CSR disclosure and state ownership: implications for earnings management and market value

Abstract

Purpose

This paper explores the relationship between earnings management and firms' value through the moderating effect of the missing elements – corporate social responsibility (CSR) disclosure and state ownership in Russian companies. The main argument of the paper is that CSR disclosure can be used as a mitigating mechanism to weaken the negative relationship between earnings manipulation and market value. Additionally test whether state ownership is an important moderating factor in this relationship are conducted as state has always played an important role in the emerging Russian market.

Design/methodology/approach

The hypotheses are tested on panel data for 223 publicly listed Russian firms for the period 2012–2018. A number of robustness tests are used to check the obtained results for consistency. Following previous research GMM method is employed to address endogeneity concerns.

Findings

Supported by stakeholder theory, it is observed that firms that disclosed more CSR information experience a weaker negative relationship between earnings management and market value because investors and other stakeholders positively evaluate a positive CSR image. This negative effect of earnings management on market value is even weaker for state-owned companies as market participants appreciate involvement of state-owned companies in CSR activities and place greater expectations on these firms to be responsible without clear understanding whether these actions are “window dressing” for this type of companies or not.

Originality/value

The study results provide new insights into the relation between earnings management, firm's value, CSR disclosure and state ownership in emerging-market firms. The paper highlight the importance of considering country-specific factors, such as state ownership, while analysing the market reaction on CSR disclosure and earnings management since the institutional peculiarities may help to explain differences in the obtained results.

Citation

Garanina, T. (2024), "CSR disclosure and state ownership: implications for earnings management and market value", Journal of Accounting in Emerging Economies, Vol. 14 No. 3, pp. 513-547. https://doi.org/10.1108/JAEE-06-2022-0175

1. Introduction

Earnings management refers to the efforts of firms' managers to report the desired levels of earnings. Companies may manipulate earnings to influence stock prices (Cheng et al., 2015; McAnally et al., 2008), build “cookie jar” reserves (Fogel-Yaari and Ronen, 2020) or meet or exceed analysts' forecasts (Athanasakou et al., 2009; Kury, 2007; Walker, 2013). Despite relatively long research interest in the effects of earnings management on firms' value, the empirical evidence remains far from conclusive.

Some papers have reported that earnings management negatively affects firms' value, concluding that the market perceives managers' opportunistic behaviour negatively when they manipulate earnings to take advantage of compensation plans (Athanasakou et al., 2009; Walker, 2013) or reduce the share price to obtain more benefits from employee stock option plans (Burgstahler and Dichev, 1997; Degeorge et al., 1999). Nevertheless, other researchers have argued for an opposite effect of earnings management on market valuation, which may occur if firms' managers apply specific accounting policies and methods to communicate a firm's specific information (Arya et al., 2003) and convey their expectations of the firm's future flows (Siregar and Utama, 2008). The market evaluates such activities positively because they decrease information asymmetry and clarify the peculiarities of diverse and complex environments (Arya et al., 2003).

Due to a lack of clear understanding about the effect of earnings management on firms' value, some authors call for further investigation of the issue (e.g. Chen et al., 2018; Jiraporn et al., 2008; Hasnan et al., 2013; Mohmed et al., 2019; Song and Rimmel, 2021; Tang and Chang, 2015). In this paper, we investigate the influence of earnings management on firms' market value in the emerging Russian market taking into consideration a company's involvement in CSR disclosure. We also assert that the moderating effect of CSR disclosure differs for state-owned enterprises (SOEs) and companies without state ownership.

The importance that stakeholders place on CSR has dramatically increased in recent decades, no longer being a concern for only large corporations but a necessity for all businesses and companies (Diab and Metwally, 2020; Grant Thornton International, 2008; Kim et al., 2012). Companies disclose information about social and environmental issues in their financial and non-financial reports to foster legitimacy, trust and transparency (Gamerschlag et al., 2011; Choa et al., 2012; Gao and Zhang, 2015; Martínez-Ferrero et al., 2016). Researchers view CSR disclosure as a useful tool for concealing managers' opportunistic behaviour, which may improve firms' value. Cespa and Cestone (2007) demonstrate that when managers engage in earnings manipulation, they implement CSR activities to protect against detrimental reactions from stakeholders. At the same time, firms involved in CSR actions behave in a more responsible way and constrain their earning management, providing more reliable information to the investors (Kim et al., 2012) that improves their value (El Ghoul et al., 2011).

Therefore, the aim of our study is to analyse whether earnings management is related to a firm's value when more information from other dimensions (e.g. CSR disclosure) is available to the market. Building on the results of previous research that claim that firms with higher involvement in CSR activities are more ethical in their behaviour (Gao and Zhang, 2015; Kim et al., 2012), we assume market evaluates more positively those companies that are involved in CSR activities disregarding the signs of existing earnings management. We extend previous literature investigating this research field (e.g. Kim et al., 2012; Gao and Zhang, 2015; Martínez-Ferrero et al., 2016) by focusing on the emerging Russian market and taking into consideration the institutional specifics of the country where state ownership plays an important role in managing companies' activities. Even though in some developed countries CSR is regulated by law, in the developing countries, it is yet not mandatory with some exceptions, e.g. in South Africa and India (see Barth et al., 2017; Manchiraju and Rajgopal, 2017 for details). CSR disclosure is not obligatory in Russia; thus, companies only voluntarily reveal information about their CSR activities. They often do this non-strategically and non-systematically, on an ad hoc basis, which is rather typical for markets with weak institutional environments (Anas et al., 2015; Boubakri et al., 2021; Mohmed et al., 2019). Such companies' CSR disclosure is often driven by the desire to mitigate existing risks, build a trustworthy image and give positive signals to critical stakeholders (Boubakri et al., 2021; Garanina and Aray, 2021). Only lately Russian companies started to pay more attention to the role of CSR disclosure as an important factor of getting legitimacy on the national and international markets and as a factor of building a transparent and responsible image (Aray et al., 2021; Garanina and Aray, 2021; Garanina and Kim, 2023). Therefore, a focus on Russian firms' reporting practices and their link to CSR disclosure and firms' valuation is likely to provide new insights due to the country's specific institutional context (Chen et al., 2018; Diab and Metwally, 2020; Garanina and Aray, 2021; Wang et al., 2018).

We focus on Russia as one of the largest emerging economies, located in Northern Eurasia (Bhatia and Makkar, 2019) with a population of 143.49m, GDP of 1.775tn USD, GDP per capita of 12,194.8 USD, population density 9 and inflation at 11.9% (World Bank data, 2021). Russia's economic growth is mainly driven by revenues from natural resources. Even though Russia is the largest exporter of natural gas and second largest exporter of petroleum, which makes it an important player in the global market, the economic development of the country has been largely volatile leading to high income inequality (Novokmet et al., 2018). Russian economy is characterised by a high level of corruption that is evidenced by the 136 rank of Russia out of 178 countries in the Transparency International Corruption Perception Index (Transparancy International, 2021), being ranked the highest in corruption relative to its peers among BRICS nations. The economic growth based on the use of natural resources comes at the cost of environmental damage and burden on the environment (Bhatia and Makkar, 2019; Fifka and Pobizhan, 2014). The country was named as one of the leading oil polluters in the world (Schwartz, 2012), and according to the Blacksmith Institute, three of the world's ten dirtiest cities are in Russia (Blacksmith Inst., 2006). According to the Emissions Gap Report (2022), Russia is ranked #2 for the indicator of GHG emissions per capita and kept this place since 1990. The development of economy and business at the expense of welfare is not good, and some companies started to undertake social service as part of their vision. That is why we think that analysing the role of CSR disclosure by Russian companies is extremely interesting and adds value to previous research.

Moreover, Russia is also characterised by high governmental intervention where one controlling shareholder (the state or one of the “oligarchs”) plays a dominating role. Government representatives also are involved in control over bank lending, and the relationships between the state and bank representatives play an important role in getting access to capital (Fifka and Pobizhan, 2014; Motohashi, 2015). That explains our interest in analysing state ownership as an important moderating factor in the relationship between earnings management, CSR disclosure and firm value.

To test our hypotheses, we use a sample of 223 publicly listed Russian companies for the period 2012–2018. Taking the absolute value of discretionary accruals as a proxy for earnings management based on a modified Jones (1991) model and real earnings management measure (Roychowdhury, 2006; Kim et al., 2012), we examine the moderating impact of CSR disclosure on the relationship between earnings management and firms' value, proxied by Tobin's Q. We calculate the CSR disclosure index using a modified methodology from Anas et al. (2015) and find that the moderating effect of CSR disclosure weakens the negative impact of earnings management on Russian firms' market value. This led us to conclude that if companies on the Russian market manipulate earnings and disclose more CSR information, investors perceive these activities as legitimate and helping to build a positive image on the market (Buchanan et al., 2018; Chen et al., 2018). According to this hypothesis, investors anticipate these actions positively eventually realising that managers use CSR disclosure as an activity to avoid negative perceptions of earnings management. At the same time, we observe that this positive effect becomes weaker for state-owned companies. Russian state-owned companies have historically been involved in CSR activities (Aray et al., 2021). Our results reflect that market may perceive CSR disclosure of SOE as “window dressing” thinking that companies with state ownership are involved in CSR only to get an easier access to government contracts and keep the link with the government.

Our paper makes several contributions to prior literature. First, it contributes to the research on earnings management and its relationship to firms' value. Motivated by the mixed results of prior research regarding the direct effect of earnings management on market value, we aim to investigate and explain this relationship through the moderating effect of CSR disclosure because it is widely used by firms as a strategic instrument to manage relationships with stakeholders more effectively and create a reliable and trustworthy image (Cespa and Cestone, 2007; Gelb and Strawser, 2001; Kury, 2007; Martínez-Ferrero et al., 2016). Our results indicate that when firms' managers use CSR disclosure to improve the firms' image and conceal opportunistic behaviour, the negative effect of earnings management on market value is weaker. This can be explained by the “reputation-building hypothesis” within stakeholder theory (Buchanan et al., 2018; Chen et al., 2018). Therefore, Russian companies use engagement in CSR activities as a mechanism to achieve better communication with stakeholders, build a transparent company image (Gamerschlag et al., 2011; Pfau et al., 2008), improve a company's reputation (Gelb and Strawser, 2001) and enhance its competitive advantage (Jensen, 2010). Thus, CSR disclosure as a moderator provides a better way to understand the consequences of a firm's earnings manipulations and market reaction.

Second, we contribute to CSR disclosure and stakeholder theory by showing that good intentions lead to positive market reactions. CSR disclosure as a tool for improving a firm's reputation and gaining credibility among stakeholders leads to positive results for companies since stakeholders perceive these actions as sincere and beneficial. A firm's enhanced reputation through CSR disclosure leads to stakeholders contributing different kinds of resources and efforts to the firm resulting in improved valuation of Russian firms. Therefore, CSR disclosure can be used as a long-term strategy for building relationships with stakeholders and improving a company's image and value, especially in cases when the overall behaviour of a firm does not contradict, but rather supports, its fair and transparent position.

Third, we investigate deeper the context of one of the largest emerging markets – Russia – by taking into consideration the role of state ownership in explaining the relationship between earnings management, CSR disclosure and market value. We further contribute to the literature that suggests that ownership structure has a differential impact on firms' engagement in earnings management and CSR disclosure and that this impact is likely to vary across countries (Nikulin et al., 2022; Aray et al., 2021). In Soviet times, SOEs were obliged to take on social responsibility to maintain some facilities for their workers, such as medical and recreation centres, kindergartens and schools (Filippov, 2012). At the same time, Nikulin et al. (2022) observes that SOEs are more involved in earnings management in Russia. Therefore, our findings provide additional insights into the moderating effect of state ownership in the relationship between earnings management, CSR disclosure and market value.

Finally, we use data from an emerging Russian market context where unfair and non-transparent actions are common. Such behaviour is shaped by the formal and informal institutions (e.g. the legal framework, norms, beliefs and values) that prevail in society (Deephouse et al., 2016). Manipulation, corruption and misleading actions are “taken for granted” in some emerging markets, including Russia (Leevik, 2017; Peters et al., 2011); therefore, firms from emerging markets are often characterised by having a poor image, prevalent corruption and use of earnings management for unfair purposes. This leads to great attention, suspiciousness and even mistrust among stakeholders towards such firms. Despite the growth in CSR reporting worldwide, the extant literature has primarily focused on developed countries, and more papers on emerging markets are needed to better understand the specifics within different institutional contexts (Boubakri et al., 2021; Chen et al., 2018; Cheng and Kung, 2016; Diab and Metwally, 2020; Garanina and Aray, 2021; Garanina and Kim, 2023). By exploring the relationship between earnings management and firms' value moderated by CSR disclosure and state ownership in the turbulent, underdeveloped institutional context of the emerging Russian market, we open new avenues for research in developing countries.

The rest of the paper is structured as follows. Section 2 presents literature overview and hypothesis development and provides the specifics of the Russian context regarding CSR and corporate governance system. Section 3 describes the data from Russian listed companies and presents the methodological approach employed for further analysis. The empirical results on the hypothesis testing and robustness tests are presented in Section 4. Section 5 we summarise our findings and concludes.

2. Literature review and hypothesis development

2.1 The relationship between earnings management and firms' value

Stakeholders are greatly concerned about the information disclosed in financial reports. The quality of the provided information is a key aspect of financial statements (Shakespeare, 2020), particularly valued by market participants (Francis et al., 2003). High-quality reporting allows investors to assess performance more effectively and make appropriate economic decisions (Bitner and Dolan, 1996; Francis et al., 2003; Hasnan et al., 2013; Kury, 2007).

Previous studies indicate that reported earnings are the key source of firm-specific information in terms of quality (e.g. Athanasakou et al., 2009; Francis et al., 2003; Viana et al., 2021; Walker, 2013), since they provide a better indicator of a firm's economic performance than cash flows (e.g. Dechow, 1994); hence, managers may be tempted to manipulate earnings. According to Schipper (1989, p. 92), earnings management involves “a purposeful intervention in the external financial reporting process, with the intent of obtaining some private gain”.

Reporting higher earnings can positively influence share prices and the results of initial public offerings (IPOs; Cheng et al., 2015; Cohen and Zarowin, 2010; Kim et al., 2012; Mangala and Dhanda, 2022). Earnings management may help managers to meet or exceed analysts' forecasts, obtain improved compensation (Athanasakou et al., 2009; Walker, 2013), and/or hide opportunistic behaviour relating to increased reward payments, promotions, and stock options (Burgstahler and Dichev, 1997; Degeorge et al., 1999). Fudenberg and Tirole (1995) observe that managers try to smooth earnings to secure their positions by increasing the current period's profits in difficult times and decreasing the current period's profits in good times. This smoothing is particularly attractive for managers whose continuing job tenures depend on reported profits. Researchers also reveal that the market ranks profitable companies more highly (Barth et al., 1999) and, moreover, companies achieve unusually strong stock market performance when they report positive earnings.

Since managers tend to manipulate earnings to improve their own wealth at the expense of stakeholders, the majority of previous studies have used an opportunistic lens, suggesting a negative relationship between earnings management and firms' value (e.g. Bitner and Dolan, 1996; Gaio and Raposo, 2011). Bitner and Dolan (1996) analyse the relationship between income smoothing and firms' valuation, proxied by Tobin's Q, and find that the US market distinguishes between managed and unmanaged earnings, thus discounting the value of the companies that manipulated their earnings. Leuz and Verrecchia (2005) investigate the role of earnings in aligning investors' goals with capital investments. They report that poor-quality reporting leads to weak coordination between firms and their investors, thereby increasing information risks for investors and, in turn, negatively influencing firms' value. By analysing the data from 38 countries from 1990 to 2003, Gaio and Raposo (2011) discover that markets tend to evaluate more highly the companies that have good earnings quality and devalue those that smooth and manipulate earnings. This result is especially significant for companies seeking investment opportunities, since managers tend to manipulate earnings to deflate stock prices prior to open market stock repurchases. When shareholders suspect that earnings are being manipulated, a firm immediately loses value on the stock market (Dechow et al., 1996). Companies may also increase their earnings prior to mergers and acquisitions. The authors therefore conclude that the increase in stock price is only short term, followed by a post-merger reversal due to the pre-merger earnings management (e.g. Louis, 2004).

The consequences of the management practices can lead to the increase in information asymmetry, reduction in company's value, its reputation and corporate image (Fombrun et al., 2000). Fombrun et al. (2000) and Roychowdhury (2006) find a negative relationship between earnings management and company valuation. It also leads to loss of support from stakeholders and damages a company's reputation (Fombrun et al., 2000).

Previous studies report that the legal environment may influence companies' involvement in earnings management. According to La Porta et al. (2002), a country's legal regime may enhance investor protection, thereby leading to more efficient market valuations; for example, in countries with strong legal environments and transparent accounting, earnings management is minimised because of high investor protection and required information transparency (Leuz et al., 2003). Conversely, Russia, as one of the emerging economies, represents a country with “a lack of norms, values of business standards caused by turbulent socio-economic and political conditions and with a weak legal system” (Peters et al., 2011, p. 433).

Massive privatisation in Russia during the 1990s could not transform the centrally planned economy into a truly dynamic open market due to the lack of proper controls to prevent self-dealing by managers and shareholders (Ahmed, 2006; Black et al., 2000; Kim, 2013). Russian companies are commonly characterised by low transparency and extensive earnings management (e.g. Goncharov and Zimmermann, 2006, 2007; Leevik, 2017), meaning that managers' personal interests may take priority over those of shareholders (Ahmed, 2006; Bataeva and Tkachenko, 2020). Moreover, Russia, as a state-controlled economy, traditionally had little desire to improve financial reporting, despite its importance for allocating goods and services in a developed market economy (Goncharov and Zimmermann, 2007; Kim, 2013).

Reforms initiated by President Putin made substantial improvements; however, unauthenticated use of discretionary earnings management has not been dealt in an effective way (Ahmed, 2006; Leevik, 2017). Based on these factors, Russia ranked 136 out of 178 countries in a recent survey of transparency and corruption (Transparancy International, 2021). Nondisclosure and non-transparency made Russia one of the riskiest countries in the world to invest in (Ahmed, 2006).

Previous research reports that Russian companies engage in earnings management to signal better earnings before an IPO, obtain better financing opportunities, or manage taxes. In their studies, Krylova (2003) and Goncharov and Zimmermann (2006) state that one of the main reasons for earnings management in Russia is that when preparing financial statements, Russian companies are likely to focus on the demands of tax authorities; hence, they tend to optimise their costs to minimise income tax rather than provide fair financial information (Goncharov and Zimmermann, 2006; Krylova, 2003). Another paper by Goncharov and Zimmermann (2007) postulates that Russian firms manage earnings to avoid showing losses when they apply for bank financing: companies are aware that banks demand positive earnings and, to facilitate financing, they “match” their income with the banks' expectations. Bataeva and Tkachenko (2020) additionally claim that the quality and transparency of financial reporting in Russian companies is not a priority for managers.

Based on the identified reasons for earnings management in Russian companies operating in a weak institutional environment depicted by previous researchers (Krylova, 2003; Goncharov and Zimmermann, 2006, 2007), we expect to find a negative relationship between earnings management and firms' valuation in the Russian market; therefore, we formulate our first hypothesis as follows:

H1.

There is a negative relationship between earnings management and the market value of Russian companies.

2.2 The relationship between CSR disclosure and firms' value

CSR engages companies in “actions that appear to further some social good, beyond the interests of the firm and that which is required by law” (McWilliams and Siegel, 2001, p. 117). Companies involved in CSR are more focused on long-term objectives, tend to foster long-term relationships with stakeholders (Baboukardos, 2017; Gelb and Strawser, 2001; Jones, 1995; Wang et al., 2018), and may enjoy an improved corporate reputation and a positive image among stakeholders (Pfau et al., 2008). According to stakeholder theory, which is fundamental to the study of business and society (Clarkson, 1995), firms might produce better results if managers adequately consider the concerns of stakeholders beyond the narrow shareholder group. This aligns with contract theory, according to which any firm consists of a bundle of contract relationships with the different groups of stakeholders that hold it accountable, such as shareholders, customers, employees, the state, etc. It postulates that the value of these contracts depends on the ability of firms to meet stakeholders' expectations (Bardos et al., 2020; Clarkson, 1995), with whom effective communication increases the value of the contracts. Civil society, the media, and other stakeholders expect companies to be transparent and open about their actions and to be accountable for the consequences of their businesses by considering their social impact and engaging in environmental protection (Brulhart et al., 2019). On the one hand, transparency is a necessary condition for CSR reporting (Pinnuck et al., 2021), but on the other, CSR reporting itself is a vehicle for transparency and a source of competitive advantage (Baboukardos, 2017; Pfau et al., 2008). According to the “reputation-building hypothesis” within stakeholder theory, there is a positive relationship between CSR disclosure and firms' value (Buchanan et al., 2018; Chen et al., 2018). Engagement in CSR activities can be used as a mechanism to achieve better communication with stakeholders and, thus, reduce conflicts of interest between managers and various non-investing stakeholders (Baboukardos, 2017; Mohmed et al., 2019). A firm's enhanced reputation through CSR disclosure leads to stakeholders contributing different kinds of resources and efforts to the firm (Jensen, 2010), resulting in the firm improving its valuation through improved relationships with stakeholders. Kimbro and Cao (2011) claim that CSR leads to better firm performance, lower cost of capital and decreases the overall firm's risk. By studying US companies Kim et al. (2012) demonstrate that operational and accounting decisions of socially responsible companies lead to more conservative decisions, less manipulative behaviour and more transparent financial reporting.

At the same time there exists another perspective in the literature. According to some results firms may use their developed stakeholder relations and legitimacy in order to hide their self-serving policies (Lins et al., 2017; Ferrell et al., 2016). Firms may use CSR activities as a corporate image management tool or reputation insurance to cover up the impact of corporate misbehaviour and can manipulate the perception of the relevant public (Hemingway and Maclagan, 2004; Choi et al., 2013). Friedman (1970) notes that CSR may signal about the presence of agency problems in a firm as “there is one and only one social responsibility of business—to use its resources and engage in activities designed to increase its profits”. The main explanation of this notion is that one of the reasons of investing in CSR activities by internal stakeholders (managers and controlling shareholders) is related to enhancement of their own reputation and careers at the expense of shareholders' wealth (Barnea and Rubin, 2010; Bénabou and Tirole, 2010). Therefore, this may lead to the situation that markets reacts negatively towards CSR disclosure or does not react at all.

Although CSR is so much a part of companies' strategies in Western societies that it is “taken for granted” (Bondy et al., 2012), it remains underdeveloped in emerging-market companies. Recently some emerging countries started to pay more attention to CSR activities, e.g. in 2010 an integrated reporting mandate was introduced in South Africa for firms listed on the Johannesburg Stock Exchange (Barth et al., 2017) and later on Indian companies were obliged to spend 2% of their income on CSR (Manchiraju and Rajgopal, 2017). However, CSR disclosure in Russian companies is uncommon due to its historical path (Filippov, 2010; Kuznetsov et al., 2009; Polishchuk, 2009) and institutional context (Aray et al., 2021; Garanina and Aray, 2021). In Russia, there are no legal requirements for companies to disclose CSR information, so the main motivation for a company's CSR disclosure is to obtain legitimacy and a positive reputation within the country of origin. This is especially relevant for companies that are large local employers/taxpayers in cities (Filippov, 2010; Kuznetsov et al., 2009).

Based on the important role of CSR information in obtaining a positive image and reputation for Russian companies, we propose our second hypothesis:

H2.

There is a positive relationship between CSR disclosure and the market value of Russian companies.

2.3 The moderating effect of CSR disclosure

Nowadays CSR plays an important role in the overall corporate strategy. Firms with effective CSR disclosure are more likely to meet the ethical expectations of stakeholders (Gelb and Strawser, 2001; Jones, 1995), since the enhanced reputation and strong corporate image create long-term value (Martínez-Ferrero et al., 2016).

Recent studies document that firms involved in CSR activities behave in a different way in their financial reporting and provide more value-relevant information (Gao and Zhang, 2015). For example, Kim et al. (2012) find that companies with higher CSR scores constrain their earnings management and deliver more transparent and reliable information to the market. Dhaliwal et al. (2012) claim that CSR disclosure improves analyst forecast accuracy. Gao and Zhang (2015) analyse the moderating effect of CSR performance on the relationship between accrual earnings management and firm's value. The authors find that the reported earnings of smoothers that are socially responsible are more value-enhancing. Shleifer (2004) observes that companies involved in CSR have less manipulative practices due to better transparency of their financial and nonfinancial disclosure. To sum up, companies involved in CSR activities have less motivation to be involved in earnings management practices and provide more value relevant financial information (Gelb and Strawser, 2001).

In their paper Martínez-Ferrero et al. (2016) highlight that CSR can be strategically used by company management against the negative perception from being involved in earnings management. The analysis is carried out on a large sample of listed companies from 26 countries and covers the period from 2006 to 2010. The authors reveal that involvement in CSR activities has a positive effect on corporate reputation and leads to lower cost of capital. Finally, the authors conclude that CSR activities can be used by companies to mask earnings management practices.

Therefore, involvement in CSR leads to improved information environment, less information asymmetry (Cui et al., 2018) and less likelihood of earnings restatement (Wans, 2020). Managers of companies performing in a socially responsible way tend to provide higher-quality information to the market, maintain a more transparent image, and achieve their objectives more efficiently by reducing information asymmetry between the firm and its stakeholders Cui et al. (2018).

At the same time, companies may use CSR disclosure to camouflage opportunistic behaviour (Cai et al., 2012). Previous research reports that managers may exaggerate their CSR activities and increase CSR investments to conceal subsequent earnings manipulation (Prior et al., 2008). Managers who are involved in earnings management have incentives to develop a socially friendly image via CSR activities, given that a good CSR image is a powerful instrument for enhancing stakeholder relationships (Prior et al., 2008; Wang et al., 2018). Following this approach, managers may aim to avoid the decrease in firms' value that is often associated with earnings management. Moreover, due to the differences existing in different CSR disclosure measures (e.g. Berg et al., 2022; Christensen et al., 2022) and as CSR ratings are developed by private organizations with limited transparency, it is difficult for the market to identify when CSR practices are used just as “window dressing” for masking earnings management (Martínez-Ferrero et al., 2016).

Building on (Martínez-Ferrero et al., 2016; Gao and Zhang, 2015) we hypothesise that companies engaged in earnings management would use CSR disclosure as a way to improve their reputation and camouflage opportunistic behaviour. CSR practices are appreciated by market participants because they generate a positive image of a company externally if they are not legally enforceable that is the case under consideration (Basu and Palazzo, 2008). Positive CSR strategy helps to enhance company's legitimacy and get support among regulators and stakeholders (Martínez-Ferrero et al., 2016).

As there is a lack of shared understanding of information disclosed about CSR by market participants (Christensen et al., 2022; Berg et al., 2022), we expect that market participants positively evaluation companies' involvement in CSR actions disregarding the existing signs of earnings management. Therefore, based on our prior hypothesis about a negative relationship between earnings management and firms' value, and the assumption that CSR disclosure helps companies to build better mutual relationships with multiple stakeholders and enhance their reputation, we propose our next hypothesis:

H3.

CSR disclosure positively moderates the relationship between earnings management and the market value of Russian companies, i.e. CSR disclosure weakens the negative effect of earnings management on the market value of Russian companies.

2.4 The moderating effect of state ownership

According to previous research (e.g. Martínez-Ferrero et al., 2016; Jamali and Karam, 2018), understanding companies' involvement in earnings management and CSR is highly dependent on country context. Therefore, in order to understand the peculiarities of market reactions on companies' involvement in CSR activities and earnings management, it is important to take into account the role of political institutions that operate in the national environment, especially in the emerging country like Russia, where due to the communist tradition the state exerts total control over all the aspects of the society (Li et al., 2014) and where the government plays a significant role in setting rules to create norms for organizations' CSR behavior (Campbell, 2007). Russian economy is characterised by a significant presence of companies controlled by the state (Abramov et al., 2017). State owned companies represent around 40% of the market value of all publicly listed companies in Russia (Abramov et al., 2017) and contribute to 70% of the total public sector (Panibratov and Klishevich, 2018).

Marquis and Qian (2014) claim that in emerging markets state ownership might be associated with an increased risk of financial misreporting as state owned companies are not inclined to adopt strong governance mechanisms. The existing research also confirms that the level of earnings management can significantly differ for companies with state ownership and without (Guo and Ma, 2015; Gaio and Pinto, 2018). Some Russian companies felt compelled to abandon attempts to be listed in the US market due to stringent financial disclosure requirements and the demands of suspicious investors (Li et al., 2014). Moreover, the recent paper of Nikulin et al. (2022) has observed that Russian SOE are more involved in earnings management than companies without state ownership. Therefore, “socialist imprints” and high share of state-owned companies may influence the relationship between earnings management, CSR disclosure and firms' market value. As the role of the links with the government is considered to be important in Russia, we expect that state ownership will moderate the relationship between earnings management and firm's value. A recent study by Garanina and Kim (2023) has also observed that Russian state-owned companies are characterised by less conservative and more aggressive financial reporting.

In Russian economy state-owned companies are more involved in CSR activities in order to reduce uncertainty and increase changes of survival (Aray et al., 2021). Therefore, involvement in CSR is considered by state companies as a “strategic action that a company takes to build, maintain or enhance the appropriateness and desirability perceived by the state through social-environmental activities based on which the company expects to access various forms of state resources” (Zhao, 2012, p. 442). The historical and economic development of Russia influenced the current situation characterised by a low level of CSR consistency (Fifka and Pobizhan, 2014; Li et al., 2010). During the Soviet times SOE took over responsibilities of maintaining facilities for their employees, for example, hospitals, kindergartens, and schools (Filippov, 2012). During the transition to the market economy the Russian government enforced companies to take over some social and environmental risks in exchange for obtaining social legitimacy and government support during the privatization period (Filippov, 2012; Kuznetsov et al., 2009).

To sum up, state-owned Russian companies have been confirmed to be more involved in earnings management practices. Managers of state-owned companies consider a possibility of formal or informal support from the government while making strategic decisions (Dikova et al., 2019). Managers of SOE try to follow economic policies of the government and, therefore, tend to be more involved in CSR activities that help them to enhance their perceived legitimacy (Cheng and Kung, 2016; Garanina and Kim, 2023). Therefore, involvement of state-owned companies in CSR activities can also be dependent on their goal of sustaining long-term relations with the government and not on their real strategic goal of being socially responsible. Following the discussion, we come up with our final hypothesis:

H4.

The positive moderating effect of CSR disclosure on the relationship between earnings management and market value is weaker for Russian companies with state ownership.

3. Methodology

3.1 Sample and data sources

Our sample includes 223 public companies listed on the Moscow Stock Exchange in 2012–2018. Financial data is obtained from Thomson Reuters Datastream, and the CSR disclosure index is calculated based on the information provided in companies' annual reports retrieved from the SKRIN (www.skrin.ru) and SPARK (www.spark.ru) databases.

The information on CSR disclosure is collected from annual reports because previous researchers have argued that annual reports are the single most important source of information about corporate activities and “can be accepted as an appropriate source of a company's attitudes towards social reporting” (Campbell, 2000, pp. 84–85). Moreover, only a few Russian publicly listed companies publish any external reports other than annual ones (Garanina and Aray, 2021; Garanina and Kim, 2023).

The sample covers a variety of industry sectors based on the Russian economic activity codes—OKVED. The sample companies belong to the following industries: agriculture (5), business services (15), power utilities (56), manufacturing (69), mining and oil (17), real estate (6), retail (17), telecommunications (14), transportation (13), and other (11). The sample does not include financial institutions and insurance companies.

3.2 Variables and models

3.2.1 Dependent Variable: Tobin's Q

We use Tobin's Q as a proxy for a firm's market value because it reflects the market's expectations of future performance (e.g. Jiraporn et al., 2008). We define Tobin's Q as the sum of the market value of equity (share price multiplied by the number of ordinary shares issued at the end of a fiscal year) and the book value of debt divided by the book value of total assets (variable TOBIN).

3.2.2 Independent variable: earnings management

Earnings management can be split into accrual-based earnings management, based on differences between accounting policies, and real earnings management, when managers transfer transactions to different accounting periods to increase earnings (Cohen and Zarowin, 2010; Hribar and Collins, 2002; Roychowdhury, 2006). Accrual-based earnings management is related to adjustments or changes in criteria or repayment systems. This type of earnings management is less visible in comparison to real earnings management decisions which affect a firm's performance, e.g. the time of writing of inventories or selling R&D projects.

3.2.2.1 Proxy for accrual-based earnings management

We calculate discretionary accruals as a proxy for earnings management to represent a collection of judgements, estimates, and allocations (e.g. Francis et al., 2003) using a modified Jones model (Dechow, 1994; Jones, 1991). These types of accruals refer to differences between the time of accounting for income and expenses and the real cash flow operations relating to those transactions (Hribar and Collins, 2002; Song and Rimmel, 2021).

First, following previous studies (e.g. Kury, 2007; Hribar and Collins, 2002), we define the level of total accruals for a particular firm (TACi,t) as the difference between net income before extraordinary items from the income statement (NIi,t) and cash flow from operations obtained from the cash flow statement (OCFi,t):

(1)TACi,t=NIi,t-OCFi,t

Second, in line with the modified Jones model, the non-discretionary accruals (NDACi,t) for each observation in the sample are assessed based on a firm's total assets (TAi,t), property, plant and equipment (PPEi,t) and change in revenue (∆REVi,t). The variables are divided by total assets from the previous period to allow us to account for differences in the size of companies:

(2)TACi,tTAi,t-1=α0+α1×(1TAi,t-1)+α2×(∆REVi,tTAi,t-1)+α3×(PPEi,tTAi,t-1)+τi,t

Finally, discretionary accruals represent a proxy for earnings management (AEMi,t), which is calculated as the difference between the overall accruals (TACi,t) and the value of non-discretionary accruals obtained using the modified Jones model (NDACi,t). In line with previous research (e.g. Song and Rimmel, 2021; Wang et al., 2018), the earnings management variable is measured according to the absolute value of accruals because we are interested in assessing the moderating effect of overall earnings management, rather than its direction (i.e. an increase or decrease in earnings).

3.2.2.2 Proxy for real-based earnings management

Following prior studies (e.g. Roychowdhury, 2006; Kim et al., 2012), sales manipulations lead to lower operating cash flows. We base our calculations on Roychowdhury (2006) to estimate the normal level of operating cash flows:

(3)CFOtAt-1=α0+α1(1At-1)+β1(StAt-1)+β2(∆StAt-1)+εt,

where

• CFO = cash flow from operations in year t;

• A = total assets;

• S = net sales; and

• ∆S=St-St-1

For every firm-year, abnormal cash flow from operations (AB_CFO) is the residual from the corresponding industry-year model and the firm-year's sales and lagged assets.

Another measure of real earnings management is abnormal production costs. Following prior studies (e.g. Roychowdhury, 2006; Kim et al., 2012) we define production costs as the sum of cost of goods sold (COGS) and change in inventory during the year, and they express expenses as a linear function of contemporaneous sales. Therefore, we estimate the following model for normal COGS:

(4)COGStAt-1=α0+α1(1At-1)+β1(StAt-1)+εt,

where COGSt = cost of goods sold in year t. Following the same approach, we estimate the model for normal inventory growth:

(5)∆INVtAt-1=α0+α1(1At-1)+β1(∆StAt-1)+β2(∆St-1At-1)+εt,

where ∆INVt is the change in inventory in year t. Following Roychowdhury (2006) and Kim et al. (2012), we define production costs as PRODt=COGSt+∆INVt. Using equations (4) and (5), we estimate normal production costs from the following model:

(6)PRODtAt-1=α0+α11At-1+β1StAt-1+β2∆StAt-1+β3∆St-1At-1+εt,

Abnormal production cost (AB_PROD) is the residual from this model.

The third measure of real activities manipulation is the abnormal discretionary expenses. We follow again Roychowdhury (2006) and Kim et al. (2012) and estimate the normal level of discretionary expenses using the following model:

(7)DISEXPtAt-1=α0+α1(1At-1)+β(St-1At-1)+εt,

where DISEXPt is the discretionary expenses in year t, defined as the sum of R&D, advertising, and SG&A expenses. For every firm-year, abnormal discretionary expenditure (AB_EXP) is the residual from the model.

Following Cohen et al. (2008), we also construct the combined measure of real earnings management by aggregating the three individual real activities manipulation proxies, AB_CFO, AB_PROD, and AB_EXP. Considering the direction of each real activities manipulation components, the combined proxy for real earnings management, REM, is calculated as AB_CFO-AB_PROD + AB_EXP.

3.2.3 Moderating variable: CSR disclosure

There are different approaches to evaluate CSR disclosure. Some authors use content analysis (e.g. Anas et al., 2015; Wiseman, 1982), whereas others use open CSR ratings (Lau et al., 2016). Since no ratings exist for Russian companies, we adopt Wiseman's (1982) approach to CSR index construction, which is later applied in Anas et al. (2015).

To capture country features that would better reflect a unique Russian setting, we make some changes to the index and increase the number of CSR items represented in Anas et al. (2015) from 17 to 22 (Appendix A). The changes were introduced in line with the Russian Regulations on Information Disclosure for Securities Issuer, and the peculiarities of Russian companies in disclosing information, especially regarding the sections on Community and Workplace (Aray et al., 2021; Garanina and Aray, 2021). In “Community”, the element “Supporting children” was divided into “Supporting children from communities”, “Supporting employees' children”, and “Supporting disabled children”. Due to the high involvement of Russian companies in infrastructure projects, we added the item “Contribution to infrastructure”. The element “Supporting retired workers” was also added to the CSR index (within “Workplace”), because Russian companies frequently reflect how they support their retired employees, given the low retirement benefits elderly people receive from the government. The last change we made to the CSR index, compared with Anas et al. (2015) is related to the “Corporate Governance” item, which we split into two separate sections – compulsory and voluntary.

We measure 22 elements of the CSR disclosure index (variable CSR) from 0 to 3, based on how fully different aspects of CSR are disclosed in the annual reports. Following Dumay and Cai (2015, p. 139), we believe that including a quality measure “might reveal new insights that may otherwise have gone unnoticed”. In line with Wiseman (1982), we assign a score of 1 if only vague CSR information is disclosed by a company, 2 if qualitative information is disclosed without supporting financial data, 3 for maximum disclosure supported by financial data, and 0 for no CSR information.

3.2.4 Control variables

To ensure that the results are not affected by firms' heterogeneity, we add control variables to cover different firms' characteristics, including firm size, company age, profitability, leverage and corporate governance characteristics. The data used for the analysis includes information from financial reports, quarterly reports to the regulator held in the SKRIN and SPARK databases (to obtain information on the corporate governance characteristics), and firm data obtained from Thomson Reuters Datastream.

Firm size is one of the most common control variables used in CSR, earnings-management, and market-performance research (Buchanan et al., 2018; Song and Rimmel, 2021). To account for the size of a company, we calculate a logarithm of total assets (variable SIZE). A firm's financial performance indicators relate to its market efficiency and market valuation; hence, we use return on assets (ROA) as a measure of firms' accounting performance (variable ROA), since this measure is commonly agreed to be a main measure of a firm's financial performance proxied by Tobin's Q (Kim et al., 2012; Song and Rimmel, 2021). Board size is a corporate governance characteristic that has a significant effect on market value (Buchanan et al., 2018; Song and Rimmel, 2021). Leverage is accounted for by dividing total debt by total assets (variable LEV), since this indicator of financial risk relates to a company's valuation on the market (Bozzolan et al., 2015; Brulhart et al., 2019; Song and Rimmel, 2021). We also control for a firm's age (variable AGE) because older companies may have stronger CSR values (Anas et al., 2015), more experience, and thus more intensive engagement in CSR activities, which may influence their market value. To control for the board size effect on Tobin's Q, we use the total number of directors on a board (variable BOARD). We also control for other corporate governance characteristics – share of women on board (variable WOMEN) and share of independent directors (variable INDEP) as these variables may also have influence on performance indicators of companies in emerging markets (Anas et al., 2015; Berezinets et al., 2017; Garanina and Muravyev, 2021; Aray et al., 2021). Finally, firms active in different industries could have different CSR activities and performance indicators (e.g. Bozzolan et al., 2015); hence, we control for industry effects.

All the variables are described in Appendix B.

3.3 Empirical methodology

Based on Baron and Kenny (1986) we test our moderating effect in three steps. First, we analyse the impact of the predictor – earnings management (EM) measured either with accrual-based proxy (AEM) or real earnings management proxy (REM) on the dependent variable (TOBIN). Then we test the relationship between the moderator (CSR) and the dependent variable. Finally, we analyse the moderating effect by including the interaction between earnings management proxies and CSR disclosure. The “moderator hypothesis is supported if the interaction is significant” (Baron and Kenny (1986, p. 1174). We then additionally add state ownership (variable STATE) as a moderator to the last model. Therefore, to address the hypotheses proposed in section 2, we assess the following empirical models:

(8)Tobini,t=α0+α1EMi,t+α2SIZEi,t+α3AGEi,t+α4ROAi,t+α5LEVi,t+α6BOARDi,t+α7WOMENεi,t+α8INDEPεi,t+εi,t

(9)Tobini,t=β0+β1CSRi,t+β2SIZEi,t+β3AGEi,t+β4ROAi,t+β5LEVi,t+β6BOARDi,t+β7WOMENεi,t+β8INDEPεi,t+ϵi,t

(10)Tobini,t=γ0+γ1EMi,t+γ2CSRi,t+γ3EMi,t*CSRi,t+γ4SIZEi,t+γ5AGEi,t+γ6ROAi,t+γ7LEVi,t+γ8BOARDi,t+γ9WOMENεi,t+γ10INDEPεi,t+θi,t

(11)Tobini,t=μ0+μ1EMi,t+μ2CSRi,t+μ3STATEi,t+μ4EMi,t*CSRi,t+μ5EMi,t*STATEi,t+μ6CSRi,t*STATEi,t+μ7EMi,t*CSRi,t*STATEi,t+μ8SIZEi,t+μ9AGEi,t+μ10ROAi,t+μ11LEVi,t+μ12BOARDi,t+μ13WOMENεi,t+ μ14INDEPεi,t+φi,t

In line with Buchanan et al. (2018), we winsorise each continuous variable at the first and ninety-fifth percentiles. As previous studies (e.g. Wang et al., 2018) suggest, the choice of the mandatory reporting firms was not random, so the setting could only be considered a quasi-natural experiment. To alleviate this concern, we include firm fixed effects to control for the effect of time-invariant firm characteristics.

4. Descriptive statistics and empirical results

4.1 Descriptive statistics

Descriptive statistics for the main variables are given in Table 1, Panel A. The mean for Tobin's Q that we use as a proxy for a firm's market value is 0.4633, showing that more than half the Russian companies in the sample are undervalued on the market. The mean for the CSR disclosure index is 17, while p75 is 28, indicating that the majority of companies in the sample do not fully disclose CSR activities or achieve the highest possible score. The AEM mean is 0.0794, with a maximum of 0.1923 and a standard deviation (SD) of 0.0932. The REM mean is −0.2126 with a SD of 1.1303. The mean for ROA is 10.32%, with an SD of 2.07. The average debt as a share of total assets is 29.1%, and the average age of our sample companies is 15 years. The average share of women on board of directors of Russian companies is 12% while the share of independent directors is about 24%.

On average Russian companies have about 3% of state ownership, and the maximum value is 91.7%. About 32% of publicly listed companies from the sample have state as a shareholder. State owned companies disclose more information about CSR in comparison to companies without state ownership and are also characterised by higher levels of earnings management (Table 1, Panel B). Our results are consistent with the finding of Gaio and Pinto (2018) and Nikulin et al. (2022) who reveal that state-owned companies tend to be more involved in earnings management in comparison to companies without state ownership. These findings can be potentially explained by the fact that state-owned companies might be more inclined to manage earnings to achieve KPIs set by the government as a shareholder. Additionally, earnings management in SOEs may be driven by the intention to obtain additional funding from the state (Nikulin et al., 2022). In general, in the weak institutional environment companies with state ownership operate with relatively soft budget constraints and experienced lower pressure from the financial markets, therefore, there is more demand for earnings management to manage the relations with the state.

4.2 Results of the empirical analysis

Before conducting the empirical analysis, we test our data for multicollinearity between the variables by calculating the variance inflation factors (VIFs), which are well below the rule-of-thumb threshold value of 10 for all variables (mean the VIFs were below 3 for all the models), indicating that multicollinearity does not affect our results. The results of the Breusch–Pagan test produced the value of 157.23 with a p-value of 0.000. The correlation matrix is shown in Table 2.

The results indicate that the CSR disclosure index is positively correlated with Tobin's Q, whereas earnings management is negatively correlated with the market value of Russian companies. Among the control variables, most of the expected correlations are observed.

Table 3 shows the main results of the empirical analysis that tests H1 – H4.

Columns (1) and (2) indicate that both proxies of earnings management – accrual-based and real earnings management – are negatively associated with Tobin's Q for Russian companies (α = −0.012, p < 0.05; α = −0.028, p < 0.01, respectively); therefore, H1 is supported. Our results are consistent with previous findings that market participants negatively evaluate involvement of Russian companies in earnings management (e.g. Bitner and Dolan, 1996; Gaio and Raposo, 2011). In other words, even though the Russian market is characterised by a “weak institutional environment” and less sophisticated investors who might consider CSR disclosure as a worthless thing, the market negatively reacts towards involvement of Russian companies in earnings management through opportunistic lens.

Column (3) of Table 3 shows the main results for H2, reflecting a positive relationship between CSR disclosure and firms' market value, proxied by Tobin's Q (β = 0.146, p < 0.01), which led us to conclude that H2 is also supported. Therefore, our results follow the stakeholder theory according to which CSR disclosure is used to achieve better communication with stakeholders and, thus, reduce conflicts of interest between managers and various non-investing stakeholders (Baboukardos, 2017; Mohmed et al., 2019). Even though CSR disclosure is not mandatory in Russia, involvement in CSR activities results in improved firm's valuation through a better and transparent image and improved relations with stakeholders (Martínez-Ferrero et al., 2016).

The results presented in Column (4) and (5) of Table 3 relate to H3. We find that CSR disclosure positively moderates the relationship between earnings management (both accrual-based earnings management and real earnings management proxies) and firms' value (γ = 0.095, p < 0.10; γ = 0.147, p < 0.01). Since the direct link between earnings management and Tobin's Q is negative, we conclude that the negative effect of earnings management on a firm's market value is weaker for those companies that disclose CSR information. In other words, firms that manipulate earnings and have a higher CSR disclosure are characterised by a weaker negative effect of earnings management on Tobin's Q compared to the companies that disclose less CSR information. This can be explained by the fact that investors and other stakeholders consider CSR disclosure as a source of a strong reputation and trust. CSR disclosure in this situation diminishes the negative reaction of the market towards earnings management, and is believed to be an effective communication tool for building a transparent company image (Gamerschlag et al., 2011; Pfau et al., 2008). At the same time if companies use CSR disclosure only for “greenwashing” or concealing purposes, the market does not capture this difference. CSR disclosure is hard to interpret (Berg et al., 2022; Christensen et al., 2022) and market in general positively reacts to higher amounts of disclosed CSR information even for companies involved in earnings management. Therefore, H3 is supported.

The results for testing H4 on whether state ownership influences the relationship between earnings management, CSR disclosure and Tobin's Q are represented in Table (3) Columns (6)–(8). We find that in general the market positively evaluates links to the government as there is a prejudice that companies with state ownership have easier access to resources (μ = 0.083, p < 0.05) that is reflected in Column (6). However, we further observe (Columns (7) and (8)) that the positive moderating effect of CSR disclosure in the relationship between earnings management and Tobin's Q is weaker for companies with state ownership involved either in accrual-based earnings management or real earnings management (μ = 0.141, p < 0.01; γ = 0.019, p < 0.01). These findings can be explained by the fact that market may consider involvement of SOEs in CSR as greenwashing as managers of state-owned companies may be more involved in CSR activities because they prioritise their own interests, and use involvement in CSR activities to develop their own careers (e.g. Boubakri et al., 2008) trying to hide involvement in earnings management. Another explanation can be that markets considers involvement of SOEs in CSR as a required action in order to get national legitimacy for an easier access to government resources (Cheng and Kung, 2016; Garanina and Kim, 2023). Therefore, managers of state-owned companies involved in earnings management consider a possibility of formal or informal support from the government while making strategic decisions (Dikova et al., 2019) to involve in CSR activities.

4.3 Robustness tests

To test for robustness, we use an alternative proxy for earnings management (Laux and Leuz, 2009; Leuz et al., 2003). Since firms may increase the reporting of future revenues in current periods or delay the reporting of costs to hide low current performance, we apply the accruals component of earnings in line with Dechow and Dichev (2002) and Scholtens and Kang (2013):

(12)Accrualsi,t=(∆CAi,t-∆Cashi,t)-(∆CLi,t-∆STDi,t-∆TPi,t)-Depi,t

where ∆CAi,t is the change in total current assets of firm i at time t; ∆Cashi,t is the change in cash; ∆CLi,t is the change in current liabilities; ∆STDi,t is the change in short-term debt; ∆TPi,t is the change in income tax payable; and Depi,t is depreciation expenses. After the accruals are calculated, we add cash flow from a firm's operations to obtain the operating earnings of the firm. This model reflects the degree to which insiders use their discretion to alter the accounting component of reported earnings. Furthermore, we assess earnings management (variable EM_proxy) via:

(13)EM=SD(operating earnings)SD(cash flows from operations)

In line with the equation, insiders use their discretion to smooth reported earnings to manipulate the volatility of operating earnings with respect to the original cash flow (Laux and Leuz, 2009; Leuz et al., 2003).

The results of testing the empirical models with the new proxy for earnings management are presented in Table 4 and bring us to the same conclusions.

Our robustness tests therefore indicate no changes in the meaningful results from the results we obtained previously (see Table 3).

In Table 5 we present additional tests for robustness. In order to improve the precision of our results, we add variables one by one to previously described models (3)–(5).

First, following Khan et al. (2013), we use a dummy variable for CSR disclosure, measured as 1 when some information on CSR is disclosed in an annual report and 0 otherwise (Table 5 variable CSR_dummy, Columns 1–6). We check the results of our model and conclude that they do not differ qualitatively from those presented in Table 3.

We further use market capitalisation (variable CAP) as another proxy for market value. The data is obtained from Eikon. The results are presented in Table 5 (Columns (7) and (8)) and do not reflect significant meaningful differences from previous findings.

Additionally, in our robustness tests we break down our entire sample into subsamples of companies with state ownership and without. The results are consistent with our previous findings and can be provided upon request.

4.4 Endogeneity test

We follow previous research (e.g. Martínez-Ferrero et al., 2016) and apply the generalised method of moments (GMM) method that helps to address the endogeneity concerns that could occur in the least squares estimators (Ogaki, 1993). GMM has benefits over two or three-stage least square estimators as it is consistent and takes into account firms' own specificity that is linked to a particular behaviour.

One of the possible GMM models used in the literature is the difference GMM model. However, the difference GMM model is criticised in research for poor estimation of performing estimators in short sample periods (which is the case of this study with the observation period of 2012–2018). Therefore, following prior research (e.g. Arioglu, 2020) we apply the system GMM model. While applying the system GMM models, various diagnostic tests are presented, e.g. AR (1) and AR (2) test values for first and second-order correlation in first-differenced residuals, with the null hypothesis stating that there is no serial correlation. Therefore, the results presented in Table 6 reflect that there is no problem with the construction of the system GMM models.

5. Conclusions

Given the inconsistent findings of the previous studies that investigate the link between earnings management and firms' value tested mainly on the developed markets, we analyse how CSR disclosure moderates this relationship. Our paper provides new insights into earnings management, market value and CSR disclosure of firms in emerging Russian market where state ownership plays an important role. Therefore, we additionally investigate whether state ownership influences the moderating effect of CSR disclosure in the relationship between earnings management and a firm's value.

Our study makes a valuable contribution to research concerning the relationship between earnings management, CSR disclosure and firms' value taking into consideration the peculiarities of the institutional environment – high involvement of state in ownership of Russian companies. First, our paper enhances the accounting literature by shedding light on the relationship between earnings management and firms' value. Building on Martínez-Ferrero et al. (2016) and Gao and Zhang (2015) who mainly focused on developed markets, we considered the missing element – CSR disclosure – to explain the relationship between the earnings management and market value on the emerging Russian market. We document that the market value of Russian companies increases when they disclose involvement in CSR activities. Adding to the debate based on the developed markets, we document that involvement in CSR disclosure can be used to mask earnings management practices in Russian companies. However, market does not clearly distinguish whether companies are really involved in CSR activities or they just disclose more information on CSR to “window dress” their actions. This conclusion is supported by recent findings that CSR disclosure, used by investors and other stakeholders, is difficult to decipher (Christensen et al., 2022; Berg et al., 2022) and companies might use this information asymmetry to improve their image, while not be really involved in CSR practices. Secondly, we contribute to previous research that highlights the importance of institutional environment (Boubakri et al., 2021; Martínez-Ferrero et al., 2016) by showing that the positive moderating effect of CSR disclosure in the relationship between earnings management and market value is weaker for state-owned companies. Therefore, our results reflect that market might consider involvement in CSR activities of Russian state-owned companies that manage their earnings as “window dressing” due to their close link with the Russian government and important role of involvement in social activities to obtain government contracts and legitimacy on the national market.

Our study has some limitations. First, there are different measures for earnings management. Other proxies for accrual-based earnings management except for the ones presented in this study can be used in future studies. Second, the measure of CSR reporting is quite subjective despite being based on previous research (e.g. Anas et al., 2015; Wiseman, 1982; Garanina and Aray, 2021; Garanina and Kim, 2023). Future studies would benefit from using other objective measures to test the relationships. However, measures such as indices and ESG rankings are largely unavailable to Russian companies. Moreover, we believe it is vital to investigate country-specific CSR issues to gain a better understanding of the context.

We urge future researchers to consider country-specific factors while analysing the moderating role of CSR disclosure in earnings management relationships since the institutional peculiarities may help to explain differences in the obtained results. Finally, although we state that our results are reliable and robust, a larger sample over a longer period could provide greater insights into the relationships between earnings management and firms' value, moderated by the defined missing element.

Despite these limitations, our study reflects the relevance of CSR disclosure and peculiarities of the institutional environment in understanding the relationship between earnings management and firms' market value, enabling some discrepancies in past empirical research to be reconciled and paving the way for a more nuanced, contextualised exploration of the topic.

Table 1

Descriptive statistics

Panel A. The table reports descriptive statistics of the main and explanatory variables. The variables are described in Appendix B
Variable	Obs	Mean	SD	Min	Max	p25	p75
TOBIN	1,275	0.4633	0.4235	0.0019	1.4562	0.1603	0.6093
CSR	1,539	17.8142	14.5084	0	62	5	28
AEM	1,327	0.0794	0.0932	0.0009	0.4654	0.0115	0.1923
REM	1,539	−0.2126	1.1303	−9.1080	1.0779	−0.0407	0
SOE_share	1,552	0.0306	0.1294	0	0.9167	0	0.2137
SIZE	1,326	17.0332	2.4095	7.4254	23.7587	15.3998	18.6296
AGE	1,555	15.5331	6.7729	0	28	10	22
ROA	1,412	0.1032	2.0722	−0.6438	1.3855	−0.0053	0.1079
LEV	1,321	0.2905	0.2290	0	0.9787	0.0882	0.4329
BOARD	1,561	8.4433	2.3046	4	15	7	10
WOMEN	1,554	0.1218	0.1351	0	0.7777	0	0.2
INDEP	1,561	0.2373	0.1744	0	0.8	0.1111	0.3636

Panel B. Differences in earnings management and CSR disclosure score between companies with and without state ownership
	Mean for non-SOE companies	Mean for SOE companies	p-value
AEM	0.0459	0.0923	0.000***
REM	−0.3464	−0.3864	0.052***
CSR	14.326	28.709	0.000***
Observations	904	423

Note(s): Variable definitions are provided in Appendix B. *, ** and *** denote statistical significance at the 10%, 5% and 1% levels, respectively

Source(s): Author's own creation/work

Table 2

Correlation matrix

	Obs	TOBIN	AEM	REM	CSR	STATE	SIZE	BOARD	AGE	ROA	WOMEN	INDEP	LEV
TOBIN	1,275	1.0000
AEM	1,327	−0.0773***	1.0000
REM	1,539	−0.1317***	0.2423***	1.0000
CSR	1,539	0.2783***	−0.3254***	0.0885***	1.0000
STATE	1,552	0.1146***	0.1703**	0.0512**	0.4220***	1.0000
SIZE	1,326	0.5205***	−0.4384***	−0.0187	0.5021***	0.3061***	1.0000
BOARD	1,561	0.1544**	0.2901 ***	0.0458*	0.5321***	0.4142***	0.3846***	1.0000
AGE	1,555	0.1214***	−0.1147***	0.0512**	0.1010***	−0.0726***	0.1620***	0.0951***	1.0000
ROA	1,412	0.0072	−0.1152	−0.0280	−0.0215	−0.0096	−0.0708***	0.0280	−0.0435	1.0000
WOMEN	1,554	−0.1066***	0.1657***	−0.0631**	−0.1605***	−0.0504**	−0.1916***	−0.1369***	−0.0512**	−0.0105	1.0000
INDEP	1,561	0.0958***	−0.1054***	0.0531**	0.2622***	0.0942***	0.2416***	0.2577***	0.1351***	0.0039	−0.0416	1.0000
LEV	1,321	−0.0428	0.0245*	−0.0139	0.1076***	0.0686**	0.1105***	0.0572**	0.0007	0.0599**	−0.0582**	0.0300	1.0000

Source(s): Author's own creation/work

The table reports Pearson correlation coefficients between variables for the observations from 2012 to 2018. Variable definitions are provided in Appendix B. *, ** and *** denote statistical significance at the 10%, 5% and 1% levels, respectively

Table 3

Results of the empirical analysis

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
AEM	−0.012**			−0.009***			−0.016***
	(−2.03)			(−2.95)			(−3.17)
REM		−0.028***			−0.044**			−0.022*
		(−3.13)			(−2.04)			(−2.74)
CSR			0.146***	0.096*	0.129**		0.213***	0.140***
			(2.55)	(1.95)	(2.25)		(6.82)	(5.51)
STATE						0.083**	0.009**	0.036**
						(2.34)	(1.93)	(1.79)
AEM×CSR				0.095*			0.010***
				(1.80)			(2.90)
REM×CSR					0.147***			0.134***
					(3.76)			(2.22)
AEM×STATE							0.429**
							(2.36)
REM×STATE								0.345*
								(1.83)
CSR×STATE							0.041*	0.049*
							(1.83)	(1.69)
AEM×CSR×STATE							0.141***
							(3.39)
REM×CSR×STATE								0.019***
								(2.59)
SIZE	−0.029***	−0.030***	−0.075***	−0.041***	−0.076**	−0.013	−0.051***	−0.0401*
	(−2.74)	(−2.62)	(−2.79)	(−3.61)	(−2.51)	(−0.34)	(−6.99)	(−9.83)
AGE	−0.051	−0.018	−0.029	−0.049	−0.029	−0.025	−0.007	−0.005
	(−1.00)	(−1.00)	(−1.24)	(−0.78)	(−0.48)	(−0.72)	(−0.28)	(−0.18)
ROA	0.026***	0.027***	0.021***	0.026***	0.014***	0.025	0.028	0.018
	(4.13)	(3.13)	(4.65)	(3.34)	(2.68)	(0.67)	(1.14)	(0.86)
LEV	0.176***	0.169***	0.165***	0.017***	0.161***	0.179***	0.070***	0.075***
	(3.26)	(2.87)	(2.69)	(2.71)	(2.64)	(4.23)	(3.51)	(3.84)
BOARD	0.008**	−0.002*	−0.056**	−0.045***	−0.051***	0.017	−0.057***	−0.080
	(1.98)	(−1.67)	(−1.99)	(−2.67)	(−2.88)	(0.49)	(−2.23)	(−3.02)
WOMEN	−0.035*	−0.034*	−0.026*	−0.033	−0.027	−0.035	−0.013	−0.042***
	(−1.67)	(−1.69)	(−1.68)	(−0.75)	(−0.49)	(−0.98)	(−0.06)	(−2.10)
INDEP	−0.019	−0.022	−0.043	−0.046	−0.036	−0.027	−0.079*	−0.059
	(−0.31)	(−0.50)	(−1.37)	(−0.58)	(−0.80)	(−0.92)	(−3.35)	(−2.61)
Industry effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Adj. R-sq	0.134	0.112	0.150	0.174	0.168	0.138	0.341	0.362
Obs	1,239	1,290	1,287	1,237	1,287	1,290	1,203	1,252

Source(s): Author's own creation/work

The table reports the results of OLS models with fixed effects for year and industry for testing Hypotheses (1)-(4), standard errors are clustered by industry. The results reflect the effect of earnings management proxies (accrual-based (AEM) and real-based (REM)) on Tobin's Q variable as well as the moderating effect of CSR disclosure (CSR) and state ownership (STATE) in the above mentioned relationship. The variables are described in Appendix B. *, ** and *** indicate statistical significance at less than the 10%, 5% and 1% levels, respectively

Table 4

Robustness tests (1)

	(1)	(2)	(3)	(4)
EM_proxy	−0.037*			−0.214**
	(−1.77)			(−2.36)
CSR		0.146***		0.159***
		(3.84)		(4.03)
STATE			0.082**	0.135***
			(2.34)	(3.43)
EM_proxy×CSR				0.071**
				(2.33)
EM_proxy×STATE				0.306
				(1.03)
CSR×STATE				0.016
				(0.96)
EM_proxy×CSR×STATE				0.059*
				(1.92)
SIZE	−0.043	−0.075*	−0.013	−0.077*
	(−1.07)	(−1.82)	(−0.34)	(−1.69)
AGE	−0.030	−0.029	−0.024	−0.051
	(−0.83)	(−0.86)	(−0.23)	(−1.41)
ROA	0.035	0.021	0.024	0.018
	(0.93)	(0.59)	(0.67)	(0.51)
LEV	0.193***	0.165***	0.179***	0.184***
	(4.37)	(3.99)	(4.23)	(4.17)
BOARD	0.012	−0.055	0.016	−0.011
	(0.36)	(−1.65)	(0.49)	(−0.29)
WOMEN	−0.027	−0.025	−0.032	−0.019
	(−0.80)	(−0.77)	(−0.97)	(−0.29)
INDEP	−0.035	−0.044	−0.027	−0.051*
	(−1.16)	(−1.38)	(−0.92)	(−1.71)
Industry effects	Yes	Yes	Yes	Yes
Year effects	Yes	Yes	Yes	Yes
Observations	1,224	1,287	1,290	1,222
Adj. R-sq	0.042	0.045	0.038	0.064

Note(s): The table reports the results of OLS models with fixed effects for year and industry for Hypotheses (1)-(4) with a new proxy for earnings management based on Laux and Leuz (2009) and Leuz et al. (2003), standard errors are clustered by industry. The results reflect the effect of the new earnings management proxy (EM_proxy) on Tobin’s Q variable as well as the moderating effect of CSR disclosure (CSR) and state ownership (STATE) in the above mentioned relationship. The variables are described in Appendix B. *, ** and *** indicate statistical significance at less than the 10%, 5% and 1% levels, respectively

Source(s): Author’s own creation

Table 5

Robustness tests (2)

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Variables	TOBIN	TOBIN	TOBIN	TOBIN	TOBIN	TOBIN	CAP	CAP
AEM	−0.007**	−0.048*			−0.028*		−0.104***
	(−2.26)	(−1.73)			(−1.89)		(−2.61)
REM			−0.044*	−0.095***		−0.093***		−0.097**
			(−1.93)	(−3.61)		(−3.52)		(−2.31)
CSR_dummy	0.176***	0.180***	0.189***	0.179***	0.170***	0.191***	0.033**	0.087***
	(4.54)	(4.62)	(5.01)	(4.77)	(3.98)	(5.17)	(2.07)	(3.51)
STATE					0.036***	0.099**	0.019*	0.042*
					(2.88)	(1.99)	(1.95)	(1.87)
AEM×CSR_dummy		0.041*			0.074*		0.180***
		(1.84)			(1.69)		(2.99)
REM×CSR_dummy				0.095***		0.075***		0.081**
				(3.68)		(2.75)		(1.82)
AEM×STATE					0.543***		0.129
					(2.69)		(0.57)
REM×STATE						0.016		0.002
						(0.67)		(0.05)
CSR_dummy×STATE					0.204**	0.137***	0.063**	0.051*
					(2.11)	(2.58)	(1.99)	(1.78)
AEM×CSR_dummy×STATE					0.297*		0.354*
					(1.83)		(1.68)
REM×CSR_dummy×STATE						0.032**		0.008**
						(2.03)		(2.12)
SIZE	−0.086**	−0.086**	−0.092**	−0.083**	−0.074	−0.085*	0.514***	0.411***
	(−1.98)	(−1.96)	(−2.27)	(−2.07)	(−1.63)	(−2.04)	(8.69)	(10.00)
AGE	−0.039	−0.038	−0.028	−0.028	−0.062*	−0.032	0.003	0.009
	(−1.13)	(−1.09)	(−0.84)	(−0.83)	(−1.78)	(−0.96)	(0.11)	(0.32)
ROA	0.022	0.022	0.016	0.012	0.011	0.003	0.026	0.01
	(0.62)	(0.63)	(0.45)	(0.35)	(0.31)	(0.03)	(1.07)	(0.99)
LEV	0.175***	0.175***	0.164***	0.161***	0.169***	0.157***	−0.078***	−0.072***
	(4.08)	(4.08)	(3.89)	(3.84)	(4.02)	(3.78)	(−3.47)	(−3.62)
BOARD	−0.058*	−0.056	−0.064*	−0.059*	−0.058	−0.051	−0.069**	−0.069***
	(−1.67)	(−1.62)	(−1.84)	(−1.71)	(−1.61)	(−1.44)	(−1.98)	(−2.60)
WOMEN	−0.029	−0.028	−0.025	−0.022	−0.020	−0.024	−0.008	−0.042**
	(−0.87)	(−0.83)	(−0.75)	(−0.69)	(−0.60)	(−0.63)	(−0.39)	(−2.08)
INDEP	−0.045	−0.046	−0.037	−0.035	−0.040	−0.032	−0.083***	−0.051**
	(−1.59)	(−1.53)	(−1.29)	(−1.21)	(−1.35)	(−1.11)	(−3.26)	(−2.45)
Industry effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Observations	1,237	1,237	1,287	1,287	1,237	1,287	1,203	1,252
Adj. R-sq	0.057	0.057	0.056	0.065	0.097	0.086	0.329	0.355

Source(s): Author's own creation/work

The table reports the results of OLS models with fixed effects for year and industry with other proxies for CSR disclosure and firm's value; standard errors are clustered by industry. The variables are described in Appendix B. *, ** and *** indicate statistical significance at less than the 10%, 5% and 1% levels, respectively

Table 6

Empirical results of applying GMM method to test for endogeneity

	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
AEM	−0.024*			−0.018*			−0.048*
	(−1.67)			(−1.65)			(−2.33)
REM		−0.024*			−0.009*			−0.005*
		(−1.78)			(−1.84)			(−2.18)
CSR			0.213***	0.132***	0.123***		0.923***	0.088**
			(3.67)	(2.99)	(2.85)		(3.95)	(2.13)
STATE						0.027*	0.355*	0.032*
						(1.76)	(1.72)	(1.94)
AEM×CSR				0.072**			0.042*
				(2.39)			(1.94)
REM×CSR					0.083***			0.129
					(2.62)			(0.42)
AEM×STATE							0.006*
							(1.73)
REM×STATE								0.009*
								(2.03)
CSR×STATE							0.178*	0.035*
							(1.81)	(1.63)
AEM×CSR×STATE							0.819**
							(2.37)
REM×CSR×STATE								0.407**
								(2.67)
SIZE	−0.022	−0.095	−0.090	−0.054	−0.046	−0.049	−0.372	−0.143
	(−0.37)	(−1.08)	(−1.45)	(−0.92)	(−0.73)	(−0.75)	(−1.26)	(−0.29)
AGE	−0.004	−0.230	−0.134	−0.031	−0.036	−0.148	−0.577*	−0.026
	(−0.06)	(−1.19)	(−1.28)	(−0.56)	(−0.64)	(−1.39)	(−1.74)	(−0.06)
ROA	0.070**	0.041	0.043	0.066**	0.069**	0.039	0.083	0.037
	(2.30)	(0.99)	(1.15)	(2.23)	(2.36)	(1.08)	(0.88)	(0.29)
LEV	0.263***	0.192***	0.233***	0.248***	0.245***	0.246***	0.121**	0.125***
	(5.55)	(2.78)	(5.05)	(5.37)	(5.28)	(5.32)	(2.02)	(3.68)
BOARD	−0.028	0.051	−0.074	−0.071	−0.073	0.002	−0.551***	−0.682**
	(−0.55)	(0.55)	(−1.37)	(−1.42)	(−1.44)	(0.03)	(−2.57)	(−2.51)
WOMEN	−0.027	−0.304	−0.127	−0.012	−0.009	−0.193*	−0.066	0.601
	(−0.66)	(−1.32)	(−1.33)	(−0.30)	(−0.23)	(−1.82)	(−0.32)	(0.52)
INDEP	−0.028	0.133	−0.014	−0.038	−0.035	0.011	−0.031	−0.351
	(−0.75)	(1.04)	(−0.37)	(−1.23)	(−0.94)	(0.28)	(−1.54)	(−1.48)
Industry effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year effects	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Arellano–Bond test AR(1) (p–value)	0.001	0.001	0.000	0.000	0.000	0.001	0.003	0.004
Arellano–Bond test AR(2) (p–value)	0.189	0.180	0.158	0.123	0.126	0.157	0.133	0.127
Observations	1,239	1,290	1,287	1,237	1,237	1,290	1,237	1,237

Source(s): Author's own creation/work

The table reports the GMM regression results for testing Hypotheses (1)-(4), where the dependent variable is Tobin's Q. The results related to the relationship between earnings management proxies and Tobin's Q are presented in Columns (1)–(3). The moderating effect of CSR disclosure in this relationship is presented in Columns (4) and (5). Columns (6) to (8) represent the results of the empirical analysis for the effect of state ownership on the above mentioned relationship. The description of all the variables is represented in Appendix B. *, ** and *** indicate statistical significance at less than the 10%, 5% and 1% levels, respectively

Table A1

Comparison of CSR indices

Category	CSR index from Anas et al. (2015)	Nuanced CSR index constructed for Russia
Environment
Efficiently using energy	X	X
Reduce emissions	X	X
Using biofuels	X	X
Measures to protect flora and fauna	X	X
Total for Environment	4 items	4 items
Community
Contributions for children	X
Contributions for children in communities (kindergartens, schools, events for children under 18)		X
Contributions for employees' children		X
Contributions for disabled children		X
Contributions to youth development	X	X
Contributions for the underprivileged	X	X
Supporting employees' community involvement	X	X
Supporting education	X	X
Contributions to infrastructure development		X
Total for Communities	5 items	8 items
Workplace
Health and safety	X	X
Human rights	X	X
Gender issues–equal employment opportunities	X	X
Quality of work environment	X	X
Supporting retired employees		X
Total for Workplace	4 items	5 items
Marketplace
Supporting green products	X	X
Ethical procurement practices	X	X
Helping to develop suppliers and other vendors	X	X
Corporate governance (CG) standards	X
CG standards and practices obligatory		X
CG standards and practices voluntary		X
Total for Marketplace	4 items	5 items
Overall CSR index	17 items	22 items

Source(s): Author's own creation/work

Table A2

Variable definitions (data sources are SKRIN and Datastream)

Notation	Variable	Definition
TOBIN	Tobin's Q	The sum of the market value of equity (share price multiplied by the number of ordinary shares issued at the end of a fiscal year) and the book value of debt divided by the book value of total assets
CAP	Market capitalization	The item represents the total market value of the company based on the year end price and number of shares outstanding
CSR	CSR disclosure index	CSR disclosure comprising 22 items based on information extracted from annual reports
AEM	Proxy for accrual-based earnings management	A measure of discretionary accruals based on the modified Jones model (Dechow, 1994; Jones, 1991)
REM	Proxy for real-based earnings management	A combined measure of real earnings management calculated by aggregating three individual real activities manipulation proxies: abnormal cash flow from operations (ABs_CFO), abnormal production cost (AB_PROD), and abnormal discretionary expenditure (AB_EXP). The approach is based on Roychowdhury (2006) and Kim et al. (2012)
EM_proxy	Proxy for accrual-based earnings management	An alternative proxy for earnings management reflecting how insiders use their discretion to smooth reported earnings to manipulate the volatility of operating earnings with respect to the original cash flow. The approach is based on Laux and Leuz (2009) and Leuz et al. (2003)
SOE_share	State ownership	The share of state ownership in total equity
STATE	State ownership	A dummy variable equal to 1 if a company has state ownership, and 0 otherwise
SIZE	Firm size	Natural log of total assets
ROA	Return on assets	Net income divided by total assets at the beginning of the year
LEV	Leverage	Total debt divided by total assets
BOARD	Board size	Natural logarithm of the amount of board members
WOMEN	Women on board	Share of women on board of directors
INDEP	Independent board members	Share of independent board members

Source(s): Author's own creation/work