Financial decentralization and corporate risk: evidence from China

Abstract

Financial decentralization grants local governments greater authority over financial resources to stimulate regional economic growth; however, it may also generate disparities. This study investigates the effect of financial decentralization on corporate risk in China. We develop a theoretical decision model that incorporates risk assessment into the credit evaluation processes of banks and firms, explicitly accounting for the influence of financial decentralization. The restructuring of rural credit cooperatives provides a quasi-natural experiment for financial decentralization, and we employ a multi-period difference-in-differences approach using data from 273 cities between 2007 and 2018. The results reveal that financial decentralization increases corporate risk, with the effect being particularly pronounced in regions with stronger fiscal decentralization, in non-key industries, and among private enterprises. Smaller and less profitable firms are especially vulnerable. Mechanism analysis further shows that financial decentralization marginalizes alternative financial institutions, such as state-owned bank branches, thereby restricting firms’ access to loans unless collateralized. These findings suggest that financial decentralization weakens financial competition and reduces the efficiency of capital allocation. Overall, this study underscores the need for stronger financial oversight and improved credit allocation policies to enhance stability and support more equitable economic development.

Introduction

In the context of China’s tax-sharing reform, local governments have actively competed for financial resources and sought fiscal authority to sustain rapid economic growth (He and Miao, 2021), a phenomenon referred to as fiscal decentralization (Shen et al., 2012). However, fiscal resources represent only part of the strategies employed to promote development. Local governments also strategically engage with financial markets, leveraging various financial instruments to mobilize capital, thereby making financial participation a critical channel in competitive growth dynamics. In parallel with fiscal decentralization, we define local governments’ pursuit of financial resource management and allocation as financial decentralization. While this framework facilitates funding for local economic development, it can also lead to credit misallocation and crowding-out effects on corporate financing, thereby amplifying corporate risk (Zhou et al., 2023). Within this context, examining the impact of financial decentralization on corporate risk holds substantial theoretical significance and constitutes the central research question of this study.

Our research is motivated by both practical and academic considerations. First, in developing countries where economic growth relies heavily on resource inputs, local governments often intervene in the process of financial marketization to promote regional development (Zhou et al., 2023). Fiscal and financial decentralization frequently direct financial resources toward sectors favored by local governments, thereby stimulating regional growth to some extent (Zhou et al., 2023; Xu, 2024; Zhao et al., 2024; Zhang et al., 2024). However, it remains uncertain whether such government-led financial allocation distorts market-based resource distribution and constrains the survival of micro-level enterprises. This study seeks to address these questions through empirical evidence from China, revealing and challenging the adverse micro-level effects of financial intervention. Second, understanding the determinants of corporate risk is a globally significant research endeavor. In the context of global economic restructuring and profound changes in financial markets, firms face unprecedented challenges (Petricevic and Teece, 2019). Increasingly complex corporate operations have heightened business risks, and the escalation of corporate risk may trigger risk contagion, ultimately exerting adverse effects on the broader economy (Qian et al., 2023). Thus, identifying the drivers of corporate risk is critical for governance and risk mitigation. Given that government policies and interventions strongly shape firms’ input–output processes and risk management, this paper investigates government practices that may mitigate corporate risk from the perspective of financial decentralization. Third, research on financial decentralization remains limited, with notable gaps in the literature. Existing studies have primarily focused on issues such as environmental outcomes, industrial upgrading, and educational equity (Li, 2017; Zheng and Wang, 2023; Xu, 2024; Zhao et al., 2024; Wang, 2025; Zhu et al., 2025). Yet enterprises—major participants in financial markets with substantial financing demands—are deeply affected by local governments’ financial decentralization. Examining how financial decentralization influences corporate risk not only enriches the financial economics literature but also provides valuable insights for financial regulation and corporate governance. The purpose of this study is to offer a comprehensive understanding of the micro-level effects of financial decentralization by analyzing, both theoretically and empirically, its impact on corporate risk while uncovering the underlying mechanisms and heterogeneous effects. Finally, this study addresses methodological deficiencies in measuring financial decentralization within existing literature. Prior research predominantly evaluates the degree of financial decentralization from the standpoint of local financial development (Xu, 2024; Zhu et al., 2025; Wang, 2025). However, financial decentralization fundamentally reflects the extent of government control and appropriation of financial resources, which is not equivalent to local financial development. Drawing on the evolution of financial decentralization with Chinese characteristics, we aim to construct indicators that more accurately capture changes in local governments’ control over regional financial resources.

To address the gap in the existing literature, we tackle the challenges of variable measurement and potential endogeneity by employing the restructuring of local rural credit cooperatives (RCCs) in China as a proxy for implicit financial decentralization. Specifically, we interpret the conversion of RCCs into rural commercial banks (RCBs) as an exogenous shock to financial decentralization. This approach provides a more precise measure than previous methods that relied on proxies such as the number of financial institution employees or the mere presence of local financial entities. Financial decentralization entails local governments consolidating control over financial resources through the establishment of local financial institutions. The transition from RCCs to RCBs substantially strengthens the capacity of local governments to access and allocate financial resources. Moreover, because the restructuring is directed by the central government, it introduces an exogenous factor that reshapes local financial landscapes.

By leveraging this structurally induced change as a quasi-natural experiment, we focus on prefecture-level restructuring to assess its implications for regional implicit financial decentralization and subsequent corporate risk. Our results show that the conversion process increases the three-year rolling standard deviation and range of corporate profits by 0.060 and 0.056 standard deviations, respectively. These findings confirm that implicit financial decentralization significantly heightens corporate risk. In addition to the baseline analysis, we conducted extensive robustness checks. We further performed heterogeneity analyses to explore how the effects of financial decentralization vary across dimensions such as firm ownership (state-owned vs. private), alignment with key industries, the degree of fiscal decentralization, and firm-specific characteristics. The evidence demonstrates that the impact of financial decentralization on corporate risk is heterogeneous across firms, underscoring the complex and differentiated effects of financial structures on corporate behaviors.

The potential marginal contributions of this study are as follows. First, we advance the literature on financial decentralization from a methodological perspective. In contrast to existing indices of financial decentralization (Li, 2017; Zheng and Wang, 2023; Zhang et al., 2024; Xu, 2024; Zhao et al., 2024; Wang, 2025; Zhu et al., 2025), we introduce a novel measurement approach based on the exogenous policy shock resulting from the transformation of RCCs into RCBs in China, which effectively captures changes in local governments’ control over financial resources. Second, we address a research gap concerning the impact of financial decentralization on micro-enterprises—a question largely unexplored in earlier studies. Theoretically, we refine corporate risk strategy models by incorporating risk selection factors into the credit decision-making processes of banks and firms, thereby developing a framework that accounts for local financial decentralization. Our empirical results demonstrate that financial decentralization disrupts bank competition and induces a crowding-out effect on corporate credit access, exacerbating financing constraints for enterprises. Moreover, our heterogeneity analysis confirms the presence of agency problems within the financial system, showing that financial decentralization enables local government intervention in the credit allocation of financial institutions, often channeling support to specific key projects or enterprises. Third, we enrich the literature that has predominantly focused on the impact of banking reforms on corporate risk (Jiang et al., 2020; Wang and Lee, 2023; Lai et al., 2023; Fungáčová et al., 2017). While previous research confirms that banking system improvements can mitigate corporate financing risks, our findings reveal that local government intervention may not reproduce these beneficial effects. Instead, the RCC reform demonstrates that financial decentralization can direct capital flows toward sectors favored by local governments, ultimately giving rise to agency problems within the banking system.

The structure of this paper is as follows. Chapter 2 reviews the policy background of financial decentralization and the related literature on financial decentralization and corporate risk. Chapter 3 provides a theoretical analysis. Chapter 4 introduces the research design of this study. Chapter 5 reports and analyzes the empirical results. The final chapter presents the conclusions and policy implications of this study.

Policy background and literature review

Policy background

Financial decentralization refers to the allocation and organization of fiscal powers between central and local governments within a country’s or region’s financial system. In promoting local economic development, local governments leverage their capacity for financial innovation by establishing platforms and guiding institutions to provide specialized services. Research categorizes China’s financial decentralization into three phases based on local governments’ appropriation of financial resources and intervention capabilities (He and Miao, 2021; Hong and Hu, 2017; Fu and Li, 2017; Xu, 2024; Zhang et al., 2024). The first phase (1994–1997) involved initial attempts, during which the central government restored financial institutions, established the independence of the People’s Bank of China from the Ministry of Finance, implemented branch leadership in four major state-owned banks, strengthened ties between institutions and local governments, delegated credit authority, and converted infrastructure grants into loans, thereby expanding local financial power. The second phase (1998–2002) featured financial re-centralization to mitigate risks such as bubbles and non-performing loans arising from the prior phase. Vertical management was instituted, including the abolition of provincial branches of the People’s Bank and centralizing loan approvals, which reduced local interference and enhanced financial independence. The third phase (2003 to present) involves implicit financial decentralization alongside an emphasis on development and risk control. In this phase, local governments compete for resources implicitly through investments in urban commercial banks, village banks, and non-bank channels such as securities firms, micro-lending, financial leasing, and shadow banking, reflecting explicit centralization alongside implicit decentralization. In 2015, local bond issuance was permitted, initiating debt swaps and platform development to balance centralization and decentralization for the real economy. During this stage, market-oriented reforms and the listings of state-owned and joint-stock banks reduced intervention; however, unequal powers continued to incentivize local pursuits of loans implicitly. A key feature of this evolution was the establishment of local institutions like urban commercial banks (UCBs) and RCBs to direct resources toward favored projects, marking the progression of implicit decentralization.

RCCs are rural cooperative institutions approved by the People’s Bank of China, formed through member shareholding with democratic management primarily to serve their members. They address the scattered, small-scale funding needs of agriculture via mutual credit derived from shares and deposits, originating in 1923 in Xianghe County, Hebei, and have proven crucial for agricultural stability. Since 1949, RCCs have evolved through five phases: farmer mutual cooperation (1952–1958), people’s commune management (1958–1978), agricultural bank management (1978–1996), people’s bank management (1996–2003), and local government management (since 2003). This transition from Agricultural Bank to People’s Bank regulation, followed by provincial-county cooperative models, and ultimately to shareholding, reflects a decentralization trend from central to local authority, aligning with the current phase of explicit centralization and implicit decentralization. As RCCs transform into RCBs, significant changes occur in structure, scope, and scale. RCCs utilized cooperative ownership with member equity and mutual aid, restricting commercial activities, whereas RCBs adopt a joint-stock model that clarifies rights, enables market operations, and enhances efficiency. While RCCs concentrated on rural services like deposits and loans for farmers, RCBs extend their offerings to peri-urban and urban areas with diverse products such as wealth management and credit cards. RCBs are larger entities, requiring a capital of 50 million yuan compared to RCCs’ 5 million, with RCCs needing over 1 billion yuan in assets and under 15% non-performing loans for reform, thus facilitating improved deposit attraction and credit expansion, empowering local governments, and aligning with decentralization goals for resilient rural banking.

The transition from RCCs to RCBs is centrally mandated and relatively independent of local influence, providing exogeneity. This study treats the establishment of RCBs as an exogenous shock to financial decentralization, assessing its risk effects on corporations through positive financial shocks. Local government bonds also signify decentralization. Historically issued by opaque entities with high risks and rates, since 2015, provinces and regions have begun issuing self-repaid bonds for lower costs, increased transparency, and enhanced accountability. This study utilizes such issuances as a proxy in robustness tests to analyze the impacts of decentralization on economies and corporate risks within the dynamics of decentralized governance.

Literature review on financial decentralization

The interrelationship among financial development, economic growth, and corporate operations has attracted considerable scholarly attention (Kendall, 2012; Han and Gao, 2024). Despite a substantial body of existing research, the dynamics governing the distribution of financial authority between local and central governments—particularly within the context of decentralization—remain underexplored. Traditional fiscal federalism theory posits that decentralization can enhance economic efficiency by enabling local governments to tailor policies to specific regional needs. Classical literature further suggests that fiscal decentralization confers greater autonomy upon local governments (Oates, 1993; Shen et al., 2012; Ding et al., 2019; Fang et al., 2023) and promotes regional economic growth (Schragger, 2010; Kalirajan and Otsuka, 2012).

Nevertheless, the fiscal constraints inherent in fiscal decentralization have compelled local governments to develop local financial markets as a means of acquiring and controlling financial resources—a phenomenon referred to as financial decentralization. Autonomy in financial management presents a double-edged sword (Xu, 2024). On one hand, some studies contend that financial decentralization provides a more nuanced and adaptable mechanism for resource coordination and allocation, thereby alleviating information asymmetry. By delegating financial authority, local governments may become more proactive in securing financial resources, fostering regional financial accumulation, and mitigating capital misallocation (Zhang et al., 2024; Xu, 2024; Zhao et al., 2024; Wang, 2025; Zhu et al., 2025). Conversely, another line of research warns that financial decentralization may exacerbate regional financial vulnerabilities. When local governments appropriate financial resources, they can crowd out the financing opportunities available to enterprises in the real sector, leading to increased financing constraints and operational risks for these firms (Zhou et al., 2023).

Fiscal decentralization has played a foundational role in China’s rapid economic ascent. Concurrently, intervention in the allocation of financial resources has become an essential strategy for local governments in China to address financial constraints arising from fiscal decentralization. Numerous studies have investigated the multifaceted impacts of financial decentralization in China, revealing varied effects across sectors such as public services (Li, 2017), sustainable development (Zheng and Wang, 2023; Xu, 2024; Zhao et al., 2024; Wang, 2025), and industrial structure (Zhu et al., 2025). For example, in the domain of public services, Li (2017) found that financial decentralization in higher education could exacerbate disparities in access to educational resources, despite its original aim to decentralize financial control and decision-making. This suggests that financial decentralization may inadvertently perpetuate inequalities in educational opportunities. In terms of sustainable development, Xu (2024) documented a U-shaped relationship between financial decentralization and the growth of the wind power industry, indicating that while early stages of decentralization might hinder industry growth, more advanced stages could stimulate expansion—including broader implementation of carbon taxation policies. This nuanced relationship underscores the interplay between financial governance structures and environmental initiatives. Furthermore, Zhao et al. (2024) identified that excessive financial decentralization could impede energy efficiency through channels such as technological stagnation, inadequate industrial upgrading, and inconsistent government policy. Regarding industrial structure, Zhu et al. (2025) found that financial decentralization positively contributes to industrial upgrading and generates beneficial spillover effects in the rationalization of industrial structures in adjacent cities. However, when fiscal decentralization exceeds a certain threshold, the positive impact of financial decentralization on industrial upgrading in non-coordinated cities becomes negative.

Beyond China, international evidence also highlights the effects of financial decentralization. Souza (1996) noted that while decentralization—including its financial dimension—can facilitate democratization, it simultaneously shapes a range of political and economic outcomes, revealing inherent limitations in its effectiveness. Using data from India, Datta and Varalakshmi (1999) demonstrated that decentralization provides an effective mechanism for grassroots financial management, strengthening both institutional and fiscal sustainability. More recently, Zhang et al. (2024) employed an enhanced autoregressive distributed lag (EADL) model to evaluate the cumulative and immediate effects of financial decentralization, green energy investment, and environmental outcomes. Their findings indicate that the combination of financial decentralization and green energy investment substantially reduces CO₂ emissions and improves air quality within the European Union. Crucially, effective governance of financial institutions moderates this relationship, shaping the extent to which decentralization and green energy investment contribute to environmental protection.

Taken together, these studies suggest that financial decentralization enables local governments to exercise greater autonomy in managing financial resources (Schragger, 2010). This autonomy fosters more efficient resource allocation, thereby reducing mismatches and promoting energy efficiency, green investment, industrial upgrading, and regional economic growth (Zheng and Wang, 2023; Zhang et al., 2024; Wang, 2025; Zhu et al., 2025). However, when fiscal decentralization is excessive (Zhu et al., 2025) or when financial institutions lack effective governance (Zhang et al., 2024), adverse consequences may arise, including the crowding out of enterprise financing and heightened financial fragility (Zhou et al., 2023). In other words, excessive intervention by local governments in directing financial resources can generate agency problems, amplifying both financial instability and capital misallocation.

Literature review on corporate risk

Currently, there is considerable scholarly interest in identifying the determinants of corporate risk. A growing body of research asserts that corporate governance—reflecting internal oversight mechanisms and strategic execution capabilities—plays a critical role in shaping corporate risk (Koirala et al., 2020). Existing studies predominantly examine the impact of corporate governance on risk through various dimensions, including board characteristics (Ji et al., 2021; Gu et al., 2025), executive attributes (Çolak and Korkeamäki, 2021; Li et al., 2023; Guo et al., 2025; Safiullah et al., 2025), and ownership structure (Chen et al., 2025; Bagh et al., 2025). Collectively, these findings indicate that internal governance mechanisms influence the severity of agency problems within firms, thereby affecting their strategic orientation and exposure to operational risks (Jiang et al., 2024).

Furthermore, some strands of literature investigate the determinants of corporate risk from the perspective of specific corporate behaviors. Prior research suggests that active engagement in corporate social responsibility (CSR) and environmental, social, and governance (ESG) practices can strengthen stakeholder relationships, enhance access to competitive resources, and ultimately mitigate overall corporate risk (Menla Ali et al., 2024; De Giuli et al., 2024; Zhuang et al., 2025). With the widespread adoption of artificial intelligence (AI) technologies in business operations, recent scholarship has increasingly focused on examining the relationship between investments in digitalization or AI and firms’ competitive capabilities (Jiang et al., 2024; Du et al., 2024; Wu and Lu, 2025).

A further strand of research highlights the influence of external factors on corporate risk (Zhou and Jiang, 2025). In particular, studies examining regional variations in financial markets are especially relevant (Fungáčová et al., 2017; Lai et al., 2023; Jiang et al., 2020; Cai et al., 2024). For example, Jiang et al. (2020) analyzed the effects of banking sector deregulation and concluded that regulatory relaxation mitigates corporate risk by reducing financing constraints. Their findings suggest that financial reforms can optimize credit allocation, improve firms’ access to capital, and lower systemic risk. Similarly, Cai et al. (2024) investigated provincial banking competition and found that greater competition encourages corporate risk-taking by easing financing barriers. These studies underscore the critical role of local banking market dynamics in shaping corporate risk; however, they primarily emphasize banks and financial markets while often overlooking the role of government intervention. Institutional theory (North, 1990) argues that financial systems are embedded within broader institutional contexts in which government policy, market regulation, and financial institutions jointly shape corporate behavior.

In summary, the academic literature has provided comprehensive assessments of the drivers of corporate risk, with research on external determinants offering important context for this study. Nevertheless, a significant gap remains concerning the direct effects of financial decentralization on corporate risk. Existing research largely focuses on aspects of financial development—such as banking expansion and competition—while neglecting the financial autonomy of local governments, that is, financial decentralization. Financial decentralization, defined as the redistribution of fiscal authority between central and local governments, encompasses both market and non-market elements and is expected to exert substantial influence on corporate risk. By incorporating risk selection criteria into the credit decision-making processes of banks and firms, we propose a corporate risk decision model that explicitly accounts for local financial decentralization, thereby addressing the theoretical gap regarding its impact on corporate risk decisions. Building on this framework, we leverage the quasi-natural experimental setting of China’s RCC reform to establish a DID model, through which we empirically evaluate the influence of financial decentralization on the risk profiles of listed companies.

Theoretical model

This section, building on the literature summary, provides a theoretical model explaining how financial decentralization affects corporate risk and presents research hypotheses, which will be validated through subsequent empirical data. A more detailed discussion of the theoretical mechanisms will be addressed in “Heterogeneity analysis”.

Model setting

This study incorporates risk selection factors into the credit decision-making processes of banks and firms, constructing a corporate risk decision model that considers local financial decentralization. Subsequently, it examines the impact of local government financial decentralization on corporate risk.

Firm

To consider a firm’s choice regarding risk, this study takes into account a firm’s revenue function based on project investment to obtain returns. Suppose there are an infinite number of projects available in the market, each with varying probabilities of success, denoted by $1-p$, where $p$ is continuously distributed in the range $[\mathrm{0,1}]$, representing the risk associated with the firm’s project investment.

Assume that the firm expects to achieve a return of ${KL}$, where $K$ is a constant and $L$ is the capital investment. When obtaining a loan, the firm needs to provide collateral $C$, which the bank will reclaim if the investment fails. This study assumes that the firm has no own capital, and the capital $L$ is entirely provided by the bank. The cost ${nR}$ represents the firm’s expense to maintain its relationship with the local government $R$, with $n$ being the unit cost. Both $n$ and $R$ are firm-specific and exhibit some degree of stickiness. Since this study focuses on short-term decision changes, these variables are assumed to be exogenous.

Therefore, the expected profit for the company is as follows:

$$\begin{array}{c}E\left(\pi \right)={KL}-\left(1-p\right){Kr}-{pC}-\frac{1}{2}\delta {p}^{2}-{nR}\end{array}$$

(1)

Government

The central government determines the loan interest rate $r$, while the local government decides the intensity of financial decentralization $\xi$. The greater the intensity of financial decentralization, the more local banks consider the relationship $R$ between the government and firms when making financial decisions.

Bank

Next, we consider the bank’s decision. The bank offers loans to meet the borrowing needs of firms at an interest rate $r$ set by the central bank, with the loan amount being $L$ and the collateral being $C$. The bank’s expected profit function consists of two parts: the expected return from the firm and the utility provided by the local government’s financial decentralization control, which we assume to be $\xi {RL}$. Here, $\xi$ in $[\mathrm{0,1}]$. When $\xi =0$, it indicates that the local government does not participate in financial decentralization, and the bank’s loans to firms are not influenced by $R$. As $\xi$ increases, the decisions of local banks are increasingly influenced by the local government. The bank’s costs are divided into two parts: ${i}^{* }L$, representing the funding cost, where ${i}^{* }$ is the deposit interest rate;$\,\frac{1}{2}\beta {C}^{2}$, representing the cost of supervising and maintaining the loan collateral $C$. The expected utility for the bank is as follows:

$$\begin{array}{c}E\left(V\right)={rKL}+p\left(-L+C\right)-{i}^{* }L-\frac{1}{2}\beta {C}^{2}+\xi {RL}\end{array}$$

(2)

Equilibrium

The decisions made by firms and banks in the short term are as follows: (1) At time ${T}_{0}$, the firm applies for a loan from the bank, and the bank grants a loan amount $L$ with collateral $C$ depending on the firm’s circumstances. (2) At time ${T}_{1}$, the firm assesses its risk by considering both the loan amount $L$ and the collateral $C$, and then chooses a project from the market. Currently, $L$ and $C$ are externally provided, enabling problem-solving. The firm’s optimal decision-making model can be expressed in the following manner:

$$\begin{array}{c}\mathop{{Max}}\limits_{p}\pi ={KL}-\left(1-p\right){Lr}-{pC}-\frac{1}{2}\delta {p}^{2}-{nR}\end{array}$$

(3)

The first-order condition for maximizing the firm’s objective function with respect to risk $p$ can be derived from Eq. (2):

$$\begin{array}{c}\frac{\partial \pi }{\partial \rho }={Lr}-C-\delta p=0\end{array}$$

(4)

Equation (3) provides the firm’s optimal risk:

$$\begin{array}{c}p=\frac{{Lr}-C}{\delta }\end{array}$$

(5)

The relationship between $p$ and $L$ is positively correlated, meaning that as $L$ increases, $p$ also increases. On the other hand, the relationship between $p$ and $C$ is negatively correlated, indicating that as $C$ drops, $p$ lowers as well. Next, let’s examine the bank’s decision to maximize utility at time $T=0$:

$$\begin{array}{c}\mathop{{Max}}\limits_{p}V={rKL}+p\left(-L+C\right)-{i}^{* }L-\frac{1}{2}\beta {C}^{2}+\xi {RL}\end{array}$$

(6)

The relationship between local financial decentralization and corporate risk can be summarized as follows:

$$\begin{array}{c}\frac{\partial p}{\partial \xi }=R > 0\end{array}$$

(7)

Based on Eq. (7), we propose the following hypothesis:

H1: Local financial decentralization increases corporate risk.

Empirical model, variables, and data

This paper adopts a theoretical and empirical approach to explore the relationship between financial decentralization and corporate risk. The corporate credit decision model presented in Section “Theoretical model” provides the theoretical explanation. Next, the paper will conduct empirical validation of the research hypotheses. This section will comprehensively outline the models, data, and research methods used in the empirical analysis.

Empirical model

Most studies on financial decentralization measure its extent using indicators such as the proportion of local bank loans (Wegner, 2024; Wang, 2025; Xu, 2024; Zhu et al., 2025), bank competition (Fungáčová et al., 2017; Lai et al., 2023; Cai et al., 2024), bank expansion (Wang and Lee, 2023; Bhukta et al., 2025), and local government debt (He et al., 2019; Zhou et al., 2023). However, these indicators face limitations in accurately capturing changes in local governments’ financial control, and they often rely on continuous endogenous variables. Most studies employ OLS regression models to estimate the causal relationship between these variables and corporate risk, which leaves results vulnerable to endogeneity concerns. In addition, Zhang et al. (2024) applied the EADL model in their analysis. While this model can capture dynamic relationships, it requires extensive historical data to ensure stability; otherwise, it risks amplifying noise. If distributed lag weights are not properly selected, the model may also face numerical instability. Moreover, the EADL model adds complexity and increases the number of parameters, thereby reducing the intuitive interpretability of results. Given these limitations and the characteristics of the dataset, the EADL model is not suitable for this study.

Following the approach of Li et al. (2016), this study treats the restructuring of RCCs into RCBs as a quasi-natural experiment and applies a DID framework to assess the impact of financial decentralization on corporate risk. RCBs have stronger lending and deposit-mobilization capacities than RCCs, while local governments retain shareholder status and institutional control post-conversion. This deepens their influence over local finance, signifying a greater degree of decentralization. Unlike indicators such as bank competition or branch expansion, which are highly sensitive to macroeconomic conditions, the restructuring process is not driven by firm-level characteristics. Accordingly, the DID framework exploits this reform as an exogenous policy shock, dividing the sample into treatment and control groups. Firms in the treatment group are exposed to the reform after implementation, while those in the control group are not, thus approximating the logic of a randomized controlled trial and effectively mitigating endogeneity concerns. The DID model is presented as follows:

$$\begin{array}{c}{{Risk}}_{i,c,t}=\alpha +\beta {{Rstru}}_{c,t}+\gamma {{\boldsymbol{X}}}_{i.t}+{\tau }_{{it}}+{\eta }_{c}+{\varepsilon }_{i.t}\end{array}$$

(8)

The subscript $i$ denotes the firm, $c$ denotes the city, and $t$ denotes the year. The variable ${{Risk}}_{i,c,t}$ represents the level of risk that firm $i$ faces in city $c$ during year $t$. ${{Rstru}}_{c,t}$ is used as a proxy variable to indicate whether a region has restructured RCCs into RCBs. Prior to the RCC reform, the variable ${{Rstru}}_{c,t}$ had a value of 0. Following the reform, ${{Rstru}}_{c,t}$ has a value of 1. The variable ${{\boldsymbol{X}}}_{i.t}$ denotes region-level and corporation-level control factors that vary over time, controlling for potential company and regional factors that might influence corporate risk. ${\tau }_{{it}}$ denotes industry-year fixed effects, controlling for all factors that vary with the industry, while ${\eta }_{c}$ denotes city fixed effects, controlling for regional-level factors that do not change over time. ${\varepsilon }_{i.t}$ represents the random error term.

The primary objective of the DID model is to estimate the causal effect of a policy by comparing the differences between the treatment and control groups before and after the implementation of the policy. A critical prerequisite for this causal inference is the assumption of parallel trends. This assumption holds that only if the treatment and control groups exhibit identical trends prior to the implementation of the policy can we reasonably attribute any post-implementation differences between the groups to the policy itself, rather than to extraneous factors. In the robustness checks, we conduct tests for parallel trends. To mitigate potential biases from data mining, omitted variables, and other unknown factors that might produce spurious results, we also employ placebo tests by constructing fictitious policy shocks or treatment group samples to observe whether the model still yields similar significant findings in the absence of an actual policy impact. Furthermore, we enhance the model’s validity by varying the metrics for financial decentralization and corporate risk, adjusting combinations of fixed effects, and modifying the sample selection.

Variables

Independent variables

The principal explanatory variable in this study is financial decentralization. A commonly used indicator is the ratio of local bank loans to the total volume of national bank loans (Wegner, 2024; Wang, 2025; Xu, 2024; Zhu et al., 2025). However, because local financial institutions represent only a small share of the overall banking sector, this ratio primarily reflects financial development rather than the extent of local government intervention in financial affairs. A second category of indicators relates to banking development, including measures of bank competition (Fungáčová et al., 2017; Lai et al., 2023; Cai et al., 2024) and bank expansion (Wang and Lee, 2023; Bhukta et al., 2025). While the expansion of joint-stock and state-owned banks may increase competition or branch density, these metrics do not adequately capture the essence of financial decentralization. Importantly, institutions most subject to local government control are UCBs and RCBs. A third set of indicators considers the scale of local government bond issuance (He et al., 2019; Zhou et al., 2023), which more directly reflects the appropriation and mobilization of financial resources by local governments. Yet, debt disclosure is often irregular and opaque, with local governments concealing off-balance-sheet liabilities to evade regulatory oversight, leading to incomplete and distorted data. Moreover, all these indicators are continuous endogenous variables, making them susceptible to reverse causality with local economic scale, thereby posing endogeneity challenges in econometric analysis.

In this study, we adopt the transformation of RCCs into RCBs as our measure of financial decentralization for several reasons. Financial decentralization fundamentally entails local governments competing for rights over financial development and direct authority over financial resources. Historically, RCCs functioned as cooperative financial institutions, mobilizing idle rural capital to support local enterprises, though constrained by their limited scale. Restructuring RCCs into RCBs significantly expanded their capacity to attract deposits and broaden operations, thereby strengthening local governments’ financial influence. Because RCB revenues are closely tied to local fiscal deposits, their operations align closely with the objectives of financial decentralization. Thus, the RCC reform accurately reflects shifts in local government control and serves as an appropriate proxy for financial decentralization. Furthermore, employing the exogenous shock of the RCC reform as the basis for constructing explanatory variables enables the application of a DID framework to examine the causal relationship between financial decentralization and corporate risk, thereby mitigating concerns over endogeneity.

In the empirical analysis, “Rstru“ is utilized as a proxy variable indicating whether a given region has implemented the RCC reform. To further ensure robustness, the proportion of RCC and RCB branches relative to the total number of local bank branches is applied as an alternative explanatory variable in regression analyses.

Dependent variables

Two primary indicators are commonly used to measure corporate risk. The first is derived from market data, capturing volatility through stock returns (Jiang et al., 2024). The second is based on accounting data, including measures such as revenue volatility (Jiang et al., 2025a; Zhuang et al., 2025; Yu et al., 2025) and Z-scores (Yu et al., 2025). This study adopts accounting-based indicators for several reasons. The speculative nature of China’s stock market often generates sharp price fluctuations that may compromise the reliability of market-based measures. In contrast, accounting indicators provide a more accurate reflection of both financial and broader operational risk dynamics. Given that this study focuses on financial decentralization, and considering that its influence on corporate risk primarily operates through financial channels such as bank lending, accounting measures are especially appropriate. The securities market, prone to speculative distortions, offers a less precise reflection of risk changes associated with financial decentralization. Accordingly, this study employs accounting indicators to provide a more stable and representative assessment of corporate risk in the context of financial decentralization. Specifically, following prior research, corporate risk is measured using the volatility of corporate profits, with higher profit volatility indicating greater risk. Following Yu et al. (2025), the calculation is as follows:

$${{ROA}}_{i,t}=\frac{{{EBIT}}_{i,t}}{{{Asset}}_{i,t}}$$

(9)

$${Adj}{{\_}}{{ROA}}_{i,t}={{ROA}}_{i,t}-\frac{1}{n}\mathop{\sum}\limits_{t=1}^{n}{{ROA}}_{i,t}$$

(10)

$${{Risk}1}_{i,t}=\sqrt{\frac{1}{T-1}\mathop{\sum}\nolimits_{t=1}^{T}{\left({{Adj}{{\_}}{ROA}}_{i,t}-\frac{1}{T}\mathop{\sum}\nolimits_{t=1}^{T}{Adj}{{\_}}{{ROA}}_{i,t}\right)}^{2}},T=3$$

(11)

$${{Risk}2}_{i,t}={Max}\left({Adj}{{\_}}{{ROA}}_{i,t}\right)-{Min}\left({Adj}{{\_}}{{ROA}}_{i,t}\right),$$

(12)

where ${ROA}$ represents the ratio of a company’s pre-interest and tax profit to total assets; ${Adj\_ROA}$ represents an industry-adjusted ROA, which is used to control for industry and cyclical variations. ${Risk}1$ denotes the rolling 3-year standard deviation of ${Adj\_ROA}$, and ${Risk}2$ represents the rolling 3-year range (the difference between the maximum and minimum values) of ${Adj\_ROA}$. For the convenience of analysis, we standardize ${Risk}1$ and ${Risk}2$.

Control variables

Certain key financial metrics can signal shifts in a corporation’s risk profile. For example, excessive indebtedness can erode operational cash flows. Firms with a high debt-to-asset ratio and limited cash reserves are more likely to experience cash flow disruptions (Chen and Jiang, 2024; Zhou et al., 2024). Tangible assets are an important proxy for collateral in financing activities (Jiang et al., 2022). An increase in tangible assets can ease financing constraints (Jiang et al., 2025b), thereby lowering operational risk. Profitability indicators capture a firm’s ability to generate cash flows and earnings (Chen et al., 2023; Du et al., 2024). Companies with weak growth prospects and low gross margins may struggle to establish a competitive position in product markets, facing elevated operational risks (Beglaryan et al., 2024). Following prior studies (Ji et al., 2021; Faccio et al., 2016), this paper incorporates several firm-level control variables: firm size (Size), firm age (Age), debt-to-asset ratio (Lev), cash-to-assets ratio (Cash), average tangible asset ratio (Tang), revenue growth rate (Growth), gross profit margin (Margin), and effective tax rate (Tax). Since corporate risk varies across years and industries, we include year-fixed, industry-fixed, and region-fixed effects.

In addition, we control for regional economic fluctuations. On one hand, enterprise risk is inevitably influenced by the regional economic environment. Economic downturns reduce demand, profitability, and tax subsidies, thereby amplifying operational risks. On the other hand, although we treat the restructuring of RCCs into RCBs as relatively exogenous, it remains necessary to account for regional economic variables to mitigate potential confounding effects arising from correlations between financial decentralization and regional conditions. Accordingly, in subsequent robustness checks, we exclude the influence of time-varying city-level factors, specifically per capita GDP (PerGDP) and total fixed asset investment (Fixed).

Data

The time period of the empirical sample in this study spans from 2007 to 2018, for two primary reasons. First, in 2007, the China Securities Regulatory Commission (CSRC) implemented new regulations on the disclosure of information by listed companies. These regulations aimed to standardize financial disclosure practices, prompting us to begin our sample period in 2007 to ensure consistency in data quality and reporting standards throughout the analysis. Second, we use the reform of China’s RCCs as a key measure to assess the extent of implicit financial decentralization across regions. After reviewing relevant documentation, we found that the reforms were largely completed by 2018, marking a natural endpoint for the sample period. Therefore, the final sample dataset ends in 2018. This timeframe was chosen based on key regulatory and reform milestones that significantly influence the scope and quality of financial data available for analysis. In the robustness checks, we extend the data period to the most recent years to ensure the robustness of the conclusions.

Data for 3508 listed companies were obtained from the CSMAR database. Using this comprehensive dataset, we constructed control variables, mechanism variables, and additional variables required for the analysis. Following the methodologies of previous research (Jiang et al., 2025b), we excluded companies designated as ST, aST, or PT, as well as firms in the financial sector, due to their distinctive financial structures and regulatory environments. We also excluded firms with extensive missing data, long-term trading suspensions, or records of financial fraud to ensure the integrity and reliability of the results. To further enhance robustness and minimize the impact of extreme values, all continuous corporate variables were winsorized at the 1% level, effectively mitigating the influence of outliers. This careful selection and treatment of data ensure that the study provides a precise and reliable assessment of the impact of financial decentralization on corporate risk.

Definitions and explanations of the variables are provided in Table A1 of the Appendix, while Table 1 reports descriptive statistics of the main variables. The risk indicators, Risk1 and Risk2, are standardized with a mean of 0 and a standard deviation of 1. The large gap between their minimum and maximum values suggests considerable variation in corporate risk levels, underscoring the importance of studying its determinants. The values of the control variables are generally consistent with those reported in existing literature (e.g., Chen and Jiang, 2024; Jiang et al., 2020), further confirming the validity of the sample selection.

Table 1 Summary statistics.

Empirical results

Benchmark regression

The findings, delineated in Table 2, reveal a statistically significant increase in corporate risk following the RCC restructuring, irrespective of the inclusion of control variables or the decision to use a 3-year rolling range or standard deviation as the measure. Initially, regression results for the entire sample, both with and without firm-specific attributes as controls, are presented in columns (1), (2), (5), and (6). Subsequent adjustments exclude data from the real estate and financial sectors due to the heightened risks associated with these industries and the distinct impact of regional financial decentralization on these firms. The adjusted regression outcomes, displayed in the third, fourth, seventh, and eighth columns of Table 2, demonstrate that the coefficients, which remain significantly positive at the 1% level, have increased in magnitude.

Table 2 Benchmark regressions: the effect of RCC reform on corporate risk.

The results from columns four and eight serve as the baseline for further analysis, showing coefficients of 0.060 and 0.056, respectively, indicating that an increase in Rstru by one unit leads to an increase in Risk1 and Risk2 by 0.060 and 0.056 standard deviations, respectively. This suggests that the establishment of RCBs in a region escalates the corporate risk for firms outside the real estate and financial sectors.

These empirical findings support our theoretical claim that greater financial decentralization led by local governments increases corporate risk, thereby confirming Hypothesis H1 of our model. This aligns with Jiang et al. (2020), who documented a negative association between bank deregulation and corporate risk. In contrast, within the context of this study, regional financial decentralization expands the scale of RCBs, reduces banking competition, and channels politically driven capital flows, thereby tightening financing constraints and elevating corporate risk. Moreover, financial decentralization enables local governments to divert financial resources toward preferred sectors such as infrastructure, often at the expense of the real economy. This reallocation worsens firms’ financing environment, intensifying the risks associated with financing constraints. These mechanisms will be examined in greater detail in the subsequent analysis.

The results for the control variables indicate that firm size, tangible asset ratio, and income tax rate are negatively associated with corporate risk, whereas leverage is positively correlated. This suggests that highly leveraged firms face greater operational risks, as debt obligations restrict cash flow and hinder business operations. By contrast, tangible assets can serve as collateral, easing financing constraints and reducing risk. Larger firms generally benefit from more stable markets and greater tangible resources, resulting in lower operational risk. Other variables, including firm age, cash-to-asset ratio, and revenue growth rate, exhibit no significant relationship with corporate risk, likely due to their complex interactions with risk dynamics and inter-variable correlations.

Robustness tests

Parallel trends test

The DID method has strict assumptions, requiring that the control group and the treatment group have consistent trends before the event. This study conducts a parallel trends test using the following model:

$${{risk}}_{i,t}=\beta +\mathop{\sum}\limits_{T=-4}^{-2}{\alpha }_{T}{{Rstru}}_{c,T}+\mathop{\sum}\limits_{T=0}^{4}{\alpha }_{T}{{Rstru}}_{c,T}+\gamma {X}_{i.t}+{\tau }_{{It}}+{\eta }_{c}+{\varepsilon }_{i.t}$$

(13)

where ${{Rstru}}_{c,T}$ is a dummy variable representing the city c in the year T before or after the restructuring of RCCs. For instance, when T = −2, it takes the value of 1 for the 2 periods before the restructuring, and 0 for other years. This parallel trend test includes the entire sample, and periods greater than 4 or less than −4 are grouped into T = 4 and T = −4. To avoid multicollinearity, the study follows the approach of Beck et al. (2010) and takes t = −1 as the base year.

The analysis incorporates consistent variable settings with the baseline regression model. Figures 1 and 2 illustrate the results of the parallel trend tests for Risk1 and Risk2, respectively. Both Figures confirm that, prior to T₀, all estimated regression coefficients are statistically insignificant, affirming that the trends between the control and treatment groups were aligned before the restructuring of RCCs, thereby showing no significant pre-existing differences. Post T₀, the regression coefficients become significantly positive, indicating a substantial increase in corporate risk within the treatment group relative to the control group. This pattern underscores the significant influences of RCC restructuring on the corporate risk of affected firms.

Fig. 1: Parallel trends test I.

The coefficient value and confidence interval of ${{Rstru}}_{c,T}$ in Model (13) establishing ${Risk}1$ as the dependent variable.

Fig. 2: Parallel trends test II.

The coefficient value and confidence interval of ${{Rstru}}_{c,T}$ in Model (13) establishing ${Risk}2$ as the dependent variable.

The lack of significance in the coefficients for the T₀ time period, as shown in Figs. 1 and 2, can be attributed to two primary factors. First, several instances of RCC restructuring in the sample occurred towards the year’s end, minimizing their immediate impact on corporate risk within that fiscal year. Second, the influence of RCC restructuring on corporate risk predominantly transpires through financial mechanisms, such as the provisioning of loans, which inherently exhibit a temporal delay. Companies typically perceive the effects only after their existing loans reach maturity. Consequently, the impact of RCC restructuring on corporate risk may not manifest until after the restructuring year, indicating a potential lag in the effect. This delayed response highlights the temporal dynamics at play in the financial repercussions of RCC restructuring on corporate risk.

5.2.2 Placebo test: shifting the event impact beforehand

This study implements a placebo test by adjusting the timeline of RCC restructurings. It hypothesizes that these restructurings took place 1–6 years earlier than they actually did and assesses the potential impact on corporate risk. The robustness of the DID methodology is tested by this approach; specifically, a robust baseline regression would exhibit insignificant coefficients when the timing of the restructuring is artificially advanced. The regression outputs, detailed in columns (1–6) of Table 3, confirm that after adjusting the timing of the restructuring event, the coefficients for proxies related to financial decentralization shocks remain statistically insignificant. This outcome supports the integrity of the baseline model, indicating that actual temporal changes associated with the restructuring events drive the observed variations in corporate risk. The above results indicate that the fabricated policy shock cannot explain the changes in firm risk, ruling out the possibility that the baseline effect is driven by other policy shocks.

Table 3 Placebo test: shifting the event impact beforehand.

Placebo test: replacing the dependent variable

Corporate risk indicators encompass both market data-based and financial data-based risks. This study focuses on indicators derived from financial data. Benchmark regression analysis reveals that the restructuring of RCCs positively influences local business risk due to alterations in the local financial environment prompted by these restructuring activities. In contrast, market data-based risks, such as stock return volatility, are generally less sensitive to local policies that do not directly relate to businesses and are predominantly influenced by corporate strategies, industry developments, and similar factors.

Consequently, it is posited that the restructuring of RCCs primarily affects operational risk rather than the financial market risk of businesses. To validate this hypothesis, the paper substitutes the previous business risk indicators with stock return volatility. If the explanatory variables’ estimated coefficients remain significant, it would imply that the observed increase in business risk might be attributable to factors other than the restructuring of RCCs, possibly indicating a broader increase in business risk. The pertinent regression results, presented in Column (1) of Table 4, demonstrate that the coefficient estimates for key variables are not significant, suggesting that the restructuring of RCCs does not influence the market risk levels of businesses. This outcome substantiates that the conclusions drawn in this paper are not influenced by extraneous variables.

Table 4 Placebo test: replacing the dependent variable.

Placebo test: randomly selecting the treatment group

In empirical studies, fully isolating all factors influencing corporate risk is a formidable task, and random variables can introduce estimation biases. To mitigate these issues, this study employs a placebo test involving the random selection of an experimental group.

The methodology applied in this analysis includes several key steps: Initially, each prefecture-level city is assigned a unique identifier (ID), and these IDs are preserved alongside the actual year of RCC restructuring. The restructuring year is maintained, while a new variable, FN, is generated randomly. Utilizing the FN sequence, a fictitious variable, Fake_id, is created and subsequently integrated with the original dataset to generate Fake_year, which represents an imagined restructuring timeline. This procedure is replicated 1000 times, and each instance undergoes regression analysis using the fictitious restructuring year while maintaining all other variables constant.

Figures 3 and 4 display the probability density distribution of the regression coefficients, with a dashed line indicating the position of the benchmark regression coefficient. The distribution of coefficients from the random sampling is approximately normal, centered around zero, and most coefficients are notably distinct from the benchmark regression coefficient. This pattern in the distribution confirms that the findings of this study are not artifacts of random factors, thereby validating the robustness of the regression results.

Fig. 3: Placebo test I.

Probability density distribution of the coefficients of ${Rstru}$ in Model (8) establishing ${Risk}1$ as the dependent variable when randomly selecting the treatment group.

Fig. 4: Placebo test II.

Probability density distribution of the coefficients of ${Rstru}$ in Model (8) establishing ${Risk}2$ as the dependent variable when randomly selecting the treatment group.

In summary, our DID model has passed the necessary prerequisite assumption tests, including the parallel trend test and the placebo test. Furthermore, we conducted a series of robustness checks, including adding more control variables, using local government debt as an alternative measure of financial decentralization, employing the industrial enterprise database, adjusting combinations of fixed effects and clustering levels, and applying machine learning analysis. The results of the robustness checks (see Tables A2–A6 in the Appendix) consistently support the validity of the baseline conclusions of this study.

Mechanism analysis

Financial institution crowding out effect

This paper argues that local governments gain effective control over financial resources through the restructuring of RCCs, thereby achieving implicit financial decentralization. The baseline regression results show that RCC restructuring increases the risk borne by local enterprises, indicating that this form of decentralization influences corporate risk. However, the underlying mechanisms remain unclear. In Section “Benchmark regression,” drawing on Jiang et al. (2020), we propose that the increase in corporate risk arises because RCC restructuring expands the scale of RCBs, crowds out other banks, and reduces banking competition, which we term the “financial institution crowding-out effect.”

The crowding-out of central banks and the reduction in competition affect corporate risk in several ways. First, reduced bank competition exacerbates financing constraints for enterprises (Jiang et al., 2020). Easing financing constraints allows firms to borrow more readily in the face of short-term shocks, thereby reducing risk. Alleviating financing constraints also mitigates moral hazard and adverse selection, encouraging firms to undertake lower-risk investments (Chen and Jiang, 2024). In contrast, financial decentralization reduces competition by crowding out other banks, thereby tightening financing channels and increasing risk. Second, bank competition promotes financial innovation, broadening access to financing methods such as supply chain finance and intellectual property pledge financing. This diversification reduces dependence on a single financing channel, lowering financial vulnerability. A decline in competition, however, weakens the alignment between corporate assets and financial services, heightening mismatches between investment and financing and raising operational risks. Third, stronger competition compels banks to improve service quality, including providing specialized risk management and hedging strategies for market and exchange rate risks. The absence of such services under a crowding-out effect deprives firms of valuable support, increasing operational risks. Finally, competition enhances fund settlement efficiency. Rapid and accurate settlement improves cash flow management, lowers transaction costs, and reduces the risk of cash flow interruptions. Conversely, reduced competition undermines settlement efficiency, further aggravating operational risks. Overall, the financial institution crowding-out effect induced by RCC restructuring contributes to higher corporate risk by constraining financing, reducing innovation, limiting risk management support, and impairing cash flow stability.

In this section, we investigate the financial institution crowding out effect at the city level. The study uses the percentage of a specific type of financial branch institution exiting, relative to the total number of all financial institutions, as the dependent variable. This approach allows for the examination of the impact of RCC reforms on other financial branch institutions within the region. This study considers the crowding out effects of RCCs reform on state-owned large banks (${Quit}1$) and joint-stock commercial banks (${Quit}2$), with the models set as follows:

$${{Quit}}_{c,t}=\alpha +\beta {{Rstru}}_{c,t}+\gamma {X}_{c,t}+{\tau }_{t}+{\eta }_{c}+{\varepsilon }_{c.t}$$

(14)

where ${{Quit}}_{c,t}$ represents the percentage of a specific type of financial branch institution exiting in city $c$ at time $t$, ${{Rstru}}_{c,t}$ indicates whether city $c$ has undergone the reform of RCCs at year $t$, ${X}_{c,t}$ represents relevant control variables for city $c$, including per capita GDP (${PerGDP}$), total fixed asset investment (${Fixed}$), and the logarithm of total population (${PoP}$), ${\tau }_{t}$ denotes time fixed effects, ${\eta }_{c}$ signifies city fixed effects, and ${\varepsilon }_{c.t}$ stands for the error term.

Our empirical evidence indicates that the reform of RCCs does not significantly affect the exit of regular joint-stock commercial bank branches. However, it has a significant positive effect on the exit of state-owned large bank branches. The regression results, presented in columns (2), (3), (5), and (6) of Table 5, show coefficients of 0.0528, 0.0489, 0.0481, and 0.0455, respectively. Specifically, after the completion of RCC reforms, an average of 8.16% of state-owned large bank branches exit the region. This finding supports the use of RCC reform as a proxy for implicit financial decentralization. However, when the dependent variable is Quit2, the estimated coefficient for Rstru is not significant, suggesting that the RCC reform did not negatively impact the market share of joint-stock banks.

Table 5 Mechanism analysis: financial institution crowding out effect.

To explain this phenomenon in the context of China’s financial development, we propose two potential reasons. Historically, China’s banking market—and, more broadly, the financial market—has been dominated by the five major state-owned banks. Consequently, the financial institutions most affected by financial decentralization were the branches of state-owned banks, which were forced to exit local markets due to increased competition. In contrast, joint-stock banks initially had fewer branches. During the sample period, the CBRC relaxed restrictions on the cross-regional establishment of bank branches, which enabled some joint-stock banks to expand rapidly. As a result, the number of joint-stock bank branches remained largely unaffected by the effects of financial decentralization. Another important reason lies in the differing responses of various types of banks to competition. Joint-stock banks, being smaller and more flexible, have leaner organizational structures and shorter decision-making processes. In response to financial decentralization, they optimize cost management and improve service quality. Conversely, state-owned banks, due to their larger scale and redundant organizational structures, often suffer from low management efficiency. Cost-control strategies, such as reducing branch networks and streamlining non-core personnel, are more effective for state-owned banks in coping with financial competition. Therefore, state-owned bank branches are more likely to exit local markets as a result of financial decentralization.

In summary, we find that the reform of RCCs crowds out central control over local finance, thereby increasing local control over regional finance. This mechanism is one way in which the reform of RCCs affects enterprise risk. Importantly, this effect does not extend to regular joint-stock commercial banks.

Corporate loan crowding out effect

Section “Fiscal decentralization” shows that RCC reform influences corporate risk by crowding out large state-owned bank branches. The relationship between financial institutions and enterprises is largely defined by borrowing and lending, meaning the most direct effect on firms stems from shifts in borrowing capacity. Financial decentralization also enables local governments to capture financial resources. With a relatively fixed financial supply, more funds are directed toward government-preferred sectors such as infrastructure and real estate, leaving fewer resources available for the real economy. This crowding effect worsens the financing environment and amplifies financing constraints.

The corporate loan crowding-out effect affects corporate risk in several ways. First, a reduction in bank loans decreases the external funds available to firms. Since operations such as purchasing raw materials and paying wages require steady cash flow, diminished financing can disrupt production processes, raising operational risk. Second, corporate expansion often demands substantial capital for building facilities, acquiring equipment, or entering new markets. A decline in bank loans deprives firms of critical financing, constraining their ability to expand production, seize market share, and enhance competitiveness (Zhou et al., 2023). Third, many firms rely on long-term investments for technological upgrades and innovation. When bank loans shrink, enterprises struggle to sustain these projects, potentially falling behind competitors and facing higher operational risks. Finally, without sufficient financing, firms cannot pursue scale expansion, weakening their price competitiveness and making them more vulnerable to losing market share. In sum, by restricting access to bank loans, RCC reform heightens corporate risk through reduced liquidity, constrained expansion, stifled innovation, and diminished competitiveness.

In this section, we examine the mechanisms through which reforms in RCCs influence corporate risk via corporate bank lending practices. We operationalize the volume of newly issued loans in the current period by calculating the difference between the current and previous total loan amounts. Given the strong association between changes in a firm’s loan amount and its inherent characteristics, we employ the ratio of newly issued loans to the previous total loan amount (nloantl) as our analytical variable. The regression findings are presented in columns (1–3) of Table 6.

Table 6 Mechanism analysis: corporate bank loan crowding out effect.

The analysis in the first column reveals that RCC reform exerts a significant negative impact on the volume of corporate newly issued loans. Results from the second and third columns illustrate that the variation in newly issued loans mediates the effect of RCC reform on corporate risk, as indicated by the risk proxy variables Risk1 and Risk2. Notably, relative to the baseline regression, the absolute value of the estimated coefficient for the restructuring variable Rstru decreases significantly, along with a reduction in its statistical significance. This trend is particularly pronounced when Risk2 is the dependent variable, where the coefficient for Rstru remains positive but becomes statistically non-significant. The diminished significance of this coefficient implies that the effect of financial decentralization on pronounced fluctuations in corporate revenue is predominantly channeled through changes in bank loan allocations. This observation suggests a partial mediation effect, though it may not be complete. Collectively, these results indicate that RCC reform impacts corporate credit channels to a notable extent, subsequently influencing corporate operational risk.

Bank loans are typically classified into four categories based on their nature: mortgage loans, pledge loans, guarantee loans, and credit loans. Our dataset comprises current balances of mortgage loans, pledge loans, and unsecured loans. We measure changes in the shares of these loan types by using the ratio of the difference between the current and previous balances to the total loan amount. Unsecured loans, which lack collateral and guarantees, represent a higher risk for banks. Consequently, during periods of reduced lending willingness, banks often curtail the provision of credit loans to enterprises. In contrast, while pledge and mortgage loans are deemed higher-quality loans by banks, they entail higher default costs for enterprises. Thus, amidst constrained credit availability, banks tend to decrease credit loan issuance, whereas enterprises are likely to lessen their applications for pledge and mortgage loans.

The regression results are displayed in columns (4–6) of Table 6, analyzing the ratios of changes in mortgage loan amounts (Dyloantl), unsecured loan amounts (Cloantl), and pledge loan amounts (Zyloantl). These results indicate that RCC reform significantly negatively impacts both mortgage and unsecured loans, suggesting shifts in both banks’ lending behaviors and enterprises’ borrowing intentions due to increased uncertainty about future borrowing conditions.

It is important to note that the regression coefficient for Rstru is not significant when Zyloantl is the dependent variable. This result suggests that financial decentralization does not negatively affect the scale of collateralized loans. The finding is consistent with the trajectory of China’s financial development, in which traditional commercial banks have long dominated the market. Given the underdeveloped state of financial technology and the limited range of financial instruments in China, commercial banks have historically relied heavily on collateralized assets to mitigate lending risks. This reliance underscores the crucial role of tangible assets in easing financing constraints for Chinese firms. When liquidity in the financial market tightens, credit loans—based on weaker credit relationships—are typically the first to be reduced, whereas collateralized loans, secured by tangible assets, remain relatively stable. The persistence of collateralized lending, even under conditions of increasing financial decentralization, illustrates its robustness and highlights the continued importance of tangible assets in China’s evolving financial landscape.

In summary, this section confirms the crowding-out effect of financial decentralization on corporate loans; however, it only applies to credit.

Heterogeneity analysis

Fiscal decentralization

Research conducted by Qian and Roland (1998) underscores that fiscal decentralization significantly influences local economic development. The extent of fiscal decentralization mirrors a region’s demand for fiscal autonomy; regions with a robust desire for autonomy are more inclined to pursue higher levels of fiscal decentralization. We hypothesize that the influence of RCC reform on corporate risk is primarily attributable to local governments’ aspirations to enhance their financial authority via such reforms. So, whether the motivation for this financial control is driven by fiscal conditions is the question we need to address in this section.

We believe that in regions with a higher degree of fiscal decentralization, financial decentralization has a stronger exacerbating effect on corporate risk. According to Qian and Roland (1998), in regions with stronger fiscal decentralization, the government has a stronger motivation to develop the local economy. Under China’s tax-sharing system, the responsibilities and tax rights of local governments are not aligned, making it difficult for local governments to support the economic management tasks assigned by the central government with their own tax revenues. The greater the degree of local fiscal decentralization, the larger the funding gap required for economic development. Therefore, in regions with higher fiscal decentralization, local governments have a stronger motivation to control financial resources to cover fiscal deficits. The crowding-out effect of financial decentralization is stronger, causing more significant negative impacts on business operations.

To explore this hypothesis, we categorized our sample into two groups based on the degree of fiscal decentralization—strong and weak. We then investigated the role of local governments’ fiscal decentralization in modulating the impact of RCC reform on corporate risk. We employed two indicators of fiscal decentralization, calculated from the ratios of prefecture-level cities’ fiscal revenue to the national fiscal expenditure.

The regression outcomes, presented in Table 7, delineate these relationships. Columns (1), (3), (5), and (7) display results from the samples with strong fiscal expenditure decentralization, whereas columns (2), (4), (6), and (8) correspond to those with weak fiscal decentralization. It was observed that the phenomenon of financial decentralization intensifying corporate risk is significant solely in samples with strong fiscal expenditure decentralization and is inconsequential in weaker samples. This pattern persists even when fiscal revenue decentralization metrics are applied to subsamples. The findings in Table 7 reveal that RCC reform significantly elevates corporate risk exclusively in regions characterized by substantial fiscal decentralization, irrespective of whether the focus is on revenue or expenditure decentralization. This variance suggests that the impact of local RCC reform is contingent upon the fiscal environment. The differential impacts likely stem from the diverse objectives pursued by local governments in regions with disparate levels of fiscal decentralization. In jurisdictions with significant fiscal decentralization, local authorities seek to consolidate financial control as a strategy to amplify their fiscal authority. This nuanced understanding of the interplay between fiscal policies and corporate risk highlights the complex dynamics at play in regions undergoing financial and fiscal reforms.

Table 7 Heterogeneity analysis: based on fiscal decentralization.

In regions with weak fiscal decentralization—as shown in columns (2), (4), (6), and (8)—local governments have a much weaker incentive to assert control over financial resources, resulting in negligible impacts on corporate risk. This finding is particularly important and warrants further discussion. The analysis so far has demonstrated that the mechanism through which financial decentralization raises corporate risk is rooted in RCC reforms. These reforms have strengthened local governments’ capacity to access and control financial resources, which can crowd out private-sector financing, tighten financing constraints, and heighten operational risks for enterprises. By contrast, in regions with lower levels of fiscal decentralization, local governments have little motivation to capture financial resources for regional development. As a result, even with RCC reforms, there is no substantial shift in local control of financial resources, and corporate risk does not significantly increase. This is reflected in the insignificance of the coefficient estimate for the restructuring variable Rstru. These findings highlight that RCC reforms alone—particularly their institutional improvements—do not inherently impose negative effects on business operations. The critical factor lies in the appropriation of financial resources by local governments. It is this implicit financial decentralization that amplifies corporate risk. This nuanced result underscores the complex interaction between fiscal decentralization, financial reforms, and corporate stability, suggesting that fiscal and financial decentralization must be carefully coordinated and managed to prevent unintended adverse consequences for the regional economy.

Industrial policies

Industrial policies are critical instruments for steering local economic development and transitioning economic growth patterns, as highlighted by Alder et al. (2016). The strategic emphasis on specific industries is especially significant for the developmental agendas of local governments. In this vein, while RCC reform is interpreted as a form of covert financial decentralization, and our baseline regression indicates that such decentralization enhances corporate risk, it is hypothesized that this impact may be mitigated for enterprises operating within industries that are designated as key priorities by local governments. In this section, we will further explain the financial inequality caused by local government intervention by utilizing industry differences.

We believe that the negative impact of financial decentralization on corporate risk may be weaker in key industries. These key industries are not merely sectors of economic focus but are also areas where enterprises receive heightened surveillance and support from local authorities. This protective oversight implies that businesses within these priority sectors are less susceptible to the adverse effects typically associated with financial decentralization. The rationale behind this hypothesis is that local governments, by channeling resources and regulatory favor towards these industries, may effectively shield them from the broader destabilizing influences of financial decentralization. Consequently, the assumption is that while financial decentralization generally escalates corporate risk through mechanisms such as increased competition for financial resources or reduced financial stability, this effect is attenuated in key industries. These sectors likely benefit from more stable financial environments and supportive policies that counterbalance the risks associated with broader financial decentralization. This nuanced interaction suggests that the impact of financial decentralization on corporate risk is not uniform across all sectors but is contingent on the strategic economic priorities of local governments.

To test this hypothesis, we categorize the entire sample into key industry enterprise samples and non-key industry enterprise samples¹. The regression results are reported in Table 8. Columns (1) and (3), which present the coefficients for enterprises in key industries, show negative but insignificant values, differing from the baseline regression. In contrast, columns (2) and (4), which report results for enterprises in non-key industries, yield coefficients consistent with the baseline regression. These findings indicate that the increase in corporate risk induced by financial decentralization is concentrated among non-key industry enterprises. This suggests that the risk effect of financial decentralization is selective, exerting limited influence on government-supported industries. The results also validate the effectiveness of using RCC reforms as a proxy for implicit financial decentralization. Enterprises in industries outside the scope of government priorities face tighter financing constraints and, consequently, heightened risks following RCC reforms. By contrast, firms in strategically important sectors remain insulated from such effects, as the government continues to ensure the flow of capital into these industries.

Table 8 Heterogeneity analysis: based on industrial policies.

Firm ownership

In Section “Policy background,” we posited that reforms of RCC effectively represent an enhancement in local government control over financial institutions, which can be characterized as covert financial decentralization. This shift implies a deeper involvement of local governments in the financial sector, ostensibly to enhance local financial autonomy and control. In this section, we reveal the differences in the impact faced by enterprises with different ownership structures when dealing with financial decentralization. This analysis allows us to further emphasize the importance of government relations in determining the relationship between financial decentralization and corporate risk.

We believe that the risk effects of financial decentralization are weaker in SOEs. SOEs, which are owned by both central or local governments, typically enjoy a degree of implicit guarantees from these entities. These guarantees often manifest as financial support during downturns or preferential treatment in regulatory and policy frameworks, providing a buffer against market fluctuations and financial instabilities. Given this backdrop, if our conjecture holds—that RCC reform symbolizes an increase in local financial decentralization—then SOEs are likely insulated from the increased corporate risks typically associated with such decentralization. This insulation stems from the implicit guarantees that local governments provide, which serve to stabilize the financial and operational environments of SOEs. In essence, these guarantees act as a risk mitigant, shielding SOEs from the immediate impacts of market and financial pressures that might otherwise be exacerbated by shifts towards more localized financial control. Thus, in the context of RCC reforms leading to increased local financial decentralization, SOEs might not only avoid heightened corporate risk but could potentially leverage this reform to secure more favorable financial and operational conditions facilitated by local government support. Therefore, the relationship between financial decentralization and corporate risk in SOEs illustrates a nuanced dynamic where the typical risks associated with decentralization are counterbalanced by governmental guarantees and support. This dynamic underscore the complex interplay between government policy, financial decentralization, and corporate stability in the state sector.

We divide the entire sample into SOE and non-SOE categories and conduct separate regression analyses for each. The results are presented in Table 9. It is evident that, whether using ${Risk}1$ or ${Risk}2$ as the proxy variable, the regression results for non-SOE samples align with the baseline regression results. In contrast, the regression coefficients for SOE samples are all negative but not significant, marking a significant difference from non-SOE samples. This result validates the hypothesis that, for SOEs benefiting from implicit guarantees provided by local governments, financial decentralization does not affect corporate risk. However, it significantly increases the risk for non-SOEs without such implicit government support. Similarly, this outcome validates that RCC reforms can represent a form of implicit financial decentralization. RCC reforms are able to reflect the financial interventions of local governments, which in turn result in the redirection of funds towards enterprises that are the focal point of local government attention. This redirection leads to reduced financing for enterprises not prioritized by local governments, thereby increasing their risk exposure.

Table 9 Heterogeneity analysis: based on firm ownership.

Firm characteristics

In this section, we reveal the differences in the impact on businesses with various operating conditions when facing financial decentralization. This analysis allows us to enrich our research conclusions, identify more businesses that need attention and support, and thereby formulate and refine detailed policy recommendations.

We believe that the risk effects of financial decentralization are weaker in larger enterprises and those with higher profit margins. In the mechanism analysis section, we delved into how financial decentralization influences corporate risk, particularly through the mechanism of crowding out enterprises’ access to bank loans. This crowding out effect results from local governments redirecting financial resources towards favored sectors or entities, thus limiting the availability of credit for other enterprises. The bargaining power of a company within the financial market is crucial in this context and is influenced by various firm characteristics, including company size, which is directly associated with financial constraints (Zheng et al., 2025). Companies that possess weaker bargaining power—typically smaller enterprises—are more likely to be adversely impacted under conditions where credit resources are constrained. These companies find it increasingly difficult to secure necessary financing, thereby heightening their risk exposure in scenarios of financial decentralization. Conversely, large enterprises often serve as pillars of local economic development and tend to maintain close relationships with government entities. This proximity to power not only provides them with better access to financial resources but also often ensures their prioritization in government policies. Previous discussions have highlighted that local governments frequently channel financial resources towards supported industries and SOEs. In a similar vein, large enterprises, which play a significant role in local economic stability and growth, often receive favorable treatment from local governments. As a result of these dynamics, the negative impacts of financial decentralization on corporate risk are substantially mitigated for large enterprises. This mitigation arises because these enterprises, due to their size and strategic importance, are less likely to face the brunt of financial resource scarcity. Instead, they benefit from continued support and preferential access to credit, facilitated by their symbiotic relationships with local governments. This differential impact underscores the importance of enterprise size and government relationships in moderating the effects of financial decentralization on corporate risk.

This study employs two indicators, firm size and firm profit margin, to represent firm characteristics and investigates the differential impact of financial decentralization on corporate risk. The regression results are presented in Table 10.

Table 10 Heterogeneity analysis: based on firm characteristics.

We divide the sample into subsamples based on firm size, measured by total assets. The results of using different size samples are presented in columns (1–4). It is evident that financial decentralization only increases risk for small-scale companies. We believe that the observed outcomes are grounded in practical reasons—listed large enterprises possess a strong local influence, making significant contributions to taxation and employment, which crucially impact local government operations. Consequently, their relationships with the government are notably closer. Driven by performance metrics, local governments may relax financial constraints on these enterprises. Therefore, even with financial decentralization, where local governments possess enhanced influence in financial markets, large listed enterprises face less financial restriction, thus their risk exposure remains minimal. In contrast, small enterprises do not maintain as close relationships with the government as their larger counterparts, leading to more significant financial constraints post-financial decentralization, which in turn increases their risk levels.

This study also categorizes the sample based on enterprise profit margin. Companies with higher profit margins generally have better future repayment capabilities, leading to lower credit constraints. Consistent with the company size grouping, financial decentralization only increases risk for companies with low profit margins. Profit margins reflect a company’s future cash flows, and enterprises with strong profitability levels are less impacted by financial decentralization, primarily for two reasons. First, financial institutions prioritize loan distribution to businesses with robust future repayment capacities. Consequently, enterprises with higher profitability experience less financing constraints compared to those with lower profitability, making them less susceptible to the effects of financial decentralization. Second, companies with higher cash flows have a stronger capacity to cope with adverse shocks. Even in the presence of financing constraints, these companies are unlikely to face significant risks. Therefore, enterprises with high profit levels are minimally affected by financial decentralization.

In summary, this section conducts a heterogeneity analysis and establishes that non-state-owned enterprises, non-priority industries, regions with high fiscal decentralization, low-profitability firms, and small enterprises are more constrained by financial decentralization. This indicates that financial decentralization reduces the efficiency of capital allocation, leading to more capital flowing toward government-preferred sectors, resulting in financial inequality and heightened financial vulnerability.

Conclusion

Findings and discussions

This study employs panel data from 273 prefecture-level cities in China over the period 2007–2018 and leverages RCC restructuring as a quasi-natural experiment to investigate the impact of financial decentralization on corporate risk. The findings demonstrate that financial decentralization increases the risk borne by local enterprises. Mechanism analysis reveals that this effect operates primarily through the crowding out of local state-owned bank branches and the restriction of corporate lending, except in cases where loans are secured by collateral. Heterogeneity analysis further shows that the effect is particularly pronounced in regions with strong fiscal decentralization, in non-key industries, and among non-state-owned enterprises. Smaller and less profitable firms are found to be especially vulnerable.

Relative to the indices of financial decentralization commonly employed in the literature (Xu, 2024; Zhao et al., 2024; Wang, 2025; Zhu et al., 2025), this study introduces a novel method of measurement and provides new evidence on how financial decentralization affects micro-enterprises. Several important insights emerge from the results. First, the crowding out of state-owned banks under financial decentralization highlights its role in reshaping the distribution of financial resources within a region. Second, the heterogeneity effects associated with fiscal decentralization, industry classification, and ownership structure suggest that financial decentralization channels resources disproportionately toward government-preferred sectors, underscoring the influence of local governments in directing financial flows. Collectively, these findings contribute to a deeper understanding of local government interventions in financial markets.

Nevertheless, this study is not without limitations. The most salient concern lies in the accuracy of causal identification. Although we incorporate control variables and multiple fixed effects, the empirical model may still suffer from omitted variable bias, raising potential endogeneity issues. Furthermore, it cannot be entirely ruled out that some local governments continue to exert considerable influence over state-owned and joint-stock banks. Variations in the number of state-owned banks may in turn affect the degree of local financial decentralization, potentially weakening the explanatory power of RCC reforms as an exogenous shock.

Building on the study’s limitations, future research could expand in methodology, data, and theoretical scope to strengthen causal identification. Methodologically, beyond employing advanced ensemble learning techniques such as Random Forests, Gradient Boosting Decision Trees, and the XGBoost model, future studies could incorporate causal inference approaches, including synthetic control methods, instrumental variable strategies integrated with machine learning, and Bayesian hierarchical models, to provide alternative perspectives and enhance robustness. From a data perspective, in addition to collaborating with local governments to obtain debt data for more precise measurement of financial decentralization, researchers could integrate multi-country firm-level datasets, financial transaction records, and geospatial economic indicators. Such integration would enable comparative studies across different institutional contexts and improve the external validity of findings. Theoretically, future work could refine the current corporate risk decision model by incorporating non-bank financial institutions and modeling the equilibrium between banks and non-bank financial players under financial decentralization. Moreover, applying perspectives from political economy, institutional theory, and network analysis could illuminate how local political incentives, regulatory structures, and interbank linkages shape the observed effects. This diversification in theory and methodology would deepen understanding of the mechanisms and boundary conditions underlying the relationship between financial decentralization and corporate risk.

Implications

In light of the research findings that local governments seize financial resources and worsen corporate risks, several concrete policy measures are proposed.

Based on the discovery that financial decentralization negatively impacts corporate risk by influencing local bank branches and loan sizes, the central government should take the lead in drafting a “Financial Decentralization Management Regulation” by the end of 2025, coordinated by the National Financial Supervision and Administration. This regulation will clearly limit local governments’ authority in banking operations. Local governments should be explicitly prohibited from interfering in core business decisions, such as credit approval and senior executive appointments in RCBs, with equity holdings capped at 30% and no disguised influence via equity pledges or subsidies allowed. All localities must submit quarterly financial resource allocation plans through a centralized regulatory platform, which require review and approval by the State Council’s Financial Stability Development Committee. Violations should result in accountability, including reduced fiscal transfers in serious cases.

To address the finding that local government intervention leads to irrational biases in credit resources, particularly in RCBs, the People’s Bank of China, together with the National Financial Supervision and Administration, should implement a comprehensive “Regulatory Big Data Platform” covering all RCBs by 2025. This platform will enable real-time monitoring of credit concentration and sector risk. If any single customer’s loan exceeds 10% of a bank’s total loans or if loans to a particular industry surpass 30%, an automatic alert should trigger immediate regulatory inspections. Annual audits, carried out by independent accounting firms, should focus on abnormal loans tied to local government projects or related parties, with results released publicly to ensure transparency.

Reflecting the finding that fiscal decentralization plays a key role in increasing risk effects of financial decentralization, the efficiency of financial resource allocation should become a significant metric in evaluating local government performance, with a minimum 15% weighting. Assessment criteria could include growth rates in inclusive finance lending and reductions in enterprise financing costs. Regions demonstrating strong performance should receive priority in refinancing quotas and special bond issuances, while lagging ones should face corrective action and supervision.

To address the impact of financial decentralization on corporate risk through loan access and allocation, the Ministry of Finance should establish a RMB 100 billion fintech development fund to subsidize up to 20% of qualified digital transformation projects in large state-owned banks. By 2026, these banks should complete a unified national “Smart Credit Platform” that integrates data from over twenty government departments, cutting credit approval times to within three working days. Regional pilots of supply chain finance blockchain platforms should also be advanced to facilitate data sharing and expand access to accounts receivable financing for SMEs.

In response to the finding that financial decentralization disproportionately affects non-key industries, non-state-owned enterprises, and small businesses, it is vital for companies to diversify their funding sources beyond traditional bank loans by using equity financing, bond issuance, and supply chain finance. Companies should establish internal risk tracking systems, regularly monitor changes in financial policy, optimize cash flows during periods of credit tightening, and use digital tools for early warning and scenario planning. Proactive communication with local authorities is crucial—firms should participate in consultations on financial resource allocation, report financing needs, and seek inclusion in local support programs. Local governments should organize regular matchmaking meetings between banks and enterprises and set up expert task forces to provide tailored assistance for firms facing severe financing constraints.

Finally, these practical measures could serve as reference points for other emerging economies experiencing similar local government intervention in financial markets, helping strengthen supervision, prevent resource misallocation, and support balanced economic development.