Discretionary Provisions as Accounting-Based Capital Buffers: Evidence from Modified Audit Opinions in the Turkish Banking Sector

Abstract

This study investigates the structural characteristics, annual trends, and qualitative rationales underlying modified audit opinions issued within the independent audit reports of Turkish banks between 2022 and 2025. Utilizing a formal content analysis of 248 bank-year independent audit reports based on a complete census approach, the study evaluates the incidence and persistence of modifications under ISA 705 and IAS 37. The empirical findings reveal a recurring pattern of qualified opinions associated with unmandated discretionary provisioning, with all qualified opinions in the sample issued by Big Four audit firms and recurring across multiple reporting periods. Interpreted through institutional decoupling and signaling perspectives, the findings are consistent with the possibility that these qualifications may reflect interactions between accounting requirements, regulatory conditions, and provisioning practices rather than providing direct evidence of financial distress or managerial intent. The study contributes to the auditing and banking literature by examining the accounting bases and longitudinal patterns underlying modified audit opinions and by discussing their implications within an institutional framework relevant to emerging economies.

1. Introduction

The conceptual baseline of contemporary corporate governance dictates that high-quality financial reporting is an essential prerequisite for the efficient allocation of capital across global financial markets (Spence, 1973). Within any market economy, financial statements function as the structural bridge that mitigates information asymmetry between an organization’s internal management and external stakeholders, including institutional investors, creditors, depositors, and regulatory oversight boards. The reliability, comparability, and transparency of these financial records are heavily dependent upon two complementary institutional structures: first, the strict adherence to a globally accepted financial reporting framework such as the International Financial Reporting Standards (IFRS) or their localized adaptations (TFRS), and second, the execution of an objective, independent external audit process.

The independent external audit report represents the ultimate formal output of this institutional verification process. It provides external financial statement users with a high level of professional assurance regarding whether an entity’s financial position is presented fairly, in all material respects, in accordance with the applicable accounting principles. When an auditor completes their investigation and obtains sufficient, appropriate audit evidence to conclude that the underlying financial statements are free from material misstatements, an unmodified or unqualified opinion is issued. However, the presence of unresolved material misstatements or severe scope limitations regarding data collection shifts the auditor’s regulatory mandate. Under the International Standard on Auditing (ISA) 705 (KGK, 2025a), Modifications to the Opinion in the Independent Auditor’s Report, external auditors are legally and professionally required to deviate from the standard unqualified report and issue a modified audit opinion (MAO). These modifications are stratified into qualified opinions, adverse opinions, and disclaimers of opinion, depending on the severity and pervasiveness of the financial anomaly.

While the determinants and financial market consequences of modified audit opinions have been extensively documented across developed Western capital markets, the institutional realities of emerging economies present a vastly different structural matrix (Habib, 2013). Developed economies benefit from highly stabilized macro-financial environments, predictable regulatory enforceability, and mature institutional mechanisms. In stark contrast, emerging market ecosystems are inherently volatile, frequently characterized by structural macroeconomic shocks, high inflation, rapid currency devaluations, and shifting macroprudential policy interventions. In these volatile settings, potential tensions may arise between the requirements of global accounting frameworks such as IAS 37 and IFRS 9 and the broader risk-management and prudential considerations faced by financial institutions. This friction is nowhere near as visible or consequential as within the banking industry. The banking sector represents the systemic engine of an emerging market’s economy. Due to their unique asset-liability structures, exceptionally high financial leverage, central role in maturity transformation, and complete reliance on public and depositor trust, banks are highly sensitive to broader economic fluctuations. Consequently, bank financial statements are subjected to rigorous double-regulatory oversight, monitored simultaneously by international accounting bodies and domestic banking regulators, such as the Banking Regulation and Supervision Agency (BRSA) in Türkiye.

When an emerging economy experiences an extended period of macroeconomic instability, banks may face increased pressure to manage financial reporting, risk, and capital considerations under changing economic and regulatory conditions. While such practices may be considered relevant from a prudential or risk-management perspective, certain accounting treatments may not be consistent with the recognition requirements of applicable financial reporting standards. Specifically, within the banking sector, while traditional credit risk provisions are governed under the expected credit loss (ECL) model of IFRS 9, discretionary or free provisions created without an underlying, past obligating credit event fall directly under the recognition parameters of IAS 37. Where such provisions do not satisfy the recognition criteria of IAS 37, auditors may identify the related accounting treatment as a basis for a qualified opinion in accordance with ISA 705.

This study explores this structural intersection by executing a thorough, qualitative, and systematic content analysis of modified audit opinions issued within the independent audit reports of banks operating in Türkiye. The empirical window of this investigation spans four critical fiscal years ending 31 December 2022, 2023, 2024, and 2025. By analyzing a comprehensive census dataset of 248 annual independent audit reports obtained directly from the official regulatory public disclosures of the BRSA, this paper identifies the accounting inconsistencies/non-compliances documented in the reports and examines the structural and longitudinal patterns underlying external audit qualifications in a volatile emerging economy.

Türkiye represents a compelling institutional setting for this line of inquiry. Throughout the 2022–2025 sample window, the Turkish economy experienced significant macroeconomic volatility, characterized by multi-year high inflation, continuous adjustments to minimum reserve requirements, complex macroprudential credit growth caps, and notable fluctuations in the local currency. Within this environment, the audit reports examined in this study document the recognition of unmandated “discretionary provisions” by several Turkish banking institutions.

This study provides three distinct contributions to the existing accounting and auditing literature. First, by examining the 2022–2025 period in an emerging market setting, it provides descriptive evidence on the distribution and temporal variation of modified audit opinions within the Turkish banking sector (addressing RQ1 and RQ2). Second, it bridges the gap between accounting standard compliance and auditing standards by unpacking the specific non-compliance mechanisms—particularly regarding discretionary provisions under IAS 37—that trigger external audit modifications under ISA 705 (addressing RQ3). Third, by analyzing longitudinal persistence within specific institutions, it contributes to the discussion of institutional decoupling in emerging banking sectors by documenting how qualified opinions and related provisioning practices recur or change across reporting periods (addressing RQ4).

2. Institutional Context and the Turkish Banking Sector (2022–2025)

To interpret the empirical outcomes of this study, it is necessary to examine the macroprudential, economic, and regulatory conditions that shaped the Turkish banking infrastructure during the 2022–2025 period. Financial reporting can be interpreted within the structural and institutional conditions of the market environment in which firms operate.

Following global economic shocks, the Turkish macro-financial landscape experienced significant adjustments. The period between 2022 and 2025 was characterized by elevated domestic inflation, which placed considerable pressure on corporate asset quality, capital adequacy ratios (CAR), and the valuation of financial instruments across the banking sector. During these years, the Central Bank of the Republic of Türkiye (CBRT) and the BRSA introduced a wide array of macroprudential regulations designed to maintain financial stability, curb credit growth, and encourage the holding of local currency-denominated assets. These policy interventions included fluctuating reserve requirement ratios, security maintenance regulations linked to deposit compositions, and strict asset ratio formulas. For bank executives, managing a banking institution under these conditions required navigating a complex regulatory environment. The rapid expansion of credit volumes, partly reflecting inflationary nominal growth, occurred alongside concerns regarding asset quality, non-performing loans (NPLs), and bank capital buffers. Although official NPL ratios remained low due to various regulatory forbearance measures and restructuring flexibilities, bank management remained highly conservative regarding potential credit risk defaults.

Consequently, the period provides a context for examining whether highly profitable periods—partially associated with high net interest margins and inflation-linked bond returns such as CPI-linkers—coincided with the recognition of discretionary provisions. This accounting choice was heavily exacerbated by the specific hyperinflationary environment of the 2022–2025 period and the asymmetrical regulatory implementation of inflation accounting under TAS 29 in Türkiye. While non-financial corporates transitioned to inflation-adjusted reporting to reflect real economic values, banking institutions were explicitly deferred from applying TAS 29 (converged with IAS 29) (KGK, 2023a) by the regulatory authorities. During this period, banks continued to report under the applicable regulatory accounting framework, while the non-application of TAS 29 created a different reporting environment from that of non-financial corporates. Transferring these artificial, nominal profits directly to equity accounts as voluntary reserves would not only trigger substantial corporate tax liabilities but would also fail to shield the income statement from potential future asset-quality reversals when macroprudential tightening eventually took place. The recognition of such amounts as current-period provision expenses under other liabilities may have reduced reported current-period income and affected the presentation of profitability and equity. However, the audit-report evidence does not establish that these accounting treatments were adopted for tax-mitigation, earnings-smoothing, or other specific managerial purposes.

However, the accounting treatment identified in the audit reports created a relevant accounting issue concerning the recognition and presentation of these amounts. Instead of allocating these funds within equity accounts as restricted retained earnings or voluntary capital reserves—which would require clear transparency and would not reduce current-year net profitability—banks chose to recognize them directly on their income statements as current-period provision expenses. This led to the recording of a corresponding liability account under Other Provisions or Expected Credit Losses on the balance sheet. This accounting treatment reduced reported current-period net income in the affected financial statements and altered the presentation of related liabilities and equity. The audit-report evidence, however, does not establish that these effects were the intended objectives of the accounting treatment. This accounting treatment was identified by external auditors as inconsistent with the applicable recognition requirements of IAS 37 and, in some cases, IFRS 9.

3. Theoretical Framework and Literature Review

3.1. Modified Audit Opinions Under the ISA 705 Framework

ISA 705, Modifications to the Opinion in the Independent Auditor’s Report, establishes the professional criteria for determining when an auditor should issue an opinion other than an unmodified opinion. The standard classifies modified opinions according to two primary dimensions: the nature of the identified matter (material misstatement or inability to obtain sufficient appropriate audit evidence) and the auditor’s assessment of whether its effects are pervasive. A qualified opinion is issued when identified misstatements, individually or collectively, are material but not pervasive to the financial statements. More severe outcomes include adverse opinions, where misstatements are both material and pervasive, and disclaimers of opinion, where insufficient audit evidence prevents the auditor from forming an opinion, and the possible effects may be pervasive.

The prior literature generally considers modified audit opinions as negative signals associated with financial reporting risk, financial distress, or uncertainty regarding the reliability of reported information (Habib, 2013; Ireland, 2003). However, emerging-market evidence suggests that audit qualifications may also reflect accounting choices, regulatory pressures, and institutional factors rather than solely indicators of corporate failure. This distinction provides the theoretical foundation for examining modified audit opinions within the Turkish banking sector.

3.2. Critical Literature Review and Synthesis

To construct a robust theoretical foundation, the existing literature can be synthesized across three closely related streams: modified audit opinions and audit quality, banking-sector provisioning discretion, and the interaction between accounting standards and institutional conditions in emerging markets. Bringing these streams together helps clarify what prior research has established and where evidence remains limited, particularly regarding discretionary provisioning and audit qualifications in the Turkish banking sector.

First, auditing research has traditionally examined modified audit opinions (MAOs) as signals associated with financial distress, reporting risk, information asymmetry, and concerns regarding financial statement reliability (Dopuch et al., 1986; DeFond & Zhang, 2014; Habib, 2013). As auditing research expanded into emerging economies, however, scholars increasingly showed that audit qualifications may also be associated with accounting choices, asset valuation difficulties, and other reporting issues rather than solely with imminent insolvency (Gajevszky, 2014; Moalla, 2017). In the Turkish context, prior empirical studies have examined audit modifications primarily in non-financial companies. Uzay and Tanç (2010) documented the relationship between the adoption of international financial reporting standards and the quality and disclosure characteristics of Turkish independent audit reports, while Akdoğan et al. (2015) and Ocak (2016) identified issues involving long-term investments, inventory valuation, and discretionary accruals among factors associated with audit qualifications. These studies provide an important foundation for understanding audit modifications in Türkiye, but their samples are largely concentrated in industrial, manufacturing, real estate, and other non-financial sectors. Consequently, comparatively less is known about how modified audit opinions arise within highly regulated banking institutions facing simultaneous accounting and macroprudential pressures.

A second stream of literature focuses on accounting discretion in the banking sector. Extensive research shows that loan-loss provisions and other discretionary accounting choices may be associated with income smoothing, cyclical risk management, capital management, and regulatory considerations (Beatty & Liao, 2014; Bushman & Williams, 2012). Cross-country evidence further indicates that the use of provisioning discretion is shaped by differences in legal, regulatory, and institutional environments, with regulatory conditions potentially affecting the extent and timing of discretionary provisioning (Fonseca & González, 2008). More broadly, research on bank provisioning emphasizes the interaction between accounting rules, prudential regulation, institutional conditions, and managerial discretion. This study indicates that provisioning is not purely a technical accounting process but may also reflect broader financial and regulatory considerations.

The literature on international financial reporting provides a further dimension to this discussion. The global convergence toward International Financial Reporting Standards (IFRS) has been associated with improvements in financial statement comparability, transparency, and market discipline across jurisdictions (Barth et al., 2008). At the same time, the implementation of principles-based standards in emerging-market environments may interact with differences in institutional structures, enforcement mechanisms, and regulatory objectives (Ball et al., 2003; Holthausen, 2009). This is particularly relevant to banking, where accounting requirements coexist with prudential regulation and where provisioning decisions may affect both reported profitability and regulatory capital. Research on discretionary loan-loss provisions likewise indicates that accounting standards and regulatory mechanisms can influence the incentives and scope for discretionary provisioning, particularly in emerging-market settings (e.g., Beatty & Liao, 2014; Fonseca & González, 2008).

Despite these established findings, the three literature streams remain only partially connected. The first gap is sectoral. Existing Turkish research on audit modifications has paid considerably less attention to banks, whose distinctive asset-liability structure, regulatory environment, and systemic importance differentiate them from non-financial firms. The second gap is standard-specific. Banking research has extensively examined expected credit losses and loan-loss provisioning, including the role of IFRS 9, but comparatively less attention has been given to discretionary provisions considered under IAS 37 and to the way such provisions are reflected in modified audit opinions. The third gap is macro-institutional. Existing research provides limited evidence on the relationship between prolonged inflation, changing macroprudential conditions, inflation-accounting requirements for financial institutions, and the application of ISA 705.

These gaps define the specific contribution of the present study. Rather than assuming that discretionary provisions necessarily represent earnings management or managerial opportunism, the study examines how such provisions are described in independent audit reports and how their persistence and subsequent reversal are associated with qualified audit opinions. The empirical setting therefore connects the established literature on banking discretion and audit modifications with the institutional conditions of an inflationary emerging market. In this respect, the study considers whether the observed coexistence of formal financial reporting requirements, prudential considerations, and discretionary provisioning practices can be interpreted through the institutional perspective developed in the subsequent theoretical discussion.

3.3. Institutional Theory and the Dynamics of Decoupling

To move beyond conventional financial distress explanations and examine why qualified audit opinions may persist across reporting periods in the banking sector, this study draws on the institutional theory (DiMaggio & Powell, 1983). The institutional theory argues that organizations operate within competing systems of expectations, including regulatory requirements, professional norms, and legitimacy pressures. Consequently, organizational practices may reflect not only technical efficiency concerns but also responses to potentially conflicting institutional demands.

In emerging-market banking environments, institutions may face simultaneous pressures from global accounting frameworks and domestic prudential objectives. Global pressures encourage compliance with IFRS and Basel III principles to maintain transparency and confidence in financial reporting, whereas domestic regulatory frameworks emphasize financial stability, capital preservation, and resilience during periods of economic uncertainty. In Türkiye, these prudential objectives are also reflected in the regulatory framework and supervisory practices of the Banking Regulation and Supervision Agency (BRSA), particularly in relation to capital adequacy, provisioning, and financial resilience. When institutional expectations conflict, organizations may engage in forms of decoupling (Meyer & Rowan, 1977), whereby formal structures coexist with internal practices shaped by different institutional priorities.

In this study, the persistence of discretionary provisioning practices is interpreted as potentially consistent with a form of institutional decoupling. Banks maintain formal financial reporting and external assurance structures under IFRS-based frameworks, while the continued occurrence of discretionary provisioning practices may reflect considerations that are not fully aligned with the recognition requirements of IAS 37. This interpretation is based on the observed coexistence of qualified audit opinions and discretionary provisioning practices across multiple reporting periods and does not establish specific managerial intentions or definitive institutional conflict.

3.4. Signaling Theory and the Potential Interpretation of Audit Modifications

To complement the institutional perspective, this study incorporates the signaling theory (Spence, 1973). The signaling theory focuses on information asymmetry between corporate insiders and external users of financial information. Within this framework, corporate actions, financial information, and independent audit opinions may serve as signals to external stakeholders. An unmodified audit report is generally associated with confidence in the reliability of financial reporting, whereas a modified audit opinion may signal reporting uncertainty, accounting issues, or other matters requiring attention.

However, the informational meaning of an audit modification may depend on the underlying reason for the qualification. In the context examined in this study, the identified qualified opinions were frequently associated with discretionary provisions that did not meet the applicable recognition requirements of IAS 37. Such provisions differ conceptually from qualifications associated with credit-loss recognition under the IFRS 9 expected credit loss framework or with financial deterioration and the understatement of liabilities. This distinction raises the possibility that external users may interpret qualifications related to discretionary over-provisioning differently from qualifications associated with financial deterioration or material reporting deficiencies.

A potential signaling mechanism can therefore be represented as follows:

Discretionary Provisions → Lower Reported Income → Potential Interpretation as Conservative Reporting.

Under this theoretical perspective, a qualified opinion associated with discretionary over-provisioning may convey information relevant to the interpretation of conservative reporting practices, rather than necessarily indicating immediate financial weakness. The persistence of such qualifications across multiple reporting periods provides a basis for considering whether the informational implications of a qualification may differ according to its underlying accounting cause.

Importantly, this interpretation is proposed as a theoretical implication of the empirical findings rather than as evidence of an observed market response. The present study does not examine stock-price reactions, credit ratings, lending costs, cost of capital, analyst assessments, or stakeholder perceptions. Therefore, no conclusion is drawn regarding whether qualified opinions related to discretionary provisions actually produce neutral or positive market effects.

This theoretical possibility provides one potential interpretation of the observed persistence, while its validity remains subject to future empirical testing. Future research could examine whether investors, rating agencies, lenders, or other external stakeholders interpret different types of modified audit opinions differently depending on the accounting basis underlying the modification.

4. Research Methodology

4.1. Research Objective and Specific Questions

The objective of this study is to investigate the structural characteristics, annual trends, and qualitative rationales underlying the modified audit opinions issued within the independent external audit reports of banks operating in Türkiye. The study aims to look beyond simple quantitative frequencies to examine the accounting issues and recurring patterns of standard inconsistency/non-compliance identified in bank audit reports during a period of macroeconomic stress.

The sample period of this study is deliberately restricted to the 2022–2025 timeframe. While data from earlier periods are accessible, the pre-2022 operational environment in Türkiye was characterized by distinct monetary policies and lower inflationary pressures, during which the structural friction between IAS 37 recognition rules and unmandated provisioning did not manifest with systemic intensity. In contrast, the 2022–2025 window represents a period characterized by persistent inflation, changing monetary conditions, and evolving macroprudential regulations, providing a relevant institutional context for examining patterns in audit modifications under ISA 705. To guide this inquiry, four research questions were formulated:

• RQ1: What specific types of modified audit opinions under ISA 705 were issued in the independent audit reports of Turkish banks during the 2022–2025 period?
• RQ2: How were these modified opinions distributed chronologically across the four-year analytical window?
• RQ3: What were the primary accounting standard inconsistency/non-compliance and regulatory inconsistencies that led external auditors to modify their professional opinions?
• RQ4: Did the underlying rationales for these qualified opinions show longitudinal institutional tracking persistence within specific banking institutions over time?

4.2. Research Design, Scope, and Universe

This study uses a qualitative and descriptive research design based on formal document analysis and content analysis techniques (Krippendorff, 2018). Content analysis is an established academic approach that allows researchers to systematically identify structural patterns and themes within textual data. The sample universe comprises all banking institutions operating within the Turkish financial sector that published audited financial statements on the official public web portal of the BRSA. The study adopts a census approach within the defined study population, meaning that all identified bank-year independent audit reports meeting the inclusion criteria were included. Every bank that operated and disclosed an independent audit report during the fiscal years ending 31 December 2022, 2023, 2024, and 2025 was included in the data panel. The final population consists of 248 independent audit reports, distributed as follows:

• Fiscal Year 2022: 57 bank audit reports analyzed.
• Fiscal Year 2023: 63 bank audit reports analyzed.
• Fiscal Year 2024: 63 bank audit reports analyzed.
• Fiscal Year 2025: 65 bank audit reports analyzed.

The dataset used in the analysis consists of 248 bank-year observations, represented by 248 annual year-end independent audit reports. The unit of observation is the bank-year independent audit report rather than the number of distinct banking institutions. These reports were prepared within the regulatory and professional framework governing the independent auditing and financial reporting of banks in Türkiye. Bank audits are conducted in accordance with the Regulation on Independent Audit of Banks, published by the Banking Regulation and Supervision Agency (BRSA) in the Official Gazette, No. 29314, dated 2 April 2015, and the Independent Auditing Standards (BDS) issued by the Public Oversight, Accounting and Auditing Standards Authority. In particular, BDS 705, which is fully aligned with ISA 705, addresses modifications to the auditor’s opinion, including qualified opinions, and therefore provides the relevant professional auditing framework for the modified opinions examined in this study. The financial statements subject to these audits are prepared in accordance with the applicable BRSA accounting and financial reporting regulations and, where relevant, Turkish Accounting and Financial Reporting Standards (TMS/TFRS). The audit opinions analyzed in this study were therefore issued within the applicable banking-specific regulatory, financial reporting, and professional auditing framework.

4.3. Data Collection and Systematic Coding Procedure

To ensure the objectivity, replicability, and scholarly rigor demanded by formal content analysis (Krippendorff, 2018), a highly structured coding protocol was developed. The qualitative texts within the basis for qualified opinion sections were evaluated using the predefined coding criteria rather than solely on subjective impressions. Instead, they were processed through a test–retest coding reliability procedure to eliminate researcher bias and establish statistical verifiability.

4.3.1. Coding Design and Operationalization

To establish a rigorous bridge between the research questions and the empirical content analysis, the study systematically operationalized the variables. Table 1 maps each of the four research questions (RQs) to its corresponding coding protocol variables and delineates the analytical approach.

Table 1. Alignment of research questions and coding variables.

Research Question (RQ)	Coding Protocol Variable(s)	Operationalization and Analytical Approach
RQ1: Specific types of modified audit opinions under ISA 705	OP_TYPE (Categorical: Qualified, Adverse, Disclaimer)	Frequency and percentage distribution of audit opinion modifications across the full sample.
RQ2: Chronological distribution of modified opinions (2022–2025)	YEAR (Interval: 2022, 2023, 2024, 2025)	Cross-tabulation of opinion types by fiscal year to identify temporal trends.
RQ3: Primary accounting standard inconsistency/non-compliances and regulatory inconsistencies	Inconsistency/Non-compliance/THEME (Nominal: TFRS 9, TFRS 15, BRSA regulations, etc.)	Thematic content analysis of the “Basis for Qualified Opinion” sections to identify specific standards breached.
RQ4: Longitudinal persistence of modified opinions within specific banks	BANK_ID + PERSISTENCE_INDEX (Longitudinal tracking over time)	Panel-like tracking of individual banks across the 4-year window to assess the recurrence of specific audit qualifications.

To ensure structural alignment between the empirical inquiry and the content analysis, each research question was systematically mapped onto specific variables within the coding protocol. Table 1 outlines this methodological alignment, detailing how the qualitative disclosures in the independent audit reports were operationalized to address the core research questions. While Table 1 establishes the macro-level alignment between the core research questions and the analytical variables, operationalizing the qualitative text of the audit reports required a more granular coding taxonomy. To achieve this, the broad variables outlined in the research design, particularly the classification of accounting standard violations (inconsistency/non-compliance_code under RQ3), were translated into a highly structured, mutually exclusive coding guide. Table 2 details this operationalized coding scheme, illustrating how qualitative narratives were systematically mapped onto specific category codes, such as DP-IAS37 or FA-IFRS9, based on explicit textual triggers and validated inclusion criteria. The open coding procedure was guided by explicit semantic and operational boundaries. Textual units such as sentences or paragraphs were classified into distinct categories based on predefined linguistic triggers and regulatory keywords extracted from the audit reports. The standardized coding scheme used to transform the unstructured text into analytical panel data is detailed in Table 2.

Table 2. Operationalized coding scheme for basis of qualification.

Category Code	Primary Accounting Inconsistency/Non-Compliance	Operational Definition/Inclusion Criteria	Explicit Textual Triggers/ Keywords
DP-IAS37	Recognition of discretionary provisions failing IAS 37 parameters	Statements indicating the creation of general precautionary reserves, macro-financial cushions, or “free” provisions without a concrete, past obligating event.	“discretionary provision”, “free provision”, “general precautionary measure”, “inconsistent with IAS 37”, “macro-financial risks”
FA-IFRS9	Irregular financial asset classifications violating IFRS 9 guidelines	Modifications stemming from the incorrect staging of non-performing loans (NPLs), expected credit loss (ECL) miscalculations, or classification errors.	“IFRS 9”, “TFRS 9”, “impairment stage”, “expected credit losses”, “credit default classification”
PR-IAS37	Reversal of previously recognized discretionary provisions affecting current-period income.	Statements targeting the dissolution or dissolution-backed reversal of unmandated reserves into current-period income statements to manage net profit.	“reversal of provisions”, “other operating income”, “IAS 37.59”
NCOMP-IFRS1	Non-compliance with initial IFRS 1 transition guidelines	Isolated historical errors or structural misalignments arising during the bank’s initial statutory transition to the international financial reporting framework.	“IFRS 1”, “TFRS 1”, “first-time adoption”, “transition adjustments”
REVAL-IAS16	Irregular fixed asset revaluation practices violating IAS 16 rules	Errors or unapproved professional valuations regarding land, buildings, or premises that artificially distort equity or revaluation surpluses.	“IAS 16”, “revaluation surplus”, “fixed asset appraisal”, “property revaluation”
TAX-ACC	Failure to recognize accrued corporate income tax liabilities	Explicit omissions of current or deferred tax obligations, creating an immediate under-reporting of financial liabilities to regulatory bodies.	“corporate income tax”, “accrued liabilities”, “IAS 12”, “tax expense understatement”

The qualitative data was processed using an open coding approach to identify common underlying drivers. This coding procedure allowed the qualitative narratives contained in the audit reports to be systematically classified into analyzable categories and compared across bank-year observations.

4.3.2. Intra-Coder Reliability Assessment

To validate the reliability and reproducibility of the qualitative text-to-data conversion process, a rigorous intra-rater test–retest reliability protocol was executed. The author coded the entire universe of 61 modified audit reports based on the operationalized protocols outlined in Table 1. To eliminate temporal bias and ensure consistency, a random subsample of approximately 33 percent of the reports, which corresponds to exactly 20 reports, was re-coded by the author after a four-week interval.

The degree of test–retest reliability was measured using two distinct statistical coefficients, namely Percent Agreement and Cohen’s Kappa. Cohen’s Kappa was selected as the primary metric because it statistically controls the probability of agreements occurring purely by random chance. The mathematical formulation utilized for the Kappa coefficient is calculated as κ = (p_o − p_e)/(1 − p_e), where p_o represents the relative observed agreement between the two coding phases, and p_e represents the hypothetical probability of chance agreement. The comparison between the initial coding pass and the subsequent test–retest verification phase yielded an exceptionally high level of systematic consistency. Out of the randomly selected subsample of 20 modified reports, the two temporal coding passes achieved a Percent Agreement of 95.0 percent, which means that exactly 19 out of the 20 test reports matched perfectly. The statistical calculation for Cohen’s Kappa achieved a value of κ = 0.92. Based on commonly used interpretive benchmarks, a Kappa score of 0.92 indicates a very high level of agreement between the two coding rounds, supporting the consistency of the coding procedure.

5. Empirical Findings

This section presents the empirical findings of the systematic content analysis conducted on 248 bank-year independent audit reports spanning the fiscal years 2022 through 2025. By operationalizing the qualitative disclosures within the auditor reports via a validated coding protocol (κ = 0.92), this study presents both a quantitative breakdown and a qualitative deconstruction of the accounting anomalies that triggered external audit modifications in the Turkish banking sector.

5.1. Chronological Distribution and Typology of Audit Modifications

To address RQ1 and RQ2, this section examines the distribution of independent audit opinions across the 248 bank-year observations and evaluates the structural characteristics of modified audit opinions. The empirical distribution of audit opinion types is presented in Table 3.

Table 3. Distribution of audit opinion types by year (2022–2025).

Fiscal Year	Qualified Opinion	Unmodified Opinion	Adverse Opinion	Disclaimer Opinion	Total Reports	Qualification Rate (%)
2022	19	38	0	0	57	33.3%
2023	9	54	0	0	63	14.3%
2024	19	44	0	0	63	30.2%
2025	14	51	0	0	65	21.5%
Total	61	187	0	0	248	24.6%

The findings presented in Table 3 provide direct evidence for RQ1 and RQ2. Among the 248 independent audit reports examined, 187 reports (75.4%) received unmodified opinions, whereas 61 reports (24.6%) contained modified audit opinions. Importantly, all modified opinions identified within the sample were classified as qualified opinions under ISA 705. No adverse opinions or disclaimers of opinion were issued during the observation period. This distribution is theoretically significant because, under ISA 705, a qualified opinion indicates that the auditor identified material misstatements or limitations that were considered significant but not pervasive to the financial statements as a whole. Therefore, the absence of adverse opinions and disclaimers suggests that the identified accounting departures were concentrated in specific areas rather than reflecting fundamental failures in the overall financial reporting system of Turkish banks. The audit qualifications were primarily associated with specific accounting treatments, particularly the recognition and measurement of discretionary provisions reported under other liabilities.

The chronological analysis provides further insights into the changing intensity of these reporting issues. In 2022, qualified opinions reached their highest level, with 19 reports representing 33.3% of all bank-year observations. This period coincided with significant macroeconomic uncertainty, inflationary pressures, and strong profitability fluctuations within the Turkish banking sector. In 2023, the number of qualified opinions declined considerably to nine reports (14.3%), suggesting a temporary reduction in provisioning-related disagreements between bank management and external auditors. However, qualifications increased again in 2024, reaching 19 reports (30.2%), before declining moderately to 14 reports (21.5%) in 2025. The observed fluctuations indicate that modified audit opinions varied across the sample period and coincided with changes in the broader macroeconomic environment and accounting policy conditions. Rather than reflecting persistent financial distress, the cyclical pattern suggests that audit qualifications were closely linked to discretionary reporting choices, particularly regarding provisioning practices under conditions of economic uncertainty.

While Table 3 establishes the temporal pattern of modified audit opinions (RQ2), further analysis is required to identify the specific accounting issues underlying these qualifications (RQ3). Accordingly, Table 4 decomposes the 61 qualified opinions according to their primary accounting inconsistency and non-compliance categories, based on the validated test–retest coding procedure (Cohen’s Kappa = 0.92).

Table 4. Longitudinal distribution of audit qualifications by category (2022–2025).

Category Code	Primary Accounting Inconsistency/ Non-Compliance	2022 (N = 57)	2023 (N = 63)	2024 (N = 63)	2025 (N = 65)	Total (N = 248)	% of Total MAOs
DP-IAS37	Discretionary provisions failing IAS 37 parameters	16	7	16	13	52	85.2%
FA-IFRS9	Irregular financial asset staging/ECL under IFRS 9	2	1	2	1	6	9.8%
PR-IAS37	Reversal of previously recognized discretionary provisions under IAS 37	0	1	0	0	1	1.6%
Other	Miscellaneous inconsistency/non-compliance (TAX-ACC/REVAL-IAS16)	1	0	1	0	2	3.4%
Total MAOs	All Audit Qualifications	19	9	19	14	61	100.0%

The data reveals an overwhelming concentration of qualifications stemming from a single accounting issue. The recognition of discretionary provisions failing TAS 37 (KGK, 2023b) parameters (DP-IAS37) accounts for 85.2% of all audit modifications (52 out of 61 reports) in the Turkish banking industry. This finding confirms that audit qualifications in this sector are not random occurrences of financial reporting errors; rather, they are driven by a highly concentrated accounting practice. Chronologically, the frequency of qualifications exhibited a distinct non-linear trend, peaking at 19 reports in both 2022 and 2024, while dropping significantly to nine reports in 2023. This fluctuation is descriptively consistent with the possibility that provisioning practices varied alongside changes in macroeconomic conditions and regulatory circumstances; however, the evidence does not establish earnings-smoothing behavior or a causal relationship between these conditions and audit qualifications:

• The 2022 Peak (19 reports): The relatively high number of qualified opinions in 2022 coincided with substantial inflationary pressures, changes in real interest rates, and strong profitability conditions in parts of the Turkish banking sector. This temporal pattern may be compatible with interpretations involving changing profitability conditions and precautionary provisioning practices; however, the audit-report evidence does not establish specific managerial incentives or deliberate earnings-smoothing behavior.
• The 2023 Contraction (nine reports): The decline in qualified opinions in 2023 coincided with changes in monetary conditions, increasing funding costs, and pressure on banks’ net interest margins (NIM). The observed decline therefore provides a temporal pattern that may be considered alongside these broader changes, without establishing a specific managerial response.
• The 2024 Rebound (19 reports): The increase in qualified opinions in 2024 coincided with continued monetary tightening and changing credit-risk conditions. During this period, discretionary provisioning remained a recurring basis for qualified opinions, although the audit-report evidence does not establish whether the observed provisioning practices were undertaken for risk-buffering, earnings-management, or other managerial purposes. By 2025, qualified opinions related to discretionary provisioning remained present, with 14 reports identified in the sample.

The observed fluctuation is descriptively compatible with the earnings-smoothing literature (Beatty & Liao, 2014), although the present study does not directly measure earnings management or establish that provisioning practices responded causally to macroeconomic conditions or profitability pressures. The 2022 peak coincided with a period of extraordinary banking sector profitability associated with inflationary dynamics and returns on CPI-linked government securities (BRSA, 2022). Consistent with the positive accounting theory (Watts & Zimmerman, 1986), banks may have had incentives to moderate unusually high reported earnings through discretionary provisioning practices. Conversely, the decline in qualified opinions during 2023 coincided with increasing funding costs and compressed net interest margins following the monetary policy transition, which may have reduced incentives for income-decreasing provisioning adjustments (Fonseca & González, 2008).

5.2. Qualitative Deconstruction of the Dominant Inconsistency/Non-Compliance: DP-IAS37

To understand the friction between bank management and external auditors, this section performs a qualitative deconstruction of the texts contained in the Basis for Qualified Opinion paragraphs for the DP-IAS37 category. Under IAS 37, Provisions, Contingent Liabilities and Contingent Assets, a provision may only be recognized when the following three cumulative conditions are met:

• An entity has a present obligation (legal or constructive) as a result of a past event.
• It is probable that an outflow of resources embodying economic benefits will be required to settle the obligation.
• A reliable estimate can be made of the amount of the obligation.

The independent audit reports examined in the DP-IAS37 category repeatedly identified discretionary or general provisions that, according to the auditors, did not satisfy the applicable recognition requirements. The following excerpt represents the typical, highly standardized language utilized by external auditors to qualify bank reports under this category: “The accompanying financial statements as of 31 December 2024 include a general discretionary provision of TRY X million, of which TRY Y million was recognized as an expense in the current period under ‘Other Operating Expenses.’ This provision has been allocated by the bank management in response to potential macroprudential risks and economic uncertainties. However, since this provision does not meet the recognition criteria specified in TAS 37 (‘Provisions, Contingent Liabilities and Contingent Assets’), net income for the year is understated by TRY Y million, and other liabilities are overstated by TRY X million.”

This qualitative evidence indicates that the central accounting issue concerned the recognition basis of discretionary provisions rather than merely the measurement of an otherwise recognized obligation. This clash aligns closely with the international accounting literature showing that the inherent subjectivity and measurement flexibility of IAS 37 frequently transform provisions into a convenient “cookie-jar” for earnings management under volatile conditions. The audit-report narratives may be interpreted as indicating that discretionary provisions functioned as precautionary or risk-buffering amounts in response to uncertainty; however, the reports do not establish whether these practices were intended as earnings-smoothing mechanisms or served another managerial purpose. Under this dynamic, the bank’s primary objective is to implement a risk-buffering, capital-preservation mechanism to mitigate earnings volatility under hyperinflationary macroeconomic pressure. By recognizing discretionary provisions in high-earning years—as evidenced by audit qualification texts—banks create an implicit buffer that can be opportunistically reversed in leaner years to artificially stabilize or boost profitability. Conversely, the auditor’s objective is to enforce strict compliance with IAS 37, ensuring that financial statements remain free from subjective, unmandated reserves that may distort the firm’s true financial position. This friction highlights a persistent trade-off in bank financial reporting between accounting conservatism driven by prudential incentives and the faithful representation of performance demanded by professional accounting standards.

5.3. Audit Market Structure and Big Four Quality Assurance

An important dimension of the empirical analysis is the structure of the external audit market in the Turkish banking sector. The full dataset comprises 248 bank-year observations covering the 2022–2025 period, and all 248 audit reports were issued by Big Four audit firms; consequently, the sample contains no non-Big Four observations. As detailed in Table 5, the distribution of these 61 qualifications is remarkably balanced among the top-tier firms.

Table 5. Distribution of qualified opinions among Big Four audit firms (2022–2025).

Independent Audit Firm	2022	2023	2024	2025	Total Qualifications	Market Share (%)
Ernst and Young (EY)	7	3	3	5	18	29.5%
PwC	5	2	7	4	18	29.5%
Deloitte	4	4	5	3	16	26.2%
KPMG	3	0	4	2	9	14.8%
Total	19	9	19	14	61	100.0%

Because all 248 observations were audited by Big Four firms, the sample does not permit a comparison between Big Four and non-Big Four auditors. The distribution of the 61 qualified opinions across EY, PwC, Deloitte, and KPMG indicates that the accounting issues identified in the qualified reports were not confined to a single audit firm. However, this distribution should not be interpreted as evidence of uniform audit quality or identical professional skepticism across the four firms.

Similarly, the composition of the sample does not allow conclusions regarding the presence or absence of opinion shopping, as there are no non-Big Four observations against which the Big Four firms can be compared. The available data also do not permit an assessment of auditor–management negotiations or comparative audit quality.

The distribution of qualified opinions across the four Big Four firms nevertheless provides useful contextual information regarding the audit market structure of the analyzed banking sector. It shows that modified opinions related to the identified accounting issues were issued by multiple major audit networks during the observation period.

5.4. Longitudinal Persistence of Qualified Opinions

To address RQ4, this section examines whether the underlying rationales for these qualified opinions persisted across multiple reporting periods within specific banking institutions. The results demonstrate a high level of longitudinal persistence within specific banking institutions.

As shown in Table 6, the six selected institutions exhibited qualified audit opinions in each year of the 2022–2025 period, despite differences in ownership and banking type. In these cases, the reported basis of the qualified opinions was consistently associated with discretionary provisions established outside the applicable regulatory requirements, although the specific accounting basis and wording varied across institutions and years.

Table 6. Multi-year qualified opinion persistence by bank classification.

Bank Code (Anonymized)	Bank Ownership Type	2022	2023	2024	2025	Total Qualified Years	Persistent Non-Compliance?
Bank A	State-Owned	Qualified (DP)	Qualified (DP)	Qualified (DP)	Qualified (DP) *	4/4	Yes (Continuous–PwC)
Bank B	Participation	Qualified (DP)	Qualified (DP)	Qualified (DP)	Qualified (DP) *	4/4	Yes (Continuous–PwC)
Bank C	Private Domestic	Qualified (DP)	Qualified (DP)	Qualified (DP)	Qualified (DP)	4/4	Yes (Continuous–Deloitte)
Bank D	Private Domestic	Qualified (DP)	Qualified (DP)	Qualified (DP)	Qualified (DP) *	4/4	Yes (Continuous–KPMG to Deloitte)
Bank E	State-Owned	Qualified (DP)	Qualified (DP)	Qualified (DP)	Qualified (DP)	4/4	Yes (Continuous–EY)
Bank F	Development-Investment	Qualified (DP)	Qualified (DP)	Qualified (DP)	Qualified (DP)	4/4	Yes (Continuous–EY to PwC)

Notes: This table presents six anonymized banking institutions (Banks A–F) selected from the 248 bank-year observations because they received qualified audit opinions continuously in each year of the 2022–2025 analytical period. The six institutions represent different banking categories, including state-owned, private domestic, participation, and development-investment banks. All six institutions were audited by Big Four firms during the period. The table traces the persistence of qualified audit opinions and the related discretionary provisions (DP) across the four-year period. Auditor changes occurred for Bank D (KPMG to Deloitte) and Bank F (EY to PwC), while the reported basis of the qualified opinion remained related to discretionary provisions. An asterisk (*) indicates the reversal in 2025 of previously accumulated free provisions, with an effect on net income.

The longitudinal persistence observed in these cases provides descriptive evidence that qualification-related accounting issues can recur across multiple reporting periods within the same banking institutions. Specifically, the continued presence of qualified opinions alongside the maintenance or subsequent reversal of discretionary provisions suggests that formal financial reporting outcomes and institutional provisioning practices may operate in parallel over time. While these patterns provide descriptive evidence of multi-year persistence, they do not imply definitive managerial intent or presuppose specific reactions from external market participants. Instead, the findings highlight that qualified opinions related to discretionary provisions can persist across multiple reporting periods within diverse institutional settings.

5.5. IFRS 9 Staging and Miscellaneous Qualifications

While discretionary provisions under IAS 37 dominate the empirical findings (representing 85.2% of modified audit opinions, as shown in Table 4), the remaining qualifications provide additional insights into financial reporting challenges faced by banks under macroeconomic stress. The FA-IFRS9 category accounted for six reports (9.8% of all qualified opinions) and primarily involved issues related to the classification, measurement, and staging of financial assets under IFRS 9.

During the 2022–2025 inflationary period, auditors identified cases involving the delayed recognition or classification of certain financial assets, including the assessment of significant increases in credit risk and the appropriate transition of exposures between Stage 1 (performing), Stage 2 (significant increase in credit risk), and Stage 3 (credit-impaired) categories. Unlike the IAS 37-related qualifications, which were associated with excessive conservatism through discretionary provisioning, IFRS 9-related qualifications reflected potential under-recognition of credit risk and insufficient alignment with expected credit loss requirements. The relatively limited occurrence of FA-IFRS9 qualifications compared with IAS 37-related modifications suggests that Turkish banks generally maintained compliance with the standardized expected credit loss framework under IFRS 9 (KGK, 2025b), while discretionary provisioning practices represented the dominant area of accounting tension during the period examined.

6. Qualitative Analysis of Accounting Inconsistency/Non-Compliance

6.1. The Technical Collision and Balance Sheet Ripple Effects of IAS 37

In the 52 analyzed reports of discretionary provisioning, the core conflict centered on the strict, non-discretionary recognition criteria of IAS 37 for Provisions, Contingent Liabilities and Contingent Assets. Under IAS 37, paragraph 14, a provision can only be recognized if a present legal or constructive obligation exists from a past event, an outflow of resources is probable, and the amount can be reliably estimated. The audit reports indicate that these discretionary provisions were not linked to specific, quantifiable default events in the explanations provided by bank management. Instead, auditors consistently flagged these balances as general precautionary cushions against future macroeconomic uncertainty. A representative independent audit qualification describes this systematic departure:

The underlying financial statements contain a discretionary provision expense recognized by bank management as a general precautionary measure against potential negative shifts in macro-financial markets. Because no past obligating event has occurred, this accounting treatment is inconsistent with IAS 37 and BRSA Accounting and Reporting Regulations (BRSA, 2019; KGK, 2023b), thereby understating current period net profit and total shareholders’ equity.

This non-compliance extends beyond a single line-item error, creating a balance sheet ripple effect. By reducing pre-tax book income, these provisions may also create temporary differences under TAS 12, which is converged with IAS 12, Income Taxes (KGK, 2025c), with potential effects on the recognition of deferred tax assets and liabilities. This accounting treatment may affect the qualitative characteristics of neutrality, verifiability, and faithful representation emphasized in the conceptual framework of financial reporting. Consequently, such accounting treatments may affect reported performance and capital-related indicators, including return on assets (ROA) and capital adequacy ratios (CAR).

6.2. The 2025 Transition: Reversals of Discretionary Provisions and IAS 37

The content analysis identifies a notable development in 2025. Three institutions, including two banks, reversed previously recognized discretionary provisions, with the amounts recognized in the income statement under “Other Operating Income.” These reversals represent a change in the treatment of provisions that had been recognized in earlier reporting periods.

The timing of these reversals coincided with important macroeconomic and regulatory developments in the Turkish banking sector. During the period, changes in monetary conditions, deposit costs, and profitability pressures occurred alongside the transition toward the application of hyperinflation accounting under TAS 29, which has converged with IAS 29, Financial Reporting in Hyperinflationary Economies (KGK, 2023a). These developments provide relevant contextual conditions for interpreting the observed reversals. However, the audit-report evidence does not establish that the reversals were undertaken in response to specific managerial incentives or that they were directly caused by macroeconomic or regulatory developments.

The accounting treatment of the reversals also raises a relevant issue concerning IAS 37, paragraph 59, which addresses the reassessment and reversal of provisions when an outflow of economic resources is no longer probable. Given that the earlier discretionary provisions were identified in the audit reports as being outside applicable regulatory requirements, the subsequent reversal of these amounts represents an important longitudinal feature of the accounting treatment. In some cases, the reversal affected reported net income and, where disclosed, related tax expenses.

The observed pattern may be consistent with earnings-smoothing incentives, particularly because discretionary provisions recognized in earlier periods were subsequently reversed when reporting conditions changed. Nevertheless, the available evidence does not permit a conclusion that these provisions were established or reversed deliberately for earnings-management purposes. Accordingly, the findings are interpreted as evidence of changing provisioning patterns over time rather than as direct evidence of managerial intent.

The 2025 reversals also coincided with changes in audit-opinion outcomes for some institutions. Where the discretionary provisions had constituted a basis for a qualified opinion in earlier periods, their subsequent reversal removed or changed that reported basis in the relevant audit report. This pattern identifies 2025 as a relevant transition point in the longitudinal analysis. Rather than establishing that qualified opinions were driven by deliberate earnings management, the findings show that modified audit opinions in the analyzed banking institutions coincided with discretionary provisioning practices whose recognition and subsequent reversal varied across reporting periods.

7. Institutional and Signaling Discussion

7.1. Institutional Decoupling, BRSA Regulations, and Auditor Rotation

The continuous presence of qualified audit opinions across state-owned, private, and participation banks provides a basis for considering the institutional decoupling perspective developed in this study (Meyer & Rowan, 1977). The findings may be interpreted within a context of potentially competing institutional expectations: international accounting and auditing frameworks emphasize consistency with IFRS/IAS requirements and transparent financial reporting, while the domestic regulatory environment places substantial emphasis on financial stability, capital preservation, and resilience under volatile economic conditions (DiMaggio & Powell, 1983).

From an institutional perspective, these potentially competing expectations may help explain why formal financial reporting and external audit structures coexist with provisioning practices that auditors identify as inconsistent with specific recognition requirements. The continued use of Big Four audit firms does not, by itself, eliminate such accounting differences. Rather, the observed pattern may be consistent with a form of organizational decoupling in which formal compliance structures coexist with practices shaped by other institutional or regulatory considerations (Meyer & Rowan, 1977; Oliver, 1991).

The regulatory environment of the Turkish banking sector provides an important context for interpreting this pattern. Under the Banking Regulation and Supervision Agency (BRSA) framework, prudential regulation places strong emphasis on maintaining adequate capital and safeguarding financial stability. Such regulatory considerations may create incentives for banks to maintain financial buffers during periods of economic uncertainty, although the present study does not establish that specific provisioning decisions were directly required or encouraged by the regulator (BRSA, 2016). The qualified opinions examined in this study instead indicate that certain discretionary provisions were recognized beyond the applicable accounting recognition requirements.

The longitudinal evidence also provides a useful context for considering the role of auditor rotation. In some cases, qualified opinions associated with discretionary provisions persisted despite changes in the audit firm. For example, the longitudinal cases reported in Table 6 include institutions that experienced changes among Big Four auditors while qualified opinions related to discretionary provisions continued to appear. This observation does not demonstrate a failure of mandatory auditor rotation or establish that the rotation is ineffective in improving audit quality. Rather, it suggests that changes in the audit firm did not necessarily eliminate the underlying accounting issue identified in these cases, a pattern that is broadly consistent with prior research questioning whether auditor rotation alone is sufficient to produce systematic changes in audit outcomes (Cameran et al., 2018).

Taken together, these findings are consistent with an institutional interpretation in which formal external assurance and regulatory reporting coexist with persistent institution-specific provisioning practices. The evidence should not be interpreted as demonstrating deliberate non-compliance, managerial intent, or a systematic prioritization of BRSA requirements over IAS 37. Instead, the observed coexistence of regulatory, accounting, and auditing requirements provides a context in which institutional decoupling may offer a useful theoretical explanation for the persistence of qualified opinions associated with discretionary provisions.

7.2. Qualified Opinions and Their Potential Informational Interpretation

When benchmarked against the prior international auditing literature, these findings both align with and diverge from established paradigms. While traditional agency and signaling frameworks in developed markets (e.g., Dopuch et al., 1986; DeFond & Zhang, 2014) treat modified audit opinions primarily as unambiguous signals of financial distress or reporting failure, our emerging-market banking findings align more closely with studies showing that audit qualifications can stem from complex managerial choices and regulatory frictions (e.g., Gajevszky, 2014; Moalla, 2017). Furthermore, unlike the broader industrial literature, the specific use of discretionary provisions observed here reflects the unique income-smoothing and regulatory capital management dynamics identified in the banking literature (e.g., Beatty & Liao, 2014; Bushman & Williams, 2012), illustrating how institutional and macroeconomic volatility may provide a relevant context for interpreting ISA 705 modifications in emerging economies.

Traditional signaling theory generally associates qualified audit opinions with negative information and potentially adverse consequences for external users of financial statements. However, the findings of this study raise the possibility that the interpretation of a qualified opinion may depend on the accounting basis underlying the modification.

In the present study, the qualified opinions identified in the audit reports were frequently associated with discretionary provisions recognized beyond applicable regulatory requirements. Such provisions differ conceptually from qualifications arising from the understatement of liabilities, recognition of losses, or other indicators of financial deterioration. The persistence of qualified opinions associated with discretionary provisions across multiple reporting periods therefore raises a theoretical possibility that external users may interpret such modifications differently from qualifications associated with financial distress or material reporting deficiencies.

This interpretation is proposed as a theoretical implication of the empirical findings rather than as an empirically demonstrated market effect. The present study does not examine stock-price reactions, credit ratings, lending costs, cost of capital, analyst assessments, or stakeholder perceptions. Consequently, the extent to which external users actually perceive qualified opinions related to discretionary provisions as neutral or positive remains an empirical question.

Accordingly, the findings generate a theoretical proposition for future research: the informational consequences of a qualified audit opinion may depend not only on the existence of the modification but also on the underlying reason for the qualification. A qualification associated with discretionary over-provisioning may potentially be interpreted differently from one associated with financial distress, loss recognition, or material reporting deficiencies. Whether such differences in interpretation translate into neutral or favorable market outcomes requires direct testing using market- and stakeholder-level evidence.

8. Conclusions

This study indicates that the application of IAS 37-related provisions in the Turkish banking sector extends beyond a purely technical accounting issue and can be considered within the broader context of financial reporting requirements, institutional conditions, and macroprudential considerations. The empirical evidence indicates that discretionary provisions identified in the audit reports were recognized in a manner that was not fully consistent with the obligation-based recognition criteria of IAS 37 and affected reported profitability during the periods examined. The main findings can be summarized as follows:

Cyclical Provisioning Patterns: The findings indicate that discretionary provisions recognized in earlier reporting periods were subsequently reversed in some institutions when reporting conditions changed. This pattern may be consistent with income-smoothing incentives; however, the audit-report evidence does not establish that the provisions were recognized or reversed deliberately for earnings-management purposes. The findings therefore indicate that IAS 37-related audit qualifications in the analyzed banking institutions coincided with changing provisioning practices over time rather than providing direct evidence of persistent financial reporting failure.

Institutional Decoupling: The observed coexistence of formal financial reporting and external audit structures with persistent discretionary provisioning practices may be interpreted through the institutional decoupling perspective. The persistence of qualified opinions across multiple reporting periods, including cases involving changes in audit firms, provides a useful context for considering how accounting requirements, regulatory considerations, and institution-specific provisioning practices may operate in parallel. Nevertheless, the evidence does not establish managerial intent or demonstrate that domestic prudential objectives systematically outweighed compliance with international accounting requirements.

Potential Informational Interpretation: The findings also generate a theoretical proposition concerning the informational interpretation of qualified audit opinions. Because the qualifications examined in this study were frequently associated with discretionary over-provisioning rather than the understatement of losses or liabilities, their informational implications may differ conceptually from qualifications associated with financial deterioration. However, the present study does not examine stock-price reactions, credit ratings, lending costs, cost of capital, analyst assessments, or stakeholder perceptions. Accordingly, no conclusion can be drawn regarding whether such qualifications are actually interpreted as neutral or positive signals by external users.

Overall, the evidence indicates that qualified audit opinions in the analyzed Turkish banking institutions coincided with discretionary provisioning practices and their changing treatment across reporting periods. The findings therefore point to an interaction between accounting recognition requirements, banking-sector regulatory conditions, and evolving provisioning practices rather than establishing a single causal explanation for the observed qualifications. Future research may investigate whether the continued application of IAS 29 hyperinflation accounting may alter discretionary provisions or encourage alternative forms of risk-buffering practices.

Policy Implications

For regulators, standard-setters, and audit oversight bodies operating within emerging economies undergoing macroeconomic volatility, these findings point toward several important regulatory and professional considerations:

Coordination Between Prudential Rules and Accounting Standards: The observed persistence of audit modifications associated with unmandated discretionary provisioning suggests a potential structural friction between macroprudential expectations and general accounting standards (IAS 37/TAS 37). Greater regulatory dialog between banking authorities and accounting standard-setters could help mitigate ambiguity, potentially reducing instances where banks face conflicting compliance pressures.

Implementation Guidance in High-Inflation Environments: As observed during the 2022–2025 study period, persistent macroeconomic volatility and inflation may increase the practical challenges associated with the subjectivity inherent in provisions and contingent liabilities. Providing more granular implementation guidance regarding the application of IAS 37 and TAS 29 might assist financial institutions in navigating institutional complexities and improving reporting consistency.

Audit Oversight and Professional Judgment: The occurrence of recurring modified opinions across the four Big Four audit firms suggests that external oversight bodies, such as the KGK, may consider closer attention to the application of professional judgment to discretionary provisions. Equipping auditors with more explicit benchmarks may help support rigorous professional skepticism when evaluating bank-level accounting treatments.