What Earnings Management Has to Do with Corporate Social Responsibility

Abstract

With the continuing rise of attention towards societal challenges like, e.g., climate change, Corporate Social Responsibility (CSR) becomes an increasingly important topic for companies. While there is no question that CSR activities are on the rise, the connection towards Earnings Management in companies is less clear. Therefore, this research paper not only aims to provide an up-to-date picture on the literature addressing this interconnection, but also provides a profound base for a more solid theoretical framework. Thus, it delivers a critical basis for further empirical analyses in this field. In order to illustrate this interconnection between those two topics, this paper presents an SLR analysis of articles published in the Chartered Association of Business Schools (CABS) or Australian Business Deans Council (ABDC), focusing on empirical analyses of CSR performance and Earnings Management. Overall, it can be stated that CSR performance has a negative influence on Accrual-Based Earnings Management, while findings on the influence on Real Earnings Management are contradicting. Furthermore, the relationship of CSR performance and Accrual-Based Earnings Management/Real Earnings Management is especially vague when the used methods are moderated by different variables such as family involvement or managerial entrenchment. While the connection between Accrual-Based Earnings Management and CSR performance is widely covered in the existing literature, the relationship between Real Earnings Management and CSR performance is clearly less outlined. This research paper makes key contributions to the existing literature, as it combines and structures results of conducted studies during the last ten years and elaborates on the differences on commonalities of the results. This analysis also suggests that other factors that possibly influence Earnings Management or CSR should be included in a future research model for upcoming analyses. It places the findings of earlier studies into the context of the Ethical Approach, creating a roadmap for the future.

1. Introduction

Financial statement fraud has always been one of the most critical issues drawing public attention to companies. Fraud, in this context, is straightforward: it is illegal. However, there are also fewer clear-cut issues to assess, such as what constitutes Earnings Management. While Earnings Management can distort a company’s balance sheet, it is not explicitly forbidden. Nevertheless, Earnings Management remains a highly controversial topic in financial accounting and financial management. This leads to another essential and legitimate question: does it impact other areas, such as Corporate Social Responsibility (CSR) performance? In this regard, perspective is crucial, especially in the context of corporate governance, where the relationship is particularly relevant. The potential correlation gives rise to risks that need addressing to ensure both a strong CSR performance and the avoidance of Earnings Management activities in presenting financial results.

Earnings Management is often referred to as a tool for smoothing earnings and presenting a company’s financial figures in a more favorable light [1,2,3]. It is a concept that consciously pushes the boundaries of what is permissible, but does not exceed them. The existence of these options within accounting regulations provides companies with the opportunity to value certain transactions in favor of their interests. However, these options entail the risk that the balance sheets may not objectively represent the current financial situation. In short, it is a design option that stirs controversy and generates more heated debate than any other instrument in financial management.

Similarly, climate change is one of the most dominant and significant challenges in today’s social discourse, significantly impacting companies’ business models. This development also means that stakeholders are exerting increasing pressure on companies to enhance their CSR performance [4]. On the one hand, publications have shown that the use of Earnings Management techniques can be reliably proven in most cases [5]. However, in practice, reliability is not enough for management levels of concerned companies, especially if it is too late to take countermeasures, and the impact is unknown. On the other hand, only a few studies have addressed the question of how closely Earnings Management and CSR are related; for example, Kim et al. [6] or Velte [7]. There are multiple interesting hypotheses emerging from this field that still need to be analyzed.

2. Theoretical Background

2.1. Earnings Management in the Non-Financial Industry

The definitions among the literature for the term Earnings Management are wide [8]. They all can be summarized as techniques meant to influence the view of the stakeholders on the financial statements of a company [1]. Since Earnings Management distorts the financial statement by stressing out the boundaries of the implemented accounting principles, it is often seen as unethical behavior [3,6,9]. For a better understanding of the importance of Earnings Management, it is crucial to review the past milestones accomplished in assessing this field of research.

When talking about Earnings Management, it is necessary to distinguish between two forms [10]: Accrual-Based Earnings and Real Earnings Management. Real Earnings Management describes activities put in place during a financial year with the aim of improving performance [10,11]. An example of Real Earnings Management is, e.g., the shortening of budgets in the field of R&D. The main threat of Real Earnings Management for the company is damage to the firm value in the long term so that short-term profits can be achieved [10,11]. Research tries to detect Real Earnings Management practices [6,7,12,13] by relying on a model introduced by Roychowdhury [11]. The model is developed to suit the non-financial industries [11], and is based on the revenues of a company, controlling operative cashflow, the production costs and discretionary expenditures.

However, Accrual-Based Earnings Management describes activities performed in the financial statement sheets, with the aim to present a financial performance that does not match reality. The literature has considered the concept of Accrual-Based Earnings Management from two main perspectives [14]: the opportunistic perspective and the signaling perspective. Researchers driving the development of detection models and explanations for the usage of Earnings Management, e.g., Healy [15], Schipper [16] and Jones [17], approached the area from an opportunistic perspective. Healy stated that managers implement discretionary accounting to maximize their private benefit at the expense of the company and its stakeholders [15]. To identify these activities, he established a model, the Healy Model, that detects the Discretionary Accruals (DA) as the difference between the total accruals (TA) and the Non-Discretionary Accruals (NDA), while taking the arbitrary chances in accounting rules into account [18]. One main weakness of the model is considered to be that accruals are reversed over years leading to a false classification of NDA as DA [19]. In 1991, Jones developed another methodology to detect Accrual-Based Earnings Management. His model is concepted as a time-series model while developing company specific expectations in form of a cross-sectional analysis [17]. The Jones Model was improved by Dechow et al. [20]. This modified Jones Model adjusts the original model by taking the control of revenue into account by systemizing it with the change in receivables during the event period [20]. This model was found to be the most accurate and effective method to detect Accrual-Based Earnings Management in the literature [10,21,22,23]. The latest improvement to the model was made by Kothari et al. [24]. They adjusted the model by considering the corporations performance, arguing that a growing company has a different level of NDA than a non-growing company, leading to a misclassification [24]. They proved their adjusted model to increase the reliability of the analysis performed [24]. Until today, the applied models to detect Earnings Management behavior in companies are based on the modified Jones Model.

2.2. Earnings Management in the Financial Industry

The Earnings Management activities of management are not limited to the non-financial sector. Within the financial sector, there are various opportunities, incentives and pressure situations in which banks decide to use the earnings management tool. This is shown in studies by, e.g., Dantas et al. [25], Flannery et al. [26] or Barth et al. [27]. Forms of Earnings Management in the financial industry include loan loss provisions, realization of gains or losses on securities or pension settlement transactions [26,28], or the focus on short-term performance in the fund management decisions [29]. In their analysis, [25] focus on the ability of banks to achieve income smoothing by influencing loan loss provisions. This area is susceptible to Earnings Management, as the quality of the loan portfolio is characterized by a high degree of information asymmetry [25], making external monitoring difficult. In addition, the creation of a loss allowance is generally associated with considerable discretionary scope.

Along with companies in the non-financial sector, banks are under increasing pressure to improve their CSR performance. This pressure is evident, for instance, in the expectation of holding banks accountable for their lending practices [30]. However, banks striving for high CSR performance face a dilemma, as studies indicate that those with the highest CSR performance tend to achieve lower short-term financial performance [30,31]. These tensions may serve as an incentive for engaging in Earnings Management practices.

On the contrary, banks, positioned as influential stakeholders, wield substantial leverage over the prioritization of Corporate Social Responsibility (CSR) performance by their clientele [30]. This influence is noticeably manifested in loan contractual terms, wherein interest rates are intricately linked to the Environmental, Social, and Governance (ESG) rating attributed to the borrower. Consequently, the augmentation of CSR performance becomes inherently more appealing, given its direct correlation with a diminished burden of debt costs.

Due to the disparities in both accounting regulations and stakeholder structures, it is unsuitable to compare earnings management activities between the financial and non-financial sectors. Therefore, including studies from both sectors in this paper would not be purposeful. As a consequence, the systematic literature review (SLR) was designed exclusively for the non-financial sector.

2.3. Corporate Social Responsibility

CSR describes a pattern of social and ethical standards that a company sets for itself [32], while there is no universal definition of CSR [33,34]. However, the various definitions have one core statement in common: CSR represents the demands of society on an enterprise, which go beyond the generation of profit and growth of the economy.

Although the concept of sustainable development is vague, studies in political science such as Manioudis and Meramveliotakis [35] show that sustainability can and must be viewed from several perspectives. Manioudis and Meramveliotakis [35] argue that the common meaning of sustainability is perceived too dominantly in the context of the profitability of capital markets, and rather needs to be considered as a developmental stage in society from a sociological perspective. This interdisciplinary view can explain the increasing pressure from stakeholders on companies to assess their sustainability performance.

Thus, CSR activities can be described as the actions a company engages in to fulfill the demands of society and stakeholders. As a part of this, a company should also disclose the activities implemented to their stakeholders.

CSR is part of the managerial discussion on performance increasing through behaving in a good way [36]. The ISO 26000 “Guidelines for Social Responsibility” [37] is one major initiative to generate standards and references that define CSR activities for companies. The ISO guideline provides a set of categories of CSR: organizational governance, human rights, labor practices, environment, fair operating practices, consumer issues and community involvement. A commitment to those CSR categories is shown to improve the long-term economic benefits as well as the social benefits [14]. Environmental social governance (ESG) is sometimes described as a part of CSR (e.g., [38]). The timeliness of a well-founded reporting of own CSR measures is also supported by several developments of standards by the US-Security Exchange Commission (SEC), the European Financial Reporting Advisory Group (EFRAG) as well as of the International Sustainability Standards Board (ISSB).

The desire to show a companies’ commitment to social and environmental problems is one major part of CSR efforts that a management takes [39]. The implementation of CSR activities, combined with the right depiction to stakeholders, not only results in a company meeting the demands of their stakeholders, but also in building a strong reputation in the public perspective [40].

In terms of CSR disclosure, it is crucial to differentiate between mandatory disclosure and voluntary disclosure. Zang [10] defines voluntary CSR disclosure as a reporting of information that is not required by the relevant government, but is still performed in accordance with the guidelines for the disclosure of sustainability information issued by the Global Reporting Initiative.

Various performance measurements to evaluate CSR measures of a company exist. On the one hand, these methods aim to assess whether the activities actually address the demands of the stakeholders. On the other hand, they are intended to provide an opinion as to whether the company realizes the activities that it declares as being implemented. The extent to which a company is involved in the right activities to address the demanded topics can be considered as a definition of CSR performance [6,7,41].

Various approaches are used in the literature to measure CSR performance. Many researchers use the scores published as the Kinder, Lydenberg, and Domini social performance index (KLD), or MSCI [42] if the research is performed in US companies (e.g., [3,6,43]). Dimitropoulos [44] and Garcia-Sanchez et al. [41] set ESG equal to CSR and measure CSR performance through the ESG score published by Refinitiv Environmental. However, there is no universal method to measure CSR performance yet. Environmental, social and governance (ESG) performance can be seen as a part of CSR, or an aggregate of CSR [7]. ESG performance focuses on the core elements of CSR, and is measured mainly through similar approaches. Velte [7] uses the KLD database to measure ESG activities.

3. Systematic Literature Review

The systematic literature review (SLR) design in this paper follows the approach of Brereton et al. [45], and includes the findings of Hiebl [46] of his review of published SLRs. Hereby, an SLR follows a clear and systematical approach by formulating research questions. These questions are formulated in the planning phase, and are then applied to the literature reviewed to identify the relevant research from published articles [47].

• Does good CSR performance reduce the activities of Earnings Management?
• Is CSR reporting used to cover up the engagement in Earnings Management activities?
• What are the important factors that moderate the relationship between Earnings Management and CSR performance?
• What are the important concerns and limitations in the existing literature?

Table 1. Overview of applied criteria.

Process	Inclusion Criterion	Exclusion Criterion
Database	Emerald
	Web of Science
	Scopus
	JSTOR
Identification	Journal Articles	Duplicates
Selection of Primary Results	Empirical Study
	Addressing Accrual-Based Earnings Management/Real Earnings Management
	Addressing CSR Performance
Quality Cut		ABDC Listing
		CABS Listing

Source: own illustration.

presents an overview of the steps performed during the identification process for the literature review. Four databases (Emerald, Web of Science, Scopus and JSTOR) were used to identify relevant studies. Green et al. [48] suggest using a minimum of two databases, whereas Hiebl [46] points out that the average number of databases used for a SLR is three. Relevant studies were selected by using the search string (‘Earnings Management’ AND ‘CSR performance’) OR (‘Earnings Management’ AND ‘ESG Performance’) OR (‘Earnings Management’ AND ‘Corporate Social Responsibility Performance’) OR (‘Earnings Management’ AND ‘CSR credibility’). The outlined approach led to a total of 205 results, of which 40 results were found on JSTOR, 24 results on Scopus, 66 results on Web of Science and 71 results on Emerald.

In the first step, all duplicate studies were removed. A total of 33 duplicate articles were identified, leading to a sample size of 172 potentially relevant results. The next step was to remove all results not being journal articles. This resulted in an exclusion of nine results. After removing duplicates and book chapters, the sample size for the selection of primary studies was narrowed down to 163 articles. The selection of primary studies was conducted in three stages. In the first step, the abstracts of the identified studies were reviewed to assess their relevance. In the review of the abstracts, the following criteria had to be met to be classified as a relevant study: The study must address either Accrual-Based Earnings Management or Earnings Management, as well as CSR performance. The CSR performance needs to be addressed specifically, meaning the article needs to state in which way the performance of CSR is addressed. The review of abstracts led to an exclusion of 78 articles. Nine out of these articles were literature reviews. The reviews did not address the relationship between Earnings Management and CSR performance; for example, Sofian et al. [49] did address the relationship between CSR and Earnings Management, but laid the focus on CSR reporting and did not address CSR performance in detail. The abstract review resulted in a total of 75 articles remaining for analysis. For these articles, a quality assessment was performed in the second step.

This quality assessment of the articles integrated in the SLR, according to prevailing opinion, is a decisive factor for a high-quality analysis [47,50,51,52]. However, there are no universal rules and guidelines to assess the quality of the literature included [46,47]. Calabro et al. [53] argue, for example, that the quality of the articles has already been addressed by including only articles from peer-reviewed journals that have already gone through a process of validation and quality checks. This study evaluates the quality in two steps. At first, the argument of Calabro et al. [53] was addressed by reviewing the remaining results, namely whether they were published in peer-reviewed journals. For the selection of journals to be included, the approach of Budhwar et al. [54] was followed, by using two different rankings: The ABDC Index Journal list by the Australian Business Deans Council’s and the Ranking by the Chartered Association of Business Schools (CABS). The journal articles are included only if they are listed either in the ABCD Index or CABS Ranking. All listed journals were included, as otherwise the sample size would be significantly below 20. Snyder [55] and Witell et al. [56] argues that limiting the results too much could affect the informative value of the analysis. In the final step of the quality assessment, the articles were evaluated based on the following considerations: whether the purpose of the analysis is accurately described, including context and approach, whether the results of the analysis are discussed and enclosed, and where the limitations are properly disclosed and discussed. The quality cut leaves a sample selection of 27 articles.

As last step, the accessibility of the remaining articles was checked, leading to an further exclusion of two studies, so the final sample includes 25 articles to be analyzed. As backward verification that all articles potentially relevant to this study were addressed by the SLR, a backward search was performed using the selected literature. However, the backward search did not reveal any further articles to be included in this analysis.

One article, a meta-analysis performed by Shi et al. [57] is part of the relevant results. However, because of the highly different design of their study compared to the empirical articles, this result has been analyzed separately from the other results.

Thus, a total of 24 empirical articles are included in the analysis.

The articles were analyzed using a review protocol developed to address the research question, including common parameters. To enable clustering into different risk factors, parameters such as the country of interest, the measurement of CSR and Earnings Management, but also the moderating factors and control variables were collected. The results of the analysis of the selected articles are presented in

Table 2. Systematic literature review.

Study	Country of Interest	CSR Measure	Earnings Management Measure	Theory	Moderator	Main Findings	Main Limitation
Almahrog et al. [9]	UK	Manual measurement of CSR	Accrual-Based Earnings Management: mod. Performance-based Jones Model by Kothari et al. [24])	Stakeholder Theory	None	-companies with a higher score of CSR engage less in Accrual-Based Earnings Management -No relation between sub-categories Customer and Other CSR Activities and Accrual-Based Earnings Management	none stated
Amar and Chakroun [14]	France	Manual measurement of CSR	Accrual-Based Earnings Management: CFO adj. Mod Jones Model	Stakeholder Theory	Corporate Governance	-CSR negatively influences Accrual-Based Earnings Management -good Corporate Governance can reduce Accrual-Based Earnings Management activities	none stated
Litt et al. [58]	USA	KLD score	Accrual-Based Earnings Management: mod. Performance-based Jones Model by Kothari et al. [24]	Multi Theoretical Approach	None	-companies with greater environmental initiatives engage less in Accrual-Based Earnings Management -notably strong pollution prevention programs and firm’s commitment to climate friendly practices are significant negatively associated with Accrual-Based Earnings Management -sub-sample for small companies losses significance for relationship between CSR and Accrual-Based Earnings Management	unable to identify the motives for companies’ engagement in environmental initiatives
Bose and Yu [59]	USA	Adj. KLD score	Accrual-Based Earnings Management: mod. Adj. Jones Model by Kothari et al. [24]	Agency Theory and Ethical Theory (managerial opportunism vs. transparent reporting)	None	-changes in CSR performance do not cause changes in earnings quality -Increase (decrease) in earnings quality causes an increase (decrease) in CSR performance -no evidence that an increase (decrease) in CSR performance also causes an increase (decrease) in earnings quality	-CSR performance is sticky over timer -cannot establish a definite cause an effect between CSR performance and earnings quality
Chen and Hung [60]	Taiwan	CSRI	Accrual-Based Earnings Management: adj. Mod Jones Model by Kothari et al. [24] Real Earnings Management: Roychowdhury (2006)	Agency Theory	Firm Value	-Fulfillment of CSR helps to prevent managers from engaging in Real Earnings Management and reduces losses in shareholder equity -promotion of environment protection, other CSR matters and CSR related authentication reduce Accrual-Based Earnings Management activities -public welfare and information transparency of CSR do not have a significant impact on Accrual-Based Earnings Management	none stated
Chen et al. [43]	USA	KLD score	Income Smoothing	Supplier-Customer Theory	Earnings Management	-companies with better CSR performance are associated with a lower level of income smoothness	none stated
Dimitropoulos [44]	EU Countries	ESG score provided by Refinitiv environmental	Accrual-Based Earnings Management: mod Jones Model Income Smoothing: Spearman correlation	Multi-Theory Approach	None	-different CSR variables create different Earnings Management motives -companies engaging in CSR activities are associated with lower Accrual-Based Earnings Management	-only includes well-developed European economies
Garcia-Sanchez et al. [41]	Multi County (USA; EMEA, Asia)	ESG score provided by Refinitiv environmental	Accrual-Based Earnings Management: mod Jones Model Dechow et al. [20] Real Earnings Management: Roychowdhury [11]	Agency Theory	Managerial Entrenchment	-CSR performance negatively related to Earnings Management, excepts when CSR interacts with managerial entrenchment -CSR performance does not significantly explain changes in Real Earnings Management -management entrenchment index positively related to Real Earnings Management	-only includes listed companies -Earnings Management measures ignore the differences in accounting standards at country level
Gerged et al. [61]	Kuwaiti	EDI	Accrual-Based Earnings Management: mod. Performance-based Jones Model (Kothari et al., 2005)	Multi Theoretical Approach	None	-companies with a high level of CSR are negatively related to Accrual-Based Earnings Management	-only 60 companies in the sample
Hwang [62]	USA	KLD score	Probity Model for the likelihood of disclosing non-GAAP earnings)	Multi Theoretical Approach	Accrual Management	-companies with a good CSR performance non-GAAP reporting decisions are less likely to be affected by their ability to manage accruals	none stated
Jian et al. [63]	China	Hexun.com database	Accrual-Based Earnings Management: revised Jones Model by Dechow et al. [20]	Legitimacy Theory, Reputational Capital Theory, Information Transfer Theory	-Firm performance -Family involvement -State ownership	-CSR scores of listed companies in China generally low -Companies will increase CSR investment while implementing Earnings Management -Family companies focus more on the socio-emotional wealth, thus engage less in Earnings Management -Incentives can restrain opportunistic behavior like CSR investment caused by Earnings Management -Companies with fund ownership and companies with higher ownership concentration engage less in Earnings Management	-Study does no address Real Earnings Management -the appear only studies A-share listed companies -The study only includes Chinese companies
Jordaan et al. [64]	South Africa	JSE SRI Index	Accrual-Based Earnings Management: adj. Mod. Jones Model Real Earnings Management: Roychowdhury [11]	Multi Theoretical Approach	None	-companies with better CSR performance are more likely to engage in Accrual-Based Earnings Management through income increasing DA -companies with better CSR performance are less likely to engage in Real Earnings Management	time period covered rather short
Kim et al. [6]	USA	KLD score	Accrual-Based Earnings Management: cross sectional adjusted mod. Jones Model Real Earnings Management: Roychowdhury [11]	Ethical Theory	Accounting and Auditing Enforcement Releases (AAERs)	-CSR companies are less likely than non-CSR companies to use Accrual-Based Earnings Management -CSR companies are less likely than non-CSR companies to use Real Earnings Management -companies substitute Accrual-Based Earnings Management and Real Earnings Management -CSR companies are less likely than non-CSR companies to be subject to SEC enforcement actions against their executives	none stated
Liu et al. [3]	USA	KLD score	Accrual-Based Earnings Management: adj. Mod Jones Model by Kothari et al. [24] Real Earnings Management: Roychowdhury [11]	Multi-Theory Approach	Family involvement	-level of Accrual-Based Earnings Management activities is not sign. Different with lower or higher overall CSR performance -neither family involvement not CSR performance is shown to have a significant influence on Real Earnings Management	-focus on S&P 500
Lu and Steven [65]	USA	KLD score	Financial Performance	Multi Theoretical Approach	None	companies with good CSR may not explicitly engage in Earnings Management, they may engage in a more subtle form of impressions management, i.e., Adjusting the tone used in their disclosures	none stated
Martinez-Martinez et al. [66]	Spain	Questionnaire	Accrual-Based Earnings Management: Jones Model [67]	Multi Theoretical Approach	SME companies	-SME companies with good CSR performance are less engaged in Accrual-Based Earnings Management -most SMEs engaged in Accrual-Based Earnings Management are younger, smaller and in industries with deceasing demand	-single country analysis -results limited to SME companies in Spain
Mohammad et al. [68]	Iran	CDI Variable (GRI Framework level)	Adjusted Modified Jones Model	Multi Theoretical Approach	None	-industries and companies that exploit the environment the most in their production have the highest environmental disclosure quality and quantity -significant negative relationship between EDQ and Accrual-Based Earnings Management	-unable to capture cross-country effect -Iran lacks a reliable database
Rahman and Zheng [69]	China	unweighted scoring method by Sareed et al. (2019	Accrual-Based Earnings Management: mod. Performance-based Jones Model by Kothari et al. [24] Real Earnings Management: Roychowdhury [11]	Multi Theoretical Approach	Family Involvement	-family companies in engage less in Accrual-Based Earnings Management and Real Earnings Management than non-family companies -CSR performance has no impact on Accrual-Based Earnings Management in family-companies -CSR performance has no impact on Real Earnings Management in family-companies	none stated
Rezaee and Tuo [70]	USA	CDI Variable (GRI Framework level)	Accrual-Based Earnings Management: mod Jones Model	Multi Theoretical Approach	None	-sustainability disclosure is associated with more Accrual-Based Earnings Management due to more managerial discretion -disclosure quality is significantly negatively correlated with Accrual-Based Earnings Management, leading to less Accrual-Based Earnings Management activities	none stated
Sial et al. [13]	China	RKS ranking	Accrual-Based Earnings Management: mod. Jones Model Real Earnings Management: Roychowdhury [11]	Stakeholder Theory	Earnings Management	-Earnings Management moderates the relationship between CSR and Firm Performance negatively	-CSR score is, on average, very low
Velte [38]	Germany	ESG Score provided by ESG-Asset4	Accrual-Based Earnings Management: mod. Performance-based Jones Model by Kothari et al. [24] Real Earnings Management: Roychowdhury [64}	Stakeholder Theory	None	-ESG performance score is negatively related to Accrual-Based Earnings Management activities -no significant effect of ESG performance on Real Earnings Management -bidirectional relationship between ESG performance and Accrual-Based Earnings Management	-limited time period covered -does not include other EQ measures, such as smoothing earnings -endogeneity concerns could not be totally neglected
Velte [7]	EU Countries	Environmental and Carbon performance Score from Thomsen Reuters Eikon database	Accrual-Based Earnings Management: mod. Performance-based Jones Model by Kothari et al. [24] Real Earnings Management: Roychowdhury [11]	Agency Theory	None	-companies with higher env. Performance have a lower amount of Accrual-Based Earnings Management -companies with higher env. Performance have a higher amount of Real Earnings Management -managers tend to opportunistically shift from Accrual-Based Earnings Management to Real Earnings Management	-comparability of carbon perf. Between specific companies and/or branches low due to variables in business practices -short time period captured (2014–2018)
Wang and Kangtao [71]	China	China Securities Market and Accounting Research index	Accrual-Based Earnings Management: mod. Performance-based Jones Model by Kothari et al. [24] mod. Growth-based Jones Model	Multi Theoretical Approach	None	-companies mandated to disclose CSR activities decrease Accrual-Based Earnings Management after mandatory CSR regulation becomes effective -impact of mandatory disclosure on Accrual-Based Earnings Management is only significant for companies with less analyst coverage	-exclusion of voluntary disclosure -companies with low CSR score may still exhibit more Accrual-Based Earnings Management than high CSR score companies
Zhang et al. [72]	China	CDI Variable (GRI Framework level)	Accrual-Based Earnings Management: adj. Mod. Jones Model by Jordaan et al. [64] Real Earnings Management: Roychowdhury [11]	Multi-Theory Approach	None	-better voluntary disclosure increases income increasing Accrual-Based Earnings Management -Companies with better disclosure are less likely to engage in Real Earnings Management -better CSR performance leads to less engagement in Accrual-Based Earnings Management -better CSR performance increases likelihood to engage in Real Earnings Management	-CSR disclosure still in developing stage -sample size limited to companies that practice voluntary disclosure -study does not consider impacts of government policies and social development indices on the firm Earnings Management motives

.

4. Findings and Discussion

4.1. Descriptive Statistics

For an overview of the articles evaluated in this paper, the descriptive statistics for the core factors of analysis follow first. The studies reviewed were all published within the last ten years. Of these, 21 of the 24 articles were published in the past five years, with the following distribution: two papers were published in 2023, three studies were from 2022, four were published in 2021, two articles were from 2020, and five papers were from 2019 and 2018 each. The remaining three studies were published in 2017, 2014 and 2012. This indicates that the field of research addressing the impact of Earnings Management on CSR and vice versa still seems to be relatively young.

Due to the recent publication years, the citation frequency is unsurprisingly very low for some studies. The article by Kim et al. [6] was cited most, with a total of 1684 citations, which is also the oldest article in the review. The average citation frequency of the articles included is below 50, with 12 out of the 22 articles. Two articles were not cited in other articles at all [62,69].

The time period covered in the reviewed articles was above 10 years in 10/24 articles (42%). The longest period was covered by Chen et al. (2019) from 1991 to 2013 (23 years), followed by [6], who used a sample of data from 19 years (1991–2004) and Dimitropoulos [44] capping 16 years (2003–2018). Nine articles reviewed (38%) covered an observation period between five and ten years, whereof three had a coverage of eight years [3,38,72] and four covered observations of five years [7,14,60,61]. Three articles enclosed really short time periods. Lu and Steven [65] used data of the years 2012–2013, Martinez-Martinez et al. [66] only used information of 2010, whereas Jordaan et al. [64] used data not of specific a period, but of single years: 2008, 2011 and 2013.

The empirical analysis focused on 21 out of the 24 articles on single-country studies, and three articles used a multi-country database. Out of the single country studies, five analyzed Chinese companies (20.8%), and eight studies used a database containing US companies (33.3%). The remaining papers focused on different countries, e.g., UK, France, Iran. Two of the three multi-country analyses used a database of EU countries.

In terms of hypothesis development, most analyses are based upon the same—or very similar—hypotheses. An overview of the hypotheses relevant to this SLR is given in

Table 3. Overview of the hypotheses.

Study	Relevant Hypothesis
Almahrog et al. [9]	H1: There is a negative relationship between the level of Earnings Management and CSR H1a: There is a negative relationship between the level of Earnings Management and community sub-score H1b: There is a negative relationship between the level of Earnings Management and employee sub-score H1c: There is a negative relationship between the level of Earnings Management and environment sub-score H1d: There is a negative relationship between the level of Earnings Management and services sub-score H1e: There is a negative relationship between the level of Earnings Management and customer sub-score H1f: There is a negative relationship between the level of Earnings Management and others sub-score
Amar and Chakroun [14]	H1: CSR has a negative impact on Earnings Management H1-1: Good corporate governance has a negative impact on earnings management H1-2: Respect for human rights has a negative impact on earnings management H1-3: Good labor relations and conditions have a negative impact on earnings management H1-4: Good environmental management has a positive impact on earnings management H1-5: The fairness of business practices has a negative impact on earnings management H1-6: Meeting consumer expectations has a negative influence on earnings management H1-7: The firm’s community involvement has a negative impact on earnings management
Litt et al. [58]	There is an inverse association between environmental incentives and Earnings Management
Bose and Yu [59]	H1: the current years CSR performance is determined by the prior year’s earnings quality, suggesting that changes in earnings quality cause changes in CSR performance. H1b: The current year’s earnings quality is determined by the prior year’s CSR performance, suggesting that changes in CSR performance cause changes in earnings quality
Chen and Hung [60]	CSR performance is positively associated with Earnings Management
Chen et al. [43]	Greater supply chain dependence motivates socially responsible companies to reduce income smoothing
Dimitropoulos [44]	CSR performance will have a negative impact on Earnings Management
Garcia-Sanchez et al. [41]	H1: with managerial entrenchment taken into consideration, CSR performance is positively related to Accrual-Based Earnings Management H2: with managerial entrenchment taken into consideration, CSR performance is positively related to Real Earnings Management
Gerged et al. [61]	Companies with higher level of CED are more likely to be less engaged in Earnings Management practices
Hwang [62]	The use of non-GAAP reporting when managers’ ability to manage accruals is constrained will differ between companies with good CSR performance and those with bad CSR performance
Jian et al. [63]	H1: The higher the degree of earnings management, the higher the social responsibility score H2: Compared with companies with better CSR performance, earnings management of companies with poor performance has a more significant positive impact on CSR score. H3: The positive effect of earnings management of non-family companies on CSR score is more significant than that of the family business. H4: Compared with state-owned companies, non-state-owned companies’ earnings management has a more significant positive impact on CSR score.
Jordaan et al. [64]	H1: companies with better CSR performance are less likely to manage earnings H2: companies with better CSR performance are more likely to manage earnings
Kim et al. [6]	H1: A CSR firm is less likely to engage in Earnings Management H2: A CSR firm is more likely to engage in Earnings Management
Liu et al. [3]	H1: companies with family involvement are more socially responsible than those without H2a: companies with family involvement engage less in Accrual-Based Earnings Management than those without H2b: Real Earnings Management is not systematically related to family involvement H3a: with family involvement taken into account, CSR performance is not systematically related to Accrual-Based Earnings Management H3b: with family involvement taken into account, CSR performance is not systematically related to Real Earnings Management
Lu and Steven [65]	CSR performance is associated with optimistic disclosure tone in earnings announcements, controlling for financial performance
Martinez-Martinez et al. [66]	H1a: the association between Earnings Management and CSR is negative H1b: the association between Earnings Management and CSR is positive H2: sector life cycle moderates the Earnings Management -CSR relationship
Mohammad et al. [68]	H1: companies with more environmental reporting have higher earnings quality H2: companies with more environmental reporting have lower earnings quality
Rahman and Zheng [69]	H1: in family companies, high CSR performance values are accompanied by low Earnings Management H2: in family companies, high CSR performance values are accompanied by high Earnings Management H3: Family ownership affects the relationship between Earnings Management and CSR
Rezaee and Tuo [70]	Positive association between ESG disclosure quality and Accrual-Based Earnings Management
Sial et al. [13]	Earnings Management negatively moderates the relationship between CSR and firm performance
Velte [38]	H1: ESG performance is linked with a decreased degree of Accrual-Based Earnings Management H2: ESG performance is linked with a decreased degree of Real Earnings Management
Velte [7]	H1: environmental performance is linked with a decreased amount of Accrual-Based Earnings Management H2: environmental performance is linked with an increased amount of Real Earnings Management
Wang and Kangtao [71]	Companies with mandatory CSR reporting requirement are less likely to engage in Earnings Management
Zhang et al. [72]	H1a: companies with better CSR performance are less likely to manage earnings H1b: companies with better CSR performance are more likely to manage earnings

Source: own illustration.

. Hypotheses that do not relate to the connection between Earnings Management and CSR performance are not listed. The analysis of Almahrog et al. [9] and Amar and Chakroun [14] formulated their hypothesis based upon the overall CSR performance, but also based upon the individual CSR dimensions. The other analysis focused on the overall CSR performance, thus testing the average CSR performance.

To analyze the impact on Earnings Management, 21 out of a total of 24 analyzed studies used the measurement of Accrual-Based Earnings Management, whereas nine studies used the measurement of Real Earnings Management and Accrual-Based Earnings Management to capture the effect of Earnings Management on CSR performance. One study measured the impact on Earnings Management through measuring the income smoothness of the analyzed companies. When analyzing Accrual-Based Earnings Management, most studies used the performance-adjusted modified Jones Model by Kothari et al. [24]. For measuring Real Earnings Management, all studies used the Roychowdhury Model [11]. For analyzing the results, this is crucial, as the comparability of the studies is high. While the method for measuring Earnings Management activities has found a consensus in the literature, the approaches to assess the CSR performance of companies differ clearly. Out of the sample, four studies used the KLD score to determine a company’s CSR performance, three used a measurement based on the application of the GRI framework and four studies used ESG scores provided for by different Thomsen Reuters databases. The other studies use either country specific providers for CSR performance or conducted a manual measurement.

For the measurement of CSR performance, 29.2% used the KLD Score provided by MSCI. Bose and Yu [59] were the only authors using an adjusted index by weighing the strengths and weaknesses of the individual scoring categories. With 16.7% usage, the method for measuring CSR performance using the ESG score provided by the Refinitiv database is the second most used measurement. With 12.5% usage each, the methods for measuring CSR performance through the GRI Framework Level, and individual assessment through own qualitative research, were equally often used. Seven articles used other methods to measure the performance. For example: Wang and Kangtao [71] used the China Securities Market and Accounting Research Index, whereas Sial et al. [13] used the RKS ranking, and Rahman and Zheng [69] measured the performance through the unweighted scoring method.

Shi et al. [57] included a sample of 35 papers and 51 independent studies in their meta-analysis. The higher sample compared to the results of this study are based on the inclusion of CSR disclosure in the meta-analysis. In total, 77% of the included studies used a CSR score as a measure, all included models used Accrual-Based Earnings Management, while only 40% used Real Earnings Management in their model when analyzing the relationship. For the analysis of the results, cultural differences and the time period covered were implemented as moderators [57].

4.2. CSR Performance Measurement

The ESG scoring systems are subject to ongoing refinement and updates as ESG criteria and best practices evolve. While it is a valuable tool for understanding a company’s sustainability and social responsibility, it is important to note that different ESG rating providers might have variations in their methodologies and emphasis on certain factors, leading to different scores for the same company. The methodology to calculate the CSR performance scores is unique for every database [73,74]. In order to interpret the results, it is important to consider the differences resulting from the methodology of the CSR scores, as well as potential limitations of databases.

The most used database, MSCI, published the KLD score. The methodology of the MSCI score uses a measurement of 33 Key Issues. A company’s score is calculated based upon two to seven of the Key Issues. The identification of pertinent Environmental and Social Key Issues for a specific company is determined by the company’s susceptibility to significant ESG risks, influenced by industry-specific and market-specific risks. The companies are ranked by comparing the companies in the respective industry.

Refinitiv comprehensively assesses ESG metrics across more than 630 measures at the company level. However, a focused subset of 186 measures is utilized, considered as the most relevant and comparable within specific industries. These metrics drive the overall evaluation and scoring process. These measures are categorized into 10 groups, thereby influencing the three core pillar scores and, ultimately, the aggregate ESG score. This composite score reflects the company’s commitment, effectiveness, and performance regarding ESG practices, all based on publicly available information. The final ESG score reflects these pillar scores, with environmental and social categories having variable weightings tailored to the specific industry. However, for governance, the weightings remain consistent across all industries.

The ESG score published by Refinitiv is subject to adjustments of historical data [75]. Thus, the usage of the ESG score allows only a snapshot interpretation at a specific point in time. The reliability of results is limited due to potential variations when using the same sample but collecting data at different points in time. Thus, the dynamic nature of data collection can yield differing outcomes, compromising result consistency and reliability.

4.3. Accrual-Based Earnings Management and CSR Performance

Interestingly, 22 out of the 24 review articles (92%) find a significant negative relationship between CSR performance and Accrual-Based Earnings Management. Only two studies [3,69] found no relationship between Accrual-Based Earnings Management and CSR performance of companies. When interpreting this result, it is important to mention that Liu et al. [3] used family involvement as a moderator of the relationship. They showed that family involvement is the main driver in the relationship between CSR performance and Accrual-Based Earnings Management, leading to no significant results regarding the overall relationship of CSR and Accrual-Based Earnings Management. Wang and Kangtao [71] found that the negative effect is only significant for larger companies, as their model loses its significance when using a sub-sample containing only small companies. Garcia-Sanchez et al. [41] showed that the negative relationship moves to a positive relationship when CSR performance interacts with management entrenchment. Bose and Yu [59] show that changes in CSR performance do not cause a shift in Earnings Management behavior, while an increase (decrease) in earnings quality does cause an increase (decrease) in CSR performance. The results indicate somewhat mixed evidence on the influence of CSR performance on Accrual-Based Earnings Management, but with the clear tendency of a negative relationship. This means that, on average, companies with a good CSR performance engage less in Accrual-Based Earnings Management compared to companies with worse CSR performance. The results also imply that there are several other factors that need to be controlled when interpreting the impact; most importantly, it shows a need for further theory development.

4.4. Real Earnings Management and CSR Performance

The results concerning the impact of Real Earnings Management on CSR performance are mixed. Ref. [6], Velte [7] and Zhang et al. [72] find a positive relationship between CSR performance and Real Earnings Management. They state that companies with a higher CSR performance also engage more in Real Earnings Management compared to non-CSR companies or companies without CSR disclosure. On the other hand, Chen and Hung [60], Kim et al. [6] and Jordaan et al. [64] found a negative relationship, stating that a fulfillment of CSR activities leads to a reduction in engagement in Real Earnings Management activities. However, Garcia-Sanchez et al. [41], Liu et al. [3] and Velte [38] could not identify a significant interaction of CSR performance and Real Earnings Management. Furthermore, Sial et al. [13] identified that a company’s Real Earnings Management activities in combination with a good CSR performance negatively affect a company’s firm value. The different findings make it necessary to take a closer look at the underlying factors of the analyses.

4.5. Moderating Factors

Approximately 33% of the sample studies were related to the USA. Out of these studies, five found a significantly negative correlation between CSR performance and Accrual-Based Earnings Management. Liu et al. [3] reported a non-significant relationship.

A total of 21% of the evaluated sample focus on China in their analyses. For Accrual-Based Earnings Management, all studies found a negative relationship with CSR performance. For Real Earnings Management, Zhang et al. [72] found a positive relationship, while Sial et al. [13] found a negative relationship, and Rahman and Zheng [69] found no relationship between CSR performance and Real Earnings Management. Only Wang and Kangtao [71] did not address Real Earnings Management in their study.

Out of the reviewed articles, three conducted a cross-country analysis. Garcia-Sanchez et al. [41] included companies belonging to the US, EMEA states and Asia in their approach. The articles of Dimitropoulos [44] and Velte [7] focused on companies listed in countries of the European Union. All of the three mentioned analyses found a negative relationship between CSR performance and Accrual-Based Earnings Management. Out of them, Garcia-Sanchez et al. [41] found no significant correlation for Real Earnings Management, whereas Velte [7] and Dimitropoulos [44] found a positive relationship between CSR performance and Real Earnings Management.

All studies, regardless of the country considered, determined either no effect, or a negative correlation between CSR performance and Real Earnings Management. It can therefore be concluded that this effect applies across countries and that a higher engagement in CSR leads to a reduction in the engagement of Accrual-Based Earnings Management on management’s side. In contrast, this statement cannot be made for the relationship between CSR performance and Real Earnings Management. On the one hand, this is due to the fact that only a few studies have integrated Real Earnings Management into their design. Above all, however, the studies that did include it in their testing produced different results. Although Shi et al. [57] interpret a cultural difference in the results of their analysis, this cannot be confirmed from the results of the analysis conducted in this article.

The study of Garcia-Sanchez et al. [41] used management entrenchment as a moderating variable when examining the relationship of CSR performance and Earnings Management. They found that the correlation of Accrual-Based Earnings Management and CSR performance switches from a significantly negative to a positive correlation when CSR performance interacts with management entrenchment. Regarding the relationship of Real Earnings Management and CSR performance, the authors could not determine a significant effect. Since the study is concepted as a cross-country analysis including USA, Asia and the EMEA states, the effect of country-specific effects on the results is already addressed within their research design. In contrast, Jian et al. [63] found a significant negative correlation between executive compensation and CSR performance. Thus, it cannot be ruled out that managers use CSR as a mask for their Earnings Management activities. This conclusion is in line with the disclosure of previous literature findings [6,38]. The results can also be interpreted in such a way that CSR performance is not necessarily presented correctly, but is distorted with greenwashing activities.

Liu et al. [3] employ their model to analyze the relationship between CSR performance and Earnings Management behavior of American companies by family involvement. They stated that family involvement was the key driver of this connection. When they controlled for family companies, they were not able to identify a significant correlation between the two variables. This indicates that the connection is moderated by many circumstances and influences. Jian et al. [63] controlled their model on Chinese companies for family involvement and found similar results. In addition, their analysis shows that fund ownership and a higher ownership concentration also have a significant negative relationship, leading to a lower engagement in Earnings Management while showing a strong CSR performance. When research tries to capture the influence of CSR performance on Earnings Management, it is crucial to consider additional factors. In contrast, Rahman and Zheng [69] detected no connection between family involvement and Accrual-Based Earnings Management. They also found that CSR performance has no effect on Real Earnings Management activities for family-companies. When analyzing CSR disclosure behavior, their results showed a lower disclosure for family-companies compared to non-family companies. In fact, they stated that family companies may not have a comparable need other than non-family companies to prove CSR activities to their shareholders.

Corporate Governance mechanisms are seen as a tool to empower transparency and to be suitable to restrain undesired behavior of the management [27]. Corporate governance has proven to have a significant negative correlation to Accrual-Based Earnings Management activities [27], and thus managers tend to use CSR performance as a masquerade for their engagement in Earnings Management activities more likely when corporate governance tends to be weak [41]. Surveys that included board independence as a control variable found a significant positive correlation to CSR performance [7] and a negative correlation to Accrual-Based Earnings Management and Real Earnings Management [38]. Not only board independence, but also board size tends to have an impact on the CSR performance of a company and the Earnings Management activities employed by management, in terms of a positive relation between board size and Accrual-Based Earnings Management, as well as Real Earnings Management activities performed [7]. A compensation system linked to CSR performance can be implemented to enhance efforts achieving good CSR performance without simultaneously increasing the incentives to pursue Earnings Management activities [7]. Findings relating to Corporate Governance moderating the impact of CSR on Earnings Management suggest that a good corporate governance system is elementary for preventing practices using CSR only as a diversion from undesirable actions.

As stated above, when analyzing the impact of CSR performance on engagement in Earnings Management activities, it is not enough to create a model that excludes other factors. So far, in order to measure the modelling effect of exogenous and endogenous influences, most of the literature reviewed added control variables to their models. It is interesting to note that the choice of control variables overlapped in many respects. For example, eight of the seventeen studies reviewed included the control variable ‘size’ in their research. The analysis of the relationship between size and Accrual-Based Earnings Management found contradictory results. In contrast, a lot of surveys found a positive correlation between size and Accrual-Based Earnings Management, like Alipour et al. [68], Garcia-Sanchez et al. [41], Almahrog et al. [9] or Dimitropoulos [44], indicating that large companies may resort to higher Earnings Management through accruals. Other studies [3,6,14,70] found size and Accrual-Based Earnings Management to be negatively correlated. These results clarify that it is not reliable to make assumptions based on the size of a company when assessing the risk of a company engaging in CSR activities to mask up their Accrual-Based Earnings Management behavior. Concerning Real Earnings Management engagement, Ref. [6], as well as Garcia-Sanchez et al. [41], detected a significant negative correlation, stating that larger companies are less likely to engage in Real Earnings Management activities compared to smaller companies. All covered studies that included Big Four as a controlling variable found that the audit being performed by one of the Big Four audit companies reduces the engagement in Accrual-Based Earnings Management activities [38,58,71].

5. Methodological Findings

5.1. Current Methodological Focus

In the reviewed articles, the majority approach the context from two different perspectives: the agency theory and the ethical theory. With the use of these two theories, hypotheses are formed that target the underlying mindset of the management.

Following agency theory, a strong CSR performance provides an opportunity for managers (agents) to behave opportunistically in their own best interest, not the stakeholders’ (principals’) interest [41,61]. Thus, CSR performance could be used as a concealing instrument to cover up the engagement in Earnings Management [6,41].

From the perspective of the ethical theory, a company having a strong CSR environment are believed to behave generally in the best interest of their stakeholders and to be focused on the long-term financial goals rather than short-term outcomes [6,44,72]. Following this theory, the management of a company with high CSR performance would also maintain a high quality of disclosures and would not present the financial situation distorted by Earnings Management techniques. When considering the ethical approach, it could be a benefit to also consider the threat that companies may engage in greenwashing activities when they engage in Earnings Management. Following the thought that companies engaging in Earnings Management activities cover up the true picture of their financial disclosure, they may also engage in greenwashing activities, covering up their true CSR engagement.

The articles mostly combine the two approaches via the development of opposing hypotheses. The possibility that management acts out of different motives is thereby reflected.

5.2. Fitting the Frame of Agency Theory

The agency theory describes the conflict between the management (agent) and the stakeholders (principals) arising from the separation of ownership and management, as well as an interest divergence between the two parties. Costs of this conflict (agency costs) result on the side of the principals through information asymmetry and adverse selection. The information asymmetry gives the agents the opportunity to behave in their own interest rather than in the interest of the shareholders and stakeholders.

Applying the model on the relationship of Earnings Management and CSR performance, the management faces the pressure to achieve a good CSR performance while maintaining a positive financial development. Thus, the management may use CSR performance and use the interest of their shareholders in this topic to cover up their engagement in Earnings Management [6,41].

The conflict of interest can be minimized by an adequate control structure upon the agent. In terms of Earnings Management and CSR performance, this could contain, for example, monitoring of the financial situation and CSR performance, or the creation of incentives to achieve a good CSR performance and, at the same time, avoiding the implementation of Earnings Management. A control structure focused on the right incentives could lead to an alignment of the interest of the agent and the principal.

The results of the analysis reviewed showed that the relationship between Earnings Management and CSR performance is not sufficiently explained by the measurement of those two factors in a nutshell. Rather, the results of the analysis of Wang and Kangtao [71] show that there is only a significant relationship for larger companies; García-Sánchez et al. [41] show that the moderator managerial entrenchment causes a shift of the relationship when CSR performance interacts with the moderator.

The assumption is that the management behaves in an ethical manner when the relationship between Earnings Management and CSR performance is negative, meaning a higher CSR performance leads to a lower engagement in Earnings Management activities, and could also be seen as a company with a well-functioning control structure minimizing their agency costs.

The results of the papers reviewed can be implemented within the framework of the agency theory, as shown in

Table 4. Control structure in the relationship of Earnings Management and CSR performance.

	Accrual-Based Earnings Management and CSR Performance		Real Earnings Management and CSR Performance
Study
	Weak	Strong	Weak	Strong
Almahrog et al. [9]		x
Amar and Chakroun [14]		x
Litt et al. [58]		x
Bose and Yu [59]		x
Chen and Hung [60]		x		x
Chen et al. [43]		x
Dimitropoulos [44]		x
Garcia-Sanchez et al. [41]	x		N/A
Gerged et al. [61]		x
Hwang [62]		x
Jian et al. [63]		x
Jordaan et al. [64]	x			x
Kim et al. [6]		x		x
Liu et al. [3]	N/A		N/A
Lu and Steven [65]		x
Martinez-Martinez et al. [66]		x
Mohammad et al. [68]		x
Rahman and Zheng [69]		x		x
Rezaee and Tuo [70]	x
Sial et al. [13]	x
Velte [38]		x	N/A
Velte [7]		x	x
Wang and Kangtao [71]		x
Zhang et al. [72]		x	x

Source: own illustration. Explanation: N/A: no significant correlation has been recorded by the authors; x: applicable form of morality approach.

. For this purpose, the results were assessed as to whether they imply a strong or weak control structure. In this context, Earnings Management is divided into Accrual-Based Earnings Management and Real Earnings Management. A positive correlation implies a structure where the interests of the agents and the principals are not sufficiently aligned and/or where the monitoring of Earnings Management activities could be improved. A negative relationship, on the other hand, implies a suitable control structure for minimizing the agency costs for Earnings Management and CSR performance. The classification is made under the assumption that the relationship has been addressed accurately by the performed analysis. In the case of analyses that used moderators, the results in which CSR performance or Earnings Management was controlled by the respective moderator were considered.

5.3. Fitting the Frame of Ethical Approaches

The articles reviewed that consider an ethical approach for their hypothesis development focus on the aspect of the management behaving in the best way for all their stakeholders, e.g., [6]. They state that management defines socially responsible behavior as an obligation within their moral principles [6].

When applying an ethical approach to the relationship of Earnings Management and CSR performance, it is important to define which perspective of ethics are considered. We implement the papers in the ethical framework of traditional morality versus market morality. Traditional morality refers to the behavior that takes place according to the concepts of obligations, duty, fairness, and care, whereas market morality is acting in its own best interest [67]. The classification of the results into these categories allows conclusions to be drawn about the moral concepts of the management.

Managers who act according to market morality would concentrate on fulfilling their own needs. These are, for example, achieving the best possible compensation. Applying market morality on the relationship of Earnings Management and CSR performance would result in management deciding to use Earnings Management techniques to meet the contradicting expectations of all shareholders and stakeholders to avoid negative outcomes for their own needs. Considering this, a management showing a good CSR performance could use this to distort the attention from their engagement in Earnings Management activities.

Managers who act according to traditional morality could use CSR to cover up Earnings Management engagement. Following the principles of fairness and care, they would not mislead their shareholders by disclosing glossed financial results. Their traditional moral concept would lead to them having the inner obligation to do the right thing.

Taking these considerations as a base, the results of the articles can be fitted in the selected ethical approaches, as shown in

Table 5. Ethical approaches.

	Accrual-Based Earnings Management and CSR Performance		Real Earnings Management and CSR Performance
Study
	Market Morality	Traditional Morality	Market Morality	Traditional Morality
Almahrog et al. [9]		x
Amar and Chakroun [14]		x
Litt et al. [58]		x
Bose and Yu [59]		x
Chen and Hung [60]		x		x
Chen et al. [43]		x
Dimitropoulos [44]		x		x
Garcia-Sanchez et al. [41]	x		N/A
Gerged et al. [61]		x
Hwang [62]		x
Jian et al. [63]		x
Jordaan et al. [64]	x			x
Kim et al. [6]		x		x
Liu et al. [3]	N/A		N/A
Lu and Steven [65]		x
Martinez-Martinez et al. [66]		x
Mohammad et al. [68]		x
Rahman and Zheng [69]		x		x
Rezaee and Tuo [70]	x
Sial et al. [13]	x
Velte [38]		x	N/A
Velte [7]		x	x
Wang and Kangtao [71]		x
Zhang et al. [72]		x	x

Source: own illustration. Explanation: N/A: no significant correlation has been recorded by the authors; x: applicable form of morality approach.

. An important limitation of the classification in the ethical approaches is that the articles did not conduct any analysis regarding the general ethical behavior of management. Thus, the conclusions about moral attitudes are drawn exclusively from the implications of the measured results.

Interestingly, the studies of Velte [7], Zhang et al. [72] and Jordaan et al. [64] show the management applied different moralities for Accrual-Based Earnings Management and Real Earnings Management in the interaction with CSR performance. This is an indicator for the need to disaggregate the CSR performance into sub-categories, in order to obtain a better understanding of the relationship.

We argue that the managements’ view towards the engagement in Earnings Management is heavily dependent on their own morals. Which path of the ethical theory being applied is a decision made upon the moral standards of the management. The meta-analysis of Shi et al. [57] also found that the link between CSR performance and the engagement in Earnings Management activities can be explained by moral track mechanisms. A measurement of the general moral views of the management could enhance the research on Earnings Management engagement and CSR performance.

6. Discussion

The results of the relationship between Accrual-Based Earnings Management and CSR performance were significantly negative for 20 articles (91%) of the conducted analysis. However, the relationship can change when other variables are added, as shown for managerial entrenchment [41] or family involvement [69]. The remaining two analyses found no significant effect of CSR performance on Accrual-Based Earnings Management or vice versa. The results for the connection between Real Earnings Management and CSR performance are less clear. Out of the nine studies, three found a positive relationship, whereas four declare a negative relationship between Real Earnings Management and CSR performance while three articles found no significant effect at all.

The studies reviewed used data reaching back to 1991, e.g., [6,43]. During this time, there was no reliable CSR measurement or even a clear definition of CSR performance. Reporting obligations for CSR have been established in the recent years. For example, this was introduced in the EU in 2017, but does not yet include all companies. The guidelines of ISO 26000 were published in November 2010. Methodologically including data from a time period without uniform reporting and an audit of stated CSR activities could be certainly problematic, as it is hardly possible to speak of robust CSR, CSP or ESG reporting. The results published do not necessarily reflect the context of CSR performance and Earnings Management.

Furthermore, seven studies employed moderators, while twelve papers did not use moderators in their models, leading to the assumption of a very simple correlation that does not exist in practice. Two papers included family involvement in their model, while three articles used Accrual-Based Earnings Management as moderator in the relationship of CSR performance. Other moderators were corporate governance, managerial entrenchment, firm value, SME companies and Accounting, Auditing and Enforcement releases. This indicates that focusing on the influence of CSR performance and Accrual-Based Earnings Management alone, to the exclusion of other factors, inaccurately portrays the relationship. The current research does not come to a clear conclusion regarding the relationship between Real Earnings Management and CSR performance, e.g., Velte [7] and Garcia-Sanchez et al. [41] found a positive relationship, while others detected a negative relationship (e.g., [6,60,64]), or no effect at all (e.g., [3,69]). These findings also indicate that the focus on the isolated connection between CSR performance and Earnings Management is not sufficient to capture the whole picture. The findings of Kim et al. [6] and Velte [7] state that companies use the two Earnings Management types as substitutes for one another, which corresponds to prior studies concerning the Earnings Management behavior of companies, e.g., [2,10,22].

The usage of the Real Earnings Management and Accrual-Based Earnings Management as substitutes, as well as the shifting in the connection between Accrual-Based Earnings Management and CSR performance when moderating for managerial entrenchment, indicate that CSR is used to cover up the engagement in Earnings Management activities. This finding is also in line with prior assumptions, e.g., [6]. Since the results are hypothesis only, further analysis is crucial. But because only two studies [7,38] analyzed their used model for reversed causality, the direction of the impact still needs more attention in future analysis. At this stage of the research, it is not possible to state whether there is a usage of substitutes and which direction the relationship of Earnings Management and CSR performance has.

The studies reviewed covered a total of ten different countries from all over the world, as well as three cross-country analyses. The different analyses did not show any significant differences in their results regarding the connection of CSR and Earnings Management. Additionally, the studies used a similar set of control variables in their models. The results regarding the influence of Big Four Auditors were all equal, stating that the Big Four reduce the engagement in Earnings Management. It would be interesting to include the control variable as a moderator in the model for Earnings Management and CSR performance as well, to test for the hypothesis whether the Big Four auditors not only reduce Earnings Management behavior, but also reduce the risk that a company uses CSR performance to cover up their engagement in Earnings Management. The evidence for the control variables ‘age’ and ‘size’ were contradictory. Thus, it is not possible yet to make a statement regarding their influence on either Earnings Management or CSR performance, but an inclusion in the model controlling for these could also lead to a deeper understanding of the drivers of the impact of Earnings Management on CSR performance.

When analyzing CSR performance, most studies used a simple proxy for their model development, taking CSR performance as one measure. Only Almahrog et al. [9] and Amar and Chakroun [14] developed hypothesis based upon individual CSR dimensions. The distinction between individual subcategories makes it possible to capture the impact of Earnings Management on CSR or vice versa more accurately than the consideration of the overall performance. The analysis taken an overall score leads to an average consideration of CSR performance, thus may not allow the actual relationship to be captured. Additionally, the measurement of Earnings Management was also conducted on average, excluding industry specific differences.

The review of the theoretical frameworks of the analyzed models shows similar approaches. By using the same or similar theoretical approaches, in particular the ethical approach and the agency theory, the resulting hypothesis are almost equal. Thus, the discussion of the results of the analysis performed tend to have one single direction. By building a more profound theoretical framework, considering further or other theories for the formulation of the hypotheses will potentially change the direction of the interpretation, but it is the only way to approach the results from a different angle in order to understand the mode of action between Earnings Management and CSR performance. With the existing theoretical focus, results can only be attributed to the ethical behavior of the management itself. Considering the existing criticism of CSR, which is characterized in particular by Friedman [31], it is not sufficient to base the theory on ethical approaches alone. He argues that the management of a company does not engage in CSR activities to do good, but to maximize their own profits [31]. By doing so, they, in fact, reduce the returns to shareholders by spending the company’s money, which reduces the profit for the year and thus the dividend to be distributed [31]. Radhakrishnan et al. [76] argue that, contrary to Friedman’s position, CSR does lead to a financial return but is only implemented by management as an investment strategy, not because of an inner obligation or a real commitment to CSR. Hanlon and Fleming [77] take both sides of the criticism by pointing out that CSR is basically a key marketing strategy. Companies use CSR and the perception that companies with good CSR performance also make ethical business decision to legitimize their business practices or to mitigate the response to controversial ethical behavior. For a deeper understanding of the correlation, this simplified conclusion must be challenged, since the current focus does not allow the connection to be examined from all necessary angles.

None of the articles paid attention to the potential for greenwashing. As several studies argue that the relationship between Earnings Management and CSR performance can be explained by ethical approaches (e.g., [57,60,63]), it is crucial to consider this possibility when interpreting the results. The analysis of CSR performance for possible greenwashing activities could cause complete dislocation in the correlations found so far. This thought is supported by the views of Jian et al. [63], questioning CSR performance to be an entrenchment strategy for the engagement in Earnings Management. The idea that a company that engages in Earnings Management in order to meet the expectations of external environmental pressure will also improve its CSR performance by engaging in greenwashing seems to emerge as well from the criticism of CSR, as well as from the analysis of the studies conducted so far.

A significant impediment across all articles and studies persists: the quality and reliability of the ESG or CSR data employed. A notable study by Berg et al. [75] or Busch et al. [78] underscores the pressing necessity to tackle issues concerning the quality of these data. The research emphasizes the paramount importance of focusing efforts on enhancing the reliability and accuracy of ESG and CSR data utilized, shedding light on the critical need for robust data sources in this research field.

This paper was intended to provide a basis for further analysis. It was possible to work out the current state of research, which shows a necessity for the usage of moderators if the impact of Earnings Management on CSR performance is to be explained. Further assessment on the usage of CSR as coverage of the engagement in Earnings Management activities could be analyzed by including corporate governance and managerial entrenchment in the model. For more advanced research, an approach of analysis in subsamples, formed according to theoretical expectations, would be to realize a cleaner and more accurate analysis of the relationship of CSR performance and Earnings Management. It is crucial to develop a clear theory framework when addressing the interaction of Earnings Management and CSR, as the analysis of empirical articles showed that the relationship is not fully explained by looking at the two variables in a nutshell.