Corporate Governance and Tax Avoidance: Evidence from Greek Service-Sector Firms

Abstract

This study investigates the relationship between corporate governance mechanisms and tax avoidance in Greek service-sector firms over the period 2014–2023. Using panel data, the analysis evaluates the influence of board characteristics, audit committees, auditor quality, and ownership structures on firms’ tax behavior. The results reveal that traditional governance mechanisms—such as board size, independence, audit committee composition, and gender diversity—do not significantly constrain tax avoidance, reflecting the formalistic rather than substantive adoption of governance practices in Greece. In contrast, external audit quality and ownership structure emerge as critical determinants. Engagement with high-quality auditors, particularly Big 4 firms, is associated with reduced tax aggressiveness, while state ownership similarly curbs avoidance, consistent with reputational and political accountability incentives. Conversely, managerial and foreign ownership are positively related to aggressive tax planning. The findings underscore the contextual nature of governance effectiveness: in weak enforcement environments, formal mechanisms serve largely symbolic roles, whereas external oversight and ownership incentives carry greater weight. This study contributes to agency and institutional theory by highlighting the limits of formal governance reforms absent substantive independence and enforcement.

1. Introduction

In recent decades, corporate governance has evolved into a central theme of both theoretical inquiry and empirical investigation, attracting the sustained interest of regulators and scholars across the globe (Tran et al., 2023; Almaharmeh et al., 2024; Arsh, 2025; Buchetti, 2025). Broadly defined, corporate governance encompasses the institutional structures, processes, and control mechanisms that steer and supervise corporate behavior, aiming to align the objectives of management with those of shareholders and other stakeholders (Omesi & Appah, 2021). This alignment becomes particularly critical in firms where ownership is widely dispersed and managerial authority is distinct from shareholder control, typical of large publicly listed corporations. Within such organizational settings, safeguarding shareholder interests and ensuring managerial accountability necessitate the design and enforcement of robust governance systems.

The widespread adoption of corporate governance practices has become not only a matter of shareholder protection but also a prerequisite for improving transparency, operational efficiency, and long-term competitiveness. Across jurisdictions such as the European Union and the United States, a series of governance codes and best-practice frameworks have been instituted to elevate corporate governance standards (Paganou, 2025). In the same vein, the Organisation for Economic Co-operation and Development (OECD) characterizes corporate governance as the network of relationships among a company’s leadership, its shareholders, its workforce, and external stakeholders, all oriented toward ensuring the firm’s sustainable performance and strategic resilience.

Academic discourse treats corporate governance both as an internal decision-making architecture and as an externally imposed institutional framework that governs the interactions among core stakeholder groups—such as equity holders, executive management, creditors, governmental bodies, employees, and other parties with vested interests (Omesi & Appah, 2021; Murni et al., 2016). This dual perspective reflects the multifaceted nature of governance, extending beyond boardroom practices to encompass the broader legal and regulatory ecosystem within which firms operate.

Turning to the Greek context, a significant institutional framework for corporate governance was established through Laws 3016/2002 and 4706/2020. These legislative acts introduced a structured set of governance obligations regarding the organizational composition of firms, the roles and duties of board members, and the responsibilities of corporate leadership toward both shareholders and the broader society (Kontogeorga et al., 2022; Paganou, 2025). The primary aim is to promote more effective and accountable management, foster transparency, and strengthen the competitiveness of Greek enterprises, thereby bridging corporate objectives with societal expectations.

Corporate governance mechanisms are typically categorized as either internal or external. Internal governance structures include the board of directors, audit committees, ownership concentration, and capital structure (e.g., leverage). At the same time, external mechanisms involve regulatory oversight, capital market pressures, and independent audit functions (Omesi & Appah, 2021). According to Mohammed (2017), internal mechanisms are primarily concerned with board efficacy and strategic decision-making designed to maximize shareholder value. External mechanisms, by contrast, encompass monitoring forces originating outside the firm, including regulatory compliance, debt covenants, and independent assurance providers.

Recent studies suggest that corporate governance reforms in Greece have had a positive but differentiated impact on firm performance, depending on firm size, sector, and ownership structure (G. Pavlou et al., 2025). A thorough understanding of corporate governance systems and their operative functions is thus essential for assessing their impact on corporate behavior, particularly in areas of strategic sensitivity such as tax planning, risk management, and compliance. Tax planning, in particular, represents a domain where governance structures can either curb or enable managerial opportunism, often under conditions of high regulatory uncertainty and limited stakeholder visibility.

The significance of this study lies in its dual contribution. First, it provides novel empirical evidence from Greece, a setting characterized by weak investor protection, concentrated ownership, and evolving governance frameworks, where the interaction between governance mechanisms and tax behavior remains underexplored. Second, it extends the international literature by showing how board attributes, audit quality, and ownership concentration operate in a post-crisis institutional environment, thereby enriching comparative insights across jurisdictions. By clarifying the channels through which governance structures influence corporate tax strategies, the study advances scholarly understanding and offers practical implications for regulators and policymakers seeking to strengthen transparency and accountability.

The article’s structure is as follows: Section 2 presents the literature review, which is analyzed in two subsections. Section 2.1 focuses on Corporate Governance and Tax Avoidance. Section 2.2 focuses on the Key Elements of Corporate Governance. In Section 3, we present the research methodology. In Section 4, we present the empirical results of the research. In Section 5, we present the conclusions, limitations, and suggestions for future research.

2. Literature Review

2.1. Corporate Governance and Tax Avoidance

The association between corporate governance and corporate tax behavior has long been a subject of academic inquiry, with theoretical roots tracing back to the seminal works of Berle and Means (1932) and Coase (1937). Over the subsequent decades, this relationship has been explored under various theoretical lenses (Fama & Jensen, 1983; Hart & Moore, 1990; Cohen et al., 2002; Aburajab et al., 2019), with the most prominent and enduring framework being agency theory (Fama & Jensen, 1983). Agency theory emphasizes the conflict of interest between management and shareholders, a conflict intensified by information and participation asymmetries. Within this context, Desai and Dharmapala (2006, 2009) argue that tax avoidance not only serves as a mechanism for reducing tax liabilities but also operates as a vehicle for managerial opportunism when governance mechanisms are weak.

Empirical studies in this area have examined multiple governance dimensions, such as the size, composition, and independence of the Board of Directors and the Audit Committee, audit quality, and capital structure (Minnick & Noga, 2010; Tran et al., 2023; Almaharmeh et al., 2024). Researchers such as Rego and Wilson (2012), Richardson et al. (2013) highlight that the strength and quality of internal governance mechanisms significantly influence the degree of tax aggressiveness, which effective governance may serve to constrain. Similarly, Mappadang (2019) and Vieira (2013) find that high-quality governance mechanisms have a positive impact on firms’ tax compliance behavior.

Gomes (2016) and Moore et al. (2017) provide further support for a systemic relationship between corporate governance characteristics and tax policy choices. Likewise, Bayar et al. (2018) confirm that robust governance mechanisms can mitigate the adverse effects of tax avoidance, especially in settings characterized by intense agency conflicts. A comprehensive review of 79 empirical studies conducted by Kovermann and Velte (2019) concluded that various governance factors—such as alignment of managerial and shareholder incentives, board composition, ownership structure, capital market environment, tax legislation and its enforcement, and the interests of other stakeholders—significantly affect corporate tax policy decisions. Effective governance structures appear to determine the optimal level of tax avoidance for each firm, balancing the benefits against the potential risks. In some cases, such mechanisms may promote aggressive tax planning if they enhance corporate profitability.

The literature presents conflicting views on whether tax avoidance can coexist with strong governance. Huseynov and Klamm (2012) suggest compatibility, whereas Hanlon and Slemrod (2009) argue that tax avoidance may harm firm value due to regulatory and reputational risks. According to Hanlon et al. (2005) and J. B. Kim et al. (2011), strong executive incentives—such as bonus-linked compensation—may discourage aggressive tax strategies. Conversely, tax avoidance is linked to an increased risk of stock price crashes. Their findings support the notion that sound governance and effective control systems reduce the likelihood of adopting aggressive tax practices and the accompanying adverse consequences.

From an incentive-based perspective, Armstrong et al. (2015) find a direct connection between high levels of tax avoidance and managerial self-interest. They conclude that the optimal level of tax avoidance is achieved when the marginal costs and benefits to management are equalized, underscoring the importance of independent and specialized boards in curbing tax aggressiveness. Similarly, Ribeiro (2015) and Al Rubaye et al. (2024) document cases of self-serving management of corporate resources through tax avoidance, particularly in environments characterized by pronounced information asymmetry.

Halioui et al. (2016) highlight the dual nature of the relationship between CEO compensation and tax aggressiveness, while Higgins et al. (2015) associate corporate tax policy choices with broader business strategies. Finally, K. P. Chen and Chu (2005) and Crocker and Slemrod (2005) stress the importance of integrating strategic motivations and governance structures into the analysis of corporate tax behavior.

2.2. Key Elements of Corporate Governance

2.2.1. Board of Directors Characteristics and Tax Avoidance

The Board of Directors (BoD) constitutes a fundamental mechanism of corporate governance, entrusted with supervisory and strategic responsibilities. Its effectiveness is associated with mitigating agency problems and safeguarding shareholders’ interests (Fama & Jensen, 1983; Jensen, 1993). The composition, size, and independence of the BoD influence accounting integrity and firms’ tax behavior (Klein, 2002; Singh et al., 2018; Khatib & Nour, 2021). Tax risk management falls under the responsibilities of the BoD, although the board may also act as a vehicle for opportunistic behavior (Erle, 2008; Tran et al., 2023). In Greece, Laws 3016/2002 and 4706/2020 establish strict eligibility criteria for board members, emphasizing independence and professional competence.

The literature highlights complex and often contradictory relationships between BoD characteristics and tax avoidance (Minnick & Noga, 2010; Lanis & Richardson, 2011; Ogbeide & Obaretin, 2018), while individual attributes of directors—such as reputation (Chyz & Gaertner, 2018), religiosity (Boone et al., 2013), political ideology (Christensen et al., 2015), personality traits (Kartikaningdyah, 2019; Bouslah et al., 2018), and financial incentives (Campbell et al., 2020)—emerge as critical determinants of tax behavior.

Concerning board size, smaller boards are often deemed more efficient in decision-making (Beasley, 1996; Yermack, 1996), while larger boards may face coordination challenges that create room for aggressive tax strategies (Pratama, 2017). Nevertheless, diversity in knowledge and experience among larger boards can enhance oversight capacity (E. A. A. Wahab et al., 2017; Osebe et al., 2019). The lack of statistically significant findings in some studies (Lanis & Richardson, 2011; Khaoula & Ali, 2012) may reflect institutional or sector-specific factors (Coles et al., 2008).

Board independence, defined by the presence of non-executive external members (Fama & Jensen, 1983; Bhagat & Bolton, 2008), is considered a proxy for governance quality and has been linked to greater transparency, improved reporting quality, and enhanced firm value (Peasnell et al., 2000; Y. Liu et al., 2015). In the Greek context, Dimitropoulos and Asteriou (2010) document a positive impact of board independence on financial reporting outcomes. However, evidence regarding its effect on tax avoidance remains mixed. While some studies suggest that independence constrains aggressive tax planning (Lanis & Richardson, 2012), others argue that robust governance may facilitate shareholder value-maximizing tax strategies (Moore et al., 2017; Flamini et al., 2021). Notably, Minnick and Noga (2010) observe that the role of board independence in tax planning may vary depending on the jurisdiction’s tax environment.

Finally, numerical independence alone is insufficient, as the functional competence of board members plays a decisive role (Sarkar et al., 2008; Zhou, 2011). Thus, the board’s influence on tax avoidance is shaped by its composition, the skillset of its members, and the broader institutional environment in which it operates.

Informed by the theoretical considerations and prior empirical findings, this research posits the following hypotheses:

H1: 

All else being equal, board of directors’ size (BSIZE) reduces the likelihood of corporate tax aggressiveness.

H2: 

All else being equal, board of directors’ independence (BIND) reduces the likelihood of corporate tax aggressiveness.

2.2.2. Audit Committee and Tax Avoidance

The Audit Committee plays a central oversight role in corporate governance, enhancing transparency, financial reporting reliability, and control of accounting and tax-related risks (Klein, 2002; Dridi & Boubaker, 2016). In Greece, Law 4449/2017 grants the committee key responsibilities, including audit supervision, risk management, and auditor independence.

Theoretical frameworks suggest that independence, size, and expertise of committee members can constrain earnings management and tax aggressiveness (Fama & Jensen, 1983; Richardson et al., 2013). Empirical results are mixed: while some find no significant impact (Yang & Krishnan, 2005; Ghosh et al., 2010; Rudiatun & Anggorowati, 2024), others emphasize the role of composition and experience (H. W. Huang & Thiruvadi, 2010; Abbott et al., 2000).

Stronger committee independence and expertise are linked to lower tax aggressiveness (Moore, 2012; Sunarsih & Oktaviani, 2016; Hendrani et al., 2020; Alqatan et al., 2024), with the committee influencing tax strategy based on corporate culture and risk orientation (Hsu et al., 2018). Tandean and Winnie (2016) and Pratama (2017) associate larger committee size with reduced tax avoidance, while García-Meca et al. (2021) show moderating effects on CEO-driven decisions, especially with larger, gender-diverse committees.

Contrarily, high expertise or size may sometimes support aggressive tax practices if not coupled with accountability (Sunarto et al., 2021; Gunawan & Resitarini, 2019). In Greece, Kourdoumpalou (2010) found no significant association between audit committees and tax evasion, possibly due to their limited institutional role during 2000–2004.

Overall, the committee’s impact depends not only on its structure but also on its effective engagement and alignment with the firm’s strategy.

Informed by the theoretical considerations and prior empirical findings, this research posits the following hypotheses:

H3: 

All else being equal, Audit Committee size (ACMEET) reduces the likelihood of corporate tax aggressiveness.

H4: 

All else being equal, Audit Committee independence (ACIND) reduces the likelihood of corporate tax aggressiveness.

2.2.3. Gender Diversity and Tax Avoidance

The inclusion of women in boards of directors and audit committees has become a core component of modern corporate governance, associated with the enhancement of skills, transparency, and social accountability (Carter et al., 2003; Suleiman & Abubakar, 2021). Many countries, including Greece, through Law 4706/2020, have adopted mandatory gender quotas for board representation.

International literature explains gender-based differences in tax behavior through three theoretical approaches: rational choice theory (Allingham & Sandmo, 1972), theories of social and psychological norms (Kirchler & Maciejovsky, 2001), and biological difference theories (Sapienza et al., 2009).

Several studies document a reduction in tax aggressiveness due to female participation (Torgler & Schneider, 2004; Hoseini et al., 2018; Rakia et al., 2024; Wang, 2024; Bataineh, 2025; Mai et al., 2025), while others find no statistically significant effect (Khaoula & Ali, 2012; Budi, 2019; Anjarwi et al., 2024), attributing these inconsistencies to institutional and cultural constraints or low female representation (Gregory-Smith et al., 2014; Oyeleke et al., 2016).

Women’s higher tax compliance has been linked to greater risk aversion (Torgler & Schneider, 2004), differences in socialization (Kastlunger et al., 2010; Matsaganis & Flevotomou, 2010), and stronger ethical values (Betz et al., 1989). Their presence on boards has been associated with increased transparency, accountability, profitability, and firm value (Francoeur et al., 2008; Adams & Ferreira, 2009; Gul et al., 2011; Srinidhi et al., 2011), with a potential contribution to reducing tax avoidance (Richardson et al., 2016).

More recent studies examine intra-gender differences, focusing on factors such as gender role and self-image (Kastlunger et al., 2010). Findings by Francis et al. (2014, 2015) suggest that replacing a male CFO with a female one is associated with reduced tax aggressiveness, particularly in more aggressive forms of avoidance. However, other studies (Adams & Funk, 2012; Bekiaris & Papanastasiou, 2020) indicate that women are not inherently more risk-averse, and their behavior may vary depending on the institutional and cultural context. In countries like Nigeria and Indonesia, limited female participation in oversight mechanisms has not yielded significant changes in firms’ tax policies (Oyeleke et al., 2016; Anjarwi et al., 2024).

In conclusion, while female representation is often linked to greater compliance and transparency, its actual impact on tax behavior depends on the institutional strength and substantive participation of women in governance. Numerical representation may be a necessary, but not sufficient, condition for enhancing corporate accountability (Torchia et al., 2011).

Informed by the theoretical considerations and prior empirical findings, this research posits the following hypotheses:

H5: 

All else being equal, female presence on the board of directors (BDIV) reduces the likelihood of corporate tax aggressiveness.

H6: 

All else being equal, female presence on the Audit Committee (ACDIV) reduces the likelihood of corporate tax aggressiveness.

2.2.4. Audit Quality and Tax Avoidance

Audit quality constitutes a significant factor in shaping corporate tax behavior, as it is associated with increased transparency, the accuracy of financial reporting, and the containment of aggressive tax strategies (Gallemore et al., 2014; Salaudeen & Abdulwahab, 2022; Lungu et al., 2023).

In Greek literature, Kourdoumpalou (2010) finds that firms audited by Big 6/5/4 firms demonstrate higher tax audit adjustments. In contrast, those audited by the national firm S.O.L. S.A. display lower levels of tax evasion. Similarly, Kourdoumpalou and Karagiorgos (2012) conclude that international audit firms prioritize financial compliance over tax legality. Fasoulas et al. (2024) show that companies with strong, independent, and diverse boards of directors are more likely to engage Big 4 auditors. In contrast, family-owned firms tend to select smaller audit firms. Additionally, Chytis et al. (2016) emphasize auditor independence as a key factor in limiting tax avoidance.

At the international level, DeAngelo (1981), and J. H. Kim and Im (2017) document that Big 4 auditors provide higher-quality audits and demonstrate greater fraud detection capabilities. In contrast, Jeong and Rho (2004) and Fauziati et al. (2018) find no significant impact of audit quality on tax behavior. Tedds (2006) identifies a positive link between external auditing and tax compliance, while Vintilă et al. (2018) report country-specific variations in the relationship between audit fees and effective tax rates.

Informed by the theoretical considerations and prior empirical findings, this research posits the following hypothesis:

H7: 

All else being equal, the type of external auditor (BIG4, EXTAUD) is associated with the likelihood of corporate tax aggressiveness.

2.2.5. Ownership Structure and Tax Avoidance

Ownership structure constitutes a critical element of corporate governance, significantly influencing firms’ strategic decisions and tax behavior. According to Gursoy and Aydogan (2002), ownership structure is defined by two key dimensions: the concentration of ownership and the identity of shareholders. Thus, the analysis considers whether ownership is concentrated or dispersed, and whether it is exercised by institutional, state, foreign, or managerial shareholders.

The literature suggests that ownership structure interacts with other governance mechanisms, shaping the effectiveness of oversight and accountability (Sartaji & Hassanzadeh, 2014). As Waluyo (2017) emphasizes, ownership configuration is associated with the quality of tax compliance, since different shareholder types promote varying levels of transparency, discipline, and willingness to engage in tax avoidance. Tedds (2006) also considers ownership a fundamental governance mechanism with the potential to influence corporate tax behavior.

Empirical tax research has explored the extent to which various ownership structures relate to firms’ tax aggressiveness (Lanis & Richardson, 2015; Kourdoumpalou, 2016; Brune et al., 2019; Budi, 2019; Chytis et al., 2019; Almaharmeh et al., 2024; Chytis et al., 2020; Chalevas et al., 2024). However, findings remain mixed. According to Velte (2023), there is no consensus as to whether and which type of owner fosters higher or lower levels of tax avoidance. These discrepancies are attributed both to institutional differences across countries and to the nature of shareholders themselves (e.g., strategic institutional investors versus passive owners).

Overall, ownership structure affects the design and implementation of tax strategies, acting either as a deterrent or as an enabler of aggressive tax planning. Further investigation into the distinct types of ownership and their interactive relationship with corporate tax policy is essential for a deeper understanding of tax avoidance behavior.

Informed by the theoretical considerations and prior empirical findings, this research posits the following hypotheses:

H8: 

All else being equal, ownership concentration (OWNCONC) is associated with the likelihood of corporate tax aggressiveness.

H9: 

All else being equal, Managerial ownership (MANOWN) is associated with the likelihood of corporate tax aggressiveness.

H10: 

All else being equal, state ownership (STATEOWN) is associated with the likelihood of corporate tax aggressiveness.

H11: 

All else being equal, institutional ownership (INSTOWN) is associated with the likelihood of corporate tax aggressiveness.

H12: 

All else being equal, foreign ownership (FOROWN) is associated with the likelihood of corporate tax aggressiveness.

3. Data and Research Model

3.1. Sample

The study sample consists of service-sector firms listed on the Athens Stock Exchange (ASE) during the period 2014–2023. Firms that were not continuously listed throughout the entire period under examination, as well as those delisted during the same timeframe, were excluded. The final sample comprises 26 service-sector firms and 260 firm-year observations, which were manually collected from the published annual financial statements of these companies. The study period was chosen to cover the decade 2014–2023, due to the significant institutional, macroeconomic, and social developments that took place in Greece, which are likely to have influenced corporate tax strategies. Between 2014 and 2018, the country underwent the final stages of the Memorandum of Understanding (MoU) on Specific Economic Policy Conditionality and the subsequent fiscal adjustment process. The period 2015–2019 was marked by the imposition of capital controls, which constrained liquidity and created difficulties in conducting transactions, particularly with foreign counterparts. From 2014 to 2021, three statutory corporate tax rate changes were enacted, while the new Income Tax Code (Law 4172/2013) was implemented, placing greater emphasis on monitoring multinational tax avoidance strategies (transfer pricing, intangible assets, etc.). In 2020, Law 4706/2020 introduced a new corporate governance framework. Moreover, from 2015 internationally, and 2021 in Greece, the strengthening of digital tax enforcement was institutionalized to combat tax evasion, including its more sophisticated forms (OECD/G20 BEPS, myDATA). Finally, from 2020 onwards, the COVID-19 pandemic led to severe financial distress for the business sector, while the government was required to provide support measures, subsidies, and tax relief.

3.2. Tax Planning Measurement

A wide range of proxies has been employed in empirical tax literature to assess the extent of tax avoidance. These proxies differ not only in their computational methodology but also in the particular dimension of avoidance they aim to capture and the limitations they entail (Badertscher et al., 2010; Dunbar et al., 2010; Hanlon & Heitzman, 2010; Lietz, 2013; Gebhart, 2017). Given that no single metric can fully encompass the complexity of corporate tax strategies, the selection of an appropriate measure is mainly dependent on the research objective and the data at hand. Ensuring construct validity and robustness in empirical findings is essential in this context (Aronmwan & Okafor, 2019).

This study employs four alternative measures to estimate tax avoidance, selected following the methodological framework and data availability.

3.2.1. Tax Avoidance Relative to Pre-Tax Income

A widely used method in empirical research, and adopted in this analysis, involves expressing tax avoidance as a ratio to pre-tax income. It estimates the fiscal benefit from tax planning by comparing expected tax liabilities (based on statutory rates) to actual tax paid, scaled by pre-tax income. Atwood et al. (2012), analyzing over 69,000 firm-year observations in 22 countries (1993–2007), showed that tax system characteristics strongly influence multinational tax avoidance. To mitigate the effects of exceptional fluctuations, estimation is typically conducted over three years.

To measure tax avoidance as a percentage of pre-tax income, we use the measurement PERCINC, which is calculated as follows:

$$PERCINC={\displaystyle \frac{\sum _{t-2}^t\left[\left(EBT\right)\times \tau \right]-{\sum }_{t-2}^t\left(Current tax expense\right)}{{\sum }_{t-2}^t\left(EBT\right)}}$$

(1)

where EBT are the Earnings Before Taxes, and τ is the statutory tax rate.

While effective, this proxy has limitations—especially when pre-tax income is negative or near zero, a common issue in ETR-based measures. Nonetheless, its application remains prevalent. Kanagaretnam et al. (2018) explored its use in the context of social trust; Na and Yan (2022) and Ni et al. (2022) examined linguistic and environmental factors; and Hasan et al. (2022) analyzed the role of foreign institutional ownership, confirming its relevance in diverse research contexts.

This study is aligned with the view expressed by Amidu et al. (2016b), who argue that tax avoidance practices do not consist solely of actions aimed at reducing taxable income, but also include the recording of transactions intended to generate accounting losses. Accordingly, the sample under examination includes loss-making firms. In calculating the selected tax avoidance measure, pre-tax losses are treated as zero in cases where firms report negative earnings before taxes. Additionally, if the numerator of the formula used to compute the measure is positive while the denominator is zero, the measure is assigned a value of 1. Conversely, if the numerator is negative—indicating the payment of taxes exceeding the expected amount based on the statutory tax rate—the measure is assigned a value of 0.

3.2.2. Total Difference Between Book and Taxable Incomes (BTD)

The Total Book–Tax Difference (BTD) is a key measure of tax avoidance, as it reflects both permanent and temporary differences between financial and taxable income (Manzon & Plesko, 2002; Wilson, 2009; Lietz, 2013). It is typically calculated as the gap between pre-tax income and estimated taxable income—based on current tax expense and statutory tax rates (Manzon & Plesko, 2002; Lee et al., 2015)—with refinements such as removing minority interests and non-taxable items.

We calculate the BTD measurement as follows:

$$BTD={\displaystyle \frac{\left[Pre-tax income-\left(\raisebox{1ex}{$Current tax expense$}\!\left/ \!\raisebox{-1ex}{$\tau $}\right.\right)\right]}{\left(Beginning total assets\right)}}$$

(2)

where τ is the statutory tax rate.

Although informative, the metric is vulnerable to earnings management (Desai & Dharmapala, 2009), prompting researchers to adjust or complement it with other measures (Martinez, 2017; Gebhart, 2017). Variants of the Total BTD scale the difference by prior-year total assets to minimize size-related bias (Brooks et al., 2016; Mocanu et al., 2020) or adjust for elements like loss carryforwards (Koumanakos et al., 2017). However, Guenther (2014) warns that the selected scaling base—be it income or assets—can significantly affect comparability of results and must be carefully considered. Empirical studies link Total BTD to firm value, capital costs, governance, and institutional ownership (Goh et al., 2016; Kolias & Koumanakos, 2022; Benkraiem et al., 2024). When critically applied, it remains a foundational tool in tax research (Hanlon & Heitzman, 2010; Chakroun & Ben Amar, 2024).

3.2.3. Temporary Differences Between Accounting and Taxable Income (Temporary BTD)

Temporary differences between accounting and taxable income give rise to deferred taxes, calculated by multiplying these discrepancies by the statutory tax rate (Manzon & Plesko, 2002; Lee et al., 2015; Aronmwan & Okafor, 2019). Key sources include tax loss carryforwards, accelerated depreciation, and asset impairments (Hanlon, 2003), while alternative metrics—like deferred tax per share—have also been proposed (Ayers et al., 2018).

Three interpretations of their significance have emerged. They may signal accrual-based earnings management (Comprix et al., 2011; Blaylock et al., 2012), reflect tax deferral strategies linked to avoidance (N. S. A. Wahab & Holland, 2015), or postpone liabilities without affecting tax permanently (Aronmwan & Okafor, 2019). Seidman (2008) stresses their strategic role, noting their impact rivals that of permanent differences.

We calculate the temporary differences between Accounting and Taxable Income as follows:

$$TEMBTD={\displaystyle \frac{\left[\left(\raisebox{1ex}{$Deferred tax expense$}\!\left/ \!\raisebox{-1ex}{$\tau $}\right.\right)\right]}{\left(Beginning total assets\right)}}$$

(3)

where τ is the statutory tax rate.

Empirical research links temporary differences to institutional ownership (Moore, 2012), audit fees (Hanlon et al., 2012), and sustained tax minimization (Dyreng et al., 2017), affirming their relevance in tax and financial reporting studies.

3.2.4. Tax Effects of Book–Tax Income Differences

Tang and Firth (2011) introduced a tax-effect approach to measuring book–tax differences (BTDs), focusing on the gap between expected tax, based on applying the statutory rate to accounting income, and the actual current tax expense. Unlike traditional methods that calculate BTDs as raw differences, this approach integrates both permanent and temporary differences, thereby addressing measurement errors tied to taxable income estimation (Salihu et al., 2013).

It also captures tax avoidance strategies that reduce tax liability without altering reported income, such as income-shifting practices. Aronmwan and Okafor (2019) consider this method theoretically superior to prior BTD proxies, though empirical validation remains limited. The measure is especially valuable in datasets with tax loss carryforwards or heterogeneous tax regimes.

We calculate the tax effects of Book–Tax Income differences as follows:

$$TAXEFFBTD={\displaystyle \frac{\left[\left(Accounting income\right)\times \tau -Current tax expenses\right]}{\left(Beginning total assets\right)}}$$

(4)

where τ is the statutory tax rate.

3.3. Research Methodology

Data processing and econometric analysis were performed with the IBM SPSS Statistics Version 29.0.1.0. The hypotheses of the study were tested using multiple logistic regression. To test the statistical significance of the models, we used the Adj.R2 and the F-test. To test correlations and possible multicollinearity issues, we used the Pearson and Spearman tests, as well as the VIF test. To test the statistical significance of the variables, the p-value (sign) was used.

The estimated regression models employed in the study are presented below:

$$\begin{gathered} {\mathrm{P}\mathrm{E}\mathrm{R}\mathrm{C}\mathrm{I}\mathrm{N}\mathrm{C}}_{it}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\times {\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+{\mathsf{\beta }}_2\times {\mathrm{R}\mathrm{O}\mathrm{A}}_{it}+{\mathsf{\beta }}_3\times {\mathrm{C}\mathrm{A}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}}_{it}+ \\ {\mathsf{\beta }}_4\times {\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{A}\mathrm{N}}_{it}+{\mathsf{\beta }}_5\times {\mathrm{E}\mathrm{T}\mathrm{A}}_{it}+{\mathsf{\beta }}_6\times {\mathrm{L}\mathrm{O}\mathrm{S}\mathrm{S}}_{it}+{\mathsf{\beta }}_7\times {\mathrm{M}\mathrm{U}\mathrm{L}\mathrm{T}}_{it}+{\mathsf{\beta }}_8\times {\mathrm{B}\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+ \\ {\mathsf{\beta }}_9\times {\mathrm{B}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+{\mathsf{\beta }}_{10}\times {\mathrm{B}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+{\mathsf{\beta }}_{11}\times {\mathrm{A}\mathrm{C}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+{\mathsf{\beta }}_{12}\times {\mathrm{A}\mathrm{C}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+ \\ {\mathsf{\beta }}_{13}\times {\mathrm{A}\mathrm{C}\mathrm{M}\mathrm{E}\mathrm{E}\mathrm{T}}_{it}+{\mathsf{\beta }}_{14}\times {\mathrm{B}\mathrm{I}\mathrm{G}4}_{it}+{\mathsf{\beta }}_{15}\times {\mathrm{E}\mathrm{X}\mathrm{T}\mathrm{A}\mathrm{U}\mathrm{D}}_{it}+{\mathsf{\beta }}_{16}\times {\mathrm{O}\mathrm{W}\mathrm{N}\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{it}+ \\ {\mathsf{\beta }}_{17}\times {\mathrm{M}\mathrm{A}\mathrm{N}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{18}\times {\mathrm{I}\mathrm{N}\mathrm{S}\mathrm{T}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{19}\times {\mathrm{S}\mathrm{T}\mathrm{A}\mathrm{T}\mathrm{E}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+ \\ {\mathsf{\beta }}_{20}\times {\mathrm{F}\mathrm{O}\mathrm{R}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\epsilon }}_{it} \end{gathered}$$

(5)

$$\begin{gathered} {\mathrm{B}\mathrm{T}\mathrm{D}}_{it}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\times {\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+{\mathsf{\beta }}_2\times {\mathrm{R}\mathrm{O}\mathrm{A}}_{it}+{\mathsf{\beta }}_3\times {\mathrm{C}\mathrm{A}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}}_{it}+{\mathsf{\beta }}_4\times {\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{A}\mathrm{N}}_{it}+ \\ {\mathsf{\beta }}_5\times {\mathrm{E}\mathrm{T}\mathrm{A}}_{it}+{\mathsf{\beta }}_6\times {\mathrm{L}\mathrm{O}\mathrm{S}\mathrm{S}}_{it}+{\mathsf{\beta }}_7\times {\mathrm{M}\mathrm{U}\mathrm{L}\mathrm{T}}_{it}+{\mathsf{\beta }}_8\times {\mathrm{B}\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+{\mathsf{\beta }}_9\times {\mathrm{B}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+ \\ {\mathsf{\beta }}_{10}\times {\mathrm{B}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+{\mathsf{\beta }}_{11}\times {\mathrm{A}\mathrm{C}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+{\mathsf{\beta }}_{12}\times {\mathrm{A}\mathrm{C}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+{\mathsf{\beta }}_{13}\times {\mathrm{A}\mathrm{C}\mathrm{M}\mathrm{E}\mathrm{E}\mathrm{T}}_{it}+ \\ {\mathsf{\beta }}_{14}\times {\mathrm{B}\mathrm{I}\mathrm{G}4}_{it}+{\mathsf{\beta }}_{15}\times {\mathrm{E}\mathrm{X}\mathrm{T}\mathrm{A}\mathrm{U}\mathrm{D}}_{it}+{\mathsf{\beta }}_{16}\times {\mathrm{O}\mathrm{W}\mathrm{N}\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{it}+{\mathsf{\beta }}_{17}\times {\mathrm{M}\mathrm{A}\mathrm{N}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+ \\ {\mathsf{\beta }}_{18}\times {\mathrm{I}\mathrm{N}\mathrm{S}\mathrm{T}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{19}\times {\mathrm{S}\mathrm{T}\mathrm{A}\mathrm{T}\mathrm{E}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{20}\times {\mathrm{F}\mathrm{O}\mathrm{R}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\epsilon }}_{it} \end{gathered}$$

(6)

$$\begin{gathered} {\mathrm{T}\mathrm{E}\mathrm{M}\mathrm{B}\mathrm{T}\mathrm{D}}_{it}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\times {\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+{\mathsf{\beta }}_2\times {\mathrm{R}\mathrm{O}\mathrm{A}}_{it}+{\mathsf{\beta }}_3\times {\mathrm{C}\mathrm{A}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}}_{it}+{\mathsf{\beta }}_4\times {\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{A}\mathrm{N}}_{it}+ \\ {\mathsf{\beta }}_5\times {\mathrm{E}\mathrm{T}\mathrm{A}}_{it}+{\mathsf{\beta }}_6\times {\mathrm{L}\mathrm{O}\mathrm{S}\mathrm{S}}_{it}+{\mathsf{\beta }}_7\times {\mathrm{M}\mathrm{U}\mathrm{L}\mathrm{T}}_{it}+{\mathsf{\beta }}_8\times {\mathrm{B}\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+{\mathsf{\beta }}_9\times {\mathrm{B}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+ \\ {\mathsf{\beta }}_{10}\times {\mathrm{B}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+{\mathsf{\beta }}_{11}\times {\mathrm{A}\mathrm{C}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+{\mathsf{\beta }}_{12}\times {\mathrm{A}\mathrm{C}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+{\mathsf{\beta }}_{13}\times {\mathrm{A}\mathrm{C}\mathrm{M}\mathrm{E}\mathrm{E}\mathrm{T}}_{it}+ \\ {\mathsf{\beta }}_{14}\times {\mathrm{B}\mathrm{I}\mathrm{G}4}_{it}+{\mathsf{\beta }}_{15}\times {\mathrm{E}\mathrm{X}\mathrm{T}\mathrm{A}\mathrm{U}\mathrm{D}}_{it}+{\mathsf{\beta }}_{16}\times {\mathrm{O}\mathrm{W}\mathrm{N}\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{it}+{\mathsf{\beta }}_{17}\times {\mathrm{M}\mathrm{A}\mathrm{N}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+ \\ {\mathsf{\beta }}_{18}\times {\mathrm{I}\mathrm{N}\mathrm{S}\mathrm{T}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{19}\times {\mathrm{S}\mathrm{T}\mathrm{A}\mathrm{T}\mathrm{E}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{20}\times {\mathrm{F}\mathrm{O}\mathrm{R}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\epsilon }}_{it} \end{gathered}$$

(7)

$$\begin{gathered} {\mathrm{T}\mathrm{A}\mathrm{X}\mathrm{E}\mathrm{F}\mathrm{F}\mathrm{B}\mathrm{T}\mathrm{D}}_{it}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\times {\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+{\mathsf{\beta }}_2\times {\mathrm{R}\mathrm{O}\mathrm{A}}_{it}+{\mathsf{\beta }}_3\times {\mathrm{C}\mathrm{A}\mathrm{P}\mathrm{I}\mathrm{N}\mathrm{T}}_{it}+ \\ {\mathsf{\beta }}_4\times {\mathrm{I}\mathrm{N}\mathrm{T}\mathrm{A}\mathrm{N}}_{it}+{\mathsf{\beta }}_5\times {\mathrm{E}\mathrm{T}\mathrm{A}}_{it}+{\mathsf{\beta }}_6\times {\mathrm{L}\mathrm{O}\mathrm{S}\mathrm{S}}_{it}+{\mathsf{\beta }}_7\times {\mathrm{M}\mathrm{U}\mathrm{L}\mathrm{T}}_{it}+{\mathsf{\beta }}_8\times {\mathrm{B}\mathrm{S}\mathrm{I}\mathrm{Z}\mathrm{E}}_{it}+ \\ {\mathsf{\beta }}_9\times {\mathrm{B}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+{\mathsf{\beta }}_{10}\times {\mathrm{B}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+{\mathsf{\beta }}_{11}\times {\mathrm{A}\mathrm{C}\mathrm{I}\mathrm{N}\mathrm{D}}_{it}+{\mathsf{\beta }}_{12}\times {\mathrm{A}\mathrm{C}\mathrm{D}\mathrm{I}\mathrm{V}}_{it}+ \\ {\mathsf{\beta }}_{13}\times {\mathrm{A}\mathrm{C}\mathrm{M}\mathrm{E}\mathrm{E}\mathrm{T}}_{it}+{\mathsf{\beta }}_{14}\times {\mathrm{B}\mathrm{I}\mathrm{G}4}_{it}+{\mathsf{\beta }}_{15}\times {\mathrm{E}\mathrm{X}\mathrm{T}\mathrm{A}\mathrm{U}\mathrm{D}}_{it}+{\mathsf{\beta }}_{16}\times {\mathrm{O}\mathrm{W}\mathrm{N}\mathrm{C}\mathrm{O}\mathrm{N}\mathrm{C}}_{it}+ \\ {\mathsf{\beta }}_{17}\times {\mathrm{M}\mathrm{A}\mathrm{N}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{18}\times {\mathrm{I}\mathrm{N}\mathrm{S}\mathrm{T}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\beta }}_{19}\times {\mathrm{S}\mathrm{T}\mathrm{A}\mathrm{T}\mathrm{E}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+ \\ {\mathsf{\beta }}_{20}\times {\mathrm{F}\mathrm{O}\mathrm{R}\mathrm{O}\mathrm{W}\mathrm{N}}_{it}+{\mathsf{\epsilon }}_{it} \end{gathered}$$

(8)

where PERCINC: A measurement of tax avoidance as a percentage of pre-tax income (Equation (1)), BTD: Total Book–Tax Difference (Equation (2)), TEMBTD: Temporary differences between accounting and taxable income (Equation (3)), TAXEFFBTD: Tax effects of Book–Tax Income Differences (Equation (4)). The four measures of tax avoidance are considered as dependent variables. Seven basic financial figures of the companies are used as control variables. The definition of these variables is: SIZE: Natural logarithm of total assets, ROA: Return on Assets, CAPINT: Capital Intensity = Total Tangible Assets/Total Assets, INTAN: Intangible asset intensity = Total Intangible Assets/Total Assets, ETA: Capital Structure of Businesses = Total Equity/Total Assets, LOSS: Dummy variable that takes the value 1 if the firm had losses and the value 0 otherwise, MULT: Dummy variable that takes the value 1 if the company is a member of a multinational group of companies and the value 0 otherwise. Finally, the key elements of corporate governance are considered as independent variables. The definition of the variables is the following: BSIZE: Natural Logarithm of the Total Number of Board Members, BIND: Percentage of independent non-executive members of the Board of Directors, BDIV: Percentage of Women on the Board of Directors, ACIND: Percentage of independent Audit Committee members, ACDIV: Percentage of Women in the Audit Committee, ACMEET: Natural Logarithm of the Total Number of Audit Committee Meetings Held During a Year, BIG4: Dummy variable that takes the value 1 if the company hires from among the BIG4 auditing firms and the value 0 otherwise, EXTAUD: Dummy variable, which takes the value 1 if the company hires between Grant Thornton LLP and Crowe Horwath International auditing firms, and the value 0 otherwise, OWNCONC: Percentage of ownership held by the largest shareholders (over 5% each).

MANOWN: Percentage of ownership held by management, INSTOWN: Percentage of ownership held by institutional investors (over 5% each), STATEOWN: Percentage of ownership controlled by the state, and FOROWN: Percentage of ownership held by foreign investors (over 5% each).

4. Empirical Results

In this section, we present the results of the empirical study. The data were processed and analyzed using the IBM SPSS Statistics Version 29.0.1.0.

In

Table 1. Descriptive Statistics.

		N	Mean	Median	Std. Deviation	Minimum	Maximum
Dependent Variables	PERCINC	260	0.1202	0.0722	0.1473	0.0000	1.0000
	BTD	260	0.0124	0.0000	0.1072	−0.2145	1.5842
	TEMBTD	260	0.0122	0.0032	0.0912	−0.1184	1.3610
	TAXEFFBTD	260	0.0041	0.0002	0.0234	−0.0472	0.3485
Control Variables	SIZE	260	18.4091	18.4201	1.9798	13.7410	22.5921
	ROA	260	0.0267	0.0204	0.1030	−0.4500	0.6107
	CAPINT	260	0.2205	0.1457	0.2181	0.0001	0.7799
	INTAN	260	0.0692	0.0056	0.1417	0.0000	0.6893
	ETA	260	0.4934	0.4728	0.2598	−0.1925	0.9752
	LOSS	260	0.3000	0.0000	0.4591	0.0000	1.0000
	MULT	260	0.6615	1.0000	0.4741	0.0000	1.0000
Independent Variables	BSIZE	260	2.0969	2.0794	0.2774	1.6094	2.5649
	BIND	260	0.3091	0.3038	0.1574	0.0000	0.9091
	BDIV	260	0.1778	0.2000	0.1387	0.0000	0.5000
	ACIND	260	0.6696	0.6667	0.3308	0.0000	1.0000
	ACDIV	260	0.1498	0.0000	0.2291	0.0000	1.0000
	ACMEET	260	1.7727	1.7918	0.7881	0.0000	3.3673
	BIG4	260	0.4231	0.0000	0.4950	0.0000	1.0000
	EXTAUD	260	0.2769	0.0000	0.4483	0.0000	1.0000
	OWNCONC	260	0.6718	0.6964	0.1856	0.0509	0.9360
	MANOWN	260	0.3225	0.3143	0.2982	0.0000	0.8965
	INSTOWN	260	0.1616	0.0538	0.2300	0.0000	0.9320
	STATEOWN	260	0.0446	0.0000	0.1677	0.0000	0.7427
	FOROWN	260	0.2820	0.1683	0.2905	0.0000	0.8360

, we present the descriptive statistics of the dataset. Our sample consists of 260 firm-year observations. The variables were divided into three categories. The four measures of tax avoidance are considered as dependent variables. Seven basic financial figures of the companies are used as control variables. Finally, the key elements of corporate governance are considered as independent variables.

In Appendix A, we present the correlation matrix where the Spearman correlation coefficients appear on the upper diagonal and the Pearson correlation coefficients on the lower diagonal. We observe a statistically significant correlation in 48.6% of cases based on the Pearson correlation coefficients and in 47.1% of cases based on the Spearman correlation coefficients. However, the correlations are moderate to low.

As can be seen from both the analytical correlation Appendix A and the summary

Table 2. Correlation Coefficients Clusters.

Correlation Coefficients	<0.8	<0.1	<0.05
all correlations	99.3%	44.7%	25.6%
SPEARMAN	99.6%	43.8%	26.1%
PEARSON	98.9%	45.7%	25.0%

with the percentages per cluster coefficient, in 99.3% of the variables, there is no significant correlation (<0.8), while typically in 44.7%, it is less than 0.1. Correspondingly, for the Spearman coefficient, the percentages are 99.6% and 43.8% and for the Pearson coefficient, 98.9% and 45.7%, respectively. Therefore, we can proceed to further analysis and control of the variables through our models.

Concerning the control variables, as can be seen from the results presented in Appendix A, the variable ROA is positively and significantly associated with the variables BTD, TEMBTD, and TAXEFFBTD, at the 1% significance level. The variable LOSS also exhibits strong positive significance across all three of these variables—at the 1% level for TEMBTD and TAXEFFBTD and the 5% level for BTD—while it shows a negative and statistically significant association at the 5% level with PERCING. The variables SIZE, CAPINT, INTAN, ETA, and MULT do not demonstrate any statistically significant association with the four dependent variables.

Table 3. Estimated Coefficients for PERCING, BTD, TEMBTD, and TEXEFFBTD.

	PERCINC	SIGN	VIF	BTD	SIGN	VIF	TEMBTD	SIGN	VIF	TAXEFF BTD	SIGN	VIF
Intercept	0.428 ***	0.006		0.023	0.830		−0.039	0.681		0.004	0.855
SIZE	−0.016 *	0.060	4.204	0.000	0.972	4.204	−0.003	0.537	4.204	0.000	0.869	4.204
ROA	0.005	0.962	1.908	0.677 ***	0.000	1.908	0.557 ***	0.000	1.908	0.144 ***	0.000	1.908
CAPINT	−0.172 ***	0.002	2.298	0.018	0.652	2.298	−0.001	0.977	2.298	0.004	0.661	2.298
INTAN	−0.112	0.110	1.551	−0.068	0.173	1.551	−0.048	0.273	1.551	−0.018 *	0.096	1.551
ETA	−0.102 **	0.013	1.790	−0.034	0.240	1.790	−0.024	0.342	1.790	−0.009	0.154	1.790
LOSS	−0.050 **	0.035	1.845	0.034 **	0.043	1.845	0.042 ***	0.004	1.845	0.010 ***	0.006	1.845
MULT	−0.061 **	0.026	2.584	−0.017	0.369	2.584	−0.007	0.659	2.584	−0.004	0.295	2.584
BSIZE	0.028	0.608	3.690	0.012	0.761	3.690	0.041	0.227	3.690	0.002	0.794
BIND	−0.107	0.151	2.159	0.106 **	0.045	2.159	0.052	0.260	2.159	0.021 *	0.065	3.690
BDIV	0.127	0.106	1.850	0.036	0.512	1.850	0.037	0.443	1.850	0.006	0.610	2.159
ACIND	0.024	0.584	3.299	−0.011	0.714	3.299	0.013	0.631	3.299	0.000	0.969	1.850
ACDIV	0.049	0.299	1.816	0.035	0.289	1.816	0.021	0.475	1.816	0.009	0.241	3.299
ACMEET	0.015	0.374	2.699	−0.011	0.335	2.699	−0.011	0.281	2.699	−0.003	0.280	1.816
BIG4	−0.078 ***	0.009	3.398	−0.054 **	0.011	3.398	−0.040 **	0.031	3.398	−0.011 **	0.014	2.699
EXTAUD	−0.030	0.190	1.685	−0.040 **	0.014	1.685	−0.028 **	0.050	1.685	−0.010 ***	0.007	3.398
OWNCONC	−0.051	0.474	2.760	0.030	0.554	2.760	0.091 **	0.040	2.760	0.011	0.322	1.685
MANOWN	0.139 ***	0.007	3.611	−0.082 **	0.023	3.611	−0.069 **	0.030	3.611	−0.023 ***	0.004	2.760
INSTOWN	0.102 *	0.064	2.495	−0.024	0.532	2.495	−0.021 **	0.545	2.495	−0.009	0.312	3.611
STATEOWN	0.025	0.733	2.461	−0.132 **	0.013	2.461	−0.094 **	0.042	2.461	−0.035 ***	0.003	2.495
FOROWN	0.162 ***	0.000	2.284	−0.019	0.509	2.284	−0.023	0.372	2.284	−0.006	0.353	2.284
Adj. R2	0.243			0.287			0.243			0.277
F-CHANGE	5.167	0.000		6.214	0.000		5.153	0.000		5.949	0.000

Notes: (1) Asterisks ***, **, * denote two-tailed statistical significance at 1%, 5%, and 10%, respectively; (2) Dependent variables: PERCINC, BTD, TEMBTD, and TEXEFFBTD; (3) Variable definitions: PERCINC: A measurement of tax avoidance as a percentage of pre-tax income, BTD: Total Book–Tax Difference, TEMBTD: Temporary differences between accounting and taxable income, TAXEFFBTD: Tax effects of Book–Tax Income Differences, SIZE: Natural logarithm of total assets, ROA: Return on Assets, CAPINT: Capital Intensity = Total Tangible Assets/Total Assets, INTAN: Intangible asset intensity = Total Intangible Assets/Total Assets, ETA: Capital Structure of Businesses = Total Equity/Total Assets, LOSS: Dummy variable that takes the value 1 if the firm had losses and the value 0 otherwise, MULT: Dummy variable that takes the value 1 if the company is a member of a multinational group of companies and the value 0 otherwise, BSIZE: Natural Logarithm of the Total Number of Board Members, BIND: Percentage of independent non-executive members of the Board of Directors, BDIV: Percentage of Women on the Board of Directors, ACIND: Percentage of independent Audit Committee members, ACDIV: Percentage of Women in the Audit Committee, ACMEET: Natural Logarithm of the Total Number of Audit Committee Meetings Held During a Year, BIG4: Dummy variable that takes the value 1 if the company hires from among the BIG4 auditing firms and the value 0 otherwise, EXTAUD: Dummy variable, which takes the value 1 if the company hires between Grant Thornton LLP and Crowe Horwath International auditing firms, and the value 0 otherwise, OWNCONC: Percentage of ownership held by the largest shareholders (over 5% each), MANOWN: Percentage of ownership held by management, INSTOWN: Percentage of ownership held by institutional investors (over 5% each), STATEOWN: Percentage of ownership controlled by the state, FOROWN: Percentage of ownership held by foreign investors (over 5% each).

presents the results of the linear regressions, where the four measures of tax avoidance serve as dependent variables and the key elements of corporate governance, along with the basic financial figures of the companies (control variables), are used as independent variables. The Adj.R2 range from 24.3% to 27.7%, satisfactory rates for our models, according to corresponding studies (Chytis et al., 2019; Koumanakos et al., 2017). At the same time, the F-change values show a p-value of 0.0, demonstrating the statistical significance of our models.

To examine potential multicollinearity problems among the independent variables included in the model under consideration, a test was performed using the Variance Inflation Factor (VIF). VIF is a widely accepted statistical indicator for estimating the intensity of the linear correlation between a variable and all the remaining independent variables. In the literature, VIF values greater than 10 are considered indicative of serious multicollinearity, while some researchers suggest a more stringent limit of 5 or even 3, especially in sensitive econometric models.

In the context of the present analysis, it was found that all variables presented VIF values significantly lower than the threshold of 3, which indicates the absence of strong or problematic linear dependence between them. Therefore, the independence of the variables concerning their correlation is documented, which enhances the reliability and statistical validity of the estimated parameters of the model. The absence of multicollinearity also makes the interpretations of the results more transparent and allows for a consistent assessment of the effects of each independent variable on the dependent variable.

As we can see in Table 3, ROA shows a statistically significant positive effect on the dependent variables BTD, TEMBTD, and TAXEFFBTD at a level of significance of 1%. LOSS shows a statistically significant positive impact on BTD, TEMBTD, and TAXEFFBTF, but a statistically significant negative effect on PERCING at a level of significance of 5%. BIG4 shows a statistically negative impact at the 1% level on PERCING and the 5% level on BTD, TEMBTD, and TAXEFFBTD. MANOWN shows a statistically significant negative impact on BTD, TEMBTD, and TAXEFFBTF, but a statistically significant positive effect on PERCING at a level of significance of 1%. Similarly, EXTAUD shows a statistically significant negative impact on the dependent variables BTD, TEMBTD, and TAXEFFBTD. In addition, INSTOWN is significant at the 10% level for PERCING and the 5% level for TEMBTD. BIND shows a statistically significant positive impact on BTD and TEXEFFBTD at the 5% and 10% level, respectively. Finally, OWNCONC and FOROWN show statistical significance in only one dependent variable (TEMBTD and PERCING, respectively), while BSIZE, BDIV, ACIND, ACDIV, and ACMEET do not show any statistical significance.

The analysis of the results is carried out per hypothesis tested as follows:

4.1. Board Size (H1) and Independence (H2)

In the regression analyses conducted to examine the effect of board characteristics—specifically, board size (BSIZE) and independence (BIND)—on the tax behavior of Greek firms operating in the services sector, both variables exhibited statistically insignificant relationships with the majority of tax avoidance proxies. Board size consistently showed no significant association with any of the four dependent variables, leading to the rejection of Hypothesis 1. Likewise, board independence was found to be statistically significant only at the 5% level for BTD, a metric that also captures earnings management practices beyond tax avoidance. Given the absence of significance across the remaining three proxies, Hypothesis 2 is also rejected.

The lack of significant relationships between both board size and board independence with corporate tax avoidance can be primarily attributed to the same institutional and operational particularities that characterize the Greek business environment. Corporate governance in Greece is frequently marked by formal compliance rather than substantive implementation, with boards of directors often composed of members with a limited supervisory role, either due to symbolic appointments or personal affiliations (Kourdoumpalou & Karagiorgos, 2012). This is particularly relevant in the context of concentrated or family-based ownership structures, where critical decisions–including those related to tax strategy–are typically made by a small group of key individuals (La Porta et al., 1999; Sartaji & Hassanzadeh, 2014).

Despite regulatory efforts to enhance board independence, such as Laws 4449/2017 and 4706/2020, most Greek firms appear to fulfill only the minimum formal requirements, without developing a culture of active oversight. As Grose et al. (2021, p. 12) note, “the majority of firms were strictly in line with the absolute minimum requirements,” while the concept of an “independent director” does not necessarily reflect substantive independence, particularly in a context where personal or ownership-related ties frequently shape board composition. The limited presence of tax-specific expertise among independent directors further constrains their ability to effectively oversee tax planning decisions (Armstrong et al., 2015).

The period under examination (2014–2023) was marked by heightened external uncertainty—including capital controls, fiscal policy volatility, and financing constraints—which led firms to prioritize short-term liquidity and viability over the formulation of long-term tax efficiency strategies. Moreover, the relatively low variation in board size, particularly in the services sector, may have further limited the explanatory power of this variable. Overall, despite the existence of a regulatory framework aligned with international standards, the practical implementation of governance reforms remained limited throughout the period under review. As highlighted in OECD (2025, p. 13), there is an ongoing need for greater transparency and the adoption of mechanisms that go beyond formal compliance to combat tax evasion and avoidance. Under these conditions, both board size and independence appear to exert limited practical influence on the tax behavior of Greek firms.

Concerning the size of the board of directors, the results are in line with prior research by Minnick and Noga (2010), Lanis and Richardson (2011), Khaoula and Ali (2012), Oyeleke et al. (2016), Chytis et al. (2020), Omesi and Appah (2021), and Salaudeen and Abdulwahab (2022). However, they contrast with the findings of Moore (2012), Ribeiro (2015), Gomes (2016), Halioui et al. (2016), Pratama (2017), E. A. A. Wahab et al. (2017), Hoseini et al. (2018), Hsu et al. (2018), Ogbeide and Obaretin (2018), Mappadang (2019), Osebe et al. (2019), Osamudiame et al. (2019), Alkurdi and Mardini (2020), Ba’aba (2020), Bashiru et al. (2020), Nwezoku and Egbunike (2020), Salhi et al. (2020), Al Lawati and Hussainey (2021), Egbunike et al. (2021) and García-Meca et al. (2021), who identify board size as a significant determinant of firms’ tax-related behavior.

As for board independence, the findings are consistent with previous studies conducted by Zhou (2011), Lanis and Richardson (2012), Chan et al. (2013), Sartaji and Hassanzadeh (2014), Halioui et al. (2016), Pratama (2017), E. A. A. Wahab et al. (2017), Chytis et al. (2019), Sunarto et al. (2021). However, they contrast with the results of Lanis and Richardson (2011, 2015), Mahenthrian and Kasipillai (2011), Khaoula and Ali (2012), Taylor and Richardson (2014), Armstrong et al. (2015), Gomes (2016), Ribeiro (2015), Oyeleke et al. (2016), Richardson et al. (2016), Sunarsih and Oktaviani (2016), Tandean and Winnie (2016), Lanis et al. (2017), Moore et al. (2017), Hsu et al. (2018), Lanis and Richardson (2018), Ogbeide and Obaretin (2018), Aburajab et al. (2019), Budi (2019), Mappadang (2019), Osebe et al. (2019), Zheng et al. (2019), Alkurdi and Mardini (2020), Chytis et al. (2020), Salhi et al. (2020), Al Lawati and Hussainey (2021), Flamini et al. (2021), Egbunike et al. (2021), Omesi and Appah (2021), Amalia and Firmansyah (2022) and C. Pavlou et al. (2025), who identify board independence as a significant determinant of firms’ tax-related behavior.

4.2. Audit Committee Size (H3) and Independence (H4)

The regression analysis revealed that neither audit committee size (ACMEET) nor audit committee independence (ACIND) demonstrated a statistically significant relationship with any of the four dependent variables across all models. Accordingly, both Hypothesis 3 and Hypothesis 4 are rejected. These findings suggest that neither the numerical composition nor the independence status of audit committees appears to influence the extent of corporate tax avoidance in Greek firms during the examined period.

This outcome may be attributed to structural characteristics of the Greek corporate environment, where corporate governance reforms have been largely formal rather than substantive. Although audit committees are legally mandated and expected to consist predominantly of independent members (Laws 4449/2017 and 4706/2020), their actual role in tax oversight often remains marginal. The establishment and operation of audit committees frequently reflect a compliance-oriented approach, lacking the functional depth necessary to exert meaningful influence over complex and strategic areas such as tax planning (OECD, 2024).

Specifically, an increase in the number of audit committee members does not automatically translate into enhanced monitoring capacity, particularly when members lack technical expertise, tax-specific knowledge, or active engagement (Armstrong et al., 2015; Warih, 2019; Delis et al., 2021). Similarly, the presence of formally independent members may not guarantee genuine oversight when appointments are influenced by personal or professional affiliations, thus undermining the principle of effective independence (Kourdoumpalou & Karagiorgos, 2012; Spanos, 2005). Under such conditions, audit committees function more as procedural formalities than as substantive control mechanisms capable of constraining tax aggressiveness.

The limitations of audit committee effectiveness are further exacerbated by the ownership structures typical of Greek firms. The prevalence of concentrated or family-controlled ownership restricts the scope of board-level influence, as key strategic decisions—including those relating to taxation—are often centralized among dominant shareholders or a narrow executive elite, bypassing formal governance channels (La Porta et al., 1999; Sartaji & Hassanzadeh, 2014; Donelson et al., 2021).

These findings are consistent with prior research reporting an insignificant relationship between audit committee characteristics and tax avoidance (Kourdoumpalou, 2010; Mahenthrian & Kasipillai, 2011; Gunawan & Resitarini, 2019; Omesi & Appah, 2021; Anjarwi et al., 2024; Rudiatun & Anggorowati, 2024).

However, they contrast with studies that identify a significant role of audit committee size and independence in curbing aggressive tax practices, particularly in institutional settings with higher enforcement and transparency standards (Moore, 2012, concerns the size of the Audit Committee; Richardson et al., 2013; Sunarsih & Oktaviani, 2016; Tandean & Winnie, 2016; Pratama, 2017; Waluyo, 2017; Hsu et al., 2018; Budi, 2019; Zheng et al., 2019; Agraha et al., 2020; Hendrani et al., 2020; Wijaya & Hasbiy, 2020; Egbunike et al., 2021; García-Meca et al., 2021; Sunarto et al., 2021; Almaharmeh et al., 2024; Anjarwi et al., 2024).

4.3. Female Presence on the Board of Directors (H5) and on the Audit Committee (H6)

In the regression analyses conducted to examine the effect of Female presence on the board of directors (BDIV) and the Audit Committee (ACDIV) on the tax behavior of Greek firms operating in the services sector, both variables exhibited statistically insignificant relationships with all four of the tax avoidance proxies, leading to the rejection of Hypotheses 5 and 6. The absence of a statistically significant association between female representation on Boards of Directors and Audit Committees and corporate tax avoidance levels in Greek firms during the period 2014–2023 may be attributed to a combination of institutional, cultural, and functional factors. Until 2020, there was no gender quota requirement for boards in Greece, resulting in low female representation, which often served a symbolic rather than substantive role. Law 4706/2020, which mandates at least 25% female participation on the boards of listed companies, came into force in 2021 and thus applies only to a small fraction of the examined period. Even in cases where women did participate in board structures, their influence over strategic areas such as tax planning may have been limited due to a lack of relevant expertise or insufficient institutional empowerment (Wahid, 2019).

The prevailing corporate governance framework in Greece is often characterized by formal compliance rather than substantive oversight mechanisms, which reduces the capacity of women in non-executive or supervisory roles to actively shape tax-related strategies (Kourdoumpalou & Karagiorgos, 2012). Furthermore, male-dominated governance structures, the prevalence of family ties among board members, and centralized decision-making processes tend to undermine the actual impact of female directors, regardless of their formal role (La Porta et al., 1999; Adams & Ferreira, 2009).

International literature suggests that the positive effect of female board presence on limiting tax avoidance becomes more evident in environments with strong regulatory frameworks, institutional maturity, and robust accountability mechanisms—conditions that do not fully characterize the Greek context (Lanis et al., 2017; C. Pavlou et al., 2025). In particular, C. Pavlou et al. (2025), based on a cross-country empirical analysis, report that female representation is associated with higher effective tax rates mainly in regulated industries, implying that the effectiveness of board gender diversity depends on the extent to which board roles are institutionally supported.

Moreover, even the recent legislative shift toward increased female participation may not have yet yielded measurable outcomes, given that the effects of such institutional reforms typically require time to materialize in observable firm behavior. Therefore, the lack of association between female board participation and corporate tax avoidance may reflect not only institutional gaps and cultural rigidities but also a time lag in the effectiveness of these policy interventions (Terjesen et al., 2009).

These results support previous studies, including those by Khaoula and Ali (2012), Oyeleke et al. (2016), Budi (2019), Gunawan and Melisa (2017), Salaudeen and Abdulwahab (2022) and Anjarwi et al. (2024), but differ from the conclusions drawn by Friedland et al. (1978), Mason and Calvin (1978), Torgler and Schneider (2004), Gërxhani (2007), Kastlunger et al. (2010), Matsaganis and Flevotomou (2010), Bazart and Pickhardt (2011), Francis et al. (2014), Oyeleke et al. (2016), Richardson et al. (2016), Lanis et al. (2017), Osebe et al. (2019), Jarboui et al. (2020), Nwezoku and Egbunike (2020), Salhi et al. (2020), García-Meca et al. (2021), Suleiman and Abubakar (2021), Rakia et al. (2024) and C. Pavlou et al. (2025).

4.4. Type of External Auditor (H7)

The empirical analysis reveals a negative and statistically significant association between the BIG4 variable and all four measures of tax avoidance, with significance levels ranging from 10% to 1% for PERCINC and BTD, respectively. Similarly, the EXTAUD variable is found to be negatively significant across all dependent variables except PERCINC. These findings provide support for Hypothesis 7 and suggest that the quality of the external auditor functions as a deterrent to the implementation of aggressive tax strategies among Greek firms.

This result reinforces the view that high audit quality can serve as a mechanism that constrains tax avoidance, primarily through the auditor’s independence, technical competence, industry-specific knowledge, professional experience, and institutional conservatism. In addition, audit quality mitigates conflicts of interest. It promotes reputational discipline, as audit firms—particularly those belonging to the Big 4—seek to protect their credibility and long-term market position (DeAngelo, 1981; Hanlon & Slemrod, 2009; Klassen et al., 2016; J. H. Kim & Im, 2017; Salaudeen & Abdulwahab, 2022). The findings are consistent with prior studies that also report an inverse relationship between audit quality and tax aggressiveness (DeAngelo, 1981; Richardson et al., 2013; Taylor & Richardson, 2014; Lanis & Richardson, 2015; Richardson et al., 2016; Sunarsih & Oktaviani, 2016; Tandean & Winnie, 2016; Lanis et al., 2017; Pratama, 2017; Waluyo, 2017; Kanagaretnam et al., 2018; Suyono, 2018; Vintilă et al., 2018; Budi, 2019; Alkurdi & Mardini, 2020; Wijaya & Hasbiy, 2020; Al Lawati & Hussainey, 2021; Omesi & Appah, 2021; Oussii & Klibi, 2024). Conversely, the results diverge from those of studies reporting either positive or statistically insignificant associations, such as Cook et al. (2008), Crabbé (2010), Mahenthrian and Kasipillai (2011), Huseynov and Klamm (2012), Lisowsky et al. (2013), McGuire et al. (2013), Donohoe and Robert Knechel (2014), Amidu et al. (2016a), Klassen et al. (2016), J. H. Kim and Im (2017).

In the specific context of Greece, the present findings stand in contrast to earlier research focused on domestic firms. More specifically, they contradict the results of Chytis et al. (2019), and Chytis et al. (2020), who found no statistically significant effect of audit quality on tax behavior, as well as the study by Kourdoumpalou (2010), which reported a positive association. This divergence can be attributed to several critical factors.

First, the temporal scope of each study is an important differentiating element. Kourdoumpalou’s (2010) study examined a pre-crisis period, during which the institutional and audit oversight framework in Greece was relatively weak, and the role of Big 4 audit firms often extended into tax advisory services. Similarly, Chytis et al. (2019) conducted their research during the years of economic crisis and institutional transition, a period marked by regulatory uncertainty. In contrast, the present study focuses on the post-2014 period, during which the institutional environment in Greece underwent a significant transformation. These include the operational independence of the Independent Authority for Public Revenue (AADE), the adoption of reforms aligned with OECD’s BEPS actions and the EU’s DAC directives, and the tightening of audit-related regulatory oversight. Collectively, these developments redefined the role of Big 4 audit firms from passive participants to active enforcers of tax compliance.

Moreover, the role of high-quality audit firms has evolved significantly in recent years. Whereas in the past there may have been more latitude for auditors to tacitly tolerate, or even facilitate, aggressive tax planning, today the context is fundamentally different. The reputational and legal risks associated with such behavior have increased considerably, prompting the Big 4 to place greater emphasis on independence, transparency, and professional accountability.

Finally, methodological differences across studies should also be acknowledged. The present research employs refined and widely accepted proxies for tax avoidance—such as book-tax differences and cash effective tax rates—while also covering a broader period and incorporating a larger sample size. These design choices enhance the reliability and generalizability of the findings.

In summary, the negative association between Big 4 audit firms and tax avoidance identified in this study reflects both structural changes in the Greek regulatory landscape and the broader international shift in the role of external auditors toward ensuring greater tax transparency and compliance.

4.5. Ownership Concentration (H8) and Institutional Ownership (H11)

The regression analysis revealed that Ownership Concentration (OWNCON) was statistically significant (positive) only at the 10% level for TEMBTD, while Institutional Ownership (INSTOWN) was significant solely at the 5% level for PERCINC. Both variables displayed statistically insignificant associations with the majority of tax avoidance proxies, leading to the rejection of Hypotheses 8 and 11.

Ownership concentration among Greek service-sector firms appears to support practices of deferring tax liabilities weakly, whereas institutional ownership is marginally associated with the avoidance of tax payments. The overall lack of statistically significant effects for both variables may be attributed to institutional specificities within the Greek corporate governance framework and structural characteristics of the domestic capital market.

The service-sector firms included in the study exhibit, in the majority, a high degree of ownership concentration. As a consequence, the examined firms demonstrate similar patterns concerning this parameter. At the same time, institutional investors in Greece rarely adopt an active monitoring role, and their participation is generally characterized as passive and short-term in nature (Aguilera & Jackson, 2003). The absence of a culture of active shareholder engagement limits the potential to influence corporate tax strategy.

Furthermore, compliance with corporate governance principles in Greece is mainly formalistic, lacking substantive mechanisms of accountability and oversight. This renders the monitoring of tax-related practices ineffective, even in cases where ownership concentration is high (Kourdoumpalou & Karagiorgos, 2012; OECD, 2021). Notably, the period 2014–2023 was marked by significant economic instability, during which the survival and liquidity preservation of firms were prioritized, often at the expense of compliance with principles of tax transparency.

Concerning the Ownership Concentration, the results are in line with prior research by Mafrolla and D’Amico (2016), Khan et al. (2017), Salaudeen and Ejeh (2018), Budi (2019), Chytis et al. (2019), Chytis et al. (2020), and Salaudeen and Abdulwahab (2022). However, they contrast with the findings of X. Liu and Cao (2007), S. Chen et al. (2010), Zhou (2011), Lanis and Richardson (2012, 2015), Richardson et al. (2016), J. Cao and Cui (2017), Lanis et al. (2017), Moore et al. (2017), Ogbeide and Obaretin (2018), Bradshaw et al. (2019), Brune et al. (2019), Osebe et al. (2019), Salhi et al. (2020), Shin and Park (2020), Flamini et al. (2021), García-Meca et al. (2021), Omesi and Appah (2021), F. Cao et al. (2023), Almaharmeh et al. (2024), Chakroun and Ben Amar (2024) and Chalevas et al. (2024) who identify Ownership Concentration as a significant determinant of firms’ tax-related behavior.

In the specific context of Greece, the present findings diverge from prior research on domestic firms (Kourdoumpalou, 2016). This divergence may be attributed to differences in the institutional and macroeconomic environment, as well as to evolving patterns in corporate culture following the period of fiscal adjustment. Moreover, inconsistencies in the operationalization and measurement of ownership concentration and tax avoidance indicators may have contributed to the observed discrepancy. Overall, our results support the view that concentrated ownership, in itself, is insufficient to constrain aggressive tax practices when effective governance mechanisms and substantive oversight do not accompany it.

As for Institutional ownership, the findings are consistent with previous studies conducted by S. Chen et al. (2010), Otieno (2014), Sartaji and Hassanzadeh (2014), Sunarsih and Oktaviani (2016), Tandean and Winnie (2016), Yuniarsih (2018), Agraha et al. (2020), Bashir and Zachariah (2020), Udisifan (2020), Hasan et al. (2022) and Rudiatun and Anggorowati (2024). However, they contrast with the results of Khurana and Moser (2010), Huseynov and Klamm (2012), Moore (2012), Hasan et al. (2022), Khan et al. (2017), Moore et al. (2017), E. A. A. Wahab et al. (2017), Hsu et al. (2018), Alkurdi and Mardini (2020), Sunarto et al. (2021), Dakhli (2022), Athira and Ramesh (2023), F. Cao et al. (2023), Benkraiem et al. (2024), Oussii and Klibi (2024) and Sarhan (2024), who identify Institutional ownership as a significant determinant of firms’ tax-related behavior.

4.6. Managerial Ownership (H9)

As can be seen from the results presented in Table 3, the variable MANOWN is statistically significant across all dependent variables. A strong positive association is observed with PERCINC at the 1% level of significance, while negative associations are recorded with the remaining three variables, with significance levels ranging from 5% to 1%. This variable exhibits a robust positive relationship with the measure that captures total tax avoidance—encompassing both compliant and non-compliant practices—and a negative relationship with the indicators that reflect only non-compliant tax avoidance (Atwood et al., 2012, p. 1838). These findings suggest that tax avoidance levels among Greek firms increase as managerial ownership rises, and more specifically, they indicate the use of compliant tax avoidance strategies, which involve the manipulation of both taxable and accounting income. Accordingly, Hypothesis 9 is accepted, while the findings do not support the view proposed by Fama and Jensen (1983) and Badertscher et al. (2013) that, when ownership and decision-making authority are concentrated in a small group of individuals, corporate tax behavior tends to be less aggressive and more compliant. However, the findings are in line with the stream of literature represented by Morck et al. (1988) and Short and Keasey (1999), who argue that low levels of managerial ownership reduce opportunities for self-serving behavior, better safeguard the interests of external shareholders, and diminish the likelihood of engaging in activities detrimental to the interests of other shareholders, including tax-reducing practices aimed at promoting personal gains (Florackis, 2008; Salaudeen & Abdulwahab, 2022). The results are consistent with prior studies conducted by Zhou (2011), Chan et al. (2013), Ribeiro (2015), Salaudeen and Ejeh (2018), and Cabello et al. (2019). However, they contrast with the findings of Zeng (2011), Moore (2012), Badertscher et al. (2013), Lanis and Richardson (2015), Kourdoumpalou (2016), Sunarsih and Oktaviani (2016), Lanis et al. (2017), Ogbeide and Obaretin (2018), Yuniarsih (2018), Alkurdi and Mardini (2020), Udisifan (2020), Salaudeen and Abdulwahab (2022), Oussii and Klibi (2024), and Sarhan (2024).

4.7. State Ownership (H10)

As can be seen from the results analyzed in Table 3, the variable STATETOWN is significantly negative for all variables except PERCINC. These findings provide support for Hypothesis 10 and suggest that State ownership functions as a deterrent to the implementation of aggressive tax strategies among Greek firms. The findings of the present study are consistent with prior empirical evidence suggesting that state-owned enterprises tend to exhibit higher levels of tax compliance, not as a result of economic rationality but due to political considerations. Their tax strategies are shaped in a way that primarily serves the interests of the dominant shareholder—the state—even at the expense of minority shareholders. In this context, tax payments operate as a form of silent dividend to the government, effectively expropriating part of the wealth of other shareholders. At the same time, these enterprises bear a significant political cost by accepting a higher tax burden, thereby contributing to redistributive policies through the transfer of wealth to broader social groups (Chan et al., 2013; Bradshaw et al., 2019).

These findings are in line with the results of prior empirical studies, such as those of Tedds (2006), Mahenthrian and Kasipillai (2011), Zhou (2011), Zhang et al. (2012), Chan et al. (2013), Wu et al. (2013), Wijaya and Hasbiy (2020), Hilling et al. (2021), Bradshaw et al. (2019). While they are not consistent with the findings of studies conducted by Nicodème (2002), Derashid and Zhang (2003), Adhikari et al. (2006), Wu et al. (2012), D. H. Huang et al. (2013), Salihu et al. (2015), C. Kim and Zhang (2016), E. A. A. Wahab et al. (2017), Long et al. (2024), Paembonan et al. (2024) and Ridwansyah (2024).

4.8. Foreign Ownership (H12)

The variable FORTOWN is strongly and positively associated with PERCINC, with the relationship being statistically significant at the 1% level. Based on this finding, Hypothesis 12 is accepted. The result suggests that foreign ownership contributes to higher levels of tax avoidance, as captured by this specific measure. A possible explanation lies in the fact that foreign investors—particularly institutional and multinational ones—tend to pursue more aggressive tax planning strategies aimed at maximizing after-tax returns (Desai & Dharmapala, 2006; Hanlon & Heitzman, 2010). During the period 2014–2023, the Greek institutional environment was characterized by fiscal instability, regulatory shifts, and relatively weak enforcement mechanisms (OECD, 2020), potentially encouraging foreign shareholders to engage in short-term return-maximization strategies with limited concern for long-term compliance. In addition, the weaker embeddedness of foreign owners in the local socio-institutional framework may reduce reputational concerns associated with aggressive tax behavior (Lanis & Richardson, 2012; Chytis et al., 2019). The use of complex cross-border structures, often involving intermediate entities located in tax-favorable jurisdictions, may further facilitate income-shifting practices and reduce transparency, ultimately intensifying firms’ capacity for tax avoidance (Gravelle, 2009; Zucman, 2014).

These findings are in line with the results of prior empirical studies, such as those of Demirgüç-Kunt and Huizinga (2001), Wilkinson et al. (2001), Langli and Saudagaran (2004), Egger et al. (2010), D. H. Huang et al. (2013), Salihu et al. (2015), and Alkurdi and Mardini (2020).

However, they contrast with the findings of Joulfaian (2000), Feeny et al. (2005), Hanlon et al. (2005), Tedds (2006), Hasan et al. (2022), Shin and Park (2020), Udisifan (2020), Hasan et al. (2022), and C. Pavlou et al. (2025).

5. Conclusions

This study contributes to the literature on corporate governance and tax avoidance by examining a large panel of Greek service-sector firms during the period 2014–2023, a context characterized by regulatory transition, fiscal instability, and evolving governance frameworks. The evidence suggests that several traditional governance mechanisms—such as board size, board independence, audit committee size and independence, and gender diversity on boards and audit committees—do not exert a significant influence on corporate tax behavior. These findings underscore the enduring gap between formal regulatory compliance and substantive governance practices in Greece, where concentrated ownership structures, family-controlled firms, and symbolic appointments undermine the supervisory role of boards.

By contrast, external audit quality and ownership structure emerge as critical determinants of firms’ tax strategies. Engagement with high-quality auditors, particularly Big 4 firms, significantly constrains tax avoidance, consistent with theories emphasizing reputational capital, monitoring capacity, and institutional conservatism (DeAngelo, 1981; Hanlon & Slemrod, 2009). Similarly, state ownership is associated with reduced aggressiveness in tax planning, reflecting the political and redistributive incentives of state-controlled enterprises. Conversely, managerial and foreign ownership are positively associated with tax avoidance, suggesting the pursuit of private benefits and short-term return maximization, often facilitated by weak institutional enforcement. These results resonate with the agency-theoretic perspective, whereby divergent shareholder and managerial incentives shape the intensity and form of tax planning activities (Fama & Jensen, 1983; Desai & Dharmapala, 2006).

Theoretically, the findings highlight the importance of contextual and institutional factors in mediating the relationship between governance mechanisms and corporate tax behavior. While agency theory and resource dependence theory predict that independent boards, diverse committees, and active institutional investors should mitigate aggressive tax strategies, such mechanisms appear ineffective in environments where independence is formal rather than substantive and where shareholder monitoring is passive. This aligns with the institutional theory perspective, which posits that governance structures embedded in weak enforcement regimes may operate as symbolic devices of legitimacy rather than effective control mechanisms (Aguilera & Jackson, 2003).

The findings of this study make a substantive contribution to the literature on corporate governance and tax behavior in developing countries by demonstrating that, within the institutional context of Greece, traditional governance mechanisms do not exert meaningful constraints on tax avoidance. This result advances existing knowledge by challenging the generalizability of agency theory predictions and prior international evidence, which often emphasize the disciplining role of internal governance features.

Instead, our results highlight the pivotal importance of external audit quality and ownership structure as determinants of tax aggressiveness, underscoring the contextual dependence of governance effectiveness. By showing that external oversight mechanisms (e.g., Big 4 auditors) and politically accountable shareholders (e.g., state ownership) are more influential in curbing tax avoidance than formal internal structures, the study extends institutional theory arguments about the limits of formal compliance in weak enforcement environments. These insights enrich comparative governance research by illustrating how governance reforms, when adopted in a largely symbolic manner, may fail to achieve their intended outcomes.

More broadly, the study informs both scholars and policymakers by identifying the conditions under which governance reforms in developing countries can effectively mitigate opportunistic tax behavior, thereby advancing a nuanced understanding of governance effectiveness in emerging and institutionally fragile markets.

The results call for a re-examination of the conditions under which governance mechanisms can constrain opportunistic practices and for a broader theoretical integration that accounts for institutional quality and cultural specificities. Future research may extend these findings through comparative analyses across jurisdictions with varying institutional maturity, thereby enriching our understanding of how governance mechanisms interact with regulatory and cultural environments in shaping corporate tax behavior.

Despite these contributions, the study is subject to certain limitations. First, the analysis is restricted to service-sector firms, which may limit the generalizability of the findings to industries with different capital structures, international exposure, or regulatory sensitivities. Second, tax avoidance is measured using accounting-based proxies (book–tax differences, effective tax rates), which, although widely adopted, may not fully capture the complexity of tax planning strategies. Third, the study is context-specific, reflecting the unique institutional, regulatory, and cultural conditions of Greece during a period of fiscal adjustment and reform; the results may differ in more mature governance environments.

Future research could address these limitations by expanding the scope across industries and countries, allowing for comparative institutional analyses of how governance mechanisms operate under varying levels of regulatory enforcement and investor protection. Additionally, the incorporation of qualitative approaches—such as interviews with board members, auditors, and tax practitioners—could offer richer insights into the informal and behavioral dimensions of governance and tax strategy. Finally, longitudinal studies beyond 2023 would allow for an assessment of whether recent reforms, such as gender quota laws and enhanced audit oversight, translate into substantive governance improvements and measurable changes in tax avoidance practices.

In conclusion, the findings emphasize that the relationship between governance and tax avoidance is highly contingent upon institutional quality. Strengthening enforcement, fostering genuine independence, and cultivating an active monitoring culture are essential for transforming governance from a symbolic exercise into a practical constraint on aggressive tax behavior.