Business Strategy, Audit Risk, and Auditor–Client Disagreement: Evidence from Korea

Abstract

This study examines the extent to which a firm’s business strategy shapes its strategic and audit risk profiles, and whether these risk characteristics ultimately manifest as measurable auditor–client disagreements. Auditor–client disagreement is operationalized using a direct, disclosure-based measure constructed as the scaled difference between unaudited preliminary net income—manually collected from mandatory timely filings disclosed through the Korea Financial Supervisory Service’s Electronic Disclosure System (DART)—and final audited net income reported in the audited financial statements. Using a sample of 6504 firm-year observations drawn from firms listed on the Korea Exchange (KOSPI and KOSDAQ) over the period 2020–2024, I find that a higher strategic score, reflecting a more innovation-oriented, prospector-type strategic posture, is consistently and significantly positively associated with the likelihood of auditor–client disagreement. Conversely, firms pursuing a cost-efficiency-oriented, defender-type strategy exhibit a significantly lower likelihood and smaller magnitude of disagreement. These findings suggest that business strategy functions as a fundamental, ex-ante determinant of inherent risk and audit risk, directly shaping auditors’ effort allocation and financial reporting outcomes. Collectively, this study contributes to the auditing literature by providing empirical evidence that a client’s strategic positioning constitutes a material, firm-level risk factor—consistent with the risk assessment framework mandated by International Standard on Auditing (ISA) 315—and should therefore be explicitly incorporated into auditors’ engagement risk assessments and the design of risk-based audit procedures.

1. Introduction

Concerns over accounting transparency and audit quality have intensified in the wake of high-profile accounting scandals and sweeping regulatory reforms. In the Republic of Korea, the landmark revision of the Act on External Audit of Stock Companies in 2018 fundamentally restructured the audit environment, driving heightened auditor conservatism and more rigorous scrutiny of management’s financial reporting assertions (Lee et al. 2020). Against this backdrop, the incidence of auditor–client disagreements has risen as firms and their auditors navigate an increasingly demanding risk landscape. Auditor–client disagreement represents a critical, directly observable outcome of the negotiation process between management’s reporting incentives and the auditor’s independent risk assessment—a phenomenon whose determinants have attracted growing scholarly attention (Amerongen and Nieuwenhuizen 2025; Choudhary et al. 2021).

While prior literature has examined auditor–client disagreements through the lenses of corporate governance, internal control quality, and regulatory intervention (Bhattacharjee et al. 2025; Chapman et al. 2023), identifying the fundamental, firm-specific antecedents of such conflicts remains a significant gap in the extant literature. A growing body of evidence demonstrates that a firm’s business strategy is a primary determinant of its inherent business and audit risks (Bentley et al. 2013; Habib et al. 2024). Yet direct empirical evidence on how a firm’s overarching strategic orientation manifests in observable conflicts with external auditors during the financial reporting process remains surprisingly scarce. This gap is consequential: under International Standard on Auditing (ISA) 315, auditors are explicitly required to obtain a thorough understanding of the client’s business model and strategic environment as the foundation for assessing the risks of material misstatement. If strategic orientation is a material driver of audit risk, its systematic exclusion from empirical models of auditor–client disagreement represents both a theoretical oversight and a practical limitation for audit engagement planning.

To address this gap, this study investigates the extent to which a firm’s business strategy shapes its strategic and audit risk profiles, and whether these risk characteristics ultimately influence the likelihood and magnitude of auditor–client disagreement. Using a sample of 6504 firm-year observations drawn from firms listed on the Korea Exchange (KOSPI and KOSDAQ) over the period 2020–2024, I quantify business strategy using the typology of Miles et al. (1978) as operationalized by Bentley et al. (2013), which classifies firms along a continuum from innovation-oriented prospectors to cost-efficiency-oriented defenders. Auditor–client disagreement is measured directly as the scaled difference between unaudited preliminary net income—disclosed through mandatory profit structure change filings in the Korea Financial Supervisory Service’s Electronic Disclosure System (DART) prior to audit completion—and final audited net income reported in the audited financial statements, thereby capturing the tangible financial outcome of audit adjustments (Sohn and Lee 2008).

This study makes three contributions to the extant literature. First, it shifts the focus of auditor–client disagreement research from traditional governance-based monitoring mechanisms—such as board structure and internal control quality—to the client’s fundamental business strategy, offering a novel, ex-ante perspective on the firm-level determinants of financial reporting outcomes. Second, it extends the business strategy literature (Habib et al. 2024; Bentley et al. 2013) by providing direct empirical evidence that strategic orientation influences not only audit effort and fee levels, but also the ultimate outcome of auditor–client negotiations as reflected in pre-audit to post-audit earnings adjustments. Third, from a practical standpoint, the findings provide empirical validation for the risk-based auditing framework mandated by ISA 315, establishing that a client’s strategic positioning constitutes a material, firm-level risk factor that auditors and regulators should explicitly incorporate into engagement risk assessments and audit planning procedures.

The remainder of this paper proceeds as follows. Section 2 reviews the relevant literature and develops the hypotheses. Section 3 describes the research design and variable measurement. Section 4 presents the empirical results, including robustness tests. Section 5 presents the conclusions and discusses the practical and theoretical implications of the findings. Section 6 discusses the limitations of the study and outlines directions for future research.

2. Literature Review and Hypothesis Development

2.1. The Dynamics of Auditor–Client Disagreement and Financial Reporting Quality

The financial reporting process is conventionally characterized as a dynamic negotiation between management, whose reporting incentives are shaped by compensation contracts, debt covenants, and capital market pressures, and external auditors, who are obligated to apply professional skepticism and challenge assertions that deviate from applicable accounting standards (Hatfield et al. 2022). Auditor–client disagreement materializes when this scrutiny results in auditors successfully challenging management’s pre-audit financial assertions, producing directly observable earnings adjustments between the preliminary and audited financial statements (Amerongen and Nieuwenhuizen 2025). While early research relied predominantly on discretionary accruals as an indirect proxy for these dynamics, more recent scholarship demonstrates that direct, disclosure-based measures of audit adjustments provide a substantially more precise characterization of both audit quality and the intensity of auditor–client conflict—circumventing the well-documented identification problems inherent in accrual-based proxies (Choudhary et al. 2021).

The institutional environment in which these disagreements unfold has grown increasingly demanding. The implementation of more stringent audit regulations—most notably the comprehensive revision of the Act on External Audit of Stock Companies in Korea—has materially elevated auditor conservatism and sharpened the scrutiny applied to management’s financial reporting assertions (Lee and Park 2016). Within this evolving regulatory context, auditor–client relationships have become progressively more sensitive to firm-specific operational complexities, corporate governance structures, and the heightened verification demands imposed by modern auditing standards (Bhattacharjee et al. 2025; Darvishi et al. 2025; Kim 2024). Chapman et al. (2023) further document that latent firm-level risks and attendant earnings reporting delays directly strain the auditor–client relationship, precipitating greater realignment of pre-audit financial positions. Notwithstanding these advances, the extant literature has focused predominantly on governance-based monitoring mechanisms—such as audit committee effectiveness and internal control quality—as determinants of auditor–client disagreement, while largely neglecting the more fundamental, firm-level antecedent that structurally shapes inherent risk across all dimensions of the financial statements: the client’s business strategy.

2.2. Business Strategy as a Determinant of Inherent and Audit Risk

A firm’s business strategy governs its operational complexity, resource allocation priorities, and competitive market positioning, thereby systematically shaping its overall business risk profile (Bentley et al. 2013; Miles et al. 1978). Within the widely adopted Miles et al. (1978) typology, firms are arrayed along a strategic continuum defined by two polar archetypes. At one end are prospectors, characterized by a commitment to continuous innovation, rapid product line expansion, and decentralized organizational structures designed to facilitate environmental responsiveness. At the opposite end are defenders, which prioritize operational cost efficiency, stable and narrowly defined product portfolios, and centralized control mechanisms oriented toward predictability and process optimization.

A growing body of empirical evidence affirms that strategic positioning drives systematic and observable variation in financial reporting outcomes and audit-related behaviors (Habib et al. 2024). Zhang et al. (2022) demonstrate that prospector-type strategies structurally incentivize aggressive tax avoidance, while Harymawan et al. (2024) document that strategic orientation materially alters corporate investment risk profiles and sustainability performance. From an auditing perspective, contemporary risk-based frameworks explicitly require auditors to evaluate an entity’s strategic objectives and associated business risks as a prerequisite for identifying and assessing the risks of material misstatement (Knechel 2007). In line with this requirement, prospector strategies—marked by heightened structural fluidity, organizational complexity, and reliance on subjective accounting estimates—consistently command substantially greater audit effort, as evidenced by higher audit fees and extended audit report lags (Habib et al. 2024; Bentley et al. 2013). More recently, Fang et al. (2025) demonstrate that prospector-oriented firms inherently elevate business risk and demand significantly greater audit effort even within highly digitalized corporate environments, while Eissa et al. (2025) corroborate that such strategic complexities systematically extend auditor report lags in emerging market contexts—collectively underscoring that business strategy remains a paramount structural driver of audit friction.

It bears emphasis that incorporating strategic analysis into the audit process neither compromises auditor independence nor introduces subjective judgment beyond the bounds of professional standards. While financial reporting is strictly governed by objective accounting rules, the underlying economic substance of transactions and the complexity of applying those rules in practice—such as the capitalization of R&D expenditures, the recognition of revenue from multi-element arrangements, or the impairment of long-lived assets—are inextricably linked to the client’s overarching strategic posture. Indeed, far from being discretionary, evaluating a client’s business strategy and strategic risks is an explicit regulatory obligation. Under International Standard on Auditing (ISA) 315, Identifying and Assessing the Risks of Material Misstatement, auditors are mandated to obtain a thorough understanding of the entity’s objectives, strategies, and related business risks as a foundation for identifying where material misstatements are most likely to arise. Accordingly, assessing a firm’s strategic risk profile is not a deviation from auditing principles; it is an indispensable component of the auditor’s objective risk identification process and a direct expression of the risk-based auditing paradigm.

2.3. Hypothesis Development: Integrating Strategic Positioning with Auditor–Client Disagreements

Drawing on the theoretical synthesis presented in Section 2.1 and Section 2.2, this study posits a structured causal chain linking business strategy to auditor–client disagreement: a firm’s strategic orientation structurally determines its inherent risk profile and internal control complexity; these risk characteristics, in turn, directly shape the external auditor’s assessment of the risks of material misstatement; and, to mitigate these elevated risks, auditors deploy expanded audit effort and propose more rigorous financial statement adjustments. Within this framework, auditor–client disagreement emerges not merely as a residual indicator of financial reporting quality, but as a direct, observable measure of the intensity and outcome of the negotiation process between management and the external auditor. The mechanisms underlying this relationship are grounded in both the incentives and the structural opportunities for financial misreporting (Hogan et al. 2008).

Prospector firms operate in highly dynamic competitive environments that necessitate continuous R&D investment and reliance on external financing to sustain growth. This trajectory frequently induces management to structure compensation packages heavily around equity-based instruments, creating robust incentives for opportunistic earnings management (Burns and Kedia 2006). Of equal significance from an audit verification standpoint, the operational complexity inherent in prospector strategies directly impairs the auditor’s capacity to corroborate management’s financial assertions. Prospectors routinely engage in novel, non-routine transactions requiring highly subjective accounting estimates—including the capitalization of development expenditures, the measurement of intangible assets arising from acquisitions, and the recognition of revenue from complex, multi-element arrangements. Simultaneously, the decentralized and structurally fluid organizational architectures characteristic of prospector firms afford management broader opportunities to exploit internal control weaknesses and exercise accounting discretion (Bentley-Goode et al. 2017). Confronted with these compounding sources of inherent risk, external auditors must apply heightened professional skepticism and commit substantially greater audit resources to obtain sufficient appropriate evidence. The resulting tension between management’s aggressive reporting incentives and the auditor’s obligation to rigorously mitigate identified risks inevitably elevates both the frequency and the magnitude of audit adjustments.

Defender firms, by contrast, operate in stable, well-defined competitive domains that emphasize operational efficiency and consistency in financial reporting (Simons 1987). Their centralized, process-driven organizational structures impose natural constraints on managerial discretion, limiting the opportunities for aggressive accounting choices and thereby suppressing inherent risk at the assertion level. From the auditor’s perspective, the predictable and systematic nature of defender operations facilitates more efficient risk assessments and reduces the scope for contested judgments over financial statement assertions. Consequently, defender firms are expected to exhibit significantly fewer instances and smaller magnitudes of auditor–client disagreement relative to their prospector counterparts. Grounded in these arguments, the following hypotheses are advanced:

Hypothesis 1 (H1).

Ceteris paribus, the higher a firm’s business strategy index score—reflecting greater proximity to a prospector-type strategic orientation—the greater the likelihood of auditor–client disagreement.

Hypothesis 1-1 (H1-1).

Ceteris paribus, firms classified as prospector-type are more likely to experience auditor–client disagreement than firms not so classified.

Hypothesis 1-2 (H1-2).

Ceteris paribus, firms classified as defender-type are less likely to experience auditor–client disagreement than firms not so classified.

Hypothesis 2 (H2).

Ceteris paribus, the higher a firm’s business strategy index score—reflecting greater proximity to a prospector-type strategic orientation—the greater the magnitude of auditor–client disagreement.

Hypothesis 2-1 (H2-1).

Ceteris paribus, the magnitude of auditor–client disagreement is greater for prospector-type firms than for non-prospector firms.

Hypothesis 2-2 (H2-2).

Ceteris paribus, the magnitude of auditor–client disagreement is smaller for defender-type firms than for non-defender firms.

3. Research Design and Sample Selection

3.1. Measuring Business Strategy

Following the established methodology of Bentley et al. (2013) and Bentley-Goode et al. (2017), which operationalize the strategic typology of Miles et al. (1978), I construct a composite business strategy index using six firm-level financial ratios, each capturing a distinct dimension of strategic orientation:

• New product development: Ratio of R&D expenditures to net sales
• Marketing effort: Ratio of selling expenses to net sales
• Growth pattern: Year-over-year percentage change in net sales
• Production efficiency: Ratio of the number of employees to net sales
• Capital intensity: Ratio of property, plant, and equipment to total assets
• Organizational stability: Standard deviation of the total number of employees

Each ratio is conceptually linked to the prospector–defender continuum in a theoretically coherent manner. Prospector firms allocate disproportionately greater resources to R&D and marketing as they continuously pursue new product development and market expansion, and their sustained growth orientation produces higher sales growth rates. Their reliance on large, flexible human capital inputs results in elevated employee-to-sales ratios, while their relative disregard for capital-intensive process optimization yields lower capital intensity scores. Finally, the decentralized and structurally fluid organizational architectures of prospector firms generate higher variability in headcount, reflecting organizational instability. Defender firms exhibit the opposite pattern across all six dimensions.

Following Bentley et al. (2013), each of the six ratios is computed annually and averaged over a five-year rolling window. This smoothing procedure is both theoretically and methodologically justified: it attenuates the influence of transient, single-year macroeconomic shocks or idiosyncratic firm-level fluctuations, thereby more accurately capturing the firm’s enduring, long-term strategic posture rather than short-term operational noise. The six components are equally weighted in constructing the composite index, ensuring that no single operational metric exerts disproportionate influence on the overall strategic classification.

Table 1. Measurement of Business Strategy.

New product development	Firm’s emphasis on new product development	Ratio of R&D expenditures to sales
Marketing effort	Marketing and sales efforts to expand new products or services	Ratio of selling expenses to sales
Growth pattern	Growth rate and investment potential	Sales growth rate
Production efficiency	The ability to efficiently produce and distribute products and services	Ratio of the number of employees to sales
Capital intensity	Degree of investment for technical efficiency	Ratio of property, plant and equipment assets to total assets
Organizational stability	Organizational instability and degree of decentralization	Standard deviation of the total number of employees

presents the measurement of business strategy.

For each firm-year observation, the six five-year rolling averages are ranked into quintiles relative to industry peers, using the Korea Standard Industry Classification (KSIC) intermediate classification as the industry grouping criterion. Each firm-year is assigned a quintile rank of 1 (lowest) to 5 (highest) for each of the six dimensions. The six quintile ranks are then summed to yield the composite STRATEGY index, which ranges from a minimum of 6 (all six dimensions in the lowest quintile: 1 × 6 = 6) to a maximum of 30 (all six dimensions in the highest quintile: 5 × 6 = 30). Higher STRATEGY scores thus indicate a more innovation-oriented, prospector-type posture—characterized by greater new product development intensity, marketing expenditure, and sales growth, combined with lower capital intensity and organizational stability—relative to industry peers. To assess the internal consistency of the six-component index, I conduct Cronbach’s alpha reliability analysis. The resulting coefficient of 0.718 exceeds the conventionally accepted threshold of 0.70, confirming adequate reliability of the composite measure.

To complement the continuous STRATEGY variable, I construct two binary classification variables following the 8-9-8 classification scheme adopted in prior Korean studies. Firms with a STRATEGY score of 23 or above are classified as prospector-type and assigned PROSPECTOR = 1 (0 otherwise); firms with a score of 13 or below are classified as defender-type and assigned DEFENDER = 1 (0 otherwise). The remaining middle tercile (scores of 14–22) serves as the reference group. The 8-9-8 scheme—as opposed to the al-ternative 7-11-7 scheme employed in some prior studies—is adopted because it yields a more balanced distribution of observations near the classification boundaries given the sample size of this study. Under Hypotheses H1 and H2, the regression coefficient on STRATEGY is predicted to be significantly positive. Under Hypotheses H1-1, H1-2, H2-1, and H2-2, the coefficient on PROSPECTOR is predicted to be significantly positive and the coefficient on DEFENDER significantly negative.

3.2. Measuring Auditor–Client Disagreement

Auditor–client disagreement is operationalized using a direct, disclosure-based measure constructed as the scaled difference between pre-audit net income and post-audit net income. Pre-audit net income reflects management’s internally finalized financial position prior to the completion of the external audit and is manually collected from mandatory profit structure change disclosures—specifically, filings categorized as “Sales Revenue or Profit Structure Variation”—submitted to the Korea Financial Supervisory Service’s Electronic Disclosure System (DART). Post-audit net income is extracted from the corresponding audited annual financial statements. This direct measurement approach overcomes the well-documented identification limitations of accrual-based proxies, which conflate management’s initial reporting assertions with the outcome of auditor–client negotiations, and has been validated and widely adopted in the Korean auditing literature (Cho et al. 2016; Lee and Park 2016; Lee et al. 2020; Shin et al. 2018; Sohn and Lee 2008).

Three complementary measures of auditor–client disagreement (ACD) are constructed. First, DUMACD is a binary indicator variable equal to one if the absolute scaled difference between pre-audit and post-audit net income meets or exceeds a materiality threshold of 0.1% of the absolute value of post-audit net income, and zero otherwise. This threshold follows Sohn and Lee (2008), who demonstrate that raw differences below this level are attributable to mechanical unit discrepancies—arising from inconsistencies between database-reported figures (denominated in Korean won) and disclosure data (denominated in millions of won)—rather than genuine audit adjustments. Second, ABSACD is the continuous absolute value of this scaled difference, capturing the magnitude of disagreement without regard to its direction. Third, REALACD retains the sign of the scaled difference, enabling analysis of the directionality of disagreement—specifically, whether the audit process results in a net upward or downward adjustment to management’s initially reported earnings.

3.3. Research Model

To test Hypotheses H1 and H2, I estimate two empirical specifications. Model (1) employs the continuous STRATEGY index as the primary variable of interest, while Model (2) replaces it with the binary classification variables PROSPECTOR and DEFENDER. The two specifications are presented below:

• Continuous variable model;

  ACDi,t (DUMACD or REALACD)
  = β0 + β1STRATEGYi,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t
  + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t
  + β10OWNi,t +β11OUTCOSTi,t + β12FORi,t + β13DATEi,t
  + β14MKi,t + ∑YEAR +∑IND + ε

  (1)
• Dummy variable model;

  ACDi,t (DUMACD or REALACD)
  = β0 + β1PROSPECTORi,t + β2DEFENDERi,t + β3SIZEi,t
  + β4GRWi,t + β5LEVi,t + β6ROAi,t + β7LOSSi,t + β8LIQi,t
  + β9CONi,t + β10BIG4i,t + β11OWNi,t + β12OUTCOSTi,t
  + β13FORi,t + β14DATEi,t + β15MKi,t
  + ∑YEAR + ∑IND + ε

  (2)

The dependent variable ACD is operationalized using three complementary measures as defined in Section 3.2: DUMACD (binary indicator of disagreement), ABSACD (magnitude of disagreement), and REALACD (signed direction of disagreement). In Model (1), the primary variable of interest is STRATEGY, a continuous index ranging from 6 to 30, where higher values reflect a more prospector-oriented strategic posture. In Model (2), the variables of interest are PROSPECTOR (equal to one for firms with STRATEGY ≥ 23, zero otherwise) and DEFENDER (equal to one for firms with STRATEGY ≤ 13, zero otherwise). Under H1 and H2, the coefficient β1 on STRATEGY is predicted to be significantly positive. Under H1-1, H1-2, H2-1, and H2-2, β1 on PROSPECTOR is predicted to be significantly positive and β2 on DEFENDER significantly negative. The definitions and measurement details of all variables are summarized in

Table 2. Variable Definitions.

Variable		Definition
ACD		=Auditor–Client Disagreement; operationalized as DUMACD, ABSACD, or REALACD
	DUMACD	=1 if (|Net Income before Audit − Net Income after Audit|/|Net Income after Audit|) ≥ 0.001, 0 otherwise
	ABSACD	=(|Net Income before Audit − Net Income after Audit|/|Net Income after Audit|)
	REALACD	=(Net Income before Audit − Net Income after Audit)/|Net Income after Audit|;
STRATEGY		Composite business strategy index (continuous, range: 6–30); higher values indicate a prospector-type orientation (Bentley et al. 2013; Bentley-Goode et al. 2017)
	PROSPECTOR	=1 if STRATEGY ≥ 23, 0 otherwise
	DEFENDER	=1 if STRATEGY ≤ 13, 0 otherwise
SIZE		Natural logarithm of total assets
GRW		=Total Asset Growth Rate (=(Total Assets for the current year − Total Assets for the previous year)/Total Assets for the previous year)
LEV		Financial leverage = Total debt/Total assets
ROA		=Return on Total Assets (=Net Income/Beginning Total Assets);
LOSS		=1 if net loss is reported, 0 otherwise
LIQ		=Current Ratio (=Current Assets/Current Liabilities)
CON		=1 if consolidated financial statements are prepared, 0 otherwise
BIG4		=1 if audited by a Big 4 audit firm, 0 otherwise
OWN		Controlling shareholder ownership ratio
OUTCOST		Audit fee intensity = External audit fees/Total assets at year-end
FOR		Foreign investor shareholding ratio
DATE		Natural logarithm of the number of days elapsed between the internal closing date and the date of the profit structure change disclosure
MK		=1 if listed on the KOSPI market, 0 if listed on the KOSDAQ market
YEAR		Year fixed effects
IND		Industry fixed effects

.

The control variables are selected on the basis of prior empirical evidence linking firm and engagement characteristics to auditor–client disagreement. Firm size (SIZE) is included because larger firms exhibit greater operational complexity and broader audit scope, which may amplify the potential for contested adjustments (Habib and Bhuiyan 2011). Asset growth (GRW) is included as a proxy for business risk, given evidence that rapidly growing firms are associated with elevated audit risk (Habib and Bhuiyan 2011). Financial leverage (LEV) controls for financial distress risk, which heightens earnings management incentives and, by extension, audit risk (Almeida 2023). Profitability (ROA) is included to control for the confounding effect of firm performance on disagreement (Darvishi et al. 2025), and the net loss indicator (LOSS) captures the incremental reporting incentives associated with loss avoidance (Hatfield et al. 2022). Liquidity (LIQ) is included following evidence that a firm’s short-term financial condition influences the scope of auditor–client negotiations (Darvishi et al. 2025). The consolidated financial statements indicator (CON) controls for the additional audit complexity arising from group-level reporting, consistent with evidence that consolidation increases audit effort and engagement duration (Dutta and Bansal 2025). Auditor size (BIG4) is included to capture the reputational disciplining effect of large audit firms, which are expected to apply more rigorous verification procedures to protect franchise value. Controlling shareholder ownership (OWN) and foreign investor ownership (FOR) serve as proxies for corporate governance quality, following evidence that ownership structure influences financial reporting outcomes and the dynamics of auditor–client negotiations (Bhattacharjee et al. 2025). Audit fee intensity (OUTCOST) controls for the level of external monitoring effort deployed by the auditor (Hatfield et al. 2022). The disclosure timing variable (DATE) controls for the interval between the internal closing date and the profit structure disclosure, which may proxy for the complexity and contentiousness of the pre-audit reporting process. Finally, the listing market indicator (MK) controls for systematic differences in regulatory standards and reporting environments between the KOSPI and KOSDAQ markets. All specifications include year (YEAR) and industry (IND) fixed effects to control for time-period-specific and industry-level confounders, and standard errors are clustered at the firm level to account for within-firm serial correlation.

3.4. Sample Selection

The initial sample comprises all firms listed on the Korea Exchange (KOSPI and KOSDAQ) over the five-year period from 2020 to 2024. To ensure comparability across observations, I apply the following sequential exclusion criteria. First, firms with fiscal year-ends other than 31 December are excluded to eliminate systematic differences in the timing of financial reporting and audit completion. Second, firms operating in the financial industry (e.g., banking, insurance, and securities) are excluded due to the fundamental differences in their financial statement structure, regulatory environment, and risk profile relative to non-financial firms. Third, firm-years with negative book equity are excluded to mitigate the confounding influence of financially distressed observations on the measurement of auditor–client disagreement and business strategy.

The construction of the primary dependent variable requires pre-audit net income, which is obtained exclusively from mandatory profit structure change disclosures filed with the Korea Financial Supervisory Service’s Electronic Disclosure System (DART). Firm-years for which such disclosures are unavailable—and thus for which pre-audit net income cannot be directly observed—are excluded from the sample. Similarly, firm-years with insufficient financial data to construct the six-component business strategy index or the full set of control variables are excluded. All continuous variables are winsorized at the 1st and 99th percentiles to mitigate the influence of extreme observations on the regression estimates. The final sample comprises 6504 firm-year observations. The sample attrition process is detailed in

Table 3. Sample Selection.

Sample Selection		Observations
Non-financial firms with December fiscal year-ends listed on the KOSPI and KOSDAQ, 2020–2024		10,368
	Less: Firms for which profit structure change disclosures are unavailable in DART	(1826)
	Less: Firms for which financial data required to construct the business strategy index are unavailable	(1701)
	Less: Firms for which financial data required to construct the control variables are unavailable	(337)
Final sample		6504

.

4. Empirical Analysis Results

4.1. Descriptive Statistics

Table 4. Descriptive Statistics (N = 6504).

Variable	Mean	Std. Dev	Min	Q1	Median	Q3	Max
DUMACD	0.623	0.475	0	0	1	1	1
ABSACD	0.182	0.481	<0.001	<0.001	0.014	0.100	3.491
REALACD	0.087	0.424	−0.923	<0.001	<0.001	0.047	2.577
STRATEGY	17.214	3.925	6	15	18	21	29
PROSPECTOR	0.082	0.234	0	0	0	0	1
DEFENDER	0.091	0.276	0	0	0	0	1
SIZE	26.213	1.464	23.683	25.213	25.918	26.913	31.115
GRW	0.085	0.273	−0.384	−0.035	0.030	0.123	1.731
LEV	0.451	0.201	0.045	0.275	0.446	0.591	0.880
ROA	0.002	0.107	−0.425	−0.025	0.015	0.047	0.294
LOSS	0.359	0.474	0.000	0.000	0.000	1.000	1.000
LIQ	2.400	2.465	0.383	0.975	1.467	2.483	16.542
CON	0.823	0.388	0.000	1.000	1.000	1.000	1.000
OUTCOST	0.001	0.001	0.000	<0.001	<0.001	0.001	0.005
BIG4	0.452	0.481	0.000	0.000	0.000	1.000	1.000
OWN	0.373	0.154	0.063	0.248	0.367	0.508	0.792
FOR	0.072	0.100	0.000	0.013	0.028	0.077	0.532
DATE	3.812	0.284	2.982	3.612	3.808	4.015	4.321
MK	0.413	0.478	0.000	0.000	0.000	1.000	1.000

reports descriptive statistics for all variables employed in the empirical analysis. With respect to the dependent variable, the mean of DUMACD is 0.623, indicating that auditor–client disagreement—defined as a scaled pre-audit to post-audit net income difference of at least 0.1%—is observed in approximately 62.3% of firm-year observations. This relatively high incidence rate is consistent with the heightened auditor conservatism documented in the post-2018 Korean audit environment (Lee et al. 2020), and corroborates prior evidence that material earnings adjustments between the preliminary and audited financial statements are a pervasive feature of Korean listed firms. The mean (median) of ABSACD is 0.182 (0.014), reflecting an average absolute adjustment equivalent to 18.2% of post-audit net income. The pronounced divergence between the mean and median suggests a right-skewed distribution, consistent with a concentration of relatively modest adjustments punctuated by a subset of firm-years with substantially larger disagreements.

Turning to the primary variables of interest, the mean (median) of the continuous STRATEGY index is 17.214 (18), positioning the average sample firm near the midpoint of the 6–30 scale and indicating that the sample is broadly distributed across the strategic continuum rather than concentrated at either extreme. Approximately 8.2% of firm-year observations are classified as prospector-type firms (PROSPECTOR mean = 0.082), while approximately 9.1% are classified as defender-type firms (DEFENDER mean = 0.091). The broadly comparable proportions of prospectors and defenders in the sample are consistent with the 8-9-8 classification scheme and provide a balanced basis for testing the hypothesized asymmetric associations between strategic orientation and auditor–client disagreement.

4.2. Correlation

Table 5. Correlation Analysis.

Variables		(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)
(1)	DUMACD	1.00
(2)	ABSACD	0.29	1.00
(3)	STRATEGY	0.06	0.03	1.00
(4)	PROSPECTOR	0.03	0.01	0.55	1.00
(5)	DEFENDER	−0.04	−0.03	−0.54	−0.08	1.00
(6)	SIZE	−0.12	−0.08	−0.05	−0.05	−0.02	1.00
(7)	GRW	0.06	0.02	0.08	0.05	−0.03	0.01	1.00
(8)	LEV	0.13	0.12	−0.07	−0.07	0.00	0.28	0.04	1.00
(9)	ROA	−0.13	−0.09	−0.06	−0.03	0.02	0.22	0.12	−0.23	1.00
(10)	LOSS	0.12	0.12	0.03	0.05	0.00	−0.21	−0.12	0.21	−0.67	1.00
(11)	LIQ	−0.09	−0.08	0.13	0.11	−0.03	−0.23	0.00	−0.65	0.08	−0.10	1.00
(12)	CON	0.03	0.02	0.06	0.03	−0.05	0.28	0.05	0.17	0.00	0.02	−0.12	1.00
(13)	OUTCOST	0.12	0.11	0.13	0.08	−0.02	−0.62	−0.06	−0.06	−0.37	0.28	0.07	−0.14	1.00
(14)	BIG4	−0.02	0.00	−0.01	−0.02	−0.01	0.46	−0.07	0.11	0.12	−0.10	−0.11	0.11	−0.15	1.00
(15)	OWN	−0.11	−0.08	−0.17	−0.10	0.07	0.23	−0.08	−0.06	0.25	−0.22	−0.01	−0.09	−0.33	0.20	1.00
(16)	FOR	−0.13	−0.08	0.08	0.02	−0.06	0.49	0.01	−0.08	0.22	−0.18	0.05	0.13	−0.22	0.28	−0.02	1.00
(17)	DATE	−0.03	0.05	0.01	−0.02	−0.03	−0.13	0.04	0.14	−0.18	0.18	−0.07	0.23	0.08	−0.12	−0.08	−0.15	1.00
(18)	MK	−0.08	−0.02	−0.17	−0.13	0.05	0.46	−0.06	0.17	0.07	−0.10	−0.11	0.07	−0.33	0.28	0.28	0.23	−0.12	1.00

Note: All reported coefficients are Pearson correlation coefficients. Bold entries denote statistical significance at the 1%, 5%, and 10% levels, respectively.

presents the Pearson correlation coefficients among the primary variables employed in the empirical analysis. The continuous STRATEGY index exhibits a statistically significant positive correlation with DUMACD, indicating that firms with more prospector-oriented strategic postures are more likely to experience auditor–client disagreement—a pattern directionally consistent with H1. The binary PROSPECTOR indicator, however, does not exhibit a statistically significant correlation with either DUMACD or ABSACD, suggesting that the association between prospector-type classification and disagreement may be better captured by the continuous STRATEGY index than by the dichotomous classification. In contrast, DEFENDER exhibits a statistically significant negative correlation with both DUMACD and ABSACD, indicating that firms classified as defender-type are associated with a lower likelihood and smaller magnitude of auditor–client disagreement relative to non-defender firms—consistent with H1-2 and H2-2.

These univariate associations, while informative, must be interpreted with caution, as correlation analysis does not control for the confounding influence of other firm and engagement characteristics. The multivariate regression results reported in Section 4.3 address this limitation by estimating the incremental associations between strategic orientation and auditor–client disagreement after controlling for the full set of covariates identified in prior literature.

To assess the potential for multicollinearity among the independent variables, Variance Inflation Factor (VIF) statistics are computed for all regression specifications. The maximum VIF across all variables is approximately 2.8, substantially below the conventional threshold of 10, confirming that multicollinearity does not pose a material threat to the validity of the regression estimates.

4.3. Regression Results and Hypothesis Testing

Table 6. Effect of Business Strategy on Auditor–Client Disagreement.

DUMACDi,t = β0 + β1STRATEGY(PROSPECTOR, DEFENDER)i,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t + β10OWNi,t +β11OUTCOSTi,t + β12FORi,t + β13DATEi,t + β14MKi,t + ∑YEAR + ∑IND + ε
PANEL A: logistic regression (Dependent = DUMACD)
	(1) Continuous variable model			(2) Dummy variable model
Variable	coefficient	z-statistic	p-value	coefficient	z-statistic	p-value
intercept	6.732 ***	6.68	<0.0001	7.452 ***	7.24	<0.0001
STRATEGY	0.042 ***	4.49	<0.0001
PROSPECTOR				0.072	0.68	0.497
DEFENDER				−0.375 ***	−3.82	<0.0001
SIZE	−0.192 ***	−5.14	<0.0001	−0.183 ***	−5.03	<0.0001
GRW	0.412 ***	3.45	<0.0001	0.442 ***	3.61	<0.0001
LEV	1.152 ***	5.49	<0.0001	1.132 ***	5.39	<0.0001
ROA	−0.487	−1.11	0.267	−0.475	−1.15	0.250
LOSS	0.181 **	2.16	0.031	0.185 **	2.17	0.030
LIQ	−0.029 *	−1.96	0.050	−0.025 *	−1.65	0.099
CON	0.393 ***	4.79	<0.0001	0.391 ***	4.78	<0.0001
BIG4	3.223	0.04	0.968	19.789	0.27	0.787
OWN	0.212 ***	3.01	0.003	0.223 ***	3.15	0.002
OUTCOST	−0.623 ***	−3.18	0.001	−0.657 ***	−3.42	<0.0001
FOR	−1.734 ***	−5.00	<0.0001	−1.705 ***	−4.89	<0.0001
DATE	−0.782 ***	−7.25	<0.0001	−0.782 ***	−7.25	<0.0001
MK	−0.048	−0.76	0.447	−0.077	−1.15	0.250
Fixed Effect	Year and Industry
N	6504
Pseudo R2	0.053			0.053
LR chi2	375.22			375.22
ABSACDi,t = β0 + β1STRATEGY(PROSPECTOR, DEFENDER)i,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t + β10OWNi,t +β11OUTCOSTi,t + β12FORi,t + β13DATEi,t + β14MKi,t + ∑YEAR + ∑IND + ε
PANEL B: OLS regression (Dependent = ABSACD)
	(1) Continuous variable model			(2) Dummy variable model
Variable	coefficient	t-value	p-value	coefficient	t-value	p-value
intercept	0.508 **	2.19	0.029	0.552 **	2.35	0.019
STRATEGY	0.003	0.93	0.352
PROSPECTOR				0.010	0.39	0.697
DEFENDER				−0.040 *	−1.77	0.077
SIZE	−0.020 **	−2.43	0.015	−0.020 **	−2.41	0.016
GRW	0.065 ***	2.62	0.009	0.066 ***	2.63	0.009
LEV	0.213 ***	4.52	<0.0001	0.212 ***	4.52	<0.0001
ROA	0.332 ***	3.72	<0.0001	0.349 ***	3.75	<0.0001
LOSS	0.119 ***	6.17	<0.0001	0.117 ***	6.22	<0.0001
LIQ	−0.003	−0.93	0.352	−0.003	−0.92	0.358
CON	0.022	1.25	0.211	0.022	1.18	0.238
BIG4	0.188 **	2.51	0.012	0.181 ***	2.53	0.011
OWN	0.043 ***	2.61	0.009	0.041 ***	2.61	0.009
OUTCOST	−37.114 ***	−4.08	<0.0001	−37.264 ***	−4.10	<0.0001
FOR	−0.275 ***	−3.43	<0.0001	−0.280 ***	−3.45	<0.0001
DATE	0.008	0.35	0.726	0.007	0.31	0.757
MK	0.039 **	2.41	0.016	0.039	2.37	0.018
Fixed Effect	Year and Industry
N	6504
Adj. R2	0.042			0.042
REALACDi,t = β0 + β1STRATEGY(PROSPECTOR, DEFENDER)i,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t + β10OWNi,t +β11OUTCOSTi,t + β12FORi,t + β13DATEi,t + β14MKi,t + ∑YEAR + ∑IND + ε
Panel C: OLS regression (Dependent = REALACD)
	(1) Continuous variable model			(2) Dummy variable model
Variable	coefficient	t-value	p-value	coefficient	t-value	p-value
intercept	0.302	1.59	0.112	0.296	1.49	0.136
STRATEGY	0.002	0.99	0.322
PROSPECTOR				0.008	0.41	0.682
DEFENDER				−0.049 **	−2.53	0.011
SIZE	−0.015 **	−2.20	0.028	−0.013 **	−2.03	0.042
GRW	0.037 *	1.82	0.069	0.036 *	1.77	0.077
LEV	0.135 ***	3.57	<0.0001	0.136 ***	3.37	<0.0001
ROA	0.006	0.08	0.936	0.015	0.23	0.818
LOSS	0.046 ***	2.85	0.004	0.044 ***	2.92	0.004
LIQ	−0.004	−1.23	0.219	−0.002	−1.34	0.180
CON	0.010	0.64	0.522	0.013	0.65	0.516
BIG4	0.025 **	2.25	0.024	0.025 **	2.00	0.046
OWN	−0.103 ***	−2.91	0.004	−0.104 ***	−2.77	0.006
OUTCOST	28.051 **	2.35	0.019	27.641 **	2.24	0.025
FOR	−0.141 **	−2.12	0.034	−0.173 **	−2.53	0.011
DATE	0.019	0.91	0.363	0.013	0.5	0.617
MK	0.026 *	1.85	0.064	0.028 **	1.95	0.051
Fixed Effect	Year and Industry
N	6504
Adj. R2	0.028			0.028

Note: *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively (two-tailed). Note for Panel A: The economic magnitude of the logistic regression coefficients can be interpreted using the odds ratio (eβ). For instance, the odds ratio for STRATEGY is e^0.042 ≈ 1.043, and for DEFENDER is e^−0.375 ≈ 0.687. The large coefficient and standard error on BIG4 in Panel A are attributable to quasi-complete separation; results are qualitatively unchanged when BIG4 is excluded from the logistic specification.

presents the multivariate regression results examining the association between business strategy and auditor–client disagreement. Results are organized across three panels corresponding to the three dependent variable specifications: Panel A reports logistic regression estimates using DUMACD as the dependent variable; Panels B and C report OLS estimates using ABSACD and REALACD, respectively. Within each panel, Column (1) presents estimates from Model (1) using the continuous STRATEGY index, and Column (2) presents estimates from Model (2) using the binary PROSPECTOR and DEFENDER indicators.

Panel A reports the logistic regression results for the likelihood of auditor–client disagreement. In Column (1), the coefficient on STRATEGY is positive and statistically significant (β = 0.042, p < 0.01), indicating that a one-unit increase in the business strategy index—reflecting a shift toward a more prospector-oriented strategic posture—increases the odds of auditor–client disagreement by approximately 4.3% (odds ratio = e^0.042≈ 1.043). This finding provides robust support for H1. Importantly, the significance of the continuous STRATEGY variable confirms that the relationship between strategic orientation and auditor–client disagreement is monotonic and pervasive across the full strategic continuum—including firms in the intermediate “analyzer” range—rather than being confined to firms at the polar extremes of the distribution.

In Column (2), the coefficient on PROSPECTOR is statistically insignificant, indicating that the binary prospector classification does not independently predict a higher likelihood of disagreement once other firm characteristics are controlled for. H1-1 is therefore not supported. By contrast, the coefficient on DEFENDER is negative and statistically significant (β = −0.375, p < 0.01), corresponding to an odds ratio of e^−0.375 ≈ 0.687. This implies that defender-type firms face a 31.3% lower likelihood of auditor–client disagreement relative to non-defender firms, providing strong empirical support for H1-2.

Panels B and C present the OLS regression results for the magnitude and directionality of auditor–client disagreement, respectively. In Column (1) of both panels, the coefficient on STRATEGY is positive in sign but does not attain statistical significance, failing to support H2. Similarly, in Column (2) of both panels, the coefficient on PROSPECTOR is statistically insignificant, indicating that prospector-type classification is not associated with a larger magnitude of earnings adjustments after controlling for other covariates; H2-1 is therefore not supported. The coefficient on DEFENDER, however, is negative and statistically significant in both Panel B (β = −0.040, p < 0.10) and Panel C (β = −0.049, p < 0.05), confirming that defender-type firms experience significantly smaller audit adjustments in both absolute magnitude and directional terms relative to non-defender firms. These results provide consistent empirical support for H2-2.

The control variable estimates are broadly consistent with prior literature. SIZE exhibits a consistently significant negative association with auditor–client disagreement across all specifications, suggesting that larger firms—despite their greater operational complexity—are associated with smaller audit adjustments, possibly reflecting their more sophisticated internal reporting processes and stronger internal controls. LEV and LOSS exhibit significant positive coefficients across all three panels. CON is significantly positive in Panel A, consistent with evidence that consolidation complexity elevates the likelihood of disagreement, but does not attain significance in Panels B and C. BIG4 is significantly positive in Panels B and C, indicating that Big 4 auditors are associated with larger magnitude adjustments—possibly reflecting their lower tolerance for deviations from applicable standards. Conversely, FOR and DATE are significantly and negatively associated with the likelihood of disagreement (Panel A), indicating that greater foreign investor oversight and longer intervals between internal closing and disclosure are associated with lower incidences of material audit adjustments. OWN exhibits a significant positive association in Panels A and B, consistent with evidence that concentrated ownership amplifies reporting incentives; however, in Panel C, the coefficient on OWN is negative and significant, suggesting that controlling shareholder concentration may constrain the directional magnitude of earnings adjustments—a finding that warrants further investigation.

4.4. Robustness Checks

To address the potential endogeneity concern that unobservable firm-specific characteristics may simultaneously influence both a firm’s strategic orientation and the incidence of auditor–client disagreement—a form of self-selection bias—I employ Propensity Score Matching (PSM) as a supplementary identification strategy. By constructing a matched control group of firms with comparable observable characteristics but differing strategic classifications, PSM enables a more credible comparison of outcomes across strategic types and mitigates the risk that the main findings are driven by systematic pre-existing differences between prospector and defender firms.

The PSM procedure is implemented in two stages. In the first stage, I estimate separate probit regression models to identify the firm-level determinants of prospector-type and defender-type classification. Each probit model includes a comprehensive set of covariates—firm size (SIZE), asset growth (GRW), financial leverage (LEV), profitability (ROA), liquidity (LIQ), auditor type (BIG4), controlling shareholder ownership (OWN), audit fee intensity (OUTCOST), foreign ownership (FOR), and listing market (MK)—along with industry and year fixed effects. The probit estimates, reported in

Table 7. Determinants of Business Strategy (Probit Regression).

PROSPECTOR or DEFENDERi,t = β0 + β1SIZEi,t + β2GRWi,t + β3LEVi,t + β4ROAi,t + β5LOSSi,t + β6LIQi,t + β7CONi,t + β8OWNi,t + β9OUTCOSTi,t + β10FORi,t + β11MKi,t + ∑YEAR + ∑IND + ε
Probit regression
	(1) Dep. = PROSPECTOR			(2) Dep. = DEFENDER
Variable	coefficient	z-statistic	p-value	coefficient	z-statistic	p-value
intercept	−2.582 ***	−2.95	0.003	−0.954	−1.12	0.263
SIZE	0.058 *	1.82	0.069	−0.062 *	−1.73	0.084
GRW	0.316 ***	3.58	<0.000	−0.415 ***	−3.55	<0.0001
LEV	−0.621 ***	−3.13	0.002	−0.032	−0.12	0.904
ROA	−0.274	−0.82	0.412	0.987 **	2.56	0.010
LOSS	0.142 *	1.96	0.050	0.082	1.08	0.280
LIQ	0.035 ***	2.63	0.009	−0.042 ***	−2.76	0.006
CON	0.135 *	1.82	0.069	−0.282 ***	−4.03	<0.0001
OWN	−0.782 ***	−3.98	<0.000	0.521 ***	2.86	0.004
OUTCOST	109.521 **	2.03	0.042	−193.921 ***	−3.12	0.002
FOR	0.763 **	2.51	0.012	−1.881 ***	−4.76	0.000
MK	−0.535 ***	−7.15	<0.000	0.232 ***	3.97	0.000
Fixed Effect	Industry and Year
N	4819
Pseudo R2	0.089			0.077
LR chi2	247.09			242.32

Note: *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively. The probit estimation sample is restricted to the 4819 firm-year observations for which the full set of PSM covariates is non-missing, consistent with the requirement that matching be performed on observed characteristics.

, reveal that strategic classification is systematically associated with firm attributes. Specifically, the likelihood of prospector classification is positively associated with asset growth (GRW) and audit fee intensity (OUTCOST), while the likelihood of defender classification is positively associated with profitability (ROA) and controlling shareholder ownership (OWN), consistent with the theoretical characterization of defender firms as stable, efficiency-oriented entities with concentrated governance structures.

In the second stage, I construct matched treatment and control samples using a one-to-one nearest-neighbor matching algorithm without replacement, based on the estimated propensity scores from the first stage. The covariate balance diagnostics reported in

Table 8. Comparison of Covariates Before and After Propensity Score Matching.

Variable	Prospector		Difference	t-Statistic
	Treatment (Mean)	Control (Mean)
SIZE	25.902	25.883	−0.019	−0.21
GRW	0.132	0.099	−0.034	−1.58
LEV	0.384	0.386	0.000	0.03
LOSS	0.403	0.390	−0.011	−0.35
LIQ	3.017	3.030	0.010	0.03
CON	0.841	0.849	0.005	0.24
OUTCOST	0.0006	0.0007	0.000	0.14
OWN	0.332	0.333	0.001	0.12
FOR	0.073	0.078	0.002	0.35
MK	0.202	0.190	−0.013	−0.51
Variable	Defender		Difference	t-Statistic
	Treatment (Mean)	Control (Mean)
SIZE	26.115	26.097	−0.014	−0.17
GRW	0.049	0.047	0.000	0.00
ROA	0.008	0.003	−0.006	−1.28
LIQ	2.018	1.951	−0.067	−0.56
CON	0.753	0.779	0.023	0.92
OUTCOST	0.0005	0.0005	0.000	1.32
OWN	0.412	0.423	0.011	0.92
FOR	0.045	0.041	−0.003	−0.91
MK	0.467	0.475	0.005	0.23

Note: Matching covariates for each strategy type are selected based on the statistically significant determinants identified in the first-stage probit estimation.

confirm that the matching procedure is effective: following matching, the mean differences in all observable characteristics between prospector (defender) firms and their respective matched control firms are statistically indistinguishable, as evidenced by insignificant t-statistics across all covariates. This balance verification supports the validity of the matched samples as a basis for causal inference.

The PSM regression results, reported in

Table 9. Regression Results of Auditor–Client Disagreement Using PSM Sample.

DUMACDi,t = β0 + β1PROSPECTOR(or DEFENDER)i,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t + β10OWNi,t + β11OUTCOSTi,t + β12FORi,t + β13DATEi,t + β14MKi,t + ∑YEAR + ∑IND + ε
PANEL A: logistic regression (Dependent = DUMACD)
	(1) PROSPECTOR			(2) DEFENDER
Variable	coefficient	z-statistic	p-value	coefficient	z-statistic	p-value
intercept	18.663 ***	5.13	<0.0001	3.784	5.74	<0.0001
PROSPECTOR	0.069	0.42	0.674
DEFENDER				−0.313 **	−2.05	0.040
SIZE	−0.503 ***	−4.24	<0.0001	0.015	0.11	0.912
GRW	0.729 **	2.28	0.023	1.033 **	2.23	0.026
LEV	1.217 *	1.96	0.050	1.377 **	2.25	0.024
ROA	0.483	0.48	0.631	1.391	1.07	0.285
LOSS	0.484 *	1.96	0.050	0.595 ***	2.85	0.004
LIQ	−0.049	−1.37	0.171	0.075	1.35	0.177
CON	0.483 *	1.95	0.051	0.368 *	1.81	0.070
BIG4	0.503 **	2.49	0.013	0.203	1.25	0.211
OWN	−1.314 **	−2.12	0.034	−1.051 **	−1.97	0.049
OUTCOST	−102.639	−0.52	0.603	279.574	1.53	0.126
FOR	−0.704	−0.88	0.379	−0.938	−0.77	0.441
DATE	−1.058 ***	−3.22	0.001	−1.316 ***	−4.73	<0.0001
MK	0.007	0.08	0.936	−0.152	−0.85	0.395
Fixed Effect	Industry and Year
N	832			853
Pseudo R2	0.113			0.098
LR chi2	104.82			120.26
ABSACDi,t = β0 + β1PROSPECTOR(or DEFENDER)i,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t + β10OWNi,t + β11OUTCOSTi,t + β12FORi,t + β13DATEi,t + β14MKi,t + ∑YEAR + ∑IND + ε
PANEL B: OLS regression (Dependent = ABSACD)
	(1) PROSPECTOR			(2) DEFENDER
Variable	coefficient	t-value	p-value	coefficient	t-value	p-value
intercept	1.403 **	1.95	0.051	0.842	1.45	0.147
PROSPECTOR	0.037	1.53	0.126
DEFENDER				−0.037	−1.32	0.187
SIZE	−0.048 **	−1.94	0.052	−0.031	−1.51	0.131
GRW	0.038	0.61	0.542	0.133 *	1.75	0.080
LEV	0.467 ***	3.72	<0.0001	0.208 *	1.86	0.063
ROA	0.707 ***	3.58	<0.0001	0.525 **	2.15	0.032
LOSS	0.178 ***	3.51	<0.0001	0.188 ***	4.57	<0.0001
LIQ	0.004	0.35	0.726	0.026 ***	2.75	0.006
CON	0.068	1.23	0.219	0.082 **	2.17	0.030
BIG4	0.087 **	2.11	0.035	0.085 **	2.56	0.011
OWN	−0.215 *	−1.72	0.085	−0.095	−0.95	0.342
OUTCOST	46.511	1.38	0.168	26.854	0.81	0.418
FOR	−0.099	−0.55	0.582	−0.101	−0.48	0.631
DATE	−0.017	−0.25	0.803	−0.038	−0.65	0.516
MK	0.068	1.11	0.267	0.012	0.59	0.555
Fixed Effect	Industry and Year
N	832			853
Adj. r-squared	0.098			0.046
REALACDi,t = β0 + β1PROSPECTOR(or DEFENDER)i,t + β2SIZEi,t + β3GRWi,t + β4LEVi,t + β5ROAi,t + β6LOSSi,t + β7LIQi,t + β8CONi,t + β9BIG4i,t + β10OWNi,t + β11OUTCOSTi,t + β12FORi,t + β13DATEi,t + β14MKi,t + ∑YEAR + ∑IND + ε
PANEL C: OLS regression (Dependent = REALACD)
	(1) PROSPECTOR			(2) DEFENDER
Variable	coefficient	t-value	p-value	coefficient	t-value	p-value
intercept	0.237	0.38	0.704	0.345	0.71	0.478
PROSPECTOR	0.017	0.68	0.497
DEFENDER				−0.056 **	−2.45	0.014
SIZE	−0.003	−0.05	0.960	−0.014	−0.87	0.384
GRW	0.025	0.44	0.660	0.018	0.25	0.803
LEV	0.315 ***	2.91	0.004	0.115	1.33	0.184
ROA	0.367 **	2.25	0.024	−0.105	−0.51	0.610
LOSS	0.121 ***	2.86	0.004	0.061 *	1.81	0.070
LIQ	0.002	0.15	0.881	0.025 ***	2.65	0.008
CON	−0.005	−0.12	0.905	0.021	0.65	0.516
BIG4	0.048	1.31	0.190	0.048	1.57	0.116
OWN	−0.046	−0.41	0.682	−0.125	−1.55	0.121
OUTCOST	42.485	1.45	0.147	15.251	0.51	0.610
FOR	−0.297 *	−1.93	0.054	0.015	0.08	0.936
DATE	−0.009	−0.15	0.881	0.013	0.25	0.803
MK	−0.055	−1.25	0.211	0.031	1.11	0.267
Fixed Effect	Industry and Year
N	832			853
Adj. r-squared	0.062			0.031
Adj. R2	0.028			0.028

Note: *, **, and *** denote statistical significance at the 10%, 5%, and 1% levels, respectively.

, are largely consistent with the main findings. In Panel A, the coefficient on DEFENDER remains negative and statistically significant (β = −0.315, p < 0.05), confirming that defender-type firms face a significantly lower likelihood of auditor–client disagreement even after accounting for potential selection bias. In Panel C, the magnitude of disagreement as captured by REALACD also remains significantly lower for defender-type firms (β = −0.059, p < 0.05). The PROSPECTOR indicator remains statistically insignificant in the categorical specification across all panels, a result that may reflect the inherently high variance in audit negotiation outcomes associated with the complex and uncertain operating environments characteristic of prospector firms. Collectively, the PSM evidence reinforces the conclusion that a defender-type strategic orientation constitutes a robust, systematic mitigating factor for auditor–client disagreement, while the prospector–disagreement association is better captured by the continuous STRATEGY index than by the binary classification.

In addition to the PSM analysis, I conduct several untabulated sensitivity tests to evaluate the robustness of the primary findings to alternative measurement choices. First, I re-estimate the baseline models using stricter materiality thresholds of 0.5% and 1.0% to define DUMACD, thereby ensuring that the binary disagreement indicator captures only more economically significant audit adjustments. Second, I employ an alternative scaling approach in which the continuous disagreement measures are deflated by lagged total assets rather than audited net income, alleviating concerns about measurement distortion for firm-years with near-zero or negative net income. Across all alternative specifications, the direction, magnitude, and statistical significance of the primary results remain qualitatively consistent with the main findings, providing additional assurance regarding the robustness of the empirical evidence.

5. Conclusions

This study examines the association between business strategy and auditor–client disagreement using a sample of 6504 firm-year observations drawn from firms listed on the Korea Exchange (KOSPI and KOSDAQ) over the period 2020–2024. The empirical results yield two principal findings. First, firms with higher strategic index scores—reflecting a more prospector-oriented posture—exhibit a significantly greater likelihood of auditor–client disagreement, as evidenced by the positive and statistically significant coefficient on the continuous STRATEGY variable. Second, firms classified as defender-type demonstrate both a significantly lower likelihood and a smaller magnitude of auditor–client disagreement relative to non-defender firms, a finding that is robust to the PSM-based endogeneity correction and alternative measurement specifications.

These findings are consistent with the theoretical proposition that a firm’s overarching strategic orientation structurally shapes its corporate risk profile and the complexity of the external audit process (Habib et al. 2024). Prospector firms, operating in highly dynamic and uncertain competitive environments, routinely engage in transactions that require highly subjective accounting estimates—such as the capitalization of R&D expenditures, the valuation of newly acquired intangibles, and complex revenue recognition—thereby elevating inherent risk at the assertion level and compelling auditors to deploy substantially greater effort. The resulting tension between management’s aggressive reporting incentives and the auditor’s obligation to apply rigorous professional skepticism manifests in a higher frequency and larger magnitude of pre-audit to post-audit earnings adjustments. Defender firms, by contrast, benefit from the inherent predictability of their centralized, process-driven organizational structures, which constrain the scope for managerial discretion, suppress inherent risk, and facilitate more transparent and less contentious audit environments.

This study makes three contributions to the extant literature and practice. First, it establishes business strategy as a fundamental, ex-ante determinant of inherent and audit risk, shifting the analytical focus of auditor–client disagreement research from reactive, governance-based monitoring mechanisms to the client’s underlying strategic posture. Second, it extends the business strategy literature (Bentley et al. 2013; Habib et al. 2024) by providing direct empirical evidence that strategic orientation influences not only audit effort and fees, but also the observable outcome of auditor–client negotiations as reflected in earnings adjustments. Third, the findings provide empirical validation for the risk-based auditing framework mandated by International Standard on Auditing (ISA) 315, establishing that a client’s strategic classification constitutes a material, firm-level risk factor that should be explicitly incorporated into engagement risk assessments and audit planning procedures. For standard-setters and regulatory bodies, the results suggest that strategic analysis should be recognized not as a supplementary consideration, but as a foundational input into the auditor’s risk identification process.

6. Limitations and Directions for Future Research

Notwithstanding its contributions, this study is subject to several limitations that warrant acknowledgment and suggest productive directions for future research.

First, the measurement of auditor–client disagreement relies on the manual collection of preliminary unaudited net income from mandatory profit structure change disclosures filed with DART. Under the Securities Market Disclosure Regulations, such filings are triggered only when revenues or net income deviate by 30% or more from the prior year (or 15% for large corporations). By construction, therefore, the sample captures only highly material audit adjustments that meet this disclosure threshold, while firm-years in which auditor–client negotiations result in smaller, sub-threshold earnings revisions are necessarily excluded. This selection mechanism implies that the estimated associations may reflect the determinants of material disagreement rather than auditor–client friction more broadly. Future research could address this limitation by employing advanced natural language processing (NLP) or text-mining techniques applied to audit committee reports, management commentary disclosures, and internal audit documentation to construct a more comprehensive and granular measure of audit adjustment activity across the full materiality spectrum.

Second, the empirical analysis is confined to firms listed on the Korea Exchange, and the findings are thus conditioned on the specific institutional, regulatory, and cultural context of the Korean capital market. Business strategy typologies and the dynamics of auditor–client negotiations are likely to vary systematically across jurisdictions with differing legal origins, auditor liability regimes, and enforcement environments. Future research should examine whether the strategic risk dynamics documented in this study generalize to other international settings—particularly across common law versus civil law jurisdictions and across developed versus emerging capital markets—to establish the external validity of these findings.

Third, this study focuses on the pre-issuance phase of the audit process, specifically the negotiation dynamic that produces observable pre-audit to post-audit earnings adjustments. Future research could substantially extend this scope by investigating the post-negotiation consequences of auditor–client disagreement. Relevant questions include whether severe or recurrent disagreements precipitate auditor turnover events—including both client-initiated dismissals and auditor-initiated resignations—and how such transitions are interpreted by capital market participants. Additionally, examining whether particular strategic orientations are systematically associated with a higher propensity for financial statement fraud or managerial collusion would provide important insights into the broader implications of strategic risk for audit quality, corporate governance, and investor protection.