of HHI10 (the sum of the squares of the top ten largest shareholders' shareholding ratios) is used to group and regress the samples.
Control variables.This paper controls for various factors that may affect a firm's accounting conservatism identified by the previous literature.The definition and measurement of these control variables are shown in Appendix A2.

Model construction
This study tests the relationship between accounting conservatism and corporate social performance to estimate the following empirical model (4): In model ( 4), the independent variable 〖CSR〗_(i,t) is SCVPS of company i in year t; the explanatory variable 〖CONACC〗_(i,t) is the accounting conservatism level of company i in year t, and the remaining variables are the control variables of the model.

Descriptive statistics
Table 3 shows the descriptive statistical results of the main variables.It can be found from Table 3 that the mean and the 50th percentile of corporate social responsibility performance (CSR) are 1.280 and 0.984, respectively.The mean of accounting conservatism (CONACC) is -0.295, the 50th percentile is -0.291, which is far from the maximum of 0.426.This shows that the sample does include a small number of companies with high accounting conservatism, but for most companies, such a high level of conservatism is not the normal situation.In addition, in terms of control variables, the mean value of company size (SIZE) is 22.13, and the 50th percentile is 21.952, which shows that the size of the sample firm conforms to the normal distribution.The average value of the asset-liability ratio (LEV) is 0.439, and the 50th percentile is 0.433.

Accounting conservatism and corporate social performance
This paper uses model (3) to test the relationship between corporate social responsibility performance and accounting conservatism in the full sample of companies.The data in this paper are panel data.A reasonable estimation model was selected through the Wald test and Hausman test, and the test results support the use of the fixed-effect model.The regression results of the model are shown in column (1) of Table 5.From the regression results, the estimated coefficient of accounting conservatism (CONACC) is significantly negative (-0.049, p < 0.01), which indicates that there is a negative correlation between accounting conservatism and social responsibility performance.
In addition, the regression results in Table 5 also show that the regression results of the con- When the asset-liability ratio (LEV) is high (0.917, p < 0.01), companies tend to perform corporate social responsibility.Moreover, in case of depleting resource base, companies are curtailing their social activity: When a company reports a loss that year, the company is more inclined to fail to perform corporate social responsibility (0.106, p < 0.01), and the same is related to the case when the operating cash flow for the year is negative (-0.020, p<0.01).

The impact of the nature of equity
Columns ( 2) and (3) of Table 5 report the test results on the nature of equity.Grouped samples of the nature of equity to perform regression analysis on the model ( 3) are used to verify whether there is a difference in the negative correlation between accounting conservatism and social responsibility performance.Among them, column (2) is the regression result of the sub-sample of SOEs: the estimated coefficient of CONACC is negative at 1% and significant.Column ( 3) is the sample regression situation of non-state-owned companies.The estimated coefficient of 1% of CONACC is significantly negative, which shows that there is a significant negative correlation between AC and CSP in the sample of state-owned and non-state-owned companies.Therefore, there is no difference in the significance of the negative correlation between accounting conservatism and corporate social responsibility when the nature of equi- ty is different, and the social responsibilities of state-owned and non-state-owned companies are easily affected by the level of accounting conservatism.

Influence of ownership concentration
Columns ( 4) and ( 5) of Table 5 report the test results of the ownership concentration.The paper uses the average value of the equity concentration index HHI10 to group the samples and use the abovegrouped samples to perform regression analysis on the model (3) to verify whether there is a difference in the relationship between AC and CSP under different equity structures.Among them, column ( 4) is the sample regression in a situation where the concentration of equity is higher than the average value.The estimated coefficient of CONACC is negative and significant (-0.040, p < 0.01).This shows that the negative correlation between accounting conservatism and corporate social responsibility has a significant negative correlation in the sample companies with concentrated equity.Column ( 5) is the sample regression in a situation where the equity concentration is lower than the mean value.The estimated coefficient of CONACC is significantly negative (-0.060, p < 0.01), which shows that there is a significant negative correlation between AC and CSP in the sample companies with relatively dispersed equity.Therefore, there is no significant difference in the negative correlation between AC and CSP in the case of different equity concentrations, and corporate social responsibility is easily affected by the level of accounting conservatism.

Robustness test
This paper inspects the robustness of its main results resorting to a host of alternative specifications and covariates and report relevant results in table 8.In this subsection, this paper uses an OLS model to replace the fixed effects model in column (1) (Table 6).This study alternatively uses CSR substantivizing data that come from the website of HEXUN (CSRH) including shareholder responsibility, employee responsibility, supplier, customer and consumer responsibility, environmental responsibility, and social responsibility in column (2).The paper finds that study's basic conclusion is unchanged from the results reported in Appendix A3.Secondly, as in the previous test, this paper uses the mean group regression of HHI10, here study further uses the mean group regression of HHI1, HHI3, and HHI5 (the sum of the squares of the shareholding ratios of the first largest shareholder, the top 3 and the top 5 largest shareholders) to repeat the regression.The empirical results (see Appendix A3) indicate that the negative relationship between CSR performance and accounting conservatism is more significant in the sample group with dispersed ownership.
In addition, to exclude potential endogeneity bias, lagged regression models are used to provide more reliable inferences in Appendix A4.The independent variable lag by one period in column (1), the independent variable lag by two periods in column (2), the independent and control variables lag by one period in column (3), and the independent and control variables lag by two periods in column (4).These are consistent with previous results, suggesting that our main results are robust to the endogeneity test.

CONCLUSION
The paper investigates the link between accounting conservatism, determined by the asymmetric timeliness of recognition of economic gains and losses, and CSP (proxied by SCVPS).The goal was to appraise the magnitude to which a commitment of firms' senior management to financial stakeholders, evident by engagement in conservative reporting, is associated with the extent of responsibility and commitment towards an array of stakeholders, manifested through a CSR engagement (SCP).This study hypothesizes that CSR active companies expending their effort and putting their resources into implementation of CSR practices are more often able to match ethical expectations of society and thus, are more likely to provide more transparent financial information characterized by a higher level of conservatism.Alternatively, it is supposed that if managers engage in CSR pursuing their self-interest just trying to mask the impact of corporate misconduct, they could mislead stakeholders with opportunistic financial reporting associated with a low level of conservatism.
On the sample of 25,490 year-company of A-share listed companies in China's Shanghai and Shenzhen stock exchanges from 2008 to 2019 retrieved from the CSMAR database, the study finds that there is a negative association between accounting conservatism and social responsibility performance.Moreover, there is no significant difference in the negative association in the case of different equity concentrations as well in state and non-state-owned enterprises.The study's results are robust on several alternative specifications and testify that current and past accounting conservatism drives current and past CSP.

Investment
Management and Financial Innovations, Volume 20, Issue 4, 2023 http://dx.doi.org/10.21511/imfi.20(4).2023.34APPENDIX A1 al., 2018; Anagnostopoulou et al., 2020; Chen & Hung, 2021; Cheng & Kung, 2016; Choi et al., 2013; Guo et al., 2020; Makarenko et al., 2020; Martinez-Martinez et al., 2021; Therefore, in this regard it is worth mentioning CSR decoupling which refers to the gap between CSR disclosure and CSR performance (García-Sánchez et al., 2021; Basu, 1997; Che 2021; Tashman et al., 2019).CSR decoupling is a propensity on part of firms to engage in symbolic CSR reporting in ways that do not correspond with their actual CSR performance (Tashman et al., 2019).CSR decoupling is believed to assist managers in achieving opportunistic objectives and can harm a firm (Anagnostopoulou et al., 2020; García-Sánchez et al., 2021; Ikram et al., 2020; Tashman et al., 2019).In line with 'window-dressing' view on CSR one may assume that CSR is utilized as a cover-up of various corporate misconduct, thus "CSR is ex- Bae et al. (2021)2011;Pasko, Zhang, et al., 2021;Tirole, 2001) 2020)ncement of the interests of a firm's financial stakeholders, as this position lowers information asymmetry between insiders and outsiders, furthers aligning the interests of management and debt and equity capital providers, alleviates agency issues, assists in balancing managerial remuneration and compensation costs, and strengthens investment efficiency(Anagnostopoulou et al., 2020;Francis et al., 2013).tomasktheirself-servingpolicieswouldbelesslikely to report accounting information conservatively, thereby getting an information edge over their stakeholders(Anagnostopoulou et al., 2020;Francis et al., 2013;Guo et al., 2020).The arguments of how firm stakeholder engagement is linked to preserving shareholder wealth have gained traction in the last decades as the debate transformed from undeniable shareholder wealth supremacy to incorporating the interests of wider stakeholders.Some authors contend that agency problems between owners and managers are exacerbated in a case when managers operate for the good of stakeholders other than shareholders(Goss & Roberts, 2011;Pasko, Zhang, et al., 2021;Tirole, 2001).Thus, CSR investments could be regarded as "a costly diversion of scarce resources.If the diversion occurs at the behest of shareholders, a firm can be viewed as a delegated philanthropic agent"(Goss & Roberts, 2011,  p. 1794).Many recent studies have indeed found a loose connection between CSR and shareholder value.Lu et al. (2021), in the sample from China, found that the effects of the mandatory CSR reporting on profitability and shareholder value are negative.Bae et al. (2021)found that "pre-crisis CSR is not effective at shielding shareholder wealth from the adverse effects of a crisis, suggesting a potential disconnect between firms' CSR orientation (ratings) and actual actions".Bae et al. (2021)give valuable advice to investors to distinguish between genuine CSR and firms engaging better through cheap talk.pected to work as a form of reputation insurance, giving managers a 'license to operate' in ways that damage shareholders"(Kim et al., 2012, p. 766). ).
Leuz et al. (2003)t al., 2020;Chih et al., 2008).The dilemma managers face is 'how to choose among multiple constituencies with competing and, in some cases, conflicting interests' (Jensen, Furthermore, according toLeuz et al. (2003), those diversion activities related to the redirection of resources often manifest themselves in accounting earnings.Given that if caught self-serving insiders could be exposed to the risk of strong legal and other disciplinary actions by outside investors, insiders are incentivized to conceal the firm's genuine economic performance seeking to reduce the likelihood of outsider interference.Thus, acknowledging the line of argument put byJensen (2001)andLeuz et al. (2003), CSR may as well worsen agency problems, thereby incentivizing insiders to lower accounting conservatism in financial reporting to camouflage their rent-seeking activities from outsiders.

Table 1 .
(Zhang et al., 2020)8)dure, year-company observationsZhang et al., 2020).SCVPS uses data from the annual financial reports and can reflect the social contribution made by the firm to its various stakeholders(Zhang et al., 2020).Furthermore, SCVPS can be used by industry peers to benchmark their so-cial performance and it is entirely reliant on the performance of firms and is not affected by market sentiments(Zhang et al., 2020), although SCVPS is informative to market(Noronha et al., 2018).Moreover, SCVPS is "balanced and powerful in a way that it condenses a great deal of crucial information into a single number"(Zhang et al., 2020).The SCVPS specific calculation formula is: NP is net profit, ITE is income tax expense, BTS is business tax and surcharges, CPE is cash paid to and for employees, EC is employee compensation payable, ESP is employee salary payable in the previous year, FE is financial expenses, DONA is donation, NS End is the number of shares at the end of this year, and NS Start is the number of shares at the start of this year.

Table 2 .
Linkage between components of SCVPS and components of stakeholder theory Source: Zhang et al. (2020).