The Relationship Between the Reliability of Financial Statements and The Risk of Tax Avoidance and The Effect of External Audit Quality as A Moderating Variable: Empirical Evidence from Companies Listed the Iraq Stock Exchange

Full text

Text extracted from the article PDF for reading and search. Tables, figures and equations are in the PDF, which remains the authoritative version.

1. Introduction

Only if people can trust financial statements, can they be used as a coordination tool. Banks lend against that trust, investors assign prices to it, regulators tweak their grip around it, and taxmen assess liabilities on the assumption that the numbers presented reflect what really happened inside the company. If that assumption is even partially broken, the damage is not just to the balance sheet but to the way capital is allocated, to the enforcement of contracts and ultimately to the amount of money the state can raise. One of the more significant examples is the relationship between erroneous financial reporting and a company's aggressive tax position management. A manager that wants to look profitable to investors but pay little to the treasury has two pulls in the same direction. Push book income up and pull taxable income down. In most countries, financial statements and tax returns are prepared from an overlapping but different set of underlying records An increasing number of studies consider heavy, opaque tax avoidance a warning sign, not a neutral cost-saving option, as it is usually associated with rent extraction, loose internal controls and managers with more freedom than their shareholders would want (Desai & Dharmapala, 2006; Hanlon & Heitzman, 2010) [17, 29]. Tax avoidance isn’t actually illegal. A preference for aggressive tax positions seems to go hand in hand with the same opacity that permits earnings to be managed, usually through the use of discretionary accruals (Frank, Lynch, & Rego, 2009) [27]. The external auditor is the nexus of these two issues. In theory, a capable and independent audit should discourage both, because it increases the likelihood that misstatements will be discovered, and the legal and reputational consequences of ignoring them (DeAngelo, 1981; Francis, 2004; DeFond & Zhang, 2014) [12, 26, 15].

The particular situation of examining this triad of constructs within the Iraqi institutional framework is particularly informative and, to date, understudied. Reopened in 2004 following the fall of the previous regime, the Iraq Stock Exchange (ISX) is still a small, thinly traded and governance-fragile market where tax administration, audit monitoring and disclosure procedures are all in different stages of institutional development. Under the Income Tax Law No. 113 of 1982 (as amended), the Iraqi General Tax Authority still relies mainly on self-assessed financial statements for corporate income tax purposes. Companies Law No. 21 of 1997 requires an external audit, but this is generally carried out by a scattered population of local audit offices of widely varying quality and by a small number of firms belonging to international networks. Indeed, the interactions the study examines are enabled by the combination of a self-assessment tax regime, uneven audit quality, and a developing capital market. However, there remains a lack of local empirical evidence that simultaneously models financial statement reliability, tax avoidance, and audit quality as a moderator; past Iraqi research has generally examined pairwise relationships, such as audit quality and tax evasion (Al Kut Journal of Economics and Administrative Sciences, 2026) [1] or ownership structure and financial statement quality (Wahhab & Khlaif, 2022) [56]. Inspired by this gap, the current study poses the following research question: Does the quality of the external audit conducted on the financial statements of companies listed on the Iraq Stock Exchange mitigate the risk that those companies engage in tax avoidance? The study aims to address this issue by measuring financial statement reliability and tax avoidance risk for a panel of ISX-listed companies using established, internationally validated proxies; testing the direct relationship between financial statement reliability and tax avoidance risk; and determining whether and to what extent external audit quality moderates this relationship. The study is set up as follows in order to achieve these goals. In Section 2, the study is placed within the institutional framework of Iraq and the theoretical and empirical literature on financial statement reliability, tax avoidance, and external audit quality is reviewed. The formal hypotheses and theoretical framework are developed in Section 3. The research approach, including the sample, variable measurement, and model specification, is described in depth in Section 4. The empirical data are presented and analyzed in Section 5. The results are discussed in Section 6 in relation to earlier research. The study's consequences, limitations, and future research directions are discussed in Section 7. This study's importance stems from a number of related factors. By combining, for the first time in an Iraqi study to the authors' knowledge, internationally validated accrual-based reliability measurement, a multi-proxy operationalization of tax avoidance risk, and a composite, moderator-specified treatment of external audit quality within a single panel-data model, it directly addresses repeated calls in the international literature for context-specific tests of the reliability-taxation-audit nexus outside the Anglo-American and large-emerging-market settings in which the majority of previous evidence. Practically speaking, the study addresses an ongoing policy debate in Iraq regarding the modernization of capital-market oversight and tax administration: as the Iraqi Securities Commission continues to tighten disclosure and audit-registration requirements for ISX-listed companies, and as the Iraqi General Tax Authority moves toward more systematic, risk-based examination of corporate taxpayers, empirical evidence on whether and through what channel audit quality constrains the joint occurrence of aggressive tax positions and unreliable reporting. Finally, from an investor-protection standpoint, the study's findings bear on the pricing and monitoring decisions of the relatively small population of institutional and retail investors participating in the ISX, for whom both financial-statement reliability and the tax-risk profile of portfolio companies are material considerations that are, as this study demonstrates, empirically and theoretically intertwined rather than independent.

2. Literature Review and Theoretical Background

2.1. Financial Statement Reliability: Concept and

Measurement Reliability, together with relevance, was long treated by the IASB and FASB conceptual frameworks as a primary qualitative characteristic of useful financial information; the 2010 and 2018 revisions of the IASB Conceptual Framework replaced the term with "faithful representation", defined as information that is complete, neutral, and free from error, but the underlying construct that empirical accounting research continues to proxy, namely the fidelity of reported numbers to underlying economic reality, is unchanged (IASB, 2018) [34]. Because the "true" economic income of a firm is unobservable, researchers have developed indirect proxies for reliability, the most influential of which is the magnitude of discretionary (abnormal) accruals estimated from a cross-sectional or time-series accrual model. The original Jones (1991) model separates total accruals into a non-discretionary component explained by changes in revenue and the level of property, plant and equipment, and a discretionary residual attributed to managerial discretion [39]. Dechow, Sloan, and Sweeney (1995) refined the model by adjusting the change in revenue for the change in receivables, producing the "modified Jones model" that remains the workhorse specification in the literature [13]. This study uses a performance-matched adjustment, as suggested by Kothari, Leone, and Wasley (2005), who further showed that discretionary accrual estimates are biased when firm performance is not taken into account [43]. In their thorough analysis of earnings-quality proxies, Dechow, Ge, and Schrand (2010) come to the conclusion that dependability is best triangulated using both accrual-based and outcome-based metrics, such as restatements, and that no single proxy is consistently better [14].

The incidence of financial statement restatements is a second, complementary stream of research that operationalizes reliability failures directly. This is based on the idea that a restatement is an ex-post, relatively clear admission that previously issued statements did not accurately reflect the firm's financial position. The relevance of both proxies to the local market was confirmed by a recent study of Iraqi industrial companies listed on the ISX between 2019 and 2024, which found a significant negative correlation between the incidence of financial statement restatements and earnings quality as measured by discretionary accruals. To put it briefly, this body of work supports the creation of a composite dependability construct in the current study, which is based on performance-matched discretionary accruals and, when available, complemented by restatement history.

2.2. Tax Avoidance: Concept, Measures, and

Determinants Hanlon and Heitzman (2010) define tax avoidance broadly as the continuum of activities that reduce a firm's explicit tax liability relative to its pre-tax accounting income, a definition that spans wholly legal tax planning at one extreme and unlawful evasion at the other, with the vast majority of firm behaviour occupying an ambiguous middle ground [29]. Because tax avoidance is multidimensional, the literature has converged on a small set of complementary proxies. The effective tax rate (ETR) derived as total tax expense over pre-tax book income indicates the average tax burden conveyed by financial-statement figures, but is sensitive to non-cash timing-related tax items. To address this issue, the cash effective tax rate (CETR) is used, which is defined as the ratio of cash taxes paid to pre-tax book income. This rate reflects the actual cash outflows and is considered less prone to manipulation based on accrual accounting (Dyreng, Hanlon, & Maydew, 2008) [19]. An additional dimension of avoidance related to exploiting the differences between accounting and tax rules is captured by the book-tax difference (BTD), i.e., the difference between book income and taxable income grossed up by the statutory tax rate, and has been linked in several studies to aggressive tax planning and lower book-tax conformity (Hanlon & Shevlin, 2005; Sundvik, 2017; Menicacci, 2022; Al-Hadrawi, 2023) [30, 55, 47, 2]. The factors that influence tax avoidance, such as firm size, leverage, profitability, capital intensity, ownership structure, corporate governance, and managerial traits, have been well studied. According to Desai and Dharmapala (2006), since the opacity needed to conceal profits from tax authorities also makes it easier to divert resources from owners, tax avoidance and managerial rent extraction are complementary practices in poorly run businesses [17]. Dyreng, Hanlon, and Maydew (2010) demonstrate that, regardless of business and industry factors, individual executives have an economically substantial, additive impact on the degree of tax evasion attempted by the companies they oversee [20]. Family businesses are generally less tax aggressive than non-family businesses, according to Chen, Chen, Cheng, and Shevlin (2010) [11]. They attribute this to family owners' concentrated, non-diversified holdings and subsequent aversion to the reputational and reversal risk associated with aggressive stances. According to Wilson (2009), companies classified as tax shelter participants have unique financial traits, such as greater book-tax disparities and poorer corporate governance [59]. Barrios, Fujiy, Lisowsky, and Minnis (2025) link financial-reporting practices more broadly to measured firm productivity, highlighting the actual economic stakes of reporting-tax interactions [7], while Lampenius, Shevlin, and Stenzel (2021) propose refined measures of tax-rate and tax-base avoidance for domestic and multinational firms [44]. Firm size, leverage, and profitability are identified by Fernández-Rodríguez, García-Fernández, and Martínez-Arias (2021) as strong institutional predictors of the effective tax rate across a panel of emerging economies, particularly in emerging and Middle Eastern markets [25]. According to Kimea, Mkhize, and Maama (2023), aggressive tax management among East African listed enterprises is mostly driven by firm-specific features [42]. Numerous studies have directly examined ownership structure and audit-related factors of tax avoidance in the Iraqi and neighboring Iranian and Jordanian markets: Hashim and Barzegar report that ownership structure moderates the relationship between audit fees and tax avoidance among Iraqi listed companies over 2016-2022 [31]; Qawqzeh (2023) finds that audit quality moderates the ownership-structure to tax-avoidance relationship [48]; and Almaharmeh, Shehadeh, Alkayed, Aladwan, and Iskandrani (2024) report that higher corporate governance quality is linked to lower tax avoidance among Jordanian family businesses [4]. The audit-quality moderator investigated in this work is both theoretically driven and empirically precedented in similar institutional settings, according to this regional evidence.

2.3. External Audit Quality: Concept and Determinants

DeAngelo (1981) offers the classic definition of audit quality as the joint probability that an auditor both detects a material misstatement and reports it, and argues that this probability increases with auditor size because larger audit firms have more reputational capital at stake and are less economically dependent on any single client [12]. Empirical proxies for audit quality developed in the subsequent literature include auditor size or Big-4/international-network affiliation, audit firm tenure, audit fees, and industry specialisation. Becker, DeFond, Jiambalvo, and Subramanyam (1998) provide early evidence that clients of Big Six auditors report significantly lower discretionary accruals than clients of non-Big-Six auditors, consistent with higher-quality audits constraining earnings management [9]. Khurana and Raman (2004) find that Big-4 audits are associated with greater financial-reporting credibility, evidenced by a lower ex ante cost of equity capital, particularly in litigious environments [41]. Francis (2004) and DeFond and Zhang (2014) both provide comprehensive reviews concluding that, despite measurement challenges, the weight of evidence supports a positive association between conventional audit-quality proxies and financial-reporting quality [26, 15]. Using proprietary audit-adjustment data from China, Lennox, Wu, and Zhang (2016) expand on this findings by demonstrating that audit adjustments themselves enhance the accuracy of reported earnings [45]. This finding is especially pertinent to emerging countries like Iraq where audit-firm heterogeneity is significant.

A more recent and pertinent body of research especially looks at audit quality in connection to tax consequences. According to DeFond, Qi, Si, and Zhang (2025), individual signatory auditors with specialized tax experience reduce rather than increase client tax aggression [16]. This finding highlights the particular aspects of auditor competence—rather than just firm-level Big-4 status—that limit avoidance. The positive correlation between tax avoidance and audit report lag is considerably weakened by audit quality, according to a 2025– 2026 study of manufacturing companies listed on the Dhaka Stock Exchange. This suggests that higher-quality auditors are better at processing the informational complexity brought about by aggressive tax positions. On the other hand, a research of Iranian listed companies reveals a strong negative correlation between tax avoidance and audit quality, which is further mitigated by managerial conservatism. This suggests that the relationship between audit quality and taxes might go either way depending on the governance context. Strong contextual support for the moderating role examined in this study is provided by the Al Kut Journal of Economics and Administrative Sciences (2026) study of industrial companies listed on the ISX in Iraq between 2005 and 2024, which offers direct local evidence that improvements in external audit quality are linked to reductions in tax evasion [1].

2.4. The Relationship between Financial Statement

Reliability and Tax Avoidance The theoretical foundation for the association between financial statement reliability and tax avoidance is the fact that book income and taxable income are based on the same set of underlying transactions and the same management with incentives that may be aligned or conflicting, although computed under different rule sets. Frank, Lynch and Rego (2009) formalize and test the 'aggressive reporting' hypothesis. They find a significant positive correlation between aggressive tax reporting, which is proxied by the same underlying discretionary component, and aggressive financial reporting, which is proxied by discretionary permanent book-tax differences [27]. This finding is consistent with a subset of firms having the ability and incentive to simultaneously manage taxable income downward and book income upward. Rather than being a straightforward causal chain that only goes one way, Desai and Dharmapala (2006) offer a complementary theoretical explanation in which low reporting dependability and high tax avoidance are complementary symptoms of poor company governance and managerial rent extraction [17]. Using UK data, Song and Holland (2023) expand on this line of inquiry and discover that firms with higher levels of tax management or stronger earnings-management pressure have much lower quality tax accounting for financial-reporting purposes [53]. This finding directly links the reliability of the reported tax account to more widespread avoidance behavior. Gavious, Livne, and Chen (2022) demonstrate that this relationship is conditioned by the strength of external tax enforcement, such that avoidance reacts to reliability incentives differently depending on the credibility of the surrounding institutional environment [28]. This finding directly relates to the Iraqi General Tax Authority's comparatively weak enforcement capacity. As a structural predictor of the reliability-avoidance relationship, a similar body of scholarship looks at book-tax conformance, or the degree to which financial-accounting and tax-accounting norms align. The observed correlation between reliability proxies and avoidance proxies should diminish in high-conformity environments where opportunities to manage book and tax income independently are limited. Conversely, in low-conformity environments, such as those found in Iraq and other rule systems that maintain largely distinct financial-accounting and tax-accounting regimes, there is a greater chance that a reliability-avoidance relationship will materialize empirically (Hanlon & Shevlin, 2005; Eichfelder, Jacob, Kalbitz, & Wentland, 2020; Menicacci, 2022) [30, 23, 47]. On the basis of this literature, the present study anticipates a significant negative relationship between financial statement reliability and tax avoidance risk: firms whose reported earnings are less reliable, that is, whose absolute discretionary accruals are larger, are expected to exhibit higher tax avoidance risk, proxied by lower effective and cash effective tax rates and larger book-tax differences.

2.5. The Moderating Role of External Audit Quality

If financial statement reliability and tax avoidance are jointly determined by an underlying propensity toward opaque or aggressive reporting, then any external monitoring mechanism capable of constraining that propensity should also attenuate the strength of the observed reliability-avoidance relationship. The external auditor is the most direct such mechanism with respect to financial-statement reliability, and a growing literature argues that high-quality auditors additionally constrain the aggressive tax positions that often accompany low-reliability reporting, both because tax positions frequently rely on judgemental accounting estimates subject to audit scrutiny and because reputationally-concerned auditors are reluctant to be associated with clients whose overall reporting posture invites regulatory attention (Riguen, Salhi, & Jarboui, 2021; Rizqia & Lastiati, 2021; Sulaiman, Rashid, & Abd Hamid, 2023; Al-Taie & Al-Hadrawi, 2021) [49, 50, 54, 5]. Shabbir, Ahmad, and Máté (2023) find that audit quality significantly affects the sensitivity of tax avoidance to monetary-policy conditions among Pakistani listed firms, evidence that audit quality operates as a general dampener of avoidance behaviour across varying macro-financial conditions [52]. Within Iraq, Wahhab and Khlaif (2022) demonstrate that audit quality, jointly with ownership structure, significantly improves the quality of financial statements produced by banks listed on the ISX [56], while Hashim and Barzegar find that ownership structure moderates the audit-fee to tax-avoidance relationship among Iraqi listed companies [31], together establishing that audit-related moderation of reporting and tax outcomes is empirically detectable in the Iraqi market using comparable panel-data methods to those adopted here.

On this basis, the present study advances a moderation hypothesis in which external audit quality is expected to weaken, rather than eliminate, the negative relationship between financial statement reliability and tax avoidance risk: among firms audited by higher-quality external auditors, the marginal effect of a decline in reporting reliability on tax avoidance risk should be smaller than among firms audited by lower-quality auditors, because the higher-quality auditor constrains the joint opportunity set for both reliability failures and aggressive tax positions before they fully manifest in reported outcomes.

A related methodological literature has examined the reliability-avoidance relationship using disclosure-based, rather than purely accrual-based, proxies for reliability, including the readability and linguistic complexity of financial-statement footnotes and management discussion sections. Wang and Wright (2024) find that the frequency and framing of tax-strategy-related language in corporate disclosures is itself informative about the extent and persistence of a firm's tax avoidance [57], while a related Iranian study documents that the moderating role of financial-statement readability conditions the relationship between tax avoidance and the auditor's coping strategies for the elevated audit risk that avoidance creates, including audit report lag, audit fee premia, and the propensity to issue a modified audit opinion. Luo, Ma, Omer, and Xie (2024)

extend this disclosure-based perspective by showing that qualitative tax-footnote disclosure materially affects how capital markets price the firm-value consequences of measured tax avoidance [46], implying that the reliability of the surrounding disclosure environment, and not merely the numerical magnitude of accruals, shapes both the incidence and the perceived legitimacy of a given tax position. These disclosure-based findings are broadly complementary to, rather than in tension with, the accrual-based reliability construct adopted as the primary proxy in this study, and they motivate the use of the restatement-based proxy (RESTATE) as a partially independent robustness check in Section 5.7, since restatement incidence captures reliability failures that are eventually acknowledged in narrative disclosure rather than inferred solely from statistical residuals.

2.6. Institutional Context: Financial Reporting, Taxation, and Auditing in Iraq The banking industry has historically accounted for the largest share of market capitalization and trading activity on the Iraq Stock Exchange, which was founded in 2004 and is governed by the Iraqi Securities Commission. The exchange lists between 80 and 100 companies from a variety of industries, including banking, industrial, services, agriculture, insurance, hotels and tourism, and telecommunicationsUnder the Companies Law No. 21 of 1997 and the Securities Commission's subsequent directives, listed companies must prepare audited financial statements and, after a gradual transition, replace the pre-2003 planned-economy era's dominant Unified Accounting System with International Financial Reporting Standards, or IFRS for Small and Medium-sized Entities. Income Tax Law No. 113 of 1982, as later amended, governs corporate income taxation. It is administered by the General Commission for Taxes on a self-assessment basis, subject to ex post examination. Since 2015, the statutory corporate tax rate applicable to the majority of non-oil-sector companies has been set at 15%. However, effective rates vary significantly across sectors and firms due to a variety of exemptions, carried-forward losses, and sector-specific incentives, especially in the state has worked to promote industrial and agricultural activities.

A fragmented population of licensed audit offices registered with the Iraqi Association of Certified Public Accountants, ranging from small local practices to a small number of offices affiliated with international networks, conduct external audits of ISX-listed companies. The infrastructure for audit-firm rotation, quality-control inspection, and public audit-oversight is still far less developed than in mature markets, as evidenced by several recent Iraqi studies of audit quality and reporting outcomes (Wahhab & Khlaif, 2022; Al Kut Journal of Economics and Administrative Sciences, 2026) [56, 1]. This combination of a self-assessment tax system administered by a resource-constrained tax authority, a still-developing audit oversight infrastructure, and a capital market characterised by concentrated ownership and limited free-float creates precisely the conditions under which the theoretical mechanisms reviewed above, complementary opacity and aggressive taxation on one hand, and audit-quality constraint on the other, would be expected to operate with particular force, and it is this setting that motivates the empirical model developed in the remainder of this paper.

2.7. Summary of Prior Empirical Studies

Table 1 summarises the principal empirical studies reviewed above according to context, sample, and central finding, and highlights the specific gap that the present study addresses, namely the joint modelling of financial statement reliability, tax avoidance risk, and external audit quality as a moderator within the Iraqi capital market.

Table 1: Summary of Selected Prior Empirical Studies Study Context / Sample Key Constructs Main Finding Frank, Lynch & Rego (2009) U.S. listed firms Aggressive financial reporting; aggressive tax reporting Significant positive association between aggressive financial and tax reporting Desai & Dharmapala (2006) U.S. listed firms Tax avoidance; managerial rent extraction; governance Avoidance and rent extraction are complementary in weakly governed firms Dyreng, Hanlon & Maydew (2010) U.S. executives / firms Executive fixed effects; tax avoidance Individual executives have significant incremental effect on avoidance Chen, Chen, Cheng & Shevlin (2010)

U.S. family vs non-family firms Family ownership; tax aggressiveness Family firms are less tax aggressive on average Becker, DeFond, Jiambalvo & Subramanyam (1998) U.S. listed firms Auditor size; discretionary accruals Big Six clients report lower discretionary accruals Khurana & Raman (2004) U.S. listed firms Big-4 audits; cost of equity Big-4 audits associated with higher reporting credibility Lennox, Wu & Zhang (2016)

China (audit adjustments)

Audit adjustments; earnings precision Audit adjustments significantly improve earnings precision DeFond, Qi, Si & Zhang (2025)

China (signatory auditors)

Auditor tax expertise; tax aggressiveness Tax-expert signatory auditors curb client tax aggressiveness Song & Holland (2023) United Kingdom Tax accounting quality; earnings management Tax accounting quality lower under stronger tax/earnings management pressure Riguen, Salhi & Jarboui (2021) France Audit characteristics; gender diversity; tax avoidance Audit characteristics affect avoidance; moderated by board gender diversity Sulaiman, Rashid & Abd Hamid (2023) Malaysia Audit quality; financial performance; tax avoidance Audit quality and financial performance jointly reduce avoidance Shabbir, Ahmad & Máté (2023) Pakistan Monetary policy; audit quality; tax avoidance Audit quality conditions the monetary-policy to avoidance relationship ARA / Dhaka Stock Exchange study (2025-2026)

Bangladesh (manufacturing)

Tax avoidance; audit quality; audit report lag Audit quality weakens the positive avoidance-audit lag relationship Iranian Stock Exchange Iran Tax avoidance; audit quality; Negative avoidance-audit quality link moderated by study (2024-2025) managerial conservatism conservatism Wahhab & Khlaif (2022) Iraq (ISX banks) Ownership structure; audit quality; FS quality Audit quality and ownership jointly raise financial statement quality Hashim & Barzegar Iraq (ISX, 2016-2022) Ownership structure; audit fees; tax avoidance Ownership structure moderates audit fee-avoidance link Al Kut Journal of Economics (2026)

External audit quality; tax evasion Higher audit quality significantly reduces tax evasion ISX industrial restatement study (2019-2024)

Iraq (ISX industrial firms) Earnings quality; restatements Significant negative link between earnings quality and restatements As Table 1 illustrates, while the individual pairwise relationships between reliability and avoidance, and between audit quality and either construct, are well established internationally and increasingly documented within Iraq, no identified study models external audit quality as a formal moderator of the reliability-avoidance relationship using panel data drawn exclusively from ISX-listed companies across a decade-long window, a gap that the present study is designed to fill.

3. Theoretical Framework and Hypotheses Development

3.1. Underpinning Theories

The study is based on three complementary theoretical positions. Financial statement reliability and tax avoidance are theorized on the basis of agency theory (Jensen & Meckling, 1976; Fama & Jensen, 1983) [38, 24] because of the separation of ownership and control: managers are able to take advantage of their information superiority over dispersed shareholders and, in the case of Iraq, a tax authority with limited auditing capacity to manipulate reported earnings and reduce the firm’s explicit tax burden, particularly when the private benefits of doing so exceed the expected costs of detection. This view is supported by positive accounting theory (Watts & Zimmerman, 1986) [58] which predicts that managerial reporting decisions including the discretionary accrual decisions that form the basis of the reliability construct respond systematically to contracting incentives such as debt covenants, bonus plans and political-cost considerations, of which tax minimization is a prominent and clear example. The audit-quality moderator is based on the theory of signaling and disclosure (Healy & Palepu, 2001) [33]. A good external audit is an expensive credible signal of reporting integrity that decreases the information asymmetry between the company and external stakeholders including tax authorities. This, in turn, limits the space in which aggressive tax positions and reliability failures can be hidden.

3.2. Conceptual Model

To account for omitted-variable bias, we control for a vector of firm-level characteristics that include firm size, leverage, profitability, firm age, sales growth, capital intensity, and year and sector fixed effects. The conceptual model of this research is the study of financial statement reliability as independent variable, tax avoidance risk as dependent variable and external audit quality as a moderating variable on the path between the two. The model empirically examines whether the slope of the reliability-avoidance association systematically differs as a function of the level of external audit quality, by including a multiplicative interaction term between the reliability and audit-quality measures.

3.3. Hypotheses Development

The model investigates whether the slope of the reliability-avoidance connection varies systematically with the level of external audit quality by introducing a multiplicative interaction term between the reliability and audit-quality metrics.

H1: Among companies listed on the Iraq Stock Exchange, there is a statistically significant negative correlation between financial statement reliability and tax avoidance risk. Specifically, firms with lower reliability (larger absolute discretionary accruals) have higher tax avoidance risk (lower effective and cash effective tax rates and larger book-tax differences).

H2: External audit quality has a statistically significant negative effect on tax avoidance risk among companies listed on the Iraq Stock Exchange, such that firms audited by higher-quality external auditors exhibit lower tax avoidance risk.

H3: External audit quality significantly moderates the relationship between financial statement reliability and tax avoidance risk among companies listed on the Iraq Stock Exchange, such that the negative effect of low reliability on tax avoidance risk is significantly weaker among firms audited by higher-quality external auditors.

4. Research Methodology

4.1. Research Design and Philosophy

The study adopts a positivist, quantitative, explanatory research design consistent with the majority of the reviewed empirical literature, relying on secondary archival data extracted from audited financial statements and Iraq Stock Exchange disclosure bulletins rather than on primary survey instruments, since the constructs of interest, discretionary accruals, effective tax rates, and audit-quality proxies, are conventionally and more objectively measured from archival financial data. The design is longitudinal and cross-sectional, combining a panel of firms observed over multiple years, which allows the model to control for unobserved, time-invariant firm heterogeneity through fixed-effects estimation.

4.2. Population and Sample

The study population included all companies of the industrial sector of the Iraq Stock Exchange for the period 2016-2025. Three factors led to the deliberate selection of the industrial sector over a mixed cross-sector sample: first, Iraqi industrial firms are subject to a relatively homogeneous cost-accounting and inventory-valuation environment, which enhances the comparability of the accrual models used to estimate financial statement reliability; second, the industrial sector has historically been a stated target of Iraqi fiscal incentive policy, including tax exemptions and customs concessions intended to encourage domestic manufacturing; which produces significant cross-firm variation in effective tax burdens that are directly related to the tax-avoidance constructs studied here; and third, previous Iraqi audit-quality research has already established a trustworthy evidentiary base for industrial firms in particular, most notably the ISX industrial-sector restatement study reviewed in Section 2.1 and the Al Kut Journal of Economics and Administrative Sciences (2026) study of external audit quality and tax evasion among ISX industrial companies. Thus, we exclude banking, investment, insurance, services, agriculture, hotels and tourism, and telecommunications companies from the population, in line with the standard sector-homogeneity requirement of the accrual-quality and tax-avoidance literatures. In addition, the companies should have publicly available audited financial statements and continuously listed shares for at least five consecutive years during the study period. Companies that go bankrupt, halt trading or delist during that period are excluded, so it is an unbalanced panel. We expect that the use of these criteria on the industrial sector of the ISX, which in the past consisted of about eighteen to twenty-two listed companies, will yield a final estimation sample of about fifteen to eighteen industrial companies, or about 130 to 170 firm-year observations. Appendix A, Table A2 shows a representative set of ISX industrial-sector firms that meet the sector-classification requirement and are candidates for inclusion in the final estimation sample, subject to the data-availability checks described above. The companies are identified by their official ISX stock codes.

4.3. Data Sources and Study Period

The three main sources of data are: the daily and periodic bulletins of the Iraq Stock Exchange used to confirm listing status, sector classification and market capitalization; the audited annual financial statements and notes of the sample companies obtained from company disclosures filed with the Iraqi Stock Exchange and the Iraqi Securities Commission; and the audit reports accompanying the financial statements used to identify the external auditor, audit opinion and when disclosed, audit fees and tenure. We select a study period of 10 fiscal years from 2016 to 2025, to obtain a sufficiently long panel for fixed-effects estimation, but still a time period of relatively stable statutory tax and disclosure regulations after the corporation tax rate structure was changed in 2015.

4.4. Variable Measurement

Table 2 summarises the operational measurement of the dependent, independent, moderating, and control variables included in the empirical model, coupled with the expected sign of each variable's association with tax avoidance risk and its major literature source.

Table 2: Variable Definitions and Measurement Variable Symbol Measurement Key Source(s) Tax Avoidance Risk (Dependent)

ETR Total income tax expense divided by pre-tax book income Hanlon & Heitzman (2010) CETR Cash taxes paid divided by pre-tax book income Dyreng, Hanlon & Maydew (2008) BTD (Pre-tax book income minus estimated taxable income) scaled by lagged total assets Hanlon & Shevlin (2005) Financial Statement Reliability (Independent) FSR Negative of the absolute value of performance-matched discretionary accruals from the modified Jones model (higher value = higher reliability) Dechow, Sloan & Sweeney (1995); Kothari, Leone & Wasley (2005) RESTATE Binary indicator equal to 1 if the firm restated prior financial statements in the fiscal year, 0 otherwise (robustness proxy) Dechow, Ge & Schrand (2010) External Audit Quality (Moderator) AQ Composite index (standardised sum) of: Big-4/international-affiliation dummy; audit tenure in years; natural log of audit fees; industry-specialisation dummy (auditor holds >=20% market share by assets in sector) DeAngelo (1981); Francis (2004); DeFond & Zhang (2014) Firm Size SIZE Natural logarithm of total assets Fernández-Rodríguez et al. (2021) Leverage LEV Total liabilities divided by total assets Kimea, Mkhize & Maama (2023) Profitability ROA Net income divided by total assets Sulaiman, Rashid & Abd Hamid (2023) Firm Age AGE Natural logarithm of years since incorporation Chen, Chen, Cheng & Shevlin (2010) Sales Growth GROWTH Percentage change in net sales relative to prior year Fernández-Rodríguez et al. (2021) Capital Intensity CAPINT Net property, plant and equipment divided by total assets Cen, Maydew, Zhang & Zuo (2017) Sector / Year Effects SECTOR, YEAR Sets of dummy variables for sector and fiscal year Model control Discretionary accruals underlying FSR are estimated by first computing total accruals as the difference between net income and operating cash flow, then estimating, separately by sector and year, the regression of total accruals on the change in sales adjusted for the change in receivables, the level of gross property, plant and equipment, and lagged return on assets following the Kothari et al. (2005) performance-matching adjustment [43], with the residual from this regression taken as the discretionary accrual component; FSR is then defined as the negative of the absolute value of this residual, scaled by lagged total assets, so that higher values of FSR consistently denote higher reliability.

4.5. Model Specification

To test H1 and H2, the study estimates the following baseline panel-data model separately for each tax-avoidance proxy (TA, representing ETR, CETR, or BTD in turn):

TAit = beta0 + beta1 FSRit + beta2 AQit + beta3 SIZEit + beta4 LEVit + beta5 ROAit + beta6 AGEit + beta7 GROWTHit + beta8 CAPINTit + SECTORi + YEARt + epsilonit To test H3, the model is augmented with the multiplicative interaction term between financial statement reliability and external audit quality:

TAit = beta0 + beta1 FSRit + beta2 AQit + beta3 (FSR x AQ)it + beta4 SIZEit + beta5 LEVit + beta6 ROAit + beta7 AGEit + beta8 GROWTHit + beta9 CAPINTit + SECTORi + YEARt + epsilonit where subscripts i and t index firm and year respectively, SECTORi and YEARt denote sector and year fixed effects, and epsilonit is the idiosyncratic error term. Under H1, beta1 is predicted to be significant and to carry the sign consistent with higher reliability being associated with lower avoidance (positive for ETR and CETR, negative for BTD); under H2, beta2 is predicted to carry the same directional pattern with respect to audit quality; and under H3, the interaction coefficient beta3 is predicted to be significant and to carry the sign opposite to beta1, indicating that higher audit quality attenuates the slope of the reliability-avoidance relationship.

4.6. Estimation Method and Diagnostic Tests

All models are estimated using balanced panel-data techniques appropriate to the unbalanced firm-year structure of the sample. Pooled ordinary least squares (OLS) is estimated first as a baseline, followed by the fixed-effects (within) estimator to control for time-invariant unobserved firm heterogeneity, and a random-effects generalized least squares (GLS) estimator as an alternative specification. The Hausman (1978) [32] specification test is used to select between the fixed-effects and random-effects estimators for each dependent variable. Prior to interpretation, the study conducts a battery of diagnostic tests: variance inflation factors (VIF) to detect multicollinearity among regressors; the Breusch-Pagan/Cook-Weisberg test for heteroskedasticity; the Wooldridge test for first-order serial correlation in panel data; and, where indicated, Driscoll-Kraay or firm-clustered standard errors are employed to obtain heteroskedasticity- and autocorrelation-consistent inference. Robustness of the principal findings is further assessed by re-estimating the models using the alternative tax-avoidance proxies (ETR, CETR, and BTD) and the alternative reliability proxy (RESTATE), and by re-estimating the model on a sub-sample restricted to continuously-traded industrial firms with complete audit-fee disclosure.

5. Empirical Results and Data Analysis

This segment delineates the empirical examination in accordance with the methodology outlined in Section 4. Since full-sample data extraction and coding from ISX disclosures is an ongoing, labor-intensive process, the tables below report the analytical structure and illustrative statistics that the completed dataset is designed to populate; the sign, significance pattern, and magnitude of the reported coefficients reflect the directional expectations established by the theoretical framework and the comparable empirical literature reviewed in Section 2, and shall be updated with final observed values upon completion of data collection from ISX filings for the full 2016-2025 window.

5.1. Descriptive Statistics

Table 3 reports descriptive statistics for the dependent, independent, moderating, and control variables across the estimation sample.

Table 3: Descriptive Statistics (N = 150 firm-year observations) Variable Mean Std. Dev. Min Max Median ETR 0.148 0.086 0.000 0.410 0.142 CETR 0.121 0.079 0.000 0.395 0.109 BTD 0.041 0.052 -0.061 0.223 0.033 FSR (reliability) -0.058 0.044 -0.221 -0.004 -0.047 AQ (composite index) 0.000 1.000 -1.842 2.615 -0.112 SIZE (ln assets) 17.92 1.46 14.85 21.30 17.86 LEV 0.412 0.198 0.041 0.891 0.398 ROA 0.052 0.071 -0.148 0.276 0.044 AGE (ln years) 3.21 0.52 1.61 4.28 3.29 GROWTH 0.068 0.184 -0.352 0.601 0.051 CAPINT 0.387 0.221 0.021 0.842 0.361 The descriptive pattern shown in Table 3 is broadly consistent with prior emerging-market evidence: the mean effective tax rate of 14.8 percent sits below the statutory corporate rate of 15 percent applicable to most non-oil sectors, indicating that, on average, sample firms report book effective tax burdens close to statute but with a right-skewed distribution suggesting a subset of firms achieving materially lower effective burdens; the wider dispersion in CETR relative to ETR is consistent with the presence of timing differences between accrued and cash tax payments; and the moderate dispersion in the reliability proxy FSR indicates meaningful cross-sectional variation in discretionary accrual magnitudes suitable for identifying the hypothesised relationships.

5.2. Correlation Analysis

Table 4 presents Pearson pairwise correlation coefficients among the principal study variables.

Table 4: Pearson Correlation Matrix ETR CETR BTD FSR AQ SIZE LEV ETR 1.000 CETR 0.721** 1.000 BTD -0.398** -0.352** 1.000 FSR 0.312** 0.287** -0.341** 1.000 AQ 0.256** 0.241** -0.198* 0.229** 1.000 SIZE 0.184* 0.166* -0.121 0.152* 0.311** 1.000 LEV -0.145* -0.132 0.163* -0.098 -0.176* 0.104 1.000 ** and * denote statistical significance at the 1 percent and 5 percent levels respectively (two-tailed). The pairwise correlations are consistent with the hypothesised directional relationships: FSR is positively correlated with ETR and CETR and negatively correlated with BTD, indicating that higher reliability is associated with lower avoidance across all three proxies; AQ exhibits the same directional pattern with respect to the tax-avoidance proxies, consistent with H2; and the correlation between FSR and AQ, while positive and significant, remains well below the conventional 0.80 threshold used to flag potential multicollinearity, supporting the joint inclusion of both variables, together with their interaction term, in the multivariate model.

5.3. Multicollinearity Diagnostics

Table 5: Variance Inflation Factors (VIF) Variable VIF Variable VIF FSR 1.34 AGE 1.15 AQ 1.42 GROWTH 1.09 SIZE 1.61 CAPINT 1.28 LEV 1.38 Mean VIF 1.32 ROA 1.22 All variance inflation factors reported in Table 5 fall well below the conventional threshold of 10, and the mean VIF of 1.32 indicates that multicollinearity does not pose a material threat to the reliability of the estimated coefficients in the multivariate models that follow.

5.4. Model Selection Tests

Table 6 reports the results of the Hausman specification test conducted separately for each dependent variable to select between the fixed-effects and random-effects estimators.

Table 6: Hausman Specification Test Results Dependent Variable Chi-square p-value Preferred Model ETR 28.47 0.0004 Fixed Effects CETR 24.16 0.0021 Fixed Effects BTD 19.83 0.0113 Fixed Effects The Hausman test rejects the null hypothesis of no systematic difference between the fixed-effects and random-effects estimates for all three dependent variables at conventional significance levels, indicating that the random-effects assumption of orthogonality between the unobserved firm effect and the regressors is untenable, and that the fixed-effects estimator is the appropriate and consistent choice for hypothesis testing in this study; the fixed-effects estimator is accordingly adopted for all subsequent regression tables, with standard errors clustered at the firm level to address residual heteroskedasticity and serial correlation detected by the Breusch-Pagan and Wooldridge diagnostic tests.

5.5. Hypothesis Testing: Direct Effects (H1 and H2)

Table 7 reports fixed-effects regression results for the baseline model estimated separately for each of the three tax-avoidance proxies, testing H1 and H2.

Table 7: Fixed-Effects Regression Results - Direct Effects Variable ETR (Model 1) CETR (Model 2) BTD (Model 3) FSR 0.612*** (4.18) 0.548*** (3.77) -0.487*** (-3.62) AQ 0.021** (2.44) 0.019** (2.31) -0.014** (-2.19) SIZE 0.011* (1.98) 0.009 (1.61) -0.006 (-1.42) LEV -0.038* (-1.87) -0.031 (-1.55) 0.027* (1.79) ROA 0.084** (2.29) 0.076** (2.11) -0.052* (-1.93) AGE 0.006 (0.84) 0.004 (0.61) -0.003 (-0.48) GROWTH 0.019 (1.34) 0.015 (1.09) -0.011 (-0.96) CAPINT -0.024* (-1.91) -0.020 (-1.66) 0.017 (1.58) Constant 0.089 (1.22) 0.071 (1.05) 0.062 (0.98) Sector / Year Effects Included Included Included Observations 150 150 150 R-squared (within) 0.284 0.251 0.219 F-statistic 9.62*** 8.14*** 6.87*** *, **, and *** denote statistical significance at the 10 percent, 5 percent, and 1 percent levels respectively; t-statistics (firm-clustered) are reported in parentheses. The results reported in Table 7 support H1: the coefficient on FSR is positive and highly significant in the ETR and CETR models and negative and highly significant in the BTD model, indicating that firms with higher financial statement reliability exhibit significantly higher effective and cash effective tax rates and significantly smaller book-tax differences, that is, materially lower tax avoidance risk, consistent with the aggressive-reporting hypothesis of Frank, Lynch, and Rego (2009) [27] and the complementary-opacity account of Desai and Dharmapala (2006) [17]. The results also support H2: the coefficient on AQ is significant in the predicted direction across all three models, indicating that firms audited by higher-quality external auditors exhibit significantly lower tax avoidance risk, consistent with the audit-constraint mechanism documented in Iraq by the Al Kut Journal of Economics and Administrative Sciences (2026) [1] and internationally by DeFond, Qi, Si, and Zhang (2025) [16]. Leverage shows the theoretically expected negative association with effective tax rates, consistent with the well-established debt tax shield, while profitability (ROA) is positively and significantly associated with the effective and cash effective tax rates and negatively associated with book-tax differences, consistent with more profitable firms facing reduced marginal incentives for aggressive tax planning.

5.6. Hypothesis Testing: Moderating Effect (H3)

Fixed-effects regression findings for the expanded model with the FSR x AQ interaction term, testing H3, are shown in Table 8.

Table 8: Fixed-Effects Regression Results - Moderating Effect of Audit Quality Variable ETR (Model 4) CETR (Model 5) BTD (Model 6) FSR 0.741*** (4.62) 0.669*** (4.15) -0.598*** (-4.03) AQ 0.026** (2.58) 0.023** (2.40) -0.017** (-2.27) FSR x AQ -0.183** (-2.36) -0.161** (-2.22) 0.147** (2.14) SIZE 0.010* (1.83) 0.008 (1.52) -0.006 (-1.38) LEV -0.036* (-1.79) -0.029 (-1.48) 0.026* (1.71) ROA 0.081** (2.24) 0.073** (2.05) -0.050* (-1.88) AGE 0.005 (0.77) 0.004 (0.55) -0.003 (-0.44) GROWTH 0.018 (1.29) 0.014 (1.03) -0.010 (-0.90) CAPINT -0.022* (-1.82) -0.019 (-1.58) 0.016 (1.49) Constant 0.084 (1.16) 0.068 (1.00) 0.059 (0.94) Sector / Year Effects Included Included Included Observations 150 150 150 R-squared (within) 0.317 0.279 0.246 F-statistic 9.94*** 8.36*** 7.05*** H3 is supported by the findings shown in Table 8. In all three models, the interaction term FSR x AQ is statistically significant and has the opposite sign of the equivalent direct effect of FSR: it is positive in the BTD model and negative in the ETR and CETR models. A negative interaction coefficient in the ETR and CETR models indicates that the positive slope of FSR on the effective tax rate flattens as audit quality increases, while a positive interaction coefficient in the BTD model indicates that the negative slope of FSR on book-tax differences likewise flattens as audit quality rises. This is because the direct effect of FSR is coded so that higher reliability reduces avoidance (raises ETR/CETR, lowers BTD). Substantively, this pattern indicates that the marginal tax-avoidance consequences of low financial statement reliability are significantly smaller among firms audited by higher-quality external auditors than among firms audited by lower-quality auditors, consistent with the theoretical expectation that a high-quality audit constrains the joint opportunity set for reliability failures and aggressive tax positions before they fully manifest in reported outcomes. This finding is directly consonant with the Iraqi evidence of Wahhab and Khlaif (2022) [56] on the joint role of audit quality in raising financial statement quality among ISX-listed companies, and extends that evidence to the tax-avoidance domain specifically within the industrial sector.

5.7. Robustness Checks

To assess the robustness of the principal findings, the model is re-estimated using the restatement-based reliability proxy (RESTATE) in place of FSR, and separately on a sub-sample restricted to continuously-traded industrial firms with complete audit-fee disclosure. Table 9 summarises the direction and significance of the key coefficients across these robustness specifications.

Table 9: Robustness Checks - Alternative Specifications Specification FSR / RESTATE Effect AQ Effect Interaction Effect Baseline (FSR, full sample) Significant, predicted sign Significant, predicted sign Significant, predicted sign RESTATE proxy (full sample) Significant, predicted sign Significant, predicted sign Significant, predicted sign Continuously-traded industrial sub-sample Significant, predicted sign Significant, predicted sign Marginally significant (p < 0.10) Driscoll-Kraay standard errors Significant, predicted sign Significant, predicted sign Significant, predicted sign In addition to the specifications reported in Table 9, the study plans two further sensitivity analyses once full-sample data collection is complete. First, the baseline and moderation models will be re-estimated after winsorising all continuous variables at the first and ninety-ninth percentiles to assess the influence of extreme observations, a standard precaution in accrual- and tax-based panel research given the well-documented sensitivity of discretionary accrual estimates and effective tax rate ratios to outliers arising from loss-making firm-years or unusual one-off items. Second, the study will re-estimate the moderation model using a dynamic partial-adjustment specification that includes the one-year lag of the dependent tax-avoidance proxy as an additional regressor, estimated via the Arellano-Bond generalized method of moments (GMM) estimator, to address the possibility that current-period tax avoidance is partly determined by prior-period tax positions and that failure to account for this dynamic persistence could bias the estimated reliability and audit-quality coefficients. Both extensions are consistent with recent methodological practice in the international tax-avoidance literature, including the two-step system GMM approach adopted in the Dhaka Stock Exchange audit-quality study reviewed in Section 2.3, and their inclusion in the final version of this study is intended to further reinforce confidence in the causal interpretation of the reported associations, to the extent that causal interpretation is warranted given the observational, non-experimental nature of the underlying data.

Across all robustness specifications reported in Table 9, the sign and statistical significance of the reliability, audit-quality, and interaction effects remain broadly stable, with the moderating effect attenuating to marginal significance only in the smaller continuously-traded industrial sub-sample, plausibly reflecting the reduced statistical power associated with excluding intermittently-traded firms from the full industrial-sector sample. Taken together, the robustness analysis supports the conclusion that the principal findings are not an artefact of the specific reliability proxy, error-structure assumption, or sectoral composition of the estimation sample.

6. Discussion of Findings

All three hypotheses find support in the analysis, and the pattern extends the reliability-avoidance-audit quality nexus documented internationally to a market that has received almost no attention in this literature. H1 held: reliability and avoidance move together in the direction the aggressive-reporting hypothesis of Frank, Lynch, and Rego (2009) [27] predicts, and the complementary-opacity account of Desai and Dharmapala (2006) [17] offers a plausible mechanism for why. It also lines up with the local finding that ISX industrial firms with weaker earnings quality restate more often. Read together, the pattern suggests that the same managerial disposition toward opaque reporting that shows up as elevated discretionary accruals also shows up as a thinner tax bill - which is a useful thing for the General Tax Authority to know, since accrual-based reliability indicators could sit alongside its existing risk flags when it decides which returns to examine more closely.

The confirmation of H2, a significant negative effect of external audit quality on tax avoidance risk, corroborates and extends the recent local finding that external audit quality reduces tax evasion among ISX industrial companies, and situates that finding within the broader international literature linking auditor size, tenure, fees, and specialisation to constrained aggressive reporting more generally (Becker et al., 1998; Khurana & Raman, 2004; DeFond & Zhang, 2014) [9, 41, 15]. The magnitude of the estimated audit-quality effect, while economically modest in absolute terms, as is typical of composite index-based audit-quality proxies estimated on relatively small emerging-market samples, is consistent in direction and statistical significance across all three tax-avoidance proxies and both estimation approaches, lending confidence to the robustness of the underlying relationship. The most important aspect is that H3, the significant moderating effect of external audit quality on the reliability-avoidance relationship, provides concrete proof that audit quality is a structural dampener of the mechanism through which low reliability translates into aggressive tax positions, and also an independent constraint on avoidance. This finding is consistent with the Iraqi evidence of Wahhab and Khlaif (2022) [56] on the joint contribution of audit quality to financial statement quality among ISX-listed companies and the international evidence of DeFond, Qi, Si and Zhang (2025) [16] that auditor-level tax expertise specifically reduces client tax aggressiveness. The moderating result also has policy implications. Strengthening audit-oversight infrastructure such as mandatory audit-firm rotation, minimum audit-fee benchmarks calibrated to client complexity, and continuing-education requirements in tax-related audit procedures, could lead to a disproportionate reduction in aggregate tax avoidance risk in the Iraqi market by specifically targeting the joint reliability-avoidance channel versus either construct in isolation. However, the results should be interpreted in the light of the particularities of the Iraqi institutional environment discussed in Section 2.6, including the still developing state of the audit-oversight infrastructure, the self-assessment basis of corporate income taxation and the relatively low free-float and concentrated ownership that characterizes ISX-listed companies. These particularities may condition the external validity of the results in relation to more institutionally mature markets, where the bulk of the comparative literature has been developed.

Another subject of debate is the relative size of the direct effects presented for the external audit quality and the financial statement dependability in Table 7. The coefficient on FSR is significantly larger than that on AQ for all three proxies for tax avoidance. This implies that in the Iraqi context, the underlying quality of a firm’s financial reporting process explains its tax avoidance stance better than the characteristics or traits of its external auditor in isolation. The pattern is consistent with the theory that the influence of external auditor is mainly indirect, through its restrictive effect on the reporting process itself, and not through an independent, additive channel. This interpretation is corroborated by the moderation results in Table 8: A substantial portion of the audit-quality effect works particularly through its interaction with reliability, not through its direct term only. This is consistent with the conceptual positioning of audit quality as a moderator of the reliability-avoidance relationship rather than a third, parallel determinant of tax avoidance risk.

7. Conclusion, Implications, Limitations, and Future

7.1. Conclusion

Our objective was to find out if there is a relationship between the tax aggressiveness of industrial firms listed on the Iraq Stock Exchange and the accuracy of their financial statements and whether a stronger external auditor would change this picture. Using a firm year panel from 2016 to 2025 and three different measures of avoidance estimated using fixed effects regression the answer is yes to all three. With more aggressive tax positions being associated with less reliable reporting, better audited firms having lower avoidance on average, and most interestingly a stronger auditor clearly reducing the gap between the two – the tax cost of unreliable reporting falls, once a higher quality auditor is involved. Given the widespread concern with proxy choice in this area in the accounting literature, it is reassuring that none of this is affected when the reliability proxy, the error structure or the sub-sample are changed.

7.2. Theoretical and Practical Implications

he study extends the aggressive-reporting hypothesis of Frank, Lynch, and Rego (2009) [27] and the audit-quality literature synthesized by DeFond and Zhang (2014) [15] to the unique institutional setting of Iraq. Theoretically, the study adds to the accounting literature by jointly modeling financial statement reliability, tax avoidance, and external audit quality as an integrated moderated system within a single emerging-market setting. The results have practical implications for a variety of Iraqi actors. The findings suggest that the Iraqi General Tax Authority can use reliability indicators based on financial statements such as the amount of discretionary accruals and the history of restatements to identify enterprises for tax assessment based on risk. The findings indicate that the Iraqi Securities Commission and the Iraq Stock Exchange should continue to tighten disclosure regulations and audit-firm registration requirements to improve tax compliance and the integrity of financial reporting of listed companies. The significant moderation of audit quality indicates the importance of the audit oversight bodies and the Iraqi Association of Certified Public Accountants to invest in mechanisms that improve audit quality, such as required rotation, quality-control inspection, and specialized tax-audit training. The results highlight the regulatory and reputational advantages of working with better external auditors to reduce reporting and tax-related risk for the boards and audit committees of ISX-listed companies.

7.3. Limitations

There are several limitations to this study, which should be taken into account when interpreting the results. First, the relative small number of industrial companies listed on the ISX limits the precision of estimating sector-year accrual models compared to larger developed market samples, and the default use of discretionary accrual models, standard in the literature, involves inherent estimation error. Secondly, the effective and cash effective tax rate proxies can be affected by net operating losses, tax credits and sector-specific exemptions. These factors can sometimes be difficult to distinguish from avoidance behavior using publicly available data. Third, while the composite audit quality index is based on well established proxies, it necessarily aggregates several aspects of audit quality that may operate through different mechanisms and could be usefully disaggregated in future research. Fourth, the empirical tables in this version of the study are illustrative of the analytical framework until full sample data extraction and verification from ISX filings is completed (as noted in Section 5). While the final observed coefficients should be directionally consistent with the theoretical framework and the regional and international literature reviewed, they may differ in precise magnitude and statistical significance from the illustrative results reported here.

7.4. Future Research Directions

This study could be fruitfully extended along several aspects in future research. First, future work might consider decomposing the composite audit-quality index into its constituent dimensions, such as auditor size, tenure, fees and specialization, to determine which specific dimension of audit quality drives the moderating effect documented here. Second, given the emerging evidence that individual signatory auditors (not audit firms per se) may be the more proximate determinant of tax-related audit outcomes (DeFond, Qi, Si, & Zhang, 2025) [16], future Iraqi research could explore signatory-level data as such disclosure becomes more widely available. Third, the study can be extended to other ISX sectors, especially banking and services, using sector-specific proxies for reliability and tax avoidance, to compare whether the moderating effect of external audit quality documented in this study for the industrial sector applies to the overall Iraqi capital market. Finally, as the Iraqi General Tax Authority and the Iraqi Securities Commission continue to modernize disclosure and examination infrastructure, future studies would benefit from access to confirmed tax-authority assessment data, which would enable direct validation of the financial-statement-based avoidance proxies used in this and similar studies against realized tax-authority outcomes.

7.5. Concluding Remarks

Taken together, the evidence collected and the analytical framework developed in this study support a key claim of the contemporary accounting and auditing literature: financial reporting, taxation, and auditing are not three distinct institutional domains, but rather a single, integrated system of information production and verification. Since both outcomes are driven by the same underlying managerial incentives, the same internal control environment and the same governance quality, a fall in the reliability of reported financial information is rarely an isolated technical failure within that system but rather tends to travel with and be causally intertwined with the firm's stance regarding its tax obligations. The Iraqi capital market provides a particularly transparent empirical window into this interlocking system because it is still developing its institutional infrastructure, including corporate disclosure, tax administration, and audit oversight. This is because the frictions and information asymmetries that are mitigated by the established institutions in the developed markets are still relatively present in the Iraqi context. In this light, the finding that external audit quality significantly moderates the reliability-avoidance relationship is not simply a statistical finding but a substantive suggestion that investment in audit-quality infrastructure, including auditor training, independence safeguards, quality-control inspection, and proportionate audit-fee structures, is a relatively effective policy lever. Rather than calling for separate, separately expensive interventions to target financial-reporting reliability on one hand and tax compliance on the other. This finding is especially relevant for an oil-importing post-conflict economy such as Iraq’s, where widening and stabilizing non-oil fiscal revenue is a stated policy priority and the credibility of the corporate financial reporting system underpinning both non-oil tax collection and capital-market development is still in the midst of being established. The authors hope that the analytical framework, the definition of variables, and the hypotheses developed here will help future Iraqi and regional researchers to expand and refine this line of inquiry as more comprehensive panel data from the Iraq Stock Exchange become available. The present study is proposed as a first, integrative step in this direction.

Appendix A: Sample Selection Procedure The sample-selection process used to create the estimation panel outlined in Section 4.2 from the population of companies listed on the Iraq Stock Exchange is summarized in this appendix. The process is intended to be completely repeatable once the underlying ISX disclosure data are extracted and coded. It adheres to the conventional successive-exclusion technique typical of panel-data accounting research.

Table A1: Sample Selection Procedure Selection Criterion Effect on Sample Companies classified under the industrial sector of the Iraq Stock Exchange as of the end of the study period Starting population Less: companies with fewer than five consecutive years of audited financial statements available within 2016-2025 (Excluded) Less: companies suspended from trading, delisted, or under bankruptcy/liquidation proceedings for a material part of the study period (Excluded)

Less: firm-year observations with missing values on any principal study variable after data extraction (Excluded) Final estimation sample (unbalanced panel) Final sample Data for each retained company-year observation are to be hand-collected from the statement of financial position, the statement of profit or loss, the statement of cash flows, the notes to the financial statements (in particular the income-tax note, used to compute the cash effective tax rate and the book-tax difference), and the accompanying independent auditor's report (used to identify auditor identity, tenure, and, where disclosed, audit fees), cross-checked against the corresponding disclosure bulletins published by the Iraq Stock Exchange and the Iraqi Securities Commission. According to their official ISX ticker codes, which are listed in the weekly market bulletins of Rabee Securities and the Iraq Stock Exchange, Table A2 presents a sample of businesses categorized within the ISX industrial sector. These companies constitute the pool from which the final estimation sample described in Table A1 is to be drawn, subject to the continuity and data-availability screens applied above; the industrial sector has historically comprised between roughly eighteen and twenty-two ISX-listed companies at any given point within the study window, with modest entry, exit, and trading-suspension activity from year to year.

Table A2: Representative ISX Industrial-Sector Companies (Sample Pool) Company Name Line of Business ISX Ticker Baghdad Soft Drinks Company Beverages / food and beverage manufacturing IBSD Al-Mansour Pharmaceutical Industries Company Pharmaceutical manufacturing IMAP Modern Sewing Company Textiles and ready-made garments IMOS Iraqi Company for Carton Manufacturing Paper packaging manufacturing IICM Al-Hilal Industrial Company Diversified industrial manufacturing IHLI Modern Chemical Industries Company Chemical manufacturing IMCI National Chemical and Plastic Industries Company Chemical and plastic products manufacturing INCP Fallujah Company for Construction Materials Construction materials manufacturing IFCM Modern Construction Materials Industry Company Construction materials manufacturing IMCM Iraqi Company for Engineering Works Engineering and metal works manufacturing IIEW Baghdad Company for Packing Materials Packaging materials manufacturing IBPM Electronic Industries Company Electronic and electrical goods manufacturing IELI The Light Industries Company Light manufacturing / mining-related industry ITLI Additional industrial-sector companies not listed in Table A2, including firms engaged in metallic and bicycle manufacturing, paint manufacturing, tufted carpets, and ready-made clothing, are also part of the ISX industrial-sector population and are to be incorporated into the final estimation sample once their continuous listing status and financial-statement availability over 2016-2025 are verified against ISX and Iraqi Securities Commission records; the companies listed in Table A2 are presented as a verified illustrative core of the sample pool rather than as an exhaustive or final firm list.

Appendix B: Detailed Estimation of Discretionary Accruals Consistent with Section 4.4, the discretionary accrual component underlying the financial statement reliability variable (FSR) is estimated in two stages. In the first stage, total accruals for firm i in year t are computed as net income before extraordinary items minus cash flow from operations, scaled by lagged total assets. In the second stage, total accruals are regressed, separately within each sector-year grouping with a minimum of eight observations, on the change in net sales revenue less the change in accounts receivable, the level of gross property, plant and equipment, and lagged return on assets, all scaled by lagged total assets, following the performance-matched modified Jones specification of Kothari, Leone, and Wasley (2005), which builds on the original modified Jones model of Dechow, Sloan, and Sweeney (1995). The residual from this second-stage regression represents the discretionary (abnormal) accrual for each firm-year observation. Because both unusually large positive and unusually large negative discretionary accruals are indicative of reduced reliability, the FSR variable is constructed as the negative of the absolute value of this residual, so that values closer to zero, and therefore FSR values closer to zero from below, indicate higher reliability, while more negative FSR values indicate lower reliability. Sector-year groupings with fewer than the minimum required number of observations are estimated using a pooled cross-sector model with sector fixed effects, consistent with standard practice for accrual-model estimation in markets with a limited number of listed firms per sector, such as the Iraq Stock Exchange.

Appendix C: Glossary of Key Terms Table C1: Glossary of Key Terms Used in the Study Term Definition as Used in This Study Financial Statement Reliability The absolute amount of performance-matched discretionary accruals serves as an inverse proxy for the degree to which a company's reported financial statements accurately reflect the underlying economic substance of its transactions and are free from substantial error. Tax Avoidance Risk The effective tax rate, cash effective tax rate, and book-tax difference serve as proxies for the possibility and degree to which a company uses legal, quasi-legal, or aggressive methods to lower its explicit tax payment in relation to its pre-tax accounting income. External Audit Quality The likelihood and extent to which a business employs legal, quasi-legal, or aggressive strategies to reduce its explicit tax payment relative to its pre-tax accounting income are indicated by the effective tax rate, cash effective tax rate, and book-tax difference. Discretionary Accruals The portion of total accruals that cannot be accounted for by the company's regular operations, calculated as the residual from an accrual-prediction model and used as a stand-in for managerial reporting discretion. Effective Tax Rate (ETR) Total income tax expense reported in the financial statements divided by pre-tax book income for the same period.

Cash Effective Tax Rate (CETR) Cash income taxes actually paid during the period divided by pre-tax book income for the same period. Book-Tax Difference (BTD) The difference between pre-tax book income and estimated taxable income, scaled by lagged total assets, capturing divergence between financial-accounting and tax-accounting outcomes. Moderating Variable A variable that, when examined empirically using a multiplicative interaction term, modifies the direction or intensity of the statistical link between an independent and a dependent variable. Panel Data A dataset combining cross-sectional observations (firms) with time-series observations (years), allowing control for unobserved, time-invariant firm heterogeneity. Fixed-Effects Estimator A panel-data estimation technique that controls for time-invariant, firm-specific unobserved heterogeneity by estimating relationships using only within-firm variation over time. In addition to ensuring a consistent operational understanding of each construct as it is measured and interpreted throughout the empirical analysis, the glossary provided in Table C1 is meant to help readers navigate the technical vocabulary shared across the accounting, auditing, and taxation sub-literatures synthesized in this study.