Internal Control Quality and Leverage Manipulation: Evidence from Chinese State-Owned Listed Companies

Abstract

Promoting structural deleveraging is a key strategy for China to reduce high debt levels and mitigate systemic financial risks. In this context, the deleveraging of state-owned enterprises (SOEs) has become a national strategic priority. This study explores whether enhancing the quality of internal control as an internal governance mechanism can facilitate the deleveraging process of SOEs. Using a sample of A-share state-owned listed companies from the Shanghai and Shenzhen stock exchanges (2009–2023) and based on resource-based theory and signaling theory, we examine the impact and mechanisms through which internal control quality influences SOE leverage reduction. Our results demonstrate that higher internal control quality significantly promotes deleveraging in SOEs, and these findings remain robust after conducting endogeneity tests and employing alternative model specifications. Improved internal control mitigates resource misallocation and encourages firms to adopt two primary strategies: debt reduction (through short-term liability repayment and retained earnings) and equity expansion. However, the positive effect diminishes as Research and Development (R&D) intensity increases, reflecting the trade-off between innovation-driven growth and financial stability. Further heterogeneity analyses reveal that the deleveraging effect is more pronounced in local SOEs and over-indebted SOEs, as enhanced internal control helps eliminate non-performing liabilities. This study contributes to the literature on the economic consequences of internal control and provides empirical insights for policymakers seeking to optimize the capital structures of SOEs.

1. Introduction

State-owned enterprises (SOEs) are the backbone of China’s economy, and reducing their leverage ratios has become a central focus of supply-side structural reforms [1]. In 2017, the National Financial Work Conference prioritized the deleveraging of SOEs, underscoring its critical role in mitigating financial risks. Similarly, the inaugural meeting of the Financial and Economic Commission of the CPC Central Committee in 2018 emphasized the importance of “structural deleveraging” and mandated that SOEs systematically reduce their leverage ratios. Furthermore, the 20th National Congress of the Communist Party of China reaffirmed the necessity of preventing and defusing financial risks, deepening SOE reforms, and ensuring that SOEs become stronger, more efficient, and larger in scale. These policy directives clearly highlight the strategic importance of SOE deleveraging within China’s broader economic agenda.

In August 2024, the “Notice on Regulating the Withdrawal of Financing Platform Companies”, jointly issued by the People’s Bank of China and three other ministries and commissions, explicitly stated that “achieving zero hidden debt” is the primary criterion for exiting the platform. Debt reduction in SOEs is indeed a central component of the economic task of “deleveraging”; however, the leverage ratio of SOEs remains persistently high. Data from the National Laboratory of Finance and Development indicate that the macro leverage ratio of government sectors was 60.3% in the third quarter of 2024, with a total debt balance of RMB 78.4 trillion. When hidden debts are accounted for, the total debt of government sectors surges to RMB 138.4 trillion, pushing the leverage ratio to 106%. This persistent high leverage ratio is primarily attributable to deleveraging strategies that often involve concealing the true debt situation through debt-to-equity swaps, which fail to resolve underlying issues and, instead, exacerbate debt risks [2]. The high leverage ratio of SOEs not only adversely affects individual firm performance but also poses significant macro-financial risks at the national level, thereby hindering the achievement of high-quality development in the state-owned economy [3]. Consequently, effectively promoting deleveraging in SOEs is a critical issue that demands urgent attention to deepen SOE reforms and support high-quality economic development in China.

Existing literature predominantly examines external drivers of deleveraging, such as tax policies [4], monetary interventions [5], and trade liberalization [6]. Some studies have also explored micro-level factors, including non-state-owned shareholder governance and corporate risk [7,8]. However, internal governance mechanisms, particularly internal control systems, remain underexplored. Internal control, as an internal governance mechanism, is centered on risk management and serves as a primary means for SOEs to effectively prevent risks and regulate the exercise of power [9]. The current literature mainly focuses on the factors influencing the quality of internal control and its economic consequences, such as corporate accounting information quality, earnings persistence, and the stability of business–customer relationships [10]. Nevertheless, systematic research examining whether improving internal control quality can reduce the leverage ratio of SOEs and how it impacts their deleveraging process is scarce. Moreover, as China is undergoing a critical phase of economic transformation driven by innovation, R&D intensity—an important internal contextual factor—also plays a crucial role in influencing the deleveraging process of SOEs and the effectiveness of internal control implementation. Therefore, this paper addresses two key questions: (1) How does the improvement in internal control quality affect SOE deleveraging? (2) How does R&D intensity, as a key innovation-driven contextual factor, moderate this relationship?

Based on the aforementioned, this article adopts the perspective of enhancing internal control quality and focuses on China’s A-share state-owned listed companies from Shanghai and Shenzhen, spanning the period from 2009 to 2023. Grounded in the resource-based view and signaling theory, the study conducts a comprehensive analysis of how internal control quality influences the deleveraging process of SOEs. It investigates the mechanisms through which internal control quality impacts SOE deleveraging, considering factors such as resource misallocation and deleveraging strategies. Furthermore, it examines the moderating role of R&D intensity in shaping the relationship between internal control quality and SOE deleveraging. The ultimate aim is to contribute to the academic literature on the economic consequences of SOE deleveraging and internal control and to offer practical recommendations for promoting deleveraging within SOEs in the context of China’s innovation-driven development strategy.

Compared to the existing literature, the potential innovations of this article are outlined below: First, it extends research on corporate deleveraging. While existing studies primarily explore deleveraging from an external macroeconomic perspective, this article adopts a focus on enhancing internal control quality to identify novel drivers of deleveraging in SOEs. Second, through the lens of resource misallocation, it offers a new framework for understanding how internal control quality influences SOE deleveraging. Furthermore, it explores shifts in deleveraging strategies employed by SOEs following improvements in internal control. These strategies are categorized into two dimensions: “debt reduction” and “capacity enhancement”, demonstrating that the quality of internal control can guide SOEs toward different deleveraging approaches, thereby expanding the research on deleveraging methods and their influencing factors. Third, it provides an in-depth analysis of the optimization mechanism underlying the impact of internal control quality on deleveraging. It verifies the moderating effect of endogenous enterprise factors—specifically R&D intensity—on this pathway and differentiates the deleveraging effects among central SOEs, local SOEs, and SOEs with varying levels of debt when internal control quality improves. This analysis offers practical policy insights for the Chinese government to further promote deleveraging and optimize deleveraging strategies.

2. Theoretical Background Research Hypothesis

2.1. Internal Control Quality and Deleveraging of SOEs

The inherent political characteristics and management structures of SOEs often give rise to phenomena such as “absentee ownership” and “soft budget constraints”, which can lead to agency conflicts and moral hazards [11]. These issues impair profitability and may trigger debt crises that hinder the deleveraging process. Excessive leverage within SOEs negatively affects various internal operations, reducing profitability and obstructing the pursuit of high-quality development. Existing research predominantly examines the factors influencing SOE deleveraging from an external macroeconomic policy perspective [12]. However, the intrinsic “political attributes” of these enterprises often render external policy factors less effective. The inadequacy of traditional internal management and control systems is a key contributor to asset losses and reduced profitability, thus affecting the deleveraging process. In response, some scholars have proposed introducing non-state shareholders at the micro level to strengthen internal supervision, reduce agency costs, and enhance endogenous financing capabilities, thereby facilitating structural deleveraging [13]. A successful deleveraging strategy requires a long-term mechanism that addresses both internal governance constraints on excessive borrowing and external macro policies. Therefore, it is essential for SOEs to establish a sustainable governance framework aimed at achieving leverage reduction. In March 2023, the State-owned Assets Supervision and Administration Commission (SASAC) issued a notice on “the construction and supervision work related to central enterprise internal control systems”, explicitly stating its goal to “further improve internal control management levels, more effectively prevent major risks, and support the achievement of strategic goals”, thus driving high-quality enterprise development. In this context, can robust internal controls—not only serve as a corporate governance tool to standardize managerial behavior but also mitigate enterprise risks and enhance operational performance, thereby supporting the deleveraging efforts of SOEs [14,15,16]?

A review of relevant literature on internal control, alongside resource-based theory and signaling theory, suggests that under conditions conducive to high-quality development, improving internal control quality may facilitate the deleveraging process [17].

Firstly, from a resource-based perspective, the heterogeneous institutional resources and managerial capabilities within SOEs are critical factors for achieving competitive advantages and sustaining long-term growth [18]. Within this framework, effective internal controls are considered vital institutional resources that influence both profitability and the sustainability of growth trajectories for SOEs [19]. High-quality internal controls foster the establishment of a sound organizational environment by assisting SOEs at the macro level in developing scientifically valid organizational structures and strategic planning systems. At the micro level, internal controls regulate departmental workflows and employee behavioral norms, reducing strategic risks associated with broader operations [15,20]. In terms of organizational resources, improving internal control quality enhances the integration of human capital, cultural elements, financial management practices, and other key resources, thereby systematizing resource utilization and improving efficiency, which ultimately contributes to achieving strategic objectives [21]. This not only strengthens governance capacities and internal operating environments but also optimizes asset structures, reducing operational and financial risks, which is conducive to advancing necessary deleveraging initiatives.

Secondly, from the perspective of signaling theory, high-quality internal controls not only help dismantle information barriers but also convey positive signals to external markets and investors, demonstrating management’s commitment to the long-term stable development of SOEs. Information barriers are prevalent in SOEs. Ko, C. et al. highlighted that management with informational advantages may engage in opportunistic behavior, potentially leading to corruption [22]. On one hand, improving internal control quality can reduce information asymmetry within SOEs, bridging the gap between internal management and external investors or creditors. This mitigates inefficiencies related to agency problems, such as suboptimal investments and financial fraud, while enhancing operational performance within SOEs [23]. On the other hand, improving internal control quality is accompanied by an enhancement in the quality of financial information, which further transmits positive signals about the authenticity and reliability of financial data externally [16]. This contributes to a stronger corporate market reputation and credibility, alleviating concerns among external investors and creditors regarding enterprise risks, lowering financing costs, and broadening financing channels [24,25]. These factors create more favorable conditions for enterprises to reduce liabilities and promote deleveraging efforts. Based on the above analysis, this paper proposes the following research hypothesis.

Hypothesis 1.

Improved internal control quality promotes deleveraging in SOEs.

2.2. Internal Control, Resource Misallocation, and State-Owned Enterprise Deleveraging

Resource misallocation within enterprises refers to the unequal distribution of production factors across departments or projects, ultimately resulting in reduced production efficiency and increased leverage ratios. The imperfections in external markets often hinder optimal resource allocation, a phenomenon that is particularly pronounced in SOEs. As the “immune system” of SOEs, improving internal control quality plays a key role in enhancing corporate governance standards, reducing internal resource “friction”, and thereby mitigating resource misallocation [26].

Upon examining the literature on internal control, the resource-based view, and signaling theory, this paper argues that in the context of high-quality development, the quality of internal control significantly influences the intermediary mechanism through which SOEs deleverage, primarily by improving resource allocation efficiency [21]. From a resource-based perspective, internal control constitutes a critical institutional and organizational resource for SOEs, and enhancing its quality strengthens their sustainable development capabilities. Firstly, high-quality internal controls regulate leadership behavior by clearly defining responsibilities and implementing effective internal audits. This deters opportunistic behaviors by both the board of directors and management, which might otherwise lead to resources being concentrated in inefficient departments or projects [14]. In essence, it “reins in power”, reducing resource misallocation. Secondly, by establishing an efficient risk assessment mechanism, organizations can quickly identify risks during the resource allocation process. This allows managers to adjust operational strategies in alignment with established objectives—reducing investments in high-risk ventures while increasing allocations to low-risk, high-return opportunities—thereby mitigating resource misallocation caused by inadequate risk assessments [15]. Lastly, internal controls promote rigorous budget management practices that ensure the rational distribution of resources across various projects and business units, while also strengthening oversight of budget execution and adjustments. This ensures that resources are utilized effectively and prevents issues related to the excessive concentration or inadequate distribution of resources.

Based on signaling theory, enhancing the quality of internal controls can improve the quality and transparency of corporate information disclosure [26]. On one hand, it helps convey positive signals to external stakeholders about “standardized business operations” and “sustained transactions”. These signals mitigate information asymmetry and strengthen investors’ and creditors’ confidence in the company’s long-term viability, thereby reducing their risk appetite [27,28]. As a result, it attracts more partners and high-quality resources for enterprises, alleviating financing pressures and securing essential resources for operational development, which aids in addressing resource misallocation [29]. On the other hand, effective information disclosure provides decision-makers with accurate and comprehensive data relevant to resource allocation [30]. It helps management understand resource needs and utilization efficiency across different business units, enabling them to make more informed decisions regarding resource distribution. In summary, improving internal control quality further standardizes decision-making processes within SOEs, ensuring that all business activities are carried out at a high standard [31]. This optimization facilitates internal resource adjustments and enhances allocation efficiency, ultimately improving overall performance and addressing issues related to resource misallocation.

When internal resources in SOEs are misallocated, it can lead to a decrease in total factor productivity [32], significantly impacting the operational efficiency and effectiveness of these enterprises, thus hindering their deleveraging efforts. The mitigation of resource misallocation’s impact on SOE deleveraging is primarily reflected in the following ways: First, by rationally allocating resources [21], SOEs can redirect resources from inefficient departments to more productive ones. This enhances production efficiency, improves product quality, and reduces costs, ultimately increasing revenue and profits for the enterprise. These improvements generate additional funds for debt repayment, thereby facilitating leverage reduction [33]. Second, by minimizing ineffective investments through a scientific decision-making mechanism for resource allocation, SOEs can more accurately evaluate the feasibility and profitability of investment projects. This helps avoid blind investments and redundant construction activities, which could worsen debt burdens due to excessive spending [34]. Third, effective resource allocation improves the operational and financial health of SOEs, leading to higher credit ratings. This improvement sends positive signals to external stakeholders, boosting investor and creditor confidence and attracting more investment. As a result, enterprises can access more advantageous financing options, optimizing their capital structure and reducing leverage ratios. Based on the above analysis, this paper proposes the following research hypothesis.

Hypothesis 2.

Enhanced Internal control quality facilitates deleveraging by reducing resource misallocation.

2.3. Deleveraging Approaches

The enhancement of internal control quality within enterprises not only influences the deleveraging behavior of SOEs through the mechanism of resource misallocation but also affects the methods employed for deleveraging, ultimately determining the extent of leverage reduction. From an “empowerment” perspective, improving internal control quality serves as a critical driver for SOEs to reduce leverage ratios by increasing owners’ equity [35]. A robust internal control system establishes strong institutional foundations in key areas such as financial management, risk prevention, and information disclosure, standardizing operational processes and significantly enhancing corporate operational efficiency and profitability [36]. This improvement facilitates the steady growth of retained earnings. As a result, investors’ confidence in the company’s credibility grows, leading to reduced financing costs for SOEs. This, in turn, enables them to expand owners’ equity through mechanisms such as issuing new shares or attracting strategic investors. These actions not only directly strengthen capital capacity and reduce leverage ratios but also optimize capital structure and improve firms’ risk resilience [15]. Importantly, high-quality internal controls ensure transparency and regulatory compliance throughout the empowerment process, mitigating potential opportunistic behavior by management that could result in excessive empowerment. This safeguards the scientific validity and effectiveness of empowerment strategies.

Moreover, when adopting a “debt reduction” approach, improvements in internal control quality play an equally indispensable role. Debt reduction strategies involve a range of measures such as debt repayment, debt restructuring, and the optimization of debt structures to achieve sustainable reductions in leverage ratios. A robust internal control system enhances the accuracy of debt management and improves monitoring efficacy within enterprises, ensuring the rational allocation of debt and the establishment of effective repayment mechanisms [37]. Specifically, internal controls help enterprises develop comprehensive debt repayment plans that mitigate default risks associated with excessive indebtedness. Furthermore, through precise cash flow management and forward-looking financial forecasts, companies can secure adequate funds to meet debt obligations while avoiding potential liquidity crises. Internal controls also foster the exploration of innovative solutions, such as debt restructuring, which helps optimize debt structures and lower financing costs, thereby reducing overall debt burdens. Consequently, SOEs can efficiently reduce leverage ratios through “debt reduction” strategies while enhancing their debt repayment capacity. Based on these observations, the following hypothesis is proposed:

Hypothesis 3.

The improvement of internal control quality can encourage enterprises to adopt two approaches—“empowerment” and “debt reduction”—thereby promoting deleveraging in state-owned enterprises.

2.4. The Moderating Role of R&D Intensity

In the context of innovation-driven, high-quality development, R&D intensity serves as a crucial internal factor that enhances the innovation capabilities of SOEs, thereby contributing to their sustained growth and advancing social progress [38]. While previous discussions have emphasized that improving internal control quality may support SOEs in their deleveraging efforts, it is essential to recognize that increased investment in innovation—manifested through higher R&D intensity—can, to some extent, exacerbate financial burdens and impede the deleveraging process. This paper explores how variations in R&D intensity, as a key internal contextual factor, influence the relationship between internal control quality and deleveraging within SOEs.

From a resource-based perspective, variations in R&D intensity within the internal context of SOEs can give rise to distinct mechanisms for resource acquisition. On one hand, firms with high R&D intensity often face greater uncertainty and volatility in their innovation and R&D investments, necessitating access to a more substantial pool of financial resources to support these endeavors [39]. As a result, the effect of improved internal control quality on deleveraging efforts may be somewhat moderated. On the other hand, enterprises with lower R&D intensity typically exhibit more stable cash flows and asset structures; their operational activities tend to be more predictable and do not require as much external financing for innovation and R&D as their high-intensity counterparts [40]. Additionally, internal control functions as a rigorous internal governance mechanism that emphasizes regulation and constraint, often conflicting with the flexible, high-risk-tolerance environment conducive to corporate innovation [41]. In such contexts, firms must strike a delicate balance between R&D intensity and internal controls to effectively achieve sustainable deleveraging outcomes. Consequently, internal control quality is likely to have a more direct influence on capital structure management and debt repayment in non-R&D-intensive firms, thereby facilitating the realization of deleveraging objectives.

Based on signaling theory, variations in R&D intensity, as an internal contextual factor, can lead to differences in how signals are transmitted among SOEs [42]. As the R&D intensity of an enterprise increases, SOEs may face heightened investment and financial pressures. This scenario acts as a prominent signal, indicating that management is placing greater emphasis on innovation and future growth prospects. In this context, investors and creditors are likely to focus more on the firm’s innovation potential rather than improvements in internal control quality. Therefore, while enhancements in internal control quality typically generate positive signals about a firm’s deleveraging efforts, under conditions of high R&D intensity, investors and creditors may prioritize innovation-related signals over those related to internal control improvements. This shift in focus could attenuate the impact of internal control quality on deleveraging outcomes. As a result, when formulating deleveraging strategies, SOEs must carefully navigate the balance between different signals to maintain investor and creditor confidence. In light of these considerations, we propose the following research hypothesis:

Hypothesis 4.

R&D intensity negatively moderates the relationship between internal control quality and deleveraging.

3. Materials and Methods

3.1. Sample Selection

This paper uses a sample of A-share state-owned listed companies from the Shanghai and Shenzhen stock exchanges, covering the period from 2009 to 2023. The selection of 2009 as the starting point is based on China’s implementation of new enterprise accounting standards in 2006 and the substantial completion of the reform of non-tradable shares in the capital market by 2007. These reforms are expected to have influenced corporate accounting practices and financing policies, which could affect the results of this study. Additionally, to mitigate the potential impact of the 2008 financial crisis, data from 2008 and earlier were excluded. Corporate financial data were sourced from the CSMAR database, while internal control data were obtained from the DIB database. The data screening and exclusion process followed these criteria: (1) Companies with ST or *ST designations were excluded. (2) Companies in the delisting consolidation phase, suspended from listing, or terminated from listing were removed. (3) Companies exhibiting significant outliers were eliminated. (4) Samples from financial institutions were excluded. (5) Companies with missing key financial indicators were omitted. After these exclusions, the final sample consists of 1177 A-share state-owned listed companies, yielding a total of 13,458 observations. To minimize the impact of extreme values on the results, all continuous variables were winsorized at the 1% level.

3.2. Variable Definition

3.2.1. Explained Variable

Degree of Deleveraging (lev). This study aims to identify the dynamic behavior of state-owned enterprises in actively adjusting their capital structure. Traditional static leverage ratios are insufficient in effectively capturing the direction and intensity of changes in a firm’s debt levels over consecutive decision-making periods. To accurately measure “deleveraging” as an active debt management behavior, this paper develops a dynamic incremental metric: the annual growth rate of the book leverage ratio. This metric is defined as the change in the leverage ratio between the current and previous periods. The annual growth rate of the book leverage ratio more precisely reflects the extent of deleveraging by isolating the impact of firm size heterogeneity through the relative change in leverage over consecutive periods, thus focusing on the marginal trend of debt adjustment. Positive values indicate a tendency for debt expansion, negative values represent efforts to reduce debt, and a value of zero corresponds to a stable debt level, effectively distinguishing between passive fluctuations in leverage and active deleveraging decisions. Additionally, by using the book value-based leverage ratio, this approach mitigates the distortionary effects of short-term capital market sentiment, liquidity shocks, and other exogenous disturbances on the debt-equity ratio, ensuring that the metric reflects the real debt management behavior at the operational level of the firm.

3.2.2. Core Explanatory Variable

The quality of internal control (ICQ) in this study is measured using the DIBO internal control index, as proposed in the works of Zhao et al. [43] and Huang et al. [44]. This index is based on the framework of five major objectives of internal control, and it constructs a multi-level evaluation system around five key dimensions: the ability to achieve strategic goals, the effectiveness of improving operational efficiency, the reliability of financial reporting, the security mechanisms for asset protection, and the level of compliance with regulatory requirements. In the design of the evaluation indicators, the initial evaluation module is developed based on the “Basic Norms for Enterprise Internal Control” and its supporting guidelines. A hierarchical and dimension-based decomposition of indicators is employed to establish quantitative evaluation standards. Additionally, a defect correction mechanism is introduced by incorporating significant internal control deficiencies disclosed by the firm as dynamic adjustment factors. These deficiencies are integrated into the evaluation system, with negative correction coefficients applied to adjust the base score downward, thereby enhancing the sensitivity of the evaluation results to substantial internal control weaknesses.

3.2.3. Control Variables

This paper draws on the research by Ma et al. [45], Li et al. [46], and Wang et al. [47] to select relevant control variables, including management ability (manage), firm age (age), asset structure (tang), the proportion of intangible assets (itang), financial liabilities (finlev), cash flow (cflow), Tobin’s Q (tobin), market ratio (mbratio), return on total assets (roa), and return on equity (roe). The specific definitions of these variables are provided in

Table 1. Variable definitions.

Variable Type	Variable Name	Variable	Variable Definitions
Explanatory variables	Leverage ratio of SOEs	lev	Total liabilities/Total assets
Core explanatory variables	Internal control quality	icq	The natural logarithm is taken after adding 1 to the DIBO internal control index
Control variables	Management ability	manage	The natural logarithm of total assets
	Enterprise age	age	The number of years the enterprise has been established plus 1 to take the logarithm
	Asset structure	tang	Net profit/Total assets
	Proportion of intangible assets	itang	The shareholding ratio of the largest shareholder
	Financial liability	finlev	Administrative expenses/total assets
	Cash flow	cflow	The difference between control and ownership

.

3.3. Model Construction

To investigate the impact of internal control on deleveraging, this study constructs the following regression model to assess how internal control quality influences the leverage ratio of SOEs.

$$le{v}_{i,t}={\gamma }_0+{\gamma }_{1}ic{q}_{i,t}+\sum \lambda X_{i,t}+{\mu }_i+{\mu }_t+{\epsilon }_{i,t}$$

(1)

In the equation, i represents the firm, t denotes the year, lev indicates the degree of deleveraging of the firm, and icq represents the internal control quality of the firm. The summation term $\sum \lambda X_{i,t}$ represents the set of control variables that affect the firm’s deleveraging degree and vary with both the firm (individual) and time. To eliminate the potential influence of omitted variables that vary across firms, the model also controls for ${\mu }_i$, representing the firm-specific fixed effects, which absorb characteristics that do not change over time at the individual level. To address the potential impact of omitted variables that change over time, the model further controls for ${\mu }_t$, representing the time-fixed effects, which capture time-specific characteristics. Finally, ϵ denotes the random error term. The primary focus of this study is the coefficient ${\gamma }_1$. If ${\gamma }_1$ is negative, it suggests that higher internal control quality is associated with a reduction in the leverage ratio of state-owned enterprises.

4. Data Analysis

4.1. Descriptive Analysis

Table 2. Descriptive statistics.

Variable	Obs	Mean	SD	Min	Median
lev	13,458	0.0324	0.3417	−0.8477	13.9027
icq	13,458	6.2210	1.3073	0.0000	6.9041
manage	13,458	0.0800	0.0800	0.0065	0.8845
age	13,458	2.7344	0.4949	0.0000	3.3673
tang	13,458	0.4068	0.1862	0.0116	0.8076
itang	13,458	0.0501	0.0580	0.0000	0.3045
finlev	13,458	0.4521	0.2526	0.0000	0.9027
cflow	13,458	0.0462	0.0700	−0.1994	0.2646
tobin	13,458	1.8059	1.1802	0.8577	8.4650
mbratio	13,458	0.6957	0.2655	0.1181	1.1660
roa	13,458	0.0277	0.0571	−0.3411	0.2749
roe	13,458	0.0426	0.1656	−1.0354	0.4716

presents the descriptive statistics for the variables used in this study. The leverage ratio (lev) has a mean value of 0.0324, with a standard deviation of 0.3417, and a range spanning from −0.8477 to 13.9027. This wide range indicates substantial variation in the debt-to-asset ratios across the sampled enterprises. The negative skewness suggests that a majority of the firms have relatively lower leverage ratios, highlighting the considerable potential for deleveraging within the sample. This variation emphasizes the importance of identifying the factors that drive deleveraging in SOEs. The descriptive statistics for the other variables are within expected ranges and are not discussed in detail here.

4.2. Benchmark Regression

Table 3. Internal control quality and deleveraging of SOEs.

Variable	(1)	(2)	(3)	(4)
	lev	lev	lev	lev
icq	−0.0349 ***	−0.0273 ***	−0.0270 ***	−0.0221 ***
	(0.0022)	(0.0023)	(0.0065)	(0.0065)
manage		−0.0320		−0.1009
		(0.0397)		(0.0984)
age		−0.0277 ***		−0.0768 ***
		(0.0060)		(0.0164)
tang		−0.0991 ***		−0.0888 ***
		(0.0173)		(0.0317)
itang		−0.0713		−0.2896 *
		(0.0526)		(0.1557)
finlev		0.0271 **		0.0620 **
		(0.0130)		(0.0254)
cflow		0.0679		−0.0553
		(0.0458)		(0.0499)
tobin		0.0175 ***		−0.0059
		(0.0043)		(0.0087)
mbratio		0.0408 **		0.0068
		(0.0195)		(0.0307)
roa		−0.6544 ***		−0.8432 ***
		(0.1015)		(0.2204)
roe		−0.0439		−0.0456
		(0.0331)		(0.0750)
_cons	0.2493 ***	0.2694 ***	0.1982 ***	0.4424 ***
	(0.0142)	(0.0308)	(0.0419)	(0.0851)
Individual fixed effects	NO	NO	YES	YES
Year fixed effects	NO	NO	YES	YES
N	13,458	13,458	13,415	13,415
r2	0.0178	0.0349	0.1483	0.1726
r2_a	0.0177	0.0342	0.0686	0.0944

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

presents the baseline regression results based on a high-dimensional fixed-effects model. To systematically assess the impact of model specifications on the estimation results, this study employs a stepwise regression approach. Model (1) includes only the core explanatory variable, while Model (2) adds firm-level control variables. Model (3) further incorporates individual fixed effects, and the final Model (4) performs a full-variable regression, simultaneously controlling for both individual and time-fixed effects. The empirical results show that the coefficient estimates for the core explanatory variable (icq) consistently range from −0.0221 to −0.0349 and are highly significant at the 1% level. This stability across different model specifications indicates that multicollinearity issues have a limited impact on the core conclusions, suggesting the robustness of the findings. In terms of economic significance, the coefficient of icq in Model (4) is −0.0221, meaning that a one-standard-deviation increase in internal control quality is associated with a decrease in the leverage ratio of state-owned enterprises by approximately 0.0221 standard deviations. This supports the theoretical expectation of a substantial effect. Although the overall explanatory power of the model (R² = 0.1726) is within a reasonable range for micro-level firm studies, it is important to note that the design of the fixed-effects model aims to effectively control for unobservable heterogeneity rather than maximizing goodness-of-fit. Its core value lies in addressing the bias introduced by omitted variables. Additionally, the significance of the control variables and their consistency with theoretical expectations further corroborates the rigor of the model specification. Based on these findings, Hypothesis 1 is supported.

4.3. Robustness Test

4.3.1. Deleveraging Regulatory Policies

To further validate the robustness and reliability of our findings, we conduct a series of robustness tests. A key factor influencing the deleveraging behavior of SOEs is the regulatory environment. Notably, the Central Economic Work Conference in December 2015 identified “deleveraging” as a critical task in China’s structural reforms. This policy shift significantly impacted SOEs, as excessive leverage could pose systemic financial risks. In light of this policy, SOEs are likely to adjust their leverage ratios to meet regulatory expectations. To control for the potential influence of this policy, we divide the sample based on the average leverage ratio of listed companies in 2015. Specifically, we classify firms into two groups: high-leverage companies (treatment group) and low-leverage companies (control group). The treatment group consists of firms with a leverage ratio higher than the sample mean for 2015, while the control group consists of firms with a lower leverage ratio. This classification is based on the idea that high-leverage firms are more likely to face greater regulatory pressure and thus have stronger incentives to adjust their leverage ratios accordingly. We include 449 firms in the treatment group and 449 firms in the control group. Additionally, we exclude firms directly affected by deleveraging policies from the analysis and focus our regression analysis on the remaining sample. The results of this robustness test are presented in

Table 4. Robustness test.

	(1)	(2)	(3)	(4)	(5)
	lev	lev	lev	lev	lev
icq	−0.0205 ***	−0.0204 ***	−0.0230 ***	−0.0249 ***	−0.0230 ***
	(0.0064)	(0.0064)	(0.0068)	(0.0077)	(0.0068)
manage	−0.1798 **	−0.1800 **	−0.1102	−0.0806	−0.1275
	(0.0768)	(0.0768)	(0.1039)	(0.1129)	(0.1042)
age	−0.0796 ***	−0.0795 ***	−0.0789 ***	−0.0788 ***	−0.0759 ***
	(0.0162)	(0.0161)	(0.0166)	(0.0212)	(0.0163)
tang	−0.0715 **	−0.0719 **	−0.0971 ***	−0.1129 ***	−0.0888 ***
	(0.0308)	(0.0307)	(0.0332)	(0.0391)	(0.0316)
itang	−0.2423	−0.2411	−0.2936 *	−0.4863 ***	−0.2754 *
	(0.1570)	(0.1568)	(0.1569)	(0.1651)	(0.1569)
finlev	0.0612 **	0.0621 **	0.0645 **	0.0531 *	0.0637 **
	(0.0251)	(0.0251)	(0.0267)	(0.0319)	(0.0254)
cflow	−0.0663	−0.0705	−0.0332	−0.0475	−0.0850 *
	(0.0489)	(0.0488)	(0.0546)	(0.0549)	(0.0487)
tobin	−0.0094	−0.0097	−0.0058	−0.0049	−0.0065
	(0.0083)	(0.0083)	(0.0088)	(0.0102)	(0.0088)
mbratio	0.0024	0.0006	0.0045	0.0629	0.0122
	(0.0282)	(0.0281)	(0.0316)	(0.0436)	(0.0309)
roa	−0.8584 ***	−0.8542 ***	−0.8276 ***	−0.9566 ***	−0.7639 ***
	(0.2049)	(0.2042)	(0.2312)	(0.2536)	(0.2257)
roe	0.0006	0.0004	−0.0523	−0.0539	−0.0351
	(0.0425)	(0.0424)	(0.0812)	(0.0925)	(0.0813)
_cons	0.4424 ***	0.4432 ***	0.4579 ***	0.4502 ***	0.4420 ***
	(0.0793)	(0.0792)	(0.0872)	(0.1039)	(0.0871)
Individual Fixed Effects	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	NO
Year × Industry Fixed Effects	NO	NO	NO	NO	YES
N	12,966	13,011	12,947	12,469	13,353
r2	0.1726	0.1717	0.1734	0.2898	0.1650
r2_a	0.0914	0.0908	0.0926	0.0725	0.0859

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

, under Regression (1).

4.3.2. Employing Different Methods to Measure Leverage Manipulation Motivation

To further mitigate selection bias within the treatment group and address potential measurement errors in variables reflecting leverage manipulation motivation, this study introduces an additional method for classifying listed companies. Specifically, we examine companies’ average book leverage ratio during the pre-policy period (2011–2015) and classify firms into treatment and control groups based on their performance relative to this five-year average. Firms are assigned to the treatment group if their book leverage ratio in 2015 exceeds the average ratio from 2011 to 2015, indicating a higher-than-usual leverage ratio. Companies whose 2015 leverage ratio is lower than or equal to their five-year average are placed in the control group. In line with the previous approach, firms directly affected by deleveraging policies are excluded from the sample to avoid confounding influences. Regression analysis is then performed on the remaining data points, and the results are presented in Regression (2) of Table 4.

4.3.3. Real Estate Industry

The real estate sector, characterized by its capital-intensive nature, relies heavily on external financing to support key activities such as land acquisition, project development, and marketing. These activities often require substantial financial resources, leading to high leverage ratios. As such, real estate companies are particularly vulnerable to fluctuations in financial markets and policy shifts. In response to deleveraging policies, firms in this sector may take proactive measures to manage financial risks, comply with regulatory standards, and optimize their capital structures. These actions may include modifying financing strategies, rebalancing asset portfolios, and enhancing internal controls. Given the high leverage typical of real estate firms, these companies face heightened financial risks and increased financing costs. As a result, they may resort to financial strategies or accounting practices to adjust their reported leverage ratios, aiming to improve their financial presentation. This study classifies real estate companies—where the incentive to manipulate leverage ratios is strong—as the treatment group, while companies from other industries serve as the control group. After excluding firms directly impacted by policy changes, the regression results are presented in Regression (3) of Table 4.

4.3.4. Omitted Variable Problem

Although this study controls for a range of covariates and includes both firm-specific and year-level fixed effects to address potential biases from time-invariant unobservable factors and time-varying unobservable, each industry is characterized by distinct operational models, market conditions, policy frameworks, technological advancements, and complex intra-industry dynamics. These industry-specific elements collectively influence both the overall development trajectory of industries and the micro-level behavior of firms, yet they often remain unaccounted for in traditional regression models. To mitigate the potential impact of these unobserved factors that could affect the deleveraging process of SOEs, this study incorporates interaction fixed effects between industry and year, alongside individual fixed effects. This approach allows for more nuanced control of unobserved heterogeneity and captures the potential influence of industry-specific factors on the deleveraging behavior of SOEs. The results of this adjustment are presented in Regression (4) of Table 4.

4.4. Endogeneity Test

4.4.1. PSM Test

As previously established, internal control quality significantly influences the deleveraging process of SOEs. However, a critical assumption underpinning this finding is the absence of self-selection bias and the unidirectional nature of the relationships between variables. Specifically, firms with superior internal control quality may already maintain an optimal leverage ratio and, therefore, may lack the incentive to deleverage. This could lead to endogeneity concerns, as firms with different levels of internal control quality might differ systematically in their baseline characteristics, thereby confounding the results. To address this issue, the study employs Propensity Score Matching (PSM) to mitigate potential self-selection bias. In this approach, the sample is first classified into two groups based on internal control quality: those with higher quality are assigned to the treatment group, and those with lower quality are assigned to the control group. Next, a k-nearest neighbor matching method with a caliper radius is applied to estimate the average treatment effect (ATE). The covariates used in the earlier analyses are retained as controls to calculate the propensity scores for both groups. By matching firms on these scores, we ensure that the treatment and control groups are nearly identical, with the sole distinction being their internal control quality. Firms that fall outside the common support range are excluded from the sample. Regression analysis is then conducted on the matched sample, and the results, presented in Regression (1) in

Table 5. Endogeneity test.

Variable	(1)	(2)	(3)	(4)
	lev	disclosed	lev	lev
z1		6.4304 ***
		(0.2217)
icq	−0.0230 ***		−0.0224 ***	−0.0273 ***
	(0.0068)		(0.0066)	(0.0023)
size		0.0873
		(0.1126)
top1		−0.0173 ***
		(0.0045)
big4		0.3573
		(0.2681)
depamo		0.0000 **
		(0.0000)
lnSale		−0.0446
		(0.0876)
tagr		−0.0352
		(0.0427)
tobin	−0.0065	0.0016	−0.0077	0.0175 ***
	(0.0088)	(0.0471)	(0.0087)	(0.0043)
mbratio	0.0122	−0.4889 *	0.0022	0.0408 **
	(0.0309)	(0.2962)	(0.0305)	(0.0195)
roa	−0.7639 ***	0.6550	−0.7589 ***	−0.6544 ***
	(0.2257)	(1.1663)	(0.2232)	(0.1015)
roe	−0.0351	−0.0530	−0.0633	−0.0439
	(0.0813)	(0.3653)	(0.0766)	(0.0331)
cflow	−0.0850 *		−0.0550	0.0679
	(0.0487)		(0.0502)	(0.0458)
imr			0.0000
			(0.0000)
manage	−0.1275		−0.0777	−0.0320
	(0.1042)		(0.1007)	(0.0397)
age	−0.0759 ***	−2.4213 ***	−0.0774 ***	−0.0277 ***
	(0.0163)	(0.1737)	(0.0158)	(0.0060)
tang	−0.0888 ***		−0.0983 ***	−0.0991 ***
	(0.0316)		(0.0302)	(0.0173)
itang	−0.2754 *		−0.2950 *	−0.0713
	(0.1569)		(0.1627)	(0.0526)
finlev	0.0637 **		0.0672 ***	0.0271 **
	(0.0254)		(0.0258)	(0.0130)
_cons	0.4420 ***	0.1011 ***	0.4507 ***	0.2694 ***
	(0.0871)	(1.6618)	(0.0810)	(0.0308)
Individual Fixed Effects	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES
N	13,353	11,304	13,182	13,458
r2_p		0.7349	0.1837
r2_a	0.0859		0.1047

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

, show that the coefficient estimate for internal control quality remains significantly negative. This outcome further strengthens our original conclusions.

4.4.2. Heckman Test

It is important to note that firms voluntarily disclosing internal audit reports are likely to exhibit higher-quality internal controls and relatively “attractive” balance sheets. Such firms, with well-regulated internal management and sound financial conditions, may be more inclined to disclose their internal control information to enhance market confidence. In contrast, firms with lower internal control quality or weaker financial conditions may avoid or delay disclosing their internal control reports to prevent revealing potential issues or damaging their market image. This selective disclosure behavior introduces a non-random sample, creating potential endogeneity concerns in the analysis. To address this sample selection bias, we employ the Heckman two-step model. The Heckman model accounts for sample selection bias by first estimating a selection equation to assess the factors influencing a firm’s decision to disclose internal control information. This step calculates the probability of selection, identifying the factors that determine whether a firm chooses to disclose its internal control status. In the second step, the model incorporates this selection probability as a control variable when estimating the relationship between internal control quality and deleveraging levels, correcting for biases resulting from non-random sampling. Specifically, we first use Probit regression to estimate the probabilities of firms falling into high, medium, and low internal control quality categories (IC, IC, and IC, respectively), as shown in Regression (2) in Table 5. The model includes various control variables, such as firm age (age), Tobin’s Q ratio (tobin), book-to-market ratio (mbratio), return on total assets (roa), return on equity (roe), company size (size), ownership concentration (top1), whether auditors are from one of the ‘Big Four” firms (big4), depreciation/amortization expenses (depamo), operating revenue (sale), total asset growth rate (tagr), along with individual fixed effects and year fixed effects. Additionally, we calculate the proportion of industry peers that disclose their internal audit reports as an instrumental variable (z1). We then include the inverse Mills ratio (λ) derived from the first-stage regression as an additional control variable in the main regression analysis. The results from this second stage are reported in Regression (3) in Table 5. As anticipated, the coefficient for internal control quality (icq) remains significantly negative with respect to leverage (lev), reinforcing the validity of our original conclusions.

4.4.3. Instrument-Free IV Estimation

In light of the difficulty in identifying a fully exogenous instrument, this study adopts an alternative approach to address endogeneity by employing the KLS method [48,49], an instrument-free estimation technique. Unlike the conventional two-stage least squares (2SLS) method, which relies on external instruments for identification and estimation, the KLS approach corrects for endogeneity biases directly within the regression framework, without requiring instrumental variables. The primary advantage of the KLS method lies in its ability to analytically adjust for biases in ordinary least squares (OLS) estimates under endogeneity assumptions, providing more reliable results when instruments are weak or unavailable. Notably, the KLS method tends to produce narrower confidence intervals compared to those from 2SLS when weak instruments are involved, thus offering more precise estimates in such scenarios [50]. The regression results reported in Table 5 (Regression 4), obtained using the kinkyreg command, demonstrate that after correcting for endogeneity concerns, the coefficient for internal control quality (icq) remains significantly negative. This reinforces the robustness of our findings, supporting the conclusion that improvements in internal control quality are strongly associated with deleveraging behavior in SOEs.

4.5. Mediation Analysis

To further validate the impact of internal control quality on resource misallocation and the choice of deleveraging strategy, this study constructs the following model for verification:

$$m_{i,t}={\beta }_0+{\beta }_{1}ic{q}_{i,t}+\sum \lambda X_{i,t}+{\mu }_i+{\mu }_t+{\epsilon }_{i,t}$$

(2)

In this equation, m represents the mediating variable, which refers to resource misallocation and the deleveraging strategy. The primary focus of this analysis is the coefficient ${\beta }_1$. If this coefficient is significant, it indicates the presence of a mediating effect, thus supporting the validity of Hypotheses 2 and 3.

4.5.1. The Role of Resource Misallocation

Previous studies have established that enhancing the quality of internal controls plays a crucial role in facilitating the deleveraging process of SOEs. Building upon Research Hypothesis 2, this paper posits that improvements in internal control quality contribute to deleveraging by reducing the degree of resource misallocation within firms. To empirically test this hypothesis, we follow the approach outlined and measure the extent of resource misallocation by examining the dispersion of firm efficiency. This is quantified using the following formula:

$$MIS=\left|TF{P}_{ijt}-TF{P}_{jt}\right|/TF{P}_{jt}$$

(3)

where TFP_ijt represents the total factor productivity of firm i in industry j during time t, and TFP_jt is the average total factor productivity within the industry. To assess the robustness of this measure, a range of estimation techniques are employed, including the OP (Oaxaca–Blinder decomposition), LP (Luenberger Productivity Indicator), OLS (Ordinary Least Squares), FE (Fixed Effects), and GMM (Generalized Method of Moments). The results are presented in

Table 6. The examination of the mediating effect of resource misallocation.

	(1)	(2)	(3)	(4)	(5)
	OP Method	LP Method	OLS Method	FE Method	GMM Method
	mis1	mis2	mis3	mis4	mis5
icq	−0.0008 *	−0.0012 ***	−0.0009 **	−0.0008 **	−0.0028 ***
	(0.0004)	(0.0004)	(0.0004)	(0.0004)	(0.0006)
manage	0.2386 ***	0.2234 ***	0.2003 ***	0.1929 ***	0.2573 ***
	(0.0131)	(0.0117)	(0.0101)	(0.0098)	(0.0159)
age	0.0096 ***	0.0023	0.0012	0.0000	0.0080 ***
	(0.0026)	(0.0023)	(0.0020)	(0.0020)	(0.0030)
tang	−0.0134 ***	−0.0037	−0.0062	−0.0065 *	−0.0146 **
	(0.0051)	(0.0046)	(0.0040)	(0.0039)	(0.0063)
itang	−0.0652 ***	−0.0758 ***	−0.0765 ***	−0.0783 ***	−0.0741 ***
	(0.0147)	(0.0143)	(0.0125)	(0.0125)	(0.0197)
finlev	−0.0179 ***	−0.0142 ***	−0.0124 ***	−0.0122 ***	−0.0196 ***
	(0.0033)	(0.0030)	(0.0027)	(0.0027)	(0.0040)
cflow	−0.0209 ***	−0.0220 ***	−0.0274 ***	−0.0272 ***	−0.0179 *
	(0.0078)	(0.0070)	(0.0062)	(0.0060)	(0.0096)
tobin	0.0027 ***	0.0039 ***	0.0046 ***	0.0047 ***	0.0021 **
	(0.0008)	(0.0008)	(0.0008)	(0.0007)	(0.0010)
mbratio	−0.0029	−0.0040	0.0002	0.0003	−0.0080
	(0.0042)	(0.0039)	(0.0035)	(0.0034)	(0.0051)
roa	−0.0180	−0.0475 ***	−0.0473 ***	−0.0496 ***	−0.0030
	(0.0188)	(0.0170)	(0.0145)	(0.0144)	(0.0254)
roe	0.0077	0.0199 ***	0.0187 ***	0.0199 ***	0.0031
	(0.0064)	(0.0056)	(0.0048)	(0.0047)	(0.0083)
_cons	0.0522 ***	0.0675 ***	0.0623 ***	0.0652 ***	0.0838 ***
	(0.0090)	(0.0084)	(0.0074)	(0.0073)	(0.0113)
Individual Fixed Effects	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	YES
N	13,415	13,415	13,415	13,415	13,415
r2	0.6953	0.7428	0.7657	0.7709	0.6610
r2_a	0.6665	0.7185	0.7435	0.7493	0.6290

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

, showing that the coefficient for internal control quality (icq) remains significantly negative across all estimation methods, with statistical significance at least at the 10% level. This consistent finding suggests that enhanced internal control quality substantially reduces the degree of resource misallocation within SOEs. Theoretically, resource misallocation—especially within capital and financial resources—represents an inefficient allocation of assets. In the context of SOEs, such inefficiency can manifest in funds being allocated to low-productivity or “zombie” firms. These entities fail to effectively utilize financial resources for productive investments, often due to factors such as implicit government guarantees, legacy issues, or poor management practices. This inefficiency directly leads to lower profitability and higher debt burdens, which in turn increase leverage ratios. The persistence of high levels of resource misallocation places considerable strain on SOEs’ deleveraging efforts. Misallocated resources exacerbate the difficulties of deleveraging by hindering the efficient allocation of capital. Specifically, SOEs facing significant inefficiencies in resource allocation struggle to secure adequate funding due to market distortions, credit mispricing, and inappropriate government interventions. These factors can make it easier for less efficient SOEs to access credit, thanks to implicit government support, while more efficient firms may struggle due to the distortionary effects of these interventions. Furthermore, inefficient credit allocation compounds the risks within financial markets, leading to higher financing costs and constrained financing channels for SOEs. Elevated financing costs restrict the ability of firms to reduce leverage through debt restructuring or asset divestiture strategies, while limited access to financing makes it difficult to optimize capital structures through equity financing or asset securitization. In conclusion, the quality of internal controls within SOEs not only enhances operational efficiency but also facilitates deleveraging by reducing resource misallocation. By improving the alignment of financial and operational resources, internal controls mitigate the inefficiencies that otherwise hinder the effective allocation of capital, thereby supporting more sustainable deleveraging behaviors within SOEs. Thus, Hypothesis 2 is supported.

4.5.2. Selection of Deleveraging Methods

To further explore how improvements in internal control quality influence the adoption of various deleveraging strategies aimed at reducing leverage ratios, this study classifies these strategies into two primary categories: debt reduction and equity expansion. Within these categories, debt reduction is further subdivided into “short-term debt reduction” and “long-term debt reduction”, while equity expansion includes strategies such as “new equity issuance” and “utilizing retained earnings”—the latter encompassing both “increasing equity” and “reinvesting retained earnings”. The following sections will provide a detailed framework for differentiating and measuring these strategies.

Debt reduction refers to the process through which a company lowers its leverage ratio by repaying outstanding debts, thereby directly reducing its total liabilities and financial risk. In this study, debt reduction is quantified using the annual growth rate of total liabilities. This method is further categorized into short-term and long-term debt reduction. Short-term debt reduction involves the repayment of liabilities with shorter maturities, which companies may pursue to optimize their debt structure, alleviate short-term repayment pressures, or improve maturity matching. This is measured using the annual growth rate of current liabilities. In contrast, long-term debt reduction focuses on reducing long-term obligations, often related to investments or capital expenditures over extended periods. A decrease in long-term debt can indicate a shift in investment strategy, capital structure optimization, or a reduction in risks associated with long-term repayments. This aspect is captured through the annual growth rate of non-current liabilities.

The concept of equity enhancement refers to the process of reducing leverage by increasing a company’s equity base. This strategy not only reduces leverage but also strengthens the firm’s capital structure, thereby providing a solid foundation for future growth. In this study, equity enhancement is quantified by the annual growth rate of the company’s owner’s equity. This approach can be further classified into two primary components: capital increase and retained earnings, along with other forms of equity enhancement. The combination of a capital increase and retained earnings forms a comprehensive deleveraging strategy. Specifically, capital increase involves raising share capital to reduce debt ratios, while retained earnings reflect improved profitability and the accumulation of undistributed net profits, both of which enhance the company’s financial strength and debt-bearing capacity. Capital increases generally involve actions such as issuing new shares or converting capital reserves into share capital, whereas retained earnings are represented by the growth in undistributed profits within the firm. To quantify capital increases, we use the annual growth rate of share capital, while the annual growth rate of retained earnings is used to measure retained income. Additionally, companies may enhance their equity through other mechanisms, such as issuing convertible bonds that eventually convert into shares or receiving additional investments from shareholders. These alternative equity enhancement strategies are captured using the annual growth rate of other forms of equity.

Table 7. Selection of deleveraging approaches (debt reduction).

	(1)	(2)	(3)
	Debt Reduction	Reduce Short-Term Debt	Reduce Long-Term Debt
	lev1	lev2	lev3
icq	−0.0053 **	−0.0047 **	0.0130
	(0.0021)	(0.0023)	(0.0101)
manage	−0.1445 ***	−0.0799	−0.0792
	(0.0506)	(0.0557)	(0.2484)
age	−0.0456 ***	−0.0455 ***	0.0005
	(0.0141)	(0.0156)	(0.0833)
tang	−0.2303 ***	−0.1237 ***	−0.8632 ***
	(0.0232)	(0.0280)	(0.1285)
itang	−0.2715 ***	−0.2341 **	0.1279
	(0.0784)	(0.0916)	(0.4441)
finlev	0.0492 ***	−0.0656 ***	0.7242 ***
	(0.0173)	(0.0199)	(0.0940)
cflow	−0.3620 ***	−0.1900 ***	−1.7693 ***
	(0.0415)	(0.0476)	(0.2093)
tobin	−0.0071 *	−0.0090 *	−0.0269
	(0.0040)	(0.0047)	(0.0213)
mbratio	0.0238	0.0362	−0.0511
	(0.0207)	(0.0242)	(0.1151)
roa	−0.2924 ***	−0.2772 **	0.2612
	(0.0952)	(0.1084)	(0.4884)
roe	0.1282 ***	0.1024 ***	0.3270 **
	(0.0253)	(0.0294)	(0.1438)
_cons	0.3723 ***	0.3550 ***	0.5313 **
	(0.0477)	(0.0531)	(0.2693)
Individual Fixed Effects	YES	YES	YES
Year Fixed Effects	YES	YES	YES
N	13,415	13,415	13,415
r2	0.1749	0.1239	0.1050
r2_a	0.0970	0.0411	0.0204

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

presents the regression results examining the impact of internal control quality on debt reduction strategies. In Regression (1), the coefficient estimate for internal control quality (icq) is significantly negative at the 5% level, suggesting that improvements in internal control quality encourage firms to reduce debt as a strategy for lowering the leverage ratios of SOEs. When comparing short-term versus long-term debt reduction strategies, Regression (2) shows that the coefficient estimate for icq in relation to short-term debt reduction is significantly negative, while Regression (3) indicates that the coefficient for icq in relation to long-term debt reduction is not statistically significant. These findings suggest that internal control improvements primarily lead firms to prioritize short-term debt repayment as a strategy for leverage reduction. From a risk management perspective, high-quality internal controls enhance a firm’s ability to manage financial risks. Short-term debt typically carries higher repayment frequency and shorter maturities compared to long-term debt, requiring more robust cash flow management and debt servicing capabilities. An improvement in internal control quality signals that firms are better equipped to manage these responsibilities, increasing their confidence in reducing leverage by repaying short-term debt. This strategy not only optimizes their capital structure but also minimizes liquidity risks. Furthermore, stronger internal control quality positively influences a company’s credibility and financing capacity. A well-established internal control system signals operational soundness and financial transparency to creditors, thereby enhancing trust and confidence. This increased credibility makes it easier for firms to negotiate favorable refinancing terms or debt extensions when facing short-term debt maturities, reducing repayment pressures. In terms of borrowing costs, it is also noteworthy that short-term debt generally carries lower interest rates and financing costs. The improvement in internal control quality allows firms to better capitalize on market opportunities and select cost-effective financing alternatives.

Table 8. Selection of deleveraging approaches (equity increase).

	(1)	(2)	(3)	(4)	(5)
	Equity Increase	Capital Increase and Profit Retention	Other Equity Increase	Capital Increase	Profit Retention
	lev4	lev5	lev6	lev7	lev8
icq	−0.0034 **	−0.0058 ***	0.0005	0.0006	−0.1299 ***
	(0.0016)	(0.0013)	(0.0006)	(0.0007)	(0.0170)
manage	−0.0825 **	0.0692 **	−0.0016	−0.0277	1.2376 ***
	(0.0369)	(0.0311)	(0.0108)	(0.0175)	(0.3612)
age	0.0153 **	−0.0300 ***	0.0085 **	−0.0385 ***	0.2013 **
	(0.0060)	(0.0100)	(0.0043)	(0.0067)	(0.1018)
tang	−0.0857 ***	−0.0457 ***	−0.0156 **	−0.0716 ***	0.5581 ***
	(0.0131)	(0.0155)	(0.0068)	(0.0095)	(0.1798)
itang	0.0055	−0.0073	−0.0431 *	0.0334	1.0765 *
	(0.0415)	(0.0526)	(0.0221)	(0.0336)	(0.5927)
finlev	0.0058	0.0347 ***	0.0088 *	−0.0011	0.4904 ***
	(0.0088)	(0.0109)	(0.0052)	(0.0067)	(0.1291)
cflow	−0.0016	0.1360 ***	0.0094	−0.0393 ***	1.1373 ***
	(0.0225)	(0.0258)	(0.0100)	(0.0150)	(0.2911)
tobin	−0.0046 *	−0.0059 **	0.0011	−0.0023	0.0319
	(0.0023)	(0.0025)	(0.0010)	(0.0016)	(0.0303)
mbratio	0.0174	−0.0535 ***	0.0073	0.0116	0.2699 *
	(0.0118)	(0.0148)	(0.0069)	(0.0088)	(0.1633)
roa	0.4800 ***	1.1727 ***	0.0208	0.2208 ***	7.5397 ***
	(0.0676)	(0.0602)	(0.0271)	(0.0305)	(0.6431)
roe	0.4696 ***	−0.0240	0.0021	−0.0131 *	1.8608 ***
	(0.0221)	(0.0156)	(0.0083)	(0.0077)	(0.1933)
_cons	0.0504 **	0.1875 ***	−0.0397 ***	0.1679 ***	−0.3930
	(0.0246)	(0.0328)	(0.0146)	(0.0210)	(0.3536)
Individual Fixed Effects	YES	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES	YES
N	13,415	13,415	13,415	13,415	13,415
r2	0.5052	0.2007	0.1936	0.1572	0.1865
r2_a	0.4585	0.1251	0.1174	0.0776	0.1097

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

presents the regression results analyzing the impact of internal control quality on empowerment strategies. Specifically, in Regression (1), the coefficient estimate for internal control quality in relation to empowerment is significantly negative at the 5% level, suggesting that improvements in internal control quality encourage firms to adopt empowerment strategies. Furthermore, in Regression (2), the coefficient estimates for internal control quality concerning the “capital increase + retained earnings” strategy is significantly negative at the 1% level, indicating that enhancing internal control quality also prompts firms to adopt this approach. In Regressions (4) and (5), the coefficient estimates indicate that internal control quality has a significantly negative effect on retained earnings. In contrast, Regression (3) shows that the coefficient estimate for internal control quality regarding other forms of empowerment is not statistically significant. These findings suggest that improvements in internal control quality led firms to prefer retained earnings as a means of empowerment, which in turn helps lower the leverage ratios of SOEs. High-quality internal controls enhance operational efficiency and reduce costs within organizations. Through the implementation of stringent management practices and rigorous internal controls, firms can optimize resource utilization, minimize waste, and ultimately increase profitability. As profitability improves, internally generated funds, including retained earnings, also grow, providing a solid foundation for retention-based strategies. In a robust internal control framework, firms are better equipped to assess and manage risks, implementing proactive measures to prevent financial disruptions. This enhanced risk management capability fosters greater confidence and resilience in navigating external economic fluctuations and financial market uncertainties, enabling firms to retain profits more effectively while strengthening their capital base and enhancing their capacity to absorb shocks. Thus, Hypothesis 2 is supported.

4.5.3. Moderating Effect of R&D Intensity

To further examine the moderating role of R&D intensity in the relationship between internal control quality and the leverage ratio of state-owned enterprises, this study constructs the following model:

$$le{v}_{i,t}={\delta }_0+{\delta }_{1}ic{q}_{i,t}+{\delta }_{2}rd{1}_{i,t}+{\delta }_{3}icq\_rd{1}_{i,t}+\sum \lambda X_{i,t}+{\mu }_i+{\mu }_t+{\epsilon }_{i,t}$$

(4)

In this equation, rd1 represents R&D intensity. The primary focus is on the coefficient estimates of ${\delta }_0$ and ${\delta }_3$. If both coefficients are statistically significant, it indicates the presence of a moderating effect. Specifically, if ${\delta }_0$ is significantly negative and ${\delta }_3$ is significantly positive, it suggests that R&D intensity mitigates the negative effect of internal control quality on the leverage ratio. In this case, Hypothesis 4 is supported.

This section explores the moderating role of R&D intensity, with the findings presented in

Table 9. Examination of the moderating effect of R&D intensity.

	(1)	(2)
	lev	lev
icq	−0.0312 ***	−0.0271 ***
	(0.0089)	(0.0072)
rd1	−6.7985 ***
	(2.1500)
rd1_icq	1.0150 ***
	(0.3283)
rd2		−0.0057 **
		(0.0025)
rd2_icq		0.0008 **
		(0.0004)
manage	−0.1062	−0.1037
	(0.0993)	(0.0989)
age	−0.0807 ***	−0.0772 ***
	(0.0164)	(0.0164)
tang	−0.0865 ***	−0.0908 ***
	(0.0317)	(0.0317)
itang	−0.3000 *	−0.3028 **
	(0.1562)	(0.1544)
finlev	0.0637 **	0.0620 **
	(0.0252)	(0.0255)
cflow	−0.0525	−0.0585
	(0.0500)	(0.0497)
tobin	−0.0054	−0.0061
	(0.0085)	(0.0086)
mbratio	0.0017	0.0032
	(0.0308)	(0.0308)
roa	−0.8624 ***	−0.8561 ***
	(0.2165)	(0.2188)
roe	−0.0379	−0.0376
	(0.0735)	(0.0743)
_cons	0.5163 ***	0.4828 ***
	(0.0980)	(0.0865)
Individual Fixed Effects	YES	YES
Year Fixed Effects	YES	YES
N	13,415	13,415
r2	0.1754	0.1737
r2_a	0.0973	0.0955

Note: Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

. In this study, R&D investment is classified into two components: human capital and physical capital. Specifically, rd1 represents the ratio of R&D investment to total assets, while rd2 denotes the ratio of R&D personnel to total employees. The interaction terms rd1_icq and rd2_icq capture the interaction between R&D investment or R&D personnel and internal control quality, respectively. From both human and physical capital perspectives, the coefficient estimates for internal control quality (icq) are significantly negative, while those for the interaction terms are significantly positive, at least at the 5% level. These findings suggest that higher levels of R&D intensity reduce the negative effect of internal control quality on leverage ratios. In other words, as R&D intensity increases, the restrictive impact of internal control quality on leverage diminishes. R&D activities typically require substantial financial outlays, often prompting firms to seek external financing to meet their funding needs. As R&D intensity rises, so does the demand for financing, which may lead to an increase in a firm’s leverage ratio. Additionally, due to the high risks associated with research endeavors—where failure can result in significant losses—the associated risk exposure increases with greater R&D activity. This heightened risk may erode creditors’ confidence in the firm’s stability, thereby raising its financing costs and further influencing its leverage ratio.

4.5.4. Heterogeneity Test

(1)

Central Enterprises vs. Local SOEs

This study further examines the differential impact of internal control quality on deleveraging between central and local SOEs. As shown in

Table 10. Heterogeneity test.

	(1)	(2)	(3)	(4)
	Central Enterprises	Local SOEs	Excessive Debt	Non-Excessive Debt
	lev	lev	lev	lev
icq	−0.0225	−0.0212 ***	−0.0196 ***	−0.0156
	(0.0297)	(0.0062)	(0.0066)	(0.0136)
manage	−0.2865 *	−0.0500	0.0508	−0.2060 *
	(0.1739)	(0.1145)	(0.1770)	(0.1167)
age	−0.0210	−0.0907 ***	−0.1036 ***	−0.1050 ***
	(0.0397)	(0.0181)	(0.0274)	(0.0211)
tang	−0.1739 **	−0.0998 ***	−0.1516 ***	−0.0657
	(0.0803)	(0.0336)	(0.0455)	(0.0497)
itang	−1.5111	−0.2597	−0.4087 **	−0.1888
	(1.2894)	(0.1595)	(0.1924)	(0.1651)
finlev	−0.0866	0.0968 ***	−0.0354	0.0842 **
	(0.1038)	(0.0258)	(0.0427)	(0.0336)
cflow	−0.0221	−0.0718	−0.0812	−0.0173
	(0.1520)	(0.0528)	(0.0677)	(0.0768)
tobin	−0.0510	0.0013	−0.0062	0.0055
	(0.0453)	(0.0070)	(0.0128)	(0.0075)
mbratio	−0.0216	0.0144	0.0181	−0.0216
	(0.0764)	(0.0335)	(0.0603)	(0.0370)
roa	−0.9921	−0.7702 ***	−0.7235 ***	−1.4800 ***
	(0.6599)	(0.2421)	(0.2554)	(0.4102)
roe	−0.0539	−0.0512	−0.0696	0.2040
	(0.1601)	(0.0874)	(0.0581)	(0.2940)
_cons	0.5546 *	0.4392 ***	0.5720 ***	0.4580 ***
	(0.3026)	(0.0805)	(0.1431)	(0.1134)
Individual Fixed Effects	YES	YES	YES	YES
Year Fixed Effects	YES	YES	YES	YES
N	1906	11,225	6826	6423
r2	0.2228	0.2009	0.3070	0.1622
r2_a	0.1000	0.1166	0.2046	0.0413

Note: Standardized significance annotations (***, **, * for 1%, 5%, 10%), respectively.

, particularly in regressions (1) and (2), the coefficient estimate for internal control quality (icq) is more pronounced in regression (2). This indicates that improving internal control quality is more effective in promoting deleveraging behavior among local SOEs. These enterprises typically operate in a more complex governance environment, characterized by challenges such as local government interference and suboptimal managerial incentives. By enhancing internal control quality, local SOEs can improve risk management, optimize resource allocation, and increase operational efficiency, thereby facilitating their deleveraging efforts.

(2)

Excessive Debt vs. Non-Excessive Debt

The preceding discussion has established that the quality of internal control can promote deleveraging behaviors in SOEs. However, it is important to note that both extremely low and high leverage ratios do not necessarily contribute to optimal operational outcomes for firms. A moderate leverage ratio enables enterprises to optimize their financial structure and enhance capital utilization efficiency. Conversely, high leverage can lead to financial distress, especially during economic downturns or unfavorable market conditions. By reducing debt levels, companies can enhance their risk resilience and maintain a healthy financial position. Nonetheless, debt financing generally incurs lower costs compared to equity financing, as interest expenses are tax-deductible, thereby reducing the effective tax rate for businesses. If the leverage ratio is too low, firms may forfeit the tax benefits associated with debt financing. In view of this, this paper draws on studies by [51,52,53] to initially calculate the target debt-to-asset ratio for enterprises. Subsequently, we assess excessive debt by examining the disparity between actual and target asset-liability ratios; a positive value indicates excessive debt, while a value less than or equal to zero suggests otherwise. Table 10 presents the regression results for models (3) and (4) across two samples, respectively. Notably, in regression model (3), the coefficient estimate for internal control quality in relation to deleveraging intensity is significantly negative, indicating that for SOEs with excessive debt levels, higher internal control quality is associated with more effective facilitation of deleveraging actions.

5. Discussion

Previous research on deleveraging in SOEs has predominantly concentrated on external factors such as macroeconomic policies, market conditions, and financial structures [54,55,56,57,58], with relatively little attention paid to internal corporate management mechanisms, particularly the quality of internal control. This study introduces a novel perspective, demonstrating that the quality of internal control plays a crucial role in facilitating deleveraging within SOEs. Specifically, the paper not only substantiates the positive effect of internal control quality on SOE deleveraging but also explores the underlying mechanisms driving this relationship. The findings reveal that enhancing internal control quality significantly promotes the deleveraging process by mitigating the misallocation of corporate resources, a critical pathway. This insight bridges the gap in existing literature regarding the connection between internal control and resource allocation efficiency, emphasizing the unique role of internal control in optimizing resource allocation and improving the efficiency of resource use within SOEs. Moreover, this study innovatively suggests that improving internal control quality encourages SOEs to adopt two primary deleveraging strategies: “debt reduction” and “equity enhancement”, which are further subdivided into “short-term debt reduction” and “profit retention” approaches. This perspective not only expands the range of deleveraging strategies available to SOEs but also provides more precise and actionable recommendations for firms seeking to reduce their leverage. Additionally, the study reveals that increased R&D intensity can partially diminish the effect of internal control quality on deleveraging. This finding suggests that, in highly innovative environments, the enhancement of internal control quality should be balanced with R&D activities to prevent adverse outcomes—an aspect that previous studies have largely overlooked, as shown in Figure 1.

Other developing countries face similar challenges, such as excessively high debt levels in SOEs and inefficiencies in resource allocation. Enhancing the quality of internal control could serve as an effective internal remedy [59]. By strengthening internal control systems, optimizing resource allocation, and encouraging the adoption of prudent deleveraging strategies, these countries can gradually reduce SOE debt levels while improving economic efficiency and stability. Additionally, the analysis in this paper regarding the interaction between R&D intensity and internal control quality offers valuable insights for other developing nations: while promoting innovation, it is crucial to ensure alignment between internal control mechanisms and R&D activities in order to sustain healthy economic development.

It is important to note that this study presents opportunities for further exploration, as deleveraging does not inherently suggest that a lower leverage ratio is always optimal [60,61]. An excessively low leverage ratio may indicate underutilization of external financing, potentially resulting in missed investment and growth opportunities. Such under-leveraging could restrict the expansion of production scale, limit technological innovation, and negatively impact economic growth. This outcome is not the goal of deleveraging policies. A balanced leverage ratio enables enterprises and individuals to access external funds for production expansion and investment, thereby enhancing capital utilization efficiency and driving economic growth. Future research should focus on identifying the optimal leverage ratio range and exploring the costs and benefits associated with maintaining such a balance.

6. Conclusions and Future

Under the strategic guidance of balancing security and development, the prevention of debt risks in SOEs is a crucial practical issue for fortifying the macroeconomic security barrier and fostering high-quality development momentum. Based on data from listed state-owned companies on the Shanghai and Shenzhen stock exchanges (2009–2023), this paper constructs a high-dimensional fixed-effects model to systematically examine the impact of internal control quality on deleveraging in enterprises and its underlying mechanisms. The main conclusions are as follows: First, improving internal control quality significantly drives deleveraging in state-owned enterprises. Second, the mechanism reveals a dual-path characteristic of “resource allocation optimization” and “debt structure adjustment”. In terms of resource allocation, improving internal control quality alleviates resource misallocation, thereby reducing the leverage ratio of SOEs. On the debt operation front, high-quality internal control encourages enterprises to adopt a proactive deleveraging strategy. On the one hand, they reduce short-term debt to achieve “debt reduction”; on the other hand, they use profit retention and equity financing to implement “equity enhancement”, forming a synergistic debt governance combination of long-term and short-term strategies, which further reduces the leverage ratio of SOEs. Third, policy effects exhibit significant heterogeneity. R&D intensity has a nonlinear moderating effect: the higher the R&D intensity, the weaker the positive impact of internal control quality on deleveraging in SOEs. Fourth, the institutional environment and debt status show significant heterogeneity. For local SOEs and highly indebted enterprises, improvements in internal control quality have a more pronounced positive effect on deleveraging.

Based on the above conclusions, this paper proposes four policy recommendations: First, improve the internal control system for state-owned enterprises. A dynamic monitoring and tiered evaluation mechanism for internal control quality should be established, incorporating the effectiveness of deleveraging as a core indicator in internal control audits. The responsibility accountability system for governance layers should be strengthened to ensure that the rigid constraints of internal control are effectively implemented. Second, establish a dynamic framework for debt structure optimization. A “debt reduction—equity enhancement” coordination operation guideline should be formulated, clearly defining early-warning thresholds for short-term debt scale and requirements for profit retention ratios. Additionally, policies to support the innovation of equity financing tools should be improved to promote the formation of a long-term, proactive debt governance mechanism. Third, implement classification-based evaluation and incentive constraints for R&D investments. For enterprises with high R&D intensity, financial stability exemption clauses should be established. A performance evaluation system that balances innovation investment and risk management should be set up, exploring the establishment of a risk reserve fund for R&D projects to prevent debt risks stemming from innovation activities. Fourth, strengthen the targeted governance of local SOEs and highly indebted entities. The effectiveness of internal control should be included in the evaluation scope of local state-owned assets supervision. A key supervisory list for internal control construction in highly indebted enterprises should be established. The configuration of risk governance responsibilities and rights within the board of local SOEs should be optimized, enhancing the collaborative effectiveness between localized supervision and micro-level governance.

Certainly, this paper also has certain limitations. The study focuses solely on A-share listed state-owned enterprises in the Shanghai and Shenzhen stock exchanges. Given that listed companies are typically larger market entities, the conclusions drawn may not be universally applicable or replicable for smaller market players. This “survivorship bias” leads to a reduction in external validity when the research conclusions are applied to unlisted SOEs or private enterprises. Non-listed entities generally exhibit characteristics such as opaque governance information and reinforced budgetary soft constraints, making their deleveraging decisions more susceptible to administrative interference and implicit guarantees. Future research could incorporate data from the National Enterprise Credit Information Publicity System to construct a comprehensive sample database of state-owned enterprises across the entire spectrum.