The Impact of Board Characteristics on Tax Avoidance: Do Industry Regulations Matter?

Abstract

This paper examines the effect of board characteristics on tax avoidance and the moderating role of industry regulation on this effect. Using a comprehensive panel of 84,153 firm-year observations from 39 countries during the period of 2000–2023, we illustrate that larger boards, higher female representation, significant foreign ownership, and the presence of independent directors are generally associated with higher effective tax rates, suggesting lower levels of tax avoidance. This study further demonstrates that the effects of board gender diversity and board independence are more pronounced in regulated industries, where stringent governance and ethical standards prevail, emphasizing the importance of regulatory oversight in mitigating aggressive tax planning. These findings are crucial for policymakers, regulators, and corporate governance practitioners aiming to align corporate practices with ethical standards and reduce the risks associated with tax avoidance.

1. Introduction

Taxes have been seen as a major expense for companies, significantly impacting their financial resources. This reduction in cash flow directly affects the amount of returns available for distribution to their shareholders. On the other hand, taxes are a critical component of government financing worldwide. According to the OECD’s Global Revenue Statistics Database, tax revenues represented an average of 34% of GDP among OECD countries in 2022, indicating the substantial role taxes play in the economic structure of these nations. According to Weller and Rao (2010) and Sjoquist et al. (2011), taxes are essential for the operation of a country in order to enhance the economic, social, and political stability. Without adequate tax revenue, governments would struggle to fulfill these roles effectively. Therefore, both corporate and individual taxpayers are required to be compliant with tax laws and obedient in fulfilling their tax duties voluntarily. Non-compliance among taxpayers might disturb nations’ budgets. One way of non-compliance is achieved by way of tax avoidance, which is a legal strategy employed by individuals and businesses to lower their tax liabilities through various means, such as exploiting loopholes in tax legislation, structuring transactions to take advantage of tax benefits, and employing complex financial instruments (Riedel, 2018; Kovermann & Velte, 2019; F. Wang et al., 2020).

Corporate governance is a significant factor that may influence tax avoidance practices (Kovermann & Velte, 2019), which may either curb or encourage aggressive tax strategies. Therefore, the link between corporate governance and tax avoidance is controversial and has attracted numerous debates in academia as well as in the business community (Stephenson & Vracheva, 2015; Kovermann & Velte, 2019; Lubis et al., 2023). A. Hasan et al. (2024) argue that corporate governance mechanisms, such as board independence and gender diversity, lead to greater transparency and reduced tax avoidance. Others suggest that certain governance structures may actually increase tax avoidance through sophisticated tax planning strategies (e.g., Yee et al., 2018; Zudana et al., 2021).

The relationship between board size and tax avoidance remains controversial regarding how tax avoidance is shaped. There are studies that argue that larger boards may increase tax avoidance due to a broader range of expertise (Ogbodo & Omonigho, 2021; Kalbuana et al., 2023). On the other hand, there are studies that believe that larger boards enhance oversight and transparency. As a result, larger boards may contribute to reducing tax avoidance (Halioui et al., 2016; Ali et al., 2024). A similar controversy exists in the literature regarding the gender diversity that makes up boards of directors. Boards of directors with a high presence of women are positively and significantly associated with greater transparency and high ethical standards. Therefore, these factors can limit tax avoidance practices (Lanis et al., 2017; Hoseini et al., 2019). On the other hand, there are studies that conclude with opposite results. That is, they conclude that the presence or absence of women on the board of directors does not affect the willingness to avoid taxes (Garcia-Blandon et al., 2022). Another topic of discussion is how independent directors influence the final decisions on boards of directors. Bodies of studies support the neutrality of independent directors as well as the objectivity they have during their supervision. Such an attitude reduces tax avoidance (Salhi et al., 2020; Ali et al., 2024). The other side argues that there is an incentive for independent directors to enhance tax avoidance strategies, since they aim to satisfy the shareholder demands (McClure et al., 2018; Chytis et al., 2020). The literature on the effect of foreign ownership on the board of directors and tax avoidance is divided. There are studies that argue that the presence of foreign investors can enhance tax avoidance (Fuest & Hemmelgarn, 2005; Egger et al., 2010) and studies that argue the opposite since the presence of foreign investors can reduce tax avoidance (Yoo & Koh, 2014; I. Hasan et al., 2022).

Corporate governance transparently aligns management decisions with the interests of shareholders in both regulated and unregulated industries (Becher & Frye, 2011). In regulated industries, there is an existing regulatory framework from which corporate governance is shaped. Regulatory mechanisms provide the opportunity for external oversight. And the combination of these aims to ensure financial stability and compliance with the relevant legislation (Joskow et al., 1993; Becher & Frye, 2011). Beyond the theory that regulation replaces governance, empirical results show the opposite. Regulation and governance work complementarily. This is because regulatory pressure leads firms to adopt stronger forms of corporate governance (Booth et al., 2002), to have larger boards of directors, and to have a higher proportion of independent supervisory members (Becher & Frye, 2011). On the other hand, deregulation often leads to a weakening of a firm’s governance mechanisms (Kole & Lehn, 1997; Becher et al., 2005). In unregulated industries, corporate governance mechanisms are mainly developed internally. Board independence and ownership concentration are key factors in reducing conflicts between management and shareholders (Shleifer & Vishny, 1997).

There are several studies that examine the effects of key features of corporate governance on tax avoidance. Among them are board size, female board representation, and the proportion of foreign ownership in tax planning (Yahaya et al., 2025). There are also some studies that examine the role of corporate governance in both regulated and unregulated sectors (Becher & Frye, 2011; Kole & Lehn, 1997). This creates a research gap that makes this study possible. In particular, the comparison of the effects of governance on tax avoidance between regulated and unregulated sectors has not been sufficiently examined. Despite the comprehensive insights from previous literature addressing governance impacts broadly, the differential regulatory contexts in which industries operate may substantially alter the effectiveness and mechanisms through which governance structures influence tax avoidance behaviors. Therefore, this manuscript seeks explicitly to bridge this gap by investigating how corporate governance mechanisms interact with regulatory environments—distinctively assessing regulated and unregulated sectors—to influence firms’ tax avoidance strategies, thus contributing an integrated perspective that expands upon and refines existing understandings in this area of research. Using a panel of 84,153 firm-year observations from 39 countries during the period 2000–2023, our findings reveal that larger board sizes, higher female representation, significant foreign ownership, and the presence of independent directors are associated with lower levels of tax avoidance. Further, the effect of board gender diversity and board independence is more pronounced in industries that are regulated, as they are characterized by stricter governance and ethical standards.

Our study offers several contributions to accounting research. First, this study corresponds to the recommendations of Kovermann and Velte (2019), Lubis et al. (2023) and Yahaya et al. (2025) to conduct a critical examination of the behaviors of board characteristics in relation to tax avoidance and to identify the factors that may influence these behaviors.

Second, unlike prior studies that have focused on the direct effect of industry regulation on tax avoidance, our study distinguishes itself by exploring whether this relationship differs between firms in regulated and unregulated industries. To the best of our knowledge, this is the first study to investigate whether the relationship between board characteristics—such as board size, independence, gender diversity, and foreign ownership—and tax avoidance is different within varying regulatory environments. Our findings reveal that board gender diversity and board independence generally lead to higher effective tax rates, indicative of lower levels of tax avoidance, with more pronounced effects in regulated industries where stricter governance and ethical standards prevail. Therefore, our study not only broadens the understanding of corporate governance’s impact on tax strategies but also highlights the differential effects of industry regulation, marking a primary and novel contribution to the academic discourse. This differentiation is essential for policymakers, regulators, and corporate governance practitioners aiming to design effective frameworks to mitigate tax avoidance practices, particularly in environments where regulatory standards vary significantly.

Third, unlike previous literature that primarily focuses on single-country analyses (e.g., Anggraenia & Kurnianto, 2020; Chytis et al., 2020; Omesi & Appah, 2021; Utaminingsih et al., 2022; Kalbuana et al., 2023), our research is the first to investigate the impact of board characteristics on tax avoidance and the moderating role of industry regulation using a global sample of 84,153 firm-year observations from 39 countries between 2000 and 2023. Our results have significant implications for policymakers and practitioners who are attempting to reduce tax avoidance in a variety of regulatory environments.

Fourth, our study extends the existing literature by jointly applying agency theory and signaling theory to offer a richer theoretical lens for interpreting our findings. While agency theory has been widely employed to explain how board characteristics such as independence and gender diversity enhance monitoring and reduce managerial opportunism, we complement this perspective with signaling theory. The latter provides additional insight into how these governance attributes communicate a firm’s ethical orientation and commitment to transparency, which may, in turn, deter tax avoidance. By combining these two theoretical frameworks, we not only clarify the direct influence of board characteristics on tax avoidance but also highlight how regulatory contexts shape these dynamics, thus offering a more nuanced understanding of corporate governance’s role in shaping firms’ tax behavior (Kovermann & Velte, 2019; Lubis et al., 2023).

We organized the remaining portions of this study as follows: We provided a literature review and developed hypotheses in Section 2. In Section 3, we described sample selection, research design, and empirical models. In Section 4, we reported the results of the empirical analysis, and in Section 5, we presented the sensitivity analysis. Finally, we presented our conclusions in Section 6.

2. Literature Review and Hypotheses Development

According to Kovermann and Velte (2019), there is recently increasing interest in examining whether (or how) specific board characteristics affect tax avoidance, whereas there are no previous studies that investigate the moderating role of industry regulation on this relationship. As a result, this study investigates the effect of specific board characteristics (board size, board independence, board gender diversity, and board foreign ownership) on tax avoidance, as well as the moderating role of industry regulation on this link.

Recent research highlights a controversial relationship between board size and tax avoidance. Some studies suggest that larger boards may increase tax avoidance. Anggraenia and Kurnianto (2020) argue that bigger boards in Indonesian firms improve oversight capabilities, enabling sophisticated tax planning. Hoseini et al. (2019) also find that larger boards in Teheran firms correlate with higher tax avoidance, attributing this to the broader range of skills and experiences. Ogbodo and Omonigho (2021) and Kalbuana et al. (2023) observe that larger boards in Nigerian and Indonesian firms, respectively, engage in more tax avoidance. On the contrary, other literature suggests that larger boards may reduce tax avoidance (Ali et al., 2024). Halioui et al. (2016) argue that increased board size in USA firms promotes transparency, reducing tax avoidance. Abdul Wahab et al. (2017) observe that smaller boards in Malaysian firms reduce tax avoidance. Lastly, using sample from USA, Australia, Indonesia, Nigeria and Tunisia, Minnick and Noga (2010), Lanis and Richardson (2011), Jamei (2017) and Omesi and Appah (2021) and Amri et al. (2023) find no significant relationship between board size and tax avoidance, suggesting that larger boards do not necessarily improve control.

Similarly, studying the international literature, the relationship between board independence and tax avoidance is ambiguous. There are those studies that conclude that there is a positive relationship between the presence of independent directors on the board of directors and tax avoidance (Richardson et al., 2015, in USA; Mulyadi & Anwar, 2015, in USA; McClure et al., 2018, in Australia; Chytis et al., 2020, in Greece). Specifically, McClure et al. (2018) and Chytis et al. (2020), examined tax avoidance in Australia and Greece, respectively, finding high levels of tax avoidance in firms with high board independence. On the contrary, there are studies that conclude in a negative relationship between the effective supervision of independent directors and tax avoidance (Lanis & Richardson, 2011; Richardson et al., 2013, in Australia; Lanis & Richardson, 2018, in USA; Salhi et al., 2020, in UK and Japan; Ali et al., 2024, in Pakistan). For instance, Lanis and Richardson (2011) suggest that having a higher proportion of independent directors on the board in Australia firms leads to a decrease of a likelihood of tax avoidance. Further, in a sample of Pakistan firms, board independence reduces tax avoidance due to stricter tax policy enforcement (A. Hasan et al., 2024). There are also some studies that do not conclude in a statistically significant relationship between board independence and tax avoidance (Minnick & Noga, 2010, in USA; Chan et al., 2013, in China; Zemzem & Ftouhi, 2013, in France; Halioui et al., 2016, in USA; Sunarsih & Oktaviani, 2016, in Indonesia; Abdul Wahab et al., 2017, in Malaysia).

Additionally, the literature presents mixed results regarding the relationship between the presence of women on the board of directors and tax avoidance. On the one hand, Kastlunger et al. (2010) argue that women in Italy are more strictly in tax compliance. Richardson et al. (2016), Lanis et al. (2017), and Hoseini et al. (2019) show that the presence of women on boards of directors enhances ethical standards and the transparency of financial statements, thus reducing tax avoidance in Australia, USA and Tehran firms, respectively. Anggraenia and Kurnianto (2020), Riguen et al. (2020) and Jarboui et al. (2020) also demonstrate that an increase in the number of women on the board is associated with a reduction in tax avoidance in Indonesian and UK firms, respectively. Using listed firms from Tunisia, Indonesia and Pakistan, Boussaidi and Hamed-Sidhom (2021), Utaminingsih et al. (2022) and Ali et al. (2024) reach the same results, reinforcing the hypothesis that the presence of women is associated with lower tax avoidance, since women with high ethical standards tend not to take high-risk decisions and put the company at risk. On the other hand, there are studies that conclude with opposite results. They find that the presence of women on the board of directors may be positively related to tax avoidance (Garcia-Blandon et al., 2022, in Norway). This is because gender inequality in Indonesia limits the effectiveness of female auditors (Zudana et al., 2021; Kalbuana et al., 2023). Finally, there are also studies that conclude with a neutral stance that gender diversity does not significantly affect tax avoidance (Budi, 2019, in Indonesia).

A previous literature review on the relationship between board foreign ownership and tax avoidance presents mixed findings. Fuest and Hemmelgarn (2005) suggest that there is a positive association between foreign ownership and tax avoidance, particularly under the assumption that corporate taxation serves primarily as a safeguard for the personal income tax system rather than as a levy on economic rent. Consistently, Egger et al. (2010) find that foreign-owned firms among 31 European countries better positioned to exploit international differences in tax rates, as well as preferential accounting standards and tax treatments available in foreign jurisdictions. These conditions enhance the attractiveness of profit and debt shifting strategies, thereby providing firms with substantial foreign involvement greater tax advantages and expanded opportunities for tax planning. Annuar et al. (2014) also argue that foreign ownership in Malaysian firms can lead to increased tax avoidance due to the advanced tax planning capabilities of foreign investors. It is supported by Salihu et al. (2015) and Suranta et al. (2020), who demonstrate similar trends in Malaysian and Indonesian firms, respectively. Additionally, Omesi and Appah (2021) highlight that in Nigerian firms, higher foreign ownership correlates with lower effective tax rates, while in South Korean firms, Choi and Park (2022) observe reduced volatility in corporate tax rates due to foreign investors’ oversight. On the contrary, Yoo and Koh (2014) argue that foreign-owned firms in the South Korean firms exhibit lower tax avoidance levels due to regulatory scrutiny. I. Hasan et al. (2022) argue that foreign ownership is negatively associated with tax avoidance across 30 countries. A. Hasan et al. (2024) also find that foreign ownership can limit tax avoidance by enhancing governance in Pakistani firms.

Based on the above analysis, the literature on the relationship between various board characteristics—such as size, independence, gender diversity, and foreign ownership—and tax avoidance presents mixed findings. In this regard, we integrate agency theory and signaling theory in order to explain these conflicted relationships. Agency theory suggests that larger boards can either enhance oversight, thereby reducing tax avoidance (Halioui et al., 2016; Ali et al., 2024), or face coordination challenges that lead to ineffective monitoring and increased tax avoidance (Ogbodo & Omonigho, 2021; Kalbuana et al., 2023). Independent directors may similarly either curb or endorse tax avoidance, depending on their effectiveness in aligning management with shareholder interests (Chytis et al., 2020; Ali et al., 2024). Gender-diverse boards, under agency theory, are posited to enhance ethical standards and oversight, reducing aggressive tax practices (Utaminingsih et al., 2022; Ali et al., 2024), while signaling theory suggests they may communicate a firm’s commitment to ethical governance, thereby boosting its reputation and investor confidence (Garcia-Blandon et al., 2022). In the context of foreign ownership, agency theory highlights the potential for either enhanced tax planning efficiency (Salihu et al., 2015; Suranta et al., 2020) or stricter compliance due to international standards and reputational concerns (Hanlon et al., 2005; Yee et al., 2018). Simultaneously, signaling theory indicates that foreign ownership can project an image of financial stability and competence or a commitment to ethical behavior, influencing market perceptions and firm valuation (Suranta et al., 2020).

Therefore, we can hypothesize that there is a correlation between board characteristics like size, independence, gender diversity, and foreign ownership and tax avoidance, integrating both theories and drawing from the contradictory findings in the literature, as previously presented. This leads to the following predictions:

H1. 

Board size is associated with tax avoidance.

H2. 

Board independence is associated with tax avoidance.

H3. 

Board gender diversity is associated with tax avoidance.

H4. 

Board foreign ownership is associated with tax avoidance.

Tax avoidance trends differ between regulated and unregulated industries due to factors such as transparency, corporate social responsibility (CSR), regulation effectiveness, and oversight (Oats & Tuck, 2019; Kerr, 2019; Wu & Zhang, 2022). In regulated industries, enhanced transparency initiatives, such as country-by-country reporting and public tax strategy disclosures, increase the visibility of tax practices, thereby deterring aggressive tax avoidance (Gribnau & Jallai, 2019; Kurniasih et al., 2023). This regulatory scrutiny, coupled with a heightened focus on CSR, pressures firms to adopt responsible tax practices, balancing societal expectations against financial benefits (Oats & Tuck, 2019). On the contrary, unregulated industries, lacking such oversight and CSR accountability, are more prone to prioritizing short-term financial gains (Unerman & O’Dwyer, 2007; Becher & Frye, 2011; Delgado-Márquez et al., 2017), leading to higher levels of tax avoidance.

The effect of board characteristics on tax avoidance, as discussed above, may be particularly salient in a regulated industry. Corporate governance in regulated industries differs significantly from that in unregulated industries (Devriese et al., 2004; Handley-Schachler et al., 2007; Gopinath, 2008; Ungureanu, 2008; Mülbert, 2013). In regulated industries, there is stringent regulatory oversight, requiring boards to comply with numerous regulations, maintain clear accountability, and ensure board members’ qualifications and independence from management (Mullineux, 2007). These boards are mandated to establish robust risk management frameworks, including separate risk committees, direct reporting lines for Chief Risk Officers, and regular risk assessments (Van Greuning & Bratanovic, 2020). The emphasis on board independence, often through the inclusion of independent directors, aims to prevent conflicts of interest (Handley-Schachler et al., 2007). Consequently, boards in regulated industries are subject to stringent requirements regarding size, independence, diversity, and ownership, ensuring effective governance and compliance. In contrast, unregulated industries enjoy greater flexibility, focusing more on strategic and operational concerns without the strict governance mandates present in regulated sectors (Mülbert, 2013). Given that board characteristics influence tax avoidance strategies across different regulatory contexts, we argue that their impact is significant in both regulated and unregulated industries. Thus, we propose the following hypotheses: We thus extend our hypotheses (H1 to H4) as follows:

H1a. 

The association between board size and tax avoidance is stronger in a regulated industry.

H2a. 

The association between board independence and tax avoidance is stronger in a regulated industry.

H3a. 

The association between board gender diversity and tax avoidance is stronger in a regulated industry.

H4a. 

The association between board foreign ownership and tax avoidance is stronger in a regulated industry.

3. Methodology

3.1. Sample Selection

For the period between 2000 and 2023, this analysis uses a sample of 39 nations all around. These nations include Bulgaria, Canada, Chile, China, Colombia, Denmark, Egypt, Finland, France, Germany, Greece, Hong Kong, India, Indonesia, Israel, Italy, Japan, Malaysia, Mexico, the Netherlands, Nigeria, Norway, Peru, the Philippines, Poland, Portugal, Romania, Saudi Arabia, Singapore, South Africa, South Korea, Spain, Sweden, Switzerland, Taiwan, Thailand, Turkey, the United Kingdom, and the United States of America. Our initial sample comes from the Thomson Reuters Asset4 database. Extracted financial data from the Datastream and Worldscope databases are matched with the sample data derived from the Thomson Reuters Asset4 database. The primary sample is 143,568 firm-year observations. We exclude from the sample firms with missing data for any of the utilized dependent, independent, and control variables (45,408 firm-year observations). We further eliminate companies whose ETR is less than zero or higher than one (14,007 firm-year observations), following Alsaadi (2020). Furthermore, 84,153 firm-year observations spanning the entire world for the period 2000–2023 constitute the final sample for the research.

Table 1. Sample selection and distribution.

Panel A: Sample Selection Process
The original sample of listed firms from 39 countries during 2000–2023					143,568
Excluding firms with missing financial information					45,408
Excluding firms whose ETR is less than zero or higher than one					14,007
Final sample					84,153
Panel B: Sample Distribution per Year
Year	No. of Firm-Year Observations	Percentage	Year	No. of Firm-Year Observations	Percentage
2000	825	0.98	2012	3779	4.49
2001	926	1.10	2013	4255	5.06
2002	1079	1.28	2014	4585	5.45
2003	1195	1.42	2015	4900	5.82
2004	1355	1.61	2016	5395	6.41
2005	1468	1.74	2017	5419	6.44
2006	1870	2.22	2018	5813	6.91
2007	2368	2.81	2019	6469	7.69
2008	2318	2.75	2020	6516	7.74
2009	2786	3.31	2021	6695	7.96
2010	3073	3.65	2022	6253	7.43
2011	3292	3.91	2023	1519	1.81
Panel C: Sample Distribution per Country
Country	No. of Firm-Year Observations	Percentage	Country	No. of Firm-Year Observations	Percentage
Bulgaria	229	0.27	Nigeria	235	0.28
Canada	891	1.06	Norway	517	0.61
Chile	757	0.90	Peru	280	0.33
China	23,715	28.18	Philippines	587	0.70
Colombia	183	0.22	Poland	1679	2.00
Denmark	435	0.52	Portugal	203	0.24
Egypt	259	0.31	Romania	185	0.22
Finland	617	0.73	Saudi Arabia	193	0.23
France	1865	2.22	Singapore	1615	1.92
Germany	2363	2.81	South Africa	184	0.22
Greece	454	0.54	South Korea	4196	4.99
Hong Kong	5954	7.08	Spain	706	0.84
India	229	0.27	Sweden	1349	1.60
Indonesia	1491	1.77	Switzerland	932	1.11
Israel	642	0.76	Taiwan	4052	4.82
Italy	1433	1.70	Thailand	1923	2.29
Japan	3775	4.49	Turkey	1126	1.34
Malaysia	2383	2.83	United Kingdom	1689	2.01
Mexico	353	0.42	United States of America	14,094	16.75
Netherlands	380	0.45
Panel D: Sample Distribution per Sector
Sector	No. of Firm-Year Observations	Percentage	Sector	No. of Firm-Year Observations	Percentage
Communication Services	3479	4.13	Industrials	20,224	24.03
Consumer Discretionary	13,328	15.84	Information Technology	11,719	13.93
Consumer Staples	7437	8.84	Materials	12,334	14.66
Energy	2396	2.85	Real Estate	2469	2.93
Financials	747	0.89	Utilities	2.590	3.08
Health Care	7430	8.83

This table reports the sample selection procedure (Panel A) and the distribution of the sample across the year (Panel B), country (Panel C), and sector groups (Panel D).

Panel A shows the sample selection process.

Table 1 shows all firm-year observations distribute across countries, sectors, and years. Table 1 Panel B indicates that the highest percentage of observations is from China (28.18%), followed by the United States of America (16.75%) and Hong Kong (7.08%). Further, Table 1 Panel C shows that the highest percentage of observations is from the Industrials sector (24.03%), followed by the Consumer Discretionary sector (15.84%) and Materials sector (14.66%). Finally, Table 1 Panel D illustrates that the highest yearly distributions of our sample are 6695 firm-year observations in 2021, 6516 firm-year observations in 2020, 6469 firm-year observations in 2019, and 6253 firm-year observations in 2022.

3.2. Measuring Tax Avoidance

Following Kovermann and Velte (2019), C. Wang et al. (2024), Hu et al. (2023), Overesch and Wolff (2021), and F. Wang et al. (2020), we use effective tax rate (ETR) as a measure of tax avoidance. ETR is calculated as the ratio of total income tax expenses to pre-tax book income, providing a clear picture of a firm’s tax burden relative to its income. A lower ETR means higher tax avoidance.

3.3. Measuring Board Characteristics

In our study, we use four measures of board characteristics—board size (BS), board independence (INDIR), board gender diversity (GENDER), and board foreign ownership (BOFODI). BS is calculated as the natural log of the total number of directors serving on a corporate board (Tanna et al., 2011; Akbar et al., 2017). INDIR is calculated as the ratio of independent directors to the total number of directors (Chen & Zhang, 2014; Akbar et al., 2017; Lanis et al., 2017). GENDER is calculated as the ratio of female board members to the total number of board members (Mateos de Cabo et al., 2012; Lanis et al., 2017). BOFODI is calculated as the proportion of shares owned by foreign investors (Miletkov et al., 2014; Kaczmarek et al., 2014; Rashid, 2020).

3.4. Measuring Industry Regulation

Consistent with Becher and Frye (2011), we measure industry regulation as a dummy variable coded one if the firm is in a regulated industry and 0 otherwise. The regulated industries include industries with an SIC code of 4900–4939 (electric and gas), 1300 (oil and gas extraction), 4000–4700 (transportation), 4800 (telecommunications), 4950–4959 (sanitary services), and all 6000s (financial firms).

3.5. Measuring Control Variables

Our empirical models also include several control variables. Therefore, we control firm-level control variables that may influence tax avoidance (Atwood et al., 2012; Kerr, 2019; Joshi et al., 2020; Argilés-Bosch et al., 2020; Overesch & Wolff, 2021; Na & Yan, 2022; Dyreng et al., 2022; Hu et al., 2023; Xie et al., 2024). Employees (EMPL) are defined as the number of full-time employees. Firm size (FS) is measured by the natural logarithm of total assets. Inventory intensity (INV) is measured as the ratio of inventories divided by total assets. Leverage (LEV) is measured as the ratio of total debt divided by total assets. Capital intensity (PPE) is measured as the ratio of property, plant, and equipment divided by total assets. R&D expenditures (R&D) are measured as the ratio of R&D expenditures divided by the total assets. Profitability (ROA) is measured as the ratio of pre-tax income less extraordinary income divided by lagged total assets. Furthermore, we control for country-level variables that may influence tax avoidance (I. Hasan et al., 2017; Na & Yan, 2022; Xie et al., 2024). Age (AGE) is measured as the natural logarithm of the median age of the residents in a country. GDP growth (GDPG) captures the growth of GDP per capita. The GINI coefficient (GINI) measures income inequality. Inflation (INFL) is defined as the percentage change in consumer prices. The statutory corporate income tax rate (STR) captures the statutory corporate income tax rate of a firm’s headquarters country. Definitions of all the variables used in this study are summarized in Appendix A.

3.6. Regression Model

To test the relationship between tax avoidance and board characteristics (H1 to H4) and the moderating role of industry regulation in this relationship (H1a to H4a), we estimate the following fixed-effect (FE) regression models with standard errors clustered at the firm level:

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{BS}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(1)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{INDIR}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+\epsilon$$

(2)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{GENDER}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(3)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{BOFODI}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(4)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{BS}_{it}+{\mathsf{\beta }}_2{REGULATED}_i+{\mathsf{\beta }}_3{REGULATED}_i\times {BS}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(5)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{INDIR}_{it}+{\mathsf{\beta }}_2{REGULATED}_i+{\mathsf{\beta }}_3{REGULATED}_i\times {INDIR}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(6)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{GENDER}_{it}+{\mathsf{\beta }}_2{REGULATED}_i+{\mathsf{\beta }}_3{REGULATED}_i\times {GENDER}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(7)

$${ETR}_{it}={\alpha }_i+{\mathsf{\beta }}_1{BOFODI}_{it}+{\mathsf{\beta }}_2{REGULATED}_i+{\mathsf{\beta }}_3{REGULATED}_i\times {BOFODI}_{it}+\gamma {\mathcal{X}}_{it}+c_t+{\delta }_j+{\epsilon }_{it}$$

(8)

where ${ETR}_{it}$ is the effective tax rate of firm i as a measure of tax avoidance (see Section 3.2) in year t. ${BS}_{it}$ is the board size, ${INDIR}_{it}$ is the board independence, ${GENDER}_{it}$ is the board gender diversity, and ${BOFODI}_{it}$ is the board foreign ownership (see Section 3.3). ${REGULATED}_j$ is the industry regulation of intusrty j (see Section 3.4). We control for several variables included in the vector ${\mathcal{X}}_{it}$ where ${\mathcal{X}}_{it}$:{ employees (EMPL), firm size (FS), inventory intensity (INV), leverage (LEV), capital intensity (PPE), R&D expenditures (R&D), profitability (ROA), age (AGE), GDP growth (GDPG), GINI coefficient (GINI), inflation (INFL), and statutory corporate income tax rate (STR)} (see Section 3.5). ${\epsilon }_{it}$ is the random error term. ${\alpha }_i$ represents firm-specific effects. Finally, we include dummy variables $c_t$ and ${\delta }_j$ to control for year and industry fixed effects, respectively. Furthermore, in order to account for potential non-linearities, we expanded our empirical model to include quadratic terms for the key explanatory variables, which allowed us to test if non-linear functional forms may offer a better fit for explaining effective tax rates. In our model comparison using the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC), the augmented models did not provide better model fit relative to the baseline linear specification. We also saw qualitatively similar results, with most of the additional terms being statistically insignificant. This suggests that the linear specification is a parsimonious and applicable model for our data.

4. Results and Discussion

4.1. Univariate Analysis

Table 2. Descriptive analysis.

	Whole Sample				Regulated Firms	Unregulated Firms
	Mean	Q25	Q75	Std. Dev	Mean	Mean
Dependent variable
ETR	0.232	0.143	0.291	0.132	0.293	0.226
Independent variable
BOFODI	5.917	3.63	7.38	3.147	5.935	5.915
BS	9.101	7	10	2.441	9.079	9.103
GENDER	15.912	6.15	23.63	12.437	16.076	15.896
INDIR	45.529	34.26	54.87	17.949	45.239	45.557
Moderate variable
REGULATED	0.090	0	1	0.286	-	-
Firm-level Controls
EMPL	6941.002	479	114,489	17,090,23	6791.251	6956.585
FS	9.826	9.041	10.581	1.094	9.810	9.827
INV	0.123	0.044	0.174	0.101	0.124	0.123
LEV	0.138	0.098	0.187	0.067	0.139	0.138
PPE	0.029	0.013	0.041	0.019	0.028	0.029
R&D	0.008	3.47 × 10−4	0.198	0.026	0.008	0.008
ROA	0.075	0.033	0.100	0.059	0.076	0.075
Country-level Controls
AGE	34.209	26.750	41.816	8.668	34.220	34.208
GDPG	4.271	3.995	4.590	0.374	4.234	4.275
GINI	0.541	0.506	0.583	0.068	0.543	0.541
INFL	2.427	1.274	2.899	2.091	2.832	2.387
STR	0.254	0.24	0.275	0.054	0.252	0.254
Observations	84,153				7621	76,532

This table reports descriptive statistics of the variables used in the regression models. Variable definitions are provided in Appendix A.

presents descriptive statistics of key variables for the final sample. The ETR exhibits a consistent mean of 0.232 across the entire sample, with values of 0.293 for regulated firms and 0.226 for unregulated firms, indicating relative uniformity in ETR. The low standard deviation (0.132) suggests limited variability, implying that most firms have an ETR close to the mean. Table 2 indicates that the mean BS across the whole sample is 9.101, with a standard deviation of 2.441, suggesting moderate variability. BS for regulated firms (mean 9.079) is nearly identical to the overall mean, while unregulated firms also show very similar statistics, indicating that BS does not differ significantly between regulated and unregulated firms. Furthermore, Table 2 shows that the mean of INDIR for the whole sample is 45.529, with a standard deviation of 17.949, indicating substantial variability. Regulated firms have a mean ratio of 45.239, closely aligning with the overall sample, whereas unregulated firms show similar statistics. This suggests that the presence of INDIR does not differ markedly between regulated and unregulated firms. Table 2 also illustrates that the mean of GENDER for the whole sample is 15.912 with a high standard deviation of 12.437, indicating significant variability. Regulated firms have a slightly higher mean (16.076) compared to the whole sample. Unregulated GENDER statistics show wide-ranging values, highlighting significant variation across different firms in terms of GENDER representation. Regarding the BOFODI variable, Table 2 shows that the whole sample shows an average of 5.917, with a relatively low standard deviation of 3.147, indicating moderate variability. Regulated firms have a slightly higher mean (5.935) compared to the whole sample, while unregulated firms show very close mean values, suggesting similar levels of BOFODI across different firm types.

Table 3. Correlation matrix.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	VIF
(1) ETR	1.000
(2) BS	0.058 ***	1.000										1.04
(3) INDIR	0.095 ***	−0.068 ***	1.000									1.06
(4) GENDER	0.015 ***	0.036 ***	0.169 ***	1.000								1.19
(5) BOFODI	0.051 ***	0.098 ***	0.007 **	0.142 ***	1.000							1.06
(6) R&D	0.017 ***	0.041 ***	0.030 ***	0.093 ***	0.099 ***	1.000						1.08
(7) LEV	0.137 ***	0.120 ***	0.055 ***	0.065 ***	0.061 ***	−0.027 ***	1.000					1.30
(8) PPE	0.036 ***	0.053 ***	−0.007 **	−0.090 ***	−0.003	−0.006 **	0.225 ***	1.000				1.14
(9) FS	0.096 ***	0.047 ***	0.041 ***	−0.323 ***	−0.173 ***	−0.263 ***	0.138 ***	0.135 ***	1.000			1.29
(10) ROA	−0.262 ***	−0.095 ***	0.007 **	0.013 ***	−0.005 *	0.042 ***	−0.337 ***	−0.113 ***	−0.154 ***	1.000		1.15
(11) EMPL	0.035 ***	0.017 ***	0.030 ***	0.012 ***	0.006 *	0.014 ***	0.016 ***	0.001	−0.000	−0.011 ***	1.000	1.00
(12) INV	0.033 ***	−0.046 ***	−0.049 ***	0.009 ***	0.027 ***	0.026 ***	0.070 ***	−0.191 ***	−0.092 ***	0.004	−0.012 ***	1.08
(13) GDPG	−0.042 ***	−0.006 *	0.037 ***	0.004	0.039 ***	−0.004	−0.187 ***	0.062 ***	−0.009	0.060 ***	0.001	1.34
(14) STR	−0.126 ***	0.001	−0.007 **	−0.003	0.007 **	−0.005	−0.043 ***	−0.002	−0.004	−0.007**	0.001	1.02
(15) GINI	0.023 ***	−0.001	−0.040 ***	−0.009 ***	−0.040 ***	0.010 ***	0.119 ***	−0.070 ***	−0.005	−0.061 ***	−0.004	1.23
(16) INFL	−0.003	0.002	0.011 ***	0.002	0.006 **	0.000	−0.035 ***	−0.004	−0.003	0.007**	−0.005	1.08
(17) AGE	−0.005	0.002	0.015 ***	0.008 **	0.014 ***	0.001	−0.052 ***	0.009 ***	−0.006*	0.014 ***	−0.002	1.02
	(12)	(13)	(14)	(15)	(16)	(17)
(12) INV	1.000
(13) GDPG	−0.134 ***	1.000
(14) STR	−0.022 ***	0.023 ***	1.000
(15) GINI	0.093 ***	−0.418 ***	−0.073 ***	1.000
(16) INFL	−0.032 ***	−0.234 ***	−0.091 ***	0.074 ***	1.000
(17) AGE	−0.032 ***	0.103 ***	0.026 ***	−0.095 ***	−0.068 ***	1.000

This table reports Pearson correlations between the variables used in the main analysis. * significant at the 10% level, ** significant at the 5% level, *** significant at the 1% level. Variable definitions are provided in Appendix A.

reports Pearson correlation coefficients among key variables for the whole sample. There is a positive correlation between ETR and BS, with a coefficient of 0.058, suggesting that larger boards are associated with slightly higher ETRs, indicative of lower tax avoidance. Similarly, INDIR has a positive correlation with ETR, with a coefficient of 0.095, suggesting that a greater presence of independent directors is associated with less aggressive tax avoidance. GENDER shows a weak positive correlation with ETR at 0.015, indicating a negligible impact on tax avoidance practices. Lastly, BOFODI has a positive correlation with ETR at 0.051, implying that firms with greater foreign ownership also exhibit higher ETRs. Overall, these findings suggest that while the relationships are generally weak, there is a tendency for firms with certain board characteristics, such as larger size and higher independence, to engage in less tax avoidance, as evidenced by higher ETRs. Finally, Table 3 shows that the variance inflation factors (VIF) for the independent and control variables are below 4.0, indicating that the inclusion of all these variables in the same regression models does not raise concerns about multicollinearity.

4.2. Multivariate Regression Analysis

Table 4. The impact of board characteristics on tax avoidance.

	(A) (Equation (1))	(B) (Equation (5))	(C) (Equation (2))	(D) (Equation (6))	(E) (Equation (3))	(F) (Equation (7))	(G) (Equation (4))	(H) (Equation (8))
BS	0.001 ***	0.001 ***
	(0.0001)	(0.0001)
INDIR			0.0006 ***	0.0006 ***
			(0.00002)	(0.00002)
GENDER					0.0003 ***	0.0003 ***
					(0.00003)	(0.00003)
BOFODI							0.002 ***	0.002 ***
							(0.0001)	(0.0001)
REGULATED		0.059 ***		0.062 ***		0.062 ***		0.055 ***
		(0.008)		(0.006)		(0.003)		(0.004)
REGULATED × BS		0.0009
		(0.0008)
REGULATED × INDIR						0.0003 *
						(0.0001)
REGULATED × GENDER								0.002 ***
								(0.0007)
REGULATED × BOFODI				0.0001
				(0.0001)
EMPL	2.44 × 10−7 ***	2.43 × 10−7 ***	2.26 × 10−7 ***	2.25 × 10−7 ***	2.44 × 10−7 ***	2.43 × 10−7 ***	2.45 × 10−7 ***	2.44 × 10−7 ***
	(2.76 × 10−8)	(2.76 × 10−8)	(2.75 × 10−8)	(2.75 × 10−8)	(2.76 × 10−8)	(2.75 × 10−8)	(2.76 × 10−8)	(2.76 × 10−8)
FS	0.008 ***	0.008 ***	0.007 ***	0.007 ***	0.009 ***	0.009 ***	0.009 ***	0.009 ***
	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0004)
INV	0.046 ***	0.046 ***	0.050 ***	0.051 ***	0.046 ***	0.047 ***	0.043 ***	0.044 ***
	(0.004)	(0.004)	(0.004)	(0.004)	(0.004)	(0.004)	(0.004)	(0.004)
LEV	0.074 ***	0.074 ***	0.064 ***	0.063 ***	0.070 ***	0.069 ***	0.069 ***	0.068 ***
	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)	(0.007)
PPE	−0.003	−0.0005	0.026	0.029	0.020	0.023	0.0008	0.004
	(0.025)	(0.025)	(0.025)	(0.025)	(0.025)	(0.025)	(0.025)	(0.025)
R&D	0.217 ***	0.219 ***	0.204 ***	0.205 ***	0.223 ***	0.224 ***	0.209 ***	0.210 ***
	(0.019)	(0.019)	(0.019)	(0.019)	(0.018)	(0.018)	(0.019)	(0.019)
ROA	−0.529 ***	−0.530 ***	−0.540 ***	−0.540 ***	−0.533 ***	−0.533 ***	−0.532 ***	−0.533 ***
	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)	(0.008)
AGE	0.00004	0.00002	0.00002	0.00001	0.00003	0.00002	0.00003	0.00001
	(0.00005)	(0.00005)	(0.00005)	(0.00005)	(0.00005)	(0.00005)	(0.00005)	(0.00005)
GDPG	−0.007 ***	−0.006 ***	−0.009 ***	−0.008 ***	−0.007 ***	−0.006 ***	−0.008 ***	−0.007 ***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)
GINI	−0.024 ***	−0.021 ***	−0.019 **	−0.017 **	−0.022 ***	−0.020 **	−0.021 ***	−0.019 **
	(0.008)	(0.007)	(0.008)	(0.007)	(0.008)	(0.007)	(0.008)	(0.007)
INFL	−0.0004 *	−0.0006 **	−0.0005 **	−0.0007 ***	−0.0004 *	−0.0006 ***	−0.0005 **	−0.0007 ***
	(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)	(0.0002)
STR	−0.022 **	−0.017 *	−0.019 **	−0.014	−0.021 **	−0.016 *	−0.022 **	−0.017 **
	(0.009)	(0.008)	(0.009)	(0.008)	(0.009)	(0.008)	(0.009)	(0.008)
Constant	0.209 ***	0.197 ***	0.200 ***	0.187 ***	0.201 ***	0.188 ***	0.201 ***	0.188 ***
	(0.011)	(0.010)	(0.011)	(0.010)	(0.011)	(0.010)	(0.011)	(0.010)
Observations	84,153	84,153	84,153	84,153	84,153	84,153	84,153	84,153
Year effects	YES	YES	YES	YES	YES	YES	YES	YES
Industry effects	YES	YES	YES	YES	YES	YES	YES	YES
F-Statistics	531.72 ***	556.61 ***	591.89 ***	561.24 ***	536.31 ***	510.70 ***	556.61 ***	510.70 ***
Adjusted R square	0.090	0.082	0.098	0.106	0.080	0.099	0.082	0.099

This table reports the FE regression results for the effect of board characteristics and industry regulation on tax avoidance and the interactions between board characteristics and industry regulation on tax avoidance. Standard errors reported in parentheses. * significant at the 10% level, ** significant at the 5% level, *** significant at the 1% level. Variable definitions are provided in Appendix A.

Columns A to H show outcomes from fixed effects regression analysis. Additionally, Columns A, C, E, and G display the baseline model excluding interaction variables, whereas Columns B, D, F, and H illustrate the complete model that includes interactions.

Regarding the effect of board size on tax avoidance, Table 4 Columns A and B present a positive and significant relationship between BS and ETR. It means that larger boards are associated with higher ETR, indicating lower tax avoidance. Hence, these results support the hypothesis that has been proposed, so H1 is accepted. The findings of this study confirm the tests conducted by Halioui et al. (2016) and Ali et al. (2024), suggesting that the presence of a larger board may enhance oversight capabilities and bring a broader range of expertise and perspectives, suggesting an increase in ETR corresponds with a decrease in tax avoidance. This finding can be explained through agency theory, which posits that larger boards enhance oversight and provide diverse expertise, thereby aligning managerial actions with shareholder interests and reducing agency costs. Additionally, signaling theory suggests that larger boards signal strong governance and ethical standards to the market, boosting investor confidence and indicating a commitment to conservative tax practices.

Regarding the effect of board independence on tax avoidance, Table 4 Columns C and D indicate that there are positive and significant coefficients related to INDIR, indicating that higher levels of board independence correlate with higher ETR, suggesting lower levels of tax avoidance. Thus, the results of this test support the hypothesis that has been proposed, so H2 is accepted. This study is consistent with Salhi et al. (2020) and Ali et al. (2024), who proved that independent directors are crucial for providing unbiased oversight, ensuring decisions align with shareholder interests, which in turn reduces the likelihood of aggressive tax strategies. The positive relationship observed can be explained through the integration of agency theory and signaling theory. According to agency theory, independent directors serve to provide impartial oversight, thereby mitigating agency problems and ensuring that management decisions align with shareholder interests, which consequently reduces tax avoidance. Signaling theory complements this perspective by positing that the presence of independent directors signals a commitment to ethical practices and strong governance to the market, enhancing investor trust and alleviating concerns regarding potential tax avoidance.

Regarding the effect of board gender diversity on tax avoidance, Table 4 Columns E and F reveal a significant positive correlation between GENDER and ETR, indicating that higher board gender is associated with higher ETR and thus lower tax avoidance. Thus, the results of this test support the hypothesis that has been proposed, so H3 is accepted. These findings are consistent with Boussaidi and Hamed-Sidhom (2021), Utaminingsih et al. (2022) and Ali et al. (2024), who suggest that female directors may contribute to higher ethical standards and greater corporate transparency. Integrating agency theory, this suggests that female directors often prioritize ethical standards and transparency, thereby reducing aggressive tax strategies and aligning corporate actions with shareholder interests. Signaling theory complements this view, proposing that gender-diverse boards enhance a firm’s ethical reputation, increase investor confidence, and lower perceived tax avoidance risk.

Regarding the effect of board foreign ownership on tax avoidance, the positive coefficients associated with BOFODI indicate that firms with significant foreign ownership experience higher ETR, suggesting lower levels of tax avoidance (see Table 4 Columns G and H). Thus, the results of this test support the hypothesis that has been proposed, so H4 is accepted. The results of this study confirm the findings from Yoo and Koh (2014) and I. Hasan et al. (2022), implying that this relationship can be attributed to foreign investors bringing international best practices and stringent governance standards. Further, these results can be explained by the application of agency theory and signaling theory. According to agency theory, foreign investors introduce rigorous governance standards and international best practices, thereby discouraging aggressive tax strategies and aligning actions with shareholder wealth maximization. Additionally, signaling theory suggests that foreign ownership serves as a signal of strong governance, indicating that a firm upholds high ethical standards and is less likely to engage in risky tax practices.

Regarding the effect of industry regulation on tax avoidance, Table 4 also shows that there are positive and significant coefficients associated with REGULATED, suggesting that firms in regulated industries tend to have higher ETR, implying lower levels of tax avoidance. These results align with the notion that regulatory oversight and stringent transparency requirements in regulated industries act as deterrents to aggressive tax strategies. According to Oats and Tuck (2019), the presence of industry-specific regulations ensures that corporate tax affairs are more visible to both tax authorities and the public, thus curbing the likelihood of tax avoidance.

More importantly, Table 4 Columns B, D, F, and H reveal the moderating role of industry regulation on the effects of board characteristics on tax avoidance. Specifically,

Table 5. The impact of board characteristics on tax avoidance—alternative measure of corporate tax avoidance.

	(A) (Equation (1))	(B) (Equation (2))	(C) (Equation (3))	(D) (Equation (4))
BS	0.001 ***
	(0.0001)
INDIR		0.0006 ***
		(0.00002)
GENDER			0.0004 ***
			(0.00003)
BOFODI				0.002 ***
				(0.0001)
Controls	INCLUDED	INCLUDED	INCLUDED	INCLUDED
Constant	0.167 ***	0.164 ***	0.159 ***	0.161 ***
	(0.009)	(0.009)	(0.009)	(0.009)
Observations	84,153	84,153	84,153	84,153
Year effects	YES	YES	YES	YES
Industry effects	YES	YES	YES	YES
F-Statistics	336.05 ***	376.57 ***	340.90 ***	363.58 ***
Adjusted R square	0.053	0.061	0.053	0.056

This table reports the FE regression results for the effect of board characteristics on different measures of tax avoidance. Standard errors reported in parentheses *** significant at the 1% level. Variable definitions are provided in Appendix A.

shows that the interaction terms REGULATED × GENDER and REGULATED × BOFODI are positively and statistically significant with ETR. The results in this study prove that the effects of board gender diversity and board foreign ownership on ETR are more pronounced in regulated industries compared to unregulated industries, meaning lower levels of tax avoidance. Thus, the results of this test support the hypotheses that have been proposed, so H3a and H4a are accepted. The findings suggest that regulated industries, characterized by stringent governance and ethical standards, benefit from the presence of female directors, who contribute to more transparent and conservative tax strategies. Additionally, the results indicate that firms with higher proportions of foreign investors on their boards engage in less aggressive tax avoidance practices. This relationship is particularly stronger within regulated industries, suggesting that regulatory oversight amplifies foreign investors’ influence on corporate tax compliance behavior. On the contrary, Table 4 Columns B and D show that there is no association between the interaction terms REGULATED × BS, REGULATED × GENDER, and ETR. Thus, the results of these tests do not support the hypothesis that has been proposed, so H1a and H2a are rejected.

The controls in Table 4 exhibit signs and significance that are in line with our expectations and with previous studies (e.g., Atwood et al., 2012; I. Hasan et al., 2017; Kerr, 2019; Joshi et al., 2020; Argilés-Bosch et al., 2020; Overesch & Wolff, 2021; Na & Yan, 2022; Dyreng et al., 2022; Hu et al., 2023; Xie et al., 2024).

5. Robustness Checks

5.1. Additional Measure of Tax Avoidance

ETRs are widely used to measure tax avoidance, but they face significant limitations, such as their inability to capture conforming tax avoidance and susceptibility to volatility and truncation bias, especially in annual measures (Aronmwan & Okaiwele, 2020; Dyreng et al., 2008). ETRs also suffer from numerator–denominator mismatches and are influenced by accounting choices, making them unreliable indicators of aggressive or sustained tax avoidance strategies (Gebhart, 2017; Hanlon & Heitzman, 2010). Furthermore, studies show that annual ETRs are poor predictors of long-term tax behavior, underscoring the need for alternative or complementary measures in tax avoidance research (Dyreng et al., 2008; Gebhart, 2017). Therefore, we use the cash effective tax rate (CETR) as an alternative measure of tax avoidance. CETR is calculated as the cash taxes paid divided by pre-tax book income (Thomsen & Watrin, 2018; F. Wang et al., 2020). Similar to the measure of ETR, the lower the CETR, the higher the degree of tax avoidance. As reported in Table 5 Columns A to D, CETR is positively and significantly associated with all board characteristics. These regression results provide supporting evidence for H1 to H4 that is consistent with the main results.

5.2. Endogeneity

According to Coles et al. (2012), Roberts and Whited (2013), and Barros et al. (2020), endogeneity arises in econometric models in finance when an explanatory variable is correlated with the error term, potentially leading to biased and inconsistent parameter estimates. The System GMM, developed by Arellano and Bover (1995) and Blundell and Bond (1998), is used in order to tackle potential endogeneity issues arising from unobserved individual effects, simultaneity, and measurement errors. Therefore, we use the first lag of the dependent variable as a regressor. The corresponding results, included in

Table 6. The impact of board characteristics on tax avoidance using System GMM.

	(A) (Equation (1))	(B) (Equation (2))	(C) (Equation (3))	(D) (Equation (4))
D_ETR	0.008	0.008	0.009	0.009
	(0.006)	(0.006)	(0.006)	(0.006)
BS	0.001 ***
	(0.0003)
INDIR		0.0007 ***
		(0.00004)
GENDER			0.0004 ***
			(0.00006)
BOFODI				0.002 ***
				(0.0002)
Controls	INCLUDED	INCLUDED	INCLUDED	INCLUDED
Constant	0.195 ***	0.185 ***	0.184 ***	0.186 ***
	(0.014)	(0.013)	(0.014)	(0.014)
Observations	84,153	84,153	84,153	84,153
AR1 test p-value	0.000	0.000	0.000	0.000
AR2 test p-value	0.432	0.546	0.561	0.571
Sargan–Hansen p-value	0.679	0.814	0.115	0.243

This table reports the System GMM results for the effect of board characteristics on tax avoidance. Standard errors reported in parentheses. *** significant at the 1% level. Variable definitions are provided in Appendix A.

Columns A to D, are similar to those found originally and presented in Table 4. Furthermore, model specification tests, including the Arellano and Bond tests of first-order (AR1) and second-order (AR2) serial correlation, as well as the Sargan–Hansen test of over-identifying restrictions, validated the results of the dynamic panel data analysis.

5.3. Regulated Versus Unregulated Industries

We conduct additional analyses to investigate whether the role of board characteristics varies across different regulatory environments.

Table 7. The impact of board characteristics on tax avoidance in regulated versus unregulated firms.

	Regulated				Unregulated
	(A) (Equation (1))	(B) (Equation (2))	(C) (Equation (3))	(D) (Equation (4))	(A) (Equation (1))	(B) (Equation (2))	(C) (Equation (3))	(D) (Equation (4))
BS	0.001 **				0.001 ***
	(0.0007)				(0.0001)
INDIR		0.0009 ***				0.0006 ***
		(0.0001)				(0.00002)
GENDER			0.0006 ***				0.0003 ***
			(0.0001)				(0.00003)
BOFODI				0.002 ***				0.002 ***
				(0.0005)				(0.0001)
Controls	YES	YES	YES	YES	YES	YES	YES	YES
Constant	0.340 ***	0.306 ***	0.321 ***	0.327 ***	0.189 ***	0.182 ***	0.181 ***	0.181 ***
	(0.041)	(0.040)	(0.041)	(0.040)	(0.011)	(0.011)	(0.011)	(0.011)
Observations	7,621	7,621	7,621	7,621	76,532	76,532	76,532	76,532
Year effects	YES	YES	YES	YES	YES	YES	YES	YES
Industry effects	YES	YES	YES	YES	YES	YES	YES	YES
F-Statistics	58.11 ***	64.39 ***	57.62 ***	60.27 ***	485.21 ***	540.16 ***	540.16 ***	509.19 ***
Adjusted R square	0.087	0.098	0.089	0.089	0.080	0.088	0.088	0.083

This table reports the FE regression results for the effect of board characteristics on tax avoidance in the regulated and unregulated firms. Standard errors reported in parentheses. ** significant at the 5% level, and *** significant at the 1% level. Variable definitions are provided in Appendix A.

reports the results of multivariate regressions conducted by splitting the sample into regulated and unregulated firms. As shown in Table 7, all board characteristics exhibit a positive and statistically significant association with tax avoidance in both regulated and unregulated firms. Additionally, our findings suggest that the impact of board gender diversity and board independence on tax avoidance is more pronounced within regulated firms, as evidenced by relatively stronger coefficients compared to their counterparts in unregulated firms.

5.4. Sample Variation: Excluding Dominant Countries

Table 8. The impact of board characteristics on tax avoidance, excluding observations from China, the USA, and Hong Kong.

	(A) (Equation (1))	(B) (Equation (2))	(C) (Equation (3))	(D) (Equation (4))
BS	0.0009 ***
	(0.0002)
INDIR		0.0007 ***
		(0.00003)
GENDER			0.0002 ***
			(0.00005)
BOFODI				0.002 ***
				(0.0002)
Controls	INCLUDED	INCLUDED	INCLUDED	INCLUDED
Constant	0.215 ***	0.195 ***	0.209 ***	0.203 ***
	(0.014)	(0.014)	(0.014)	(0.014)
Observations	40,390	40,390	40,390	40,390
Year effects	YES	YES	YES	YES
Industry effects	YES	YES	YES	YES
F-Statistics	271.25 ***	308.71 ***	271.88 ***	282.62 ***
Adjusted R square	0.095	0.106	0.096	0.098

This table reports the FE regression results for the effect of board characteristics on tax avoidance, excluding observations from China, the USA, and Hong Kong. Standard errors reported in parentheses. *** significant at the 1% level. Variable definitions are provided in Appendix A.

Columns A to D present the results analyzing the effect of board characteristics on corporate tax avoidance, excluding China, the USA, and Hong Kong, which present 50% of the sample observations. As a result, Table 8 shows that all board characteristics are positively and statistically significantly linked to tax avoidance. This is in line with the results shown in Table 4, which supports our main hypotheses (H1 to H4).

6. Conclusions

Our paper examines the effect of board characteristics on tax avoidance and the moderating role of industry regulation in this relationship. Using comprehensive panel data of 84.153 firm-year observations from 39 countries during the period of 2000–2023 and multiple regression models, our analysis confirms that larger board sizes, higher female representation, significant foreign ownership, and the presence of independent directors are associated with higher effective tax rates, indicative of reduced tax avoidance. These findings support the hypotheses that larger boards enhance oversight capabilities and that diverse board composition, including gender diversity and foreign ownership, brings higher ethical standards and governance practices that align with reduced tax avoidance. Furthermore, our results indicate that the effect of board gender diversity and board foreign ownership on tax avoidance is more pronounced in regulated industries, where stringent governance and ethical standards prevail.

The findings of this study provide substantial practical implications, offering critical insights for policymakers, regulators, and corporate governance practitioners. The evidence demonstrates that promoting diverse board compositions, particularly through increased gender diversity and the inclusion of independent directors and foreign ownership, can effectively mitigate tax avoidance, especially within regulated sectors. Policymakers and regulators can leverage these results to inform targeted governance reforms by developing and enforcing regulations that incentivize or mandate enhanced board diversity and independence. Furthermore, the pronounced effects observed in regulated industries underscore the necessity for stringent regulatory oversight, suggesting that robust external monitoring mechanisms can significantly enhance the effectiveness of governance practices. Corporate boards and stakeholders are thus encouraged to proactively adopt these governance strategies not only to comply with ethical standards and regulatory expectations but also to strengthen their firms’ reputations, reduce reputational and financial risks associated with aggressive tax strategies, and enhance long-term sustainability. Overall, these findings support the development and implementation of regulatory frameworks that foster transparency and accountability, aligning corporate conduct with broader societal expectations and contributing to more equitable and sustainable economic outcomes.

We acknowledge that, despite its contributions, this study is subject to certain limitations. Primarily, our analysis does not directly incorporate the complexities arising from frequent changes in tax codes and regulatory environments across different jurisdictions, which could significantly impact firms’ tax avoidance behaviors. Future research could benefit from explicitly examining how changes in tax codes and varying regulatory frameworks influence corporate governance and tax avoidance practices. Moreover, extending the research to include qualitative methodologies, such as interviews with board members or regulatory officials, might yield deeper insights into the practical implications of regulatory dynamics. Clearly outlining these areas for further exploration ensures our conclusions remain appropriately contextualized, focused, and informative for policymakers and practitioners.