Taxation is the primary revenue source for the Indonesian government, contributing to over 70% of the total state income. In 2023, the Directorate General of Taxes reported that tax revenue reached IDR 1,869.2 trillion, which was approximately 102.8% of the state budget target. Despite this achievement, corporate tax avoidance remains a major challenge that undermines optimal tax revenue collections. The Tax Justice Network’s Financial Secrecy Index 2022 ranked Indonesia 66th, highlighting the persistent gaps in corporate financial transparency that facilitate tax avoidance.
Hanlon and Heitzman (2010) define tax avoidance as any legal action taken by firms to reduce their tax liabilities. Although legal, such practices generate future tax risks, significantly reduce state revenue, and facilitate tunneling activities that harm minority shareholders (Falbo & Fletcher, 2025). Within tax accounting, three key factors drive corporate tax avoidance: tax planning, deferred tax expenses, and earnings management. These factors are closely linked to book-tax differences, which are discrepancies between accounting and taxable income that serve as indicators of tax avoidance.
Tax planning involves the strategic arrangement of financial transactions to minimize tax burdens (Karayan & Swenson, 2006). Alfadin and Ernandi (2024) used SmartPLS on LQ45 companies and found that tax planning significantly affects earnings management, which contributes to tax avoidance. Deferred tax expenses arising from temporary differences between accounting and taxable income can also signal earnings management aimed at reducing the tax burden (Midiastuty et al., 2023). This relationship is particularly relevant in Indonesia, where complex fiscal regulations provide room for the strategic exploitation of temporary differences.
Earnings management, through both accrual and real approaches, is frequently observed in Indonesian corporate reports. Kurniyawati et al. (2025) found that real earnings management increases book-tax differences (BTD) among IDX manufacturing firms (2021–2023), although it does not significantly affect the effective tax rate (ETR). Delgado et al. (2023) used a neural network approach and confirmed a significant earnings management–tax avoidance relationship among large European firms. These mixed findings underscore the complexity of this relationship and the need for further investigation in emerging markets, such as Indonesia.
Corporate governance functions as a monitoring mechanism that constrains managerial opportunistic behavior, including aggressive tax avoidance. Salehi et al. (2024) showed that board size, board independence, and audit quality influence tax avoidance levels. In Indonesia, corporate governance reduces tax avoidance in IDX-listed companies. However, prior research yields mixed results on whether governance effectively moderates the tax accounting factors–tax avoidance relationship, indicating a research gap.
This study addresses this gap with three distinct contributions that differentiate it from previous research. First, unlike previous studies that examined tax planning, deferred tax expense, or earnings management in isolation or in pairs, this study integrates all three tax accounting variables within a single structural model with corporate governance as a moderating variable, a combination that has not been tested within the unified SEM-PLS framework. This multivariable approach captures the simultaneous and interactive effects of these factors on tax avoidance, providing a more comprehensive picture than single-variable designs. Second, SmartPLS 4 enables the simultaneous testing of direct relationships and moderating effects with greater statistical rigor than conventional regression methods. Third, the observation period of 2020–2024 captures post-COVID-19 pandemic dynamics that altered fiscal policy and corporate tax behavior in Indonesia, a contextual dimension that is absent from earlier studies that relied on pre-pandemic data.
Based on these gaps and contributions, this study aims to: (1) examine the effect of tax planning on tax avoidance; (2) examine the effect of deferred tax expense on tax avoidance; (3) examine the effect of earnings management on tax avoidance; (4) test the moderating role of corporate governance in the tax planning–tax avoidance relationship; (5) test the moderating role of corporate governance in the deferred tax expense–tax avoidance relationship; and (6) test the moderating role of corporate governance in the earnings management–tax avoidance relationship.

2.1. Agency Theory
Agency theory, developed by Jensen and Meckling (1976), explains the contractual relationship between principals (shareholders) and agents (management). In the taxation context, agency conflicts arise when management has incentives to engage in tax avoidance to increase a firm's net income, which, in turn, may enhance managerial compensation. However, this behavior may not align with long-term shareholder interests, as it can increase tax and reputational risks (Al-Faryan, 2024). Agency theory provides an important foundation for understanding why corporate governance mechanisms are needed to monitor and constrain managerial opportunistic behavior regarding tax avoidance. Within this framework, information asymmetry between principals and agents creates space for management to make tax decisions that benefit themselves at the shareholders’ expense.

2.2. Tax Compliance Theory
Tax compliance theory posits that taxpayer compliance is influenced not only by law enforcement factors but also by perceptions of fairness, trust in tax institutions, and reputation. Within this framework, internal tax administration practices, such as transparency, board oversight, and internal risk controls, function as institutional mechanisms that reinforce compliant behavior and reduce the likelihood of aggressive tax planning. Recent developments in tax compliance theory also emphasize the role of technology, including artificial intelligence and big data analytics, in enhancing tax authorities' capacity to detect patterns of corporate tax avoidance (Belahouaoui & Attak, 2024).
Critically, agency and tax compliance theories are complementary rather than competing frameworks for explaining tax avoidance. Agency theory identifies the source of the problem, information asymmetry, and misaligned incentives that motivate managers to pursue aggressive tax strategies, while tax compliance theory addresses the institutional and behavioral mechanisms that can counteract this tendency. Corporate governance sits at the intersection of both theories: it functions as a monitoring tool that reduces agency conflict (agency theory) while simultaneously reinforcing the norms of transparency and compliance that promote voluntary adherence to tax obligations (tax compliance theory). This integrative perspective explains why governance mechanisms are hypothesized to moderate the relationship between tax accounting factors and tax avoidance, as they simultaneously reduce opportunistic behavior and strengthen compliance incentives.

2.3. Tax Avoidance
Tax avoidance is broadly defined as any legal action taken by firms to reduce tax liabilities (Hanlon & Heitzman, 2010). In the accounting literature, tax avoidance is measured using various proxies, including the effective tax rate (ETR), cash effective tax rate (CETR), and book-tax differences (BTD). A lower ETR indicates tax avoidance. Recent research by Fransiska et al. (2025) demonstrates that tax avoidance negatively affects firm value, meaning markets penalize firms engaged in aggressive tax avoidance. Additionally, Falbo and Fletcher (2025) found that tax avoidance can facilitate tunneling activities detrimental to minority shareholders in Indonesia, strengthening the argument that tax avoidance is not merely a fiscal issue but also a corporate governance concern.
In the Indonesian context, a systematic literature review by Andrianto et al. (2025), covering 66 articles in Sinta-1 and Sinta-2 accredited journals during 2015–2025, found that the governance pillar dominates 87.6% of proxies used in Indonesian tax avoidance research, with institutional ownership, audit committees, and independent commissioners as the most common variables. This finding underscores the importance of governance mechanisms within Indonesia’s tax-avoidance research framework.

2.4. Prior Research
Several relevant prior studies provide an empirical foundation for this research. Alfadin and Ernandi (2024) examined the effect of tax planning and deferred tax expenses on earnings management with tax rates as a moderating variable using SmartPLS 3 on LQ45 companies for 2017–2019, finding that tax planning significantly influences earnings management. Salehi et al. (2024) tested various governance mechanisms against tax avoidance on the Tehran Stock Exchange and found that board size, board independence, and audit quality are significant factors. Kurniyawati et al. (2025) analyzed the moderating role of board gender diversity on the earnings management–tax avoidance relationship in IDX manufacturing firms, finding that real earnings management increases BTD, but the proportion of female directors (averaging 10.9%) fails to moderate the relationship.
Haris and Ramadhan (2025) examined the effects of profitability and sales growth on tax avoidance with earnings management as a moderating variable in food and beverage companies for 2014–2023, finding that the interaction between sales growth and earnings management is significant. Chen et al. (2024) explored the impact of tax planning and earnings management on earnings persistence, finding significant implications for earnings quality.

2.5. Hypotheses Development

2.5.1. The Effect of Tax Planning on Tax Avoidance
Effective tax planning enables companies to legally minimize their tax burdens. Based on agency theory, management has the incentive to pursue tax planning to enhance net income and firm performance. Tax planning encompasses investment, financing, and operational decisions that strategically consider tax implications (Karayan & Swenson, 2006). Alfadin and Ernandi (2024) demonstrate that tax planning influences tax-avoidance practices. The more effective the tax planning, the greater the firm's ability to reduce its tax burden. Therefore:
H1: Tax planning positively affects tax avoidance.

2.5.2. The Effect of Deferred Tax Expense on Tax Avoidance
A higher deferred tax expense reflects larger temporary differences between accounting and taxable income. Under PSAK 46 (the Indonesian equivalent of IAS 12), these temporary differences create future tax obligations that can be exploited to defer tax payment. Research by Midiastuty et al. (2023) and Hanifah et al. (2025) found associations between deferred tax expenses and earnings management practices, with implications for tax avoidance. Companies with substantial deferred tax expenses have greater flexibility in managing the timing of income and expense recognition for tax purposes. Accordingly:
H2: Deferred tax expenses positively affect tax avoidance.

2.5.3. The Effect of Earnings Management on Tax Avoidance
Earnings management through discretionary accruals enables companies to manipulate their taxable income figures. Kurniyawati et al. (2025) found that real earnings management increases book-tax differences, which serve as indicators of tax avoidance. Delgado et al. (2023) confirmed a significant relationship between earnings management and tax avoidance using a neural network approach. Managers can leverage accrual policies to lower taxable income while maintaining a high accounting income, creating substantial book-tax differences. The third hypothesis is formulated as follows:
H3: Earnings management positively affects tax avoidance.

2.5.4. The Moderating Role of Corporate Governance
Strong corporate governance, particularly through oversight by independent commissioners, audit committees, and external audit quality, is expected to constrain aggressive tax planning and earnings management for tax avoidance. Kerr et al. (2024) found that governance reforms lead to a reduction in tax avoidance levels. Baatwah et al. (2025) demonstrated that audit committee chair expertise effectively reduces tax avoidance in emerging markets. Kovermann and Velte (2019), through a comprehensive literature review, affirmed that internal and external governance mechanisms play important roles in controlling corporate tax aggressiveness. Corporate governance reduces tax-avoidance practices in Indonesia. Based on these arguments:
H4: Corporate governance weakens the effect of tax planning on tax avoidance.
H5: Corporate governance weakens the effect of deferred tax expenses on tax avoidance.
H6: Corporate governance weakens the effect of earnings management on tax avoidance.

3.1. Research Design
This study employed a quantitative approach with a causal research design to test the cause-and-effect relationships between the independent and dependent variables. The analytical method used is Structural Equation Modeling-Partial Least Squares (SEM-PLS) using SmartPLS 4. The selection of SEM-PLS is based on the consideration that this method can handle complex research models with moderation variables, does not require normally distributed data, and is suitable for relatively small to medium sample sizes (Hair et al., 2019). Furthermore, SEM-PLS enables the simultaneous testing of measurement and structural models, providing more comprehensive results than conventional regression methods.

3.2. Population and Sample
The research population comprises all manufacturing companies listed on the Indonesia Stock Exchange between 2020 and 2024. The manufacturing sector was selected because of its considerable variation in tax policies, capital structures, and financial reporting practices. Purposive sampling was applied with the following criteria: (1) consistently listed on the IDX during 2020–2024; (2) presenting complete audited financial statements; (3) reporting no losses during the observation period, as ETR cannot be meaningfully calculated for loss-making firms; and (4) providing complete data on all variables. Based on these criteria, 52 companies were selected, yielding 260 firm-year observations for the final sample. Data were obtained from annual financial reports accessed through the official IDX website (www.idx.co.id).

3.3. Variable Operationalization
The dependent variable, tax avoidance (Y), is measured using the Effective Tax Rate (ETR), which is the ratio of income tax expense to pre-tax income. A lower ETR indicates tax avoidance. The first independent variable, tax planning (X1), is measured using the Tax Retention Rate (TRR), the ratio of net income after tax to pre-tax income. The second independent variable, deferred tax expense (X2), is measured as the ratio of deferred tax expenses to total assets. The third independent variable, earnings management (X3), is measured using discretionary accruals based on the Modified Jones Model (Dechow et al., 1995).
The moderating variable, corporate governance (Z), is measured using a composite index comprising three indicators: (1) the proportion of independent commissioners (Z1) to the total number of board members; (2) audit committee size (Z2); and (3) audit quality (Z3), measured as a dummy variable (1 if audited by a Big Four firm, 0 otherwise). These three indicators were reflected in a single latent construct within the SEM-PLS model. Two control variables are included: firm size (natural logarithm of total assets) and leverage (debt-to-assets ratio) to account for the influence of firm characteristics on tax avoidance.

3.4. Data Analysis Technique
The data analysis was conducted in several stages. First, the descriptive statistics characterized the data distribution. Second, the measurement model evaluation (outer model) assesses convergent validity (outer loading > 0.70; AVE > 0.50), discriminant validity (HTMT criterion < 0.90), and reliability (Cronbach’s Alpha and Composite Reliability > 0.70). Third, the structural model evaluation (inner model) examines the coefficient of determination (R²), predictive relevance (Q²) through blindfolding, effect size (f²), and hypothesis testing through bootstrapping with 5,000 subsamples at the 5% significance level (t-statistic > 1.96; p-value < 0.05). Moderation effects were tested using the product indicator approach in SmartPLS 4.

4.1. Results

4.1.1. Descriptive Statistics
Based on purposive sampling criteria, the final sample comprised 52 manufacturing companies with 260 observations from 2020 to 2024. The descriptive statistics are presented in Table 1.

Table 1. Descriptive Statistics

Source: Data processed with SmartPLS 4, 2026

Table 1 shows that the mean ETR of manufacturing companies is 0.243 (24.3%), approximating the Indonesian corporate tax rate of 22%, indicating moderate levels of tax avoidance. The relatively low ETR standard deviation (0.062) masks a wide range between the minimum (0.082) and maximum (0.418), revealing heterogeneous tax practices across countries. The mean tax retention rate of 0.757 indicates that companies retain 75.7% of their pre-tax income after tax deductions. The average discretionary accruals are near zero (0.012), with a standard deviation of 0.074, indicating diverse earnings management practices. The mean proportion of independent commissioners (39.7%) exceeds the minimum OJK requirement of 30%, and 36.5% of the companies are audited by Big Four firms.

4.1.2. Measurement Model Evaluation
The convergent validity assessment confirmed that all corporate governance indicators exhibited outer loadings above 0.70, with an AVE of 0.614 exceeding the 0.50 threshold. Discriminant validity was verified through HTMT values below 0.90 across all construct pairs. Reliability was established with a Cronbach’s alpha of 0.718 and a Composite Reliability of 0.827. The results are presented in Table 2.

Table 2. Measurement Model Results

Source: SmartPLS 4 Output, 2026

All measurement model criteria were satisfied, confirming that the corporate governance construct was valid and reliable. For single-item indicators, outer model evaluation is unnecessary, as outer loading automatically equals 1.00.

4.1.3. Structural Model Evaluation and Hypothesis Testing
The R² value for tax avoidance is 0.412, meaning that 41.2% of the variation in tax avoidance is explained by the model. The Adjusted R² of 0.394 is classified as moderate according to Hair et al. (2019). The Q² value from blindfolding is 0.287 (> 0), confirming good predictive relevance. The bootstrapping results with 5,000 subsamples are presented in Table 3.

Table 3. Hypothesis Testing Results (Bootstrapping)

Source: SmartPLS 4 Output, Bootstrapping 5,000 subsamples, 2026
Note: Negative coefficients against ETR indicate higher tax avoidance (lower ETR)

4.2. Discussion

4.2.1. The Effect of Tax Planning on Tax Avoidance (H1: Supported)
The results demonstrate that tax planning has a significant negative effect on ETR (coefficient = -0.327; t = 4.218; p = 0.000; f² = 0.124). The negative coefficient of ETR means that higher tax planning effectiveness (higher TRR) leads to lower effective tax rates, indicating greater tax avoidance. An effect size of 0.124 is classified as medium, demonstrating that tax planning contributes substantially to the model. This finding is consistent with those of Alfadin and Ernandi (2024) and supports agency theory's prediction that management leverages tax planning expertise to optimize the corporate tax burden. From a practical standpoint, companies with high TRR demonstrate a superior ability to retain after-tax income through effective yet potentially aggressive planning strategies. This implies that tax authorities should pay particular attention to firms with consistently high tax retention rates as potential indicators of aggressive tax planning.

4.2.2. The Effect of Deferred Tax Expense on Tax Avoidance (H2: Supported)
Deferred tax expense exhibits a significant negative effect on ETR (coefficient = -0.186; t = 2.542; p = 0.011; f² = 0.042). This finding indicates that companies with higher deferred tax expenses tend to exhibit higher levels of tax avoidance. Temporary differences underlying deferred tax expenses can be exploited as instruments for tax deferral. This is consistent with the findings of Midiastuty et al. (2023) and Hanifah et al. (2025), who found an association between deferred tax expenses and tax avoidance practices. In the Indonesian regulatory context, flexibility in income and expense recognition between financial accounting standards (SAK) and tax regulations provides room for companies to strategically manage temporary differences to lower their current-period effective tax burden. The relatively small effect size (0.042) suggests that, while significant, deferred tax expense plays a more supplementary role than tax planning in driving tax avoidance.

4.2.3. The Effect of Earnings Management on Tax Avoidance (H3: Supported)
Earnings management, measured through discretionary accruals, shows a significant negative effect on ETR (coefficient = -0.214; t = 3.017; p = 0.003; f² = 0.056). This demonstrates that companies engaging in more aggressive earnings management tend to pay lower effective tax rates than others. This finding is consistent with Kurniyawati et al. (2025), who found that real earnings management increases book-tax differences, and Delgado et al. (2023), who confirmed the earnings management–tax avoidance nexus in European economies. Discretionary accruals provide managers with flexibility in timing of income and expense recognition, which can be leveraged to lower taxable income. Managers may accelerate expense recognition or defer revenue recognition to reduce fiscal income, thus lowering the effective tax burden. This finding has important implications for tax authorities, suggesting that the analysis of discretionary accrual patterns can serve as an effective tool for identifying potential tax-avoidance activities.

4.2.4. The Moderating Role of Corporate Governance (H4, H5, H6)
Corporate governance effectively moderates (weakens) the effect of tax planning on tax avoidance (H4: interaction coefficient = 0.152; t = 2.108; p = 0.035; f² = 0.028). The positive interaction coefficient of ETR indicates that stronger governance increases ETR (reduces tax avoidance) in companies with aggressive tax planning. Independent commissioners and effective audit committees can oversee management's tax policies, thereby constraining overly aggressive tax planning. This finding aligns with that of Kerr et al. (2024), who found that governance reforms reduce tax avoidance.
Corporate governance also moderates the effect of earnings management on tax avoidance (H6: coefficient = 0.168, t = 2.274, p = 0.023, f² = 0.033). Strong internal oversight mechanisms, including audits by Big Four firms, can detect and constrain earnings management practices aimed at reducing the tax burden. The Big Four auditors are known for their stricter audit standards and higher independence, which reduces management's ability to engage in earnings management for tax avoidance. The combined evidence from H4 and H6 demonstrates that corporate governance serves as an effective safeguard against the two primary behavioral mechanisms tax planning and earnings manipulation through which firms reduce their tax obligations.
However, the most notable finding is that corporate governance does not significantly moderate the effect of deferred tax expenses on tax avoidance (H5: coefficient = 0.098; t = 1.347; p = 0.178; f² = 0.011). This result is unexpected, given the theoretical expectation that governance should broadly constrain all tax avoidance mechanisms, and warrants deeper interpretation. We propose that the divergent moderating effect of H4/H6 versus H5 reflects a fundamental distinction between the behavioral and technical mechanisms of tax avoidance. Tax planning (H4) and earnings management (H6) are behavioral decisions made by managers that are visible in strategic choices and discretionary accrual adjustments and are thus detectable through board oversight. In contrast, deferred tax expenses arise from technical accounting treatments governed by PSAK 46 and involve complex estimates of temporary differences between financial and tax reporting. These technical judgments require specialized knowledge of both accounting standards and tax regulations, making them less accessible to conventional governance and oversight.
This interpretation is consistent with Baatwah et al. (2025), who emphasized that the audit committee chair’s specific tax expertise, rather than general governance quality, is the crucial differentiating factor in effective tax oversight. Without adequate tax accounting expertise, boards and audit committees may lack the capacity to critically evaluate deferred tax policies involving complex temporary difference calculations, asset recognition thresholds, and reversal timing assumptions. The small effect size (f² = 0.011) further supports this interpretation, suggesting that conventional governance mechanisms have a near-zero practical impact on deferred tax-related avoidance. This finding has important implications for regulatory reform: requiring audit committees to include members with specific tax accounting qualifications could bridge the expertise gap and extend governance effectiveness to technical tax areas that currently remain beyond its reach.
Among the control variables, leverage exhibits a significant negative effect on ETR (coefficient = -0.143; t = 2.016; p = 0.044), confirming that more highly leveraged firms benefit from the tax shield of debt through interest expense deductions that reduce taxable income. Firm size does not significantly affect tax avoidance (p = 0.147), indicating that tax avoidance is not confined to large firms but occurs across various scales of operation.

This study yielded several significant findings. First, tax planning has a significant positive effect on tax avoidance, with a medium effect size (f² = 0.124), making it the strongest determinant in the model. More effective tax planning leads to lower tax rates. Second, deferred tax expenses significantly increase tax avoidance, confirming that temporary differences are strategically exploited to defer tax payments. Third, earnings management through discretionary accruals significantly increases tax avoidance, demonstrating that income manipulation reduces the effective tax burden.
Fourth, corporate governance effectively moderates (weakens) the effects of tax planning and earnings management on tax avoidance, affirming the important role of oversight mechanisms in reducing tax aggressiveness. Fifth, corporate governance fails to moderate the effect of deferred tax expenses on tax avoidance, indicating the need for specific tax expertise within governance mechanisms to oversee complex deferred tax policies.
Theoretically, this study provides empirical evidence that agency theory and tax compliance theory are relevant in explaining tax avoidance in Indonesia and strengthens the argument that governance mechanisms serve as essential instruments in mitigating agency conflicts related to taxation. Practically, this study recommends that the Directorate General of Taxes prioritize the surveillance of companies exhibiting high tax retention rates and large discretionary accruals. For companies, these findings underscore the importance of strengthening commissioner independence and audit committee effectiveness and considering the addition of audit committee members with tax expertise.
This study has several limitations. First, the sample was limited to the manufacturing sector, requiring generalizability testing across other industries. Second, governance measurement employs only three indicators that may not fully represent the comprehensive governance quality. Third, the observation period encompasses the COVID-19 pandemic, which may influence financial reporting patterns. Future research should expand to multiple industry sectors, incorporate ESG proxies for governance (Andrianto et al., 2025), consider transfer pricing and political connection variables, and employ longitudinal methods to capture the temporal dynamics of tax avoidance behavior.