1 Introduction

“Zombie enterprises” are companies that are unable to pay their debt servicing costs with their current profits over an extended period. Zombie enterprises have attracted increasing interest in academic and policy arenas (Banerjee & Hofmann, 2018)—in part because they slow down productivity and squeeze out entrepreneurship (McGowan et al., 2017a), thereby causing a persistent economic downturn (Caballero et al., 2008). The current massive COVID-19-related government programs in many countries might cause the “zombification” of post-COVID-19 economies (Hoshi et al., 2023), characterized by a decline in business dynamism and barriers to entry for small businesses that may be disruptive (Naudé, 2022). In this context, an urgent question for countries worldwide is what policies governments can adopt to avoid “ossified economies.”

China provides an ideal setting to study this issue. The administrative monopoly system has seriously distorted the allocation of entrepreneurship, resulting in many zombie enterprises and a worsening “ossified economy.” In 2004, China launched an administrative reform aimed at reducing the government’s power to allocate resources and breaking administrative monopolies to eliminate zombie enterprises’ unfair competitive advantages and expose them to market selection. Can this administrative reform effectively remedy the problem of zombie enterprises in China? How will China’s administrative reform reshape the entrepreneurial behavior and business model of zombie enterprises? By what channels will it impact zombie enterprises? How heterogeneous will the reform’s effects be in the face of the dominance of government-owned and private enterprises? These essential theoretical and practical questions have not yet been clarified. Furthermore, existing assessments of the policy’s effects remain largely theoretical; rigorous empirical evidence is needed.

We use data from the Annual Survey of Industrial Enterprises (ASIF) from 1998 to 2007 and adopt the difference-in-difference approach (DID) to examine these questions empirically. The results show that the administrative reform increases the propensity for entrepreneurship by breaking down industry barriers and institutional transaction costs. Moreover, the competition effect triggered by this policy stimulates enterprises to reallocate entrepreneurial efforts from rent-seeking through regime commitment to productive entrepreneurship. Increased investment autonomy breaks the barrier to firm growth that results from excessive government intervention, generates scale effects, and increases total factor productivity (TFP). Eventually, zombie enterprises can enhance their viability, and the proportion of zombie enterprises decreases significantly. However, the unique status of state-owned enterprises (SOEs) leads to ineffective treatment.

The remainder of this paper is as follows: Section 2 presents the study’s theoretical background. The institutional backdrop and causal identification approach of the reform are discussed in the third section. The basic empirical result is shown in the fourth part. The fifth section examines the mechanism and heterogeneity of the effect of administrative reform on zombie enterprises. Finally, we present the conclusions, discussion, and policy implications.

2 Theoretical background

Entrepreneurship is essential for the long-term sustained growth of a country’s economy (Acs et al., 2014; Baumol & Strom, 2007; Lafuente et al., 2020; Prieger et al., 2016; Schumpeter, 1934). The “entrepreneurial economy” is characterized by the development and increased innovation of new small firms, while old large firms decline and fail. However, global economies are experiencing a prolonged decline in entrepreneurship and innovation—the number of non-viable zombie enterprises that rely on loans and subsidies to avoid bankruptcy is growing, inhibiting dynamic capitalism and gradually ossifying the economy (Naudé, 2022).

Since zombie enterprises take up many social resources and generate inefficient supply, which has a significant crowding-out effect on healthy business investment (Caballero et al., 2008), it is essential to understand why the economy is continuing to stagnate and what can be done to revitalize the entrepreneurial economy. Existing studies attribute the persistence of zombie enterprises to banks’ efforts to cover up bad debt losses (Caballero et al., 2008) and government pressures, such as job preservation (Jaskowski, 2015), a weak rule of law (McGowan et al., 2017b), and inappropriate government bailouts of firms (Hoshi et al., 2023). Therefore, corporate restructuring (Carreira et al., 2022; Fukuda & Nakamura, 2011), particularly downsizing in zombie enterprises and corporate fixed asset sales, and government funding for production and innovation subsidies (Yang et al., 2021) are essential to eradicate zombie enterprises.

The institution determines whether entrepreneurship is unproductive, productive, or destructively creative (Baumol, 1990) and impacts national economic outcomes (Lafuente et al., 2020). The key to stimulating entrepreneurship and eliminating an “ossified economy” lies in removing the obstacles to the national entrepreneurial system and building an institutional environment conducive to the creation and growth of new businesses and the efficient allocation of resources (Acs et al., 2014; Baumol & Strom, 2007). Given the importance of institutions, some scholars have attempted to examine how bankruptcy system reform can reduce the proportion of zombie enterprises (McGowan & Andrews, 2018; Nieto-Carrillo et al., 2022). However, they have ignored the impact of fundamental political and institutional changes, such as administrative monopolies; this prevents us from analyzing the disposal of zombie enterprises and effectively stimulating entrepreneurship from an institutional root cause perspective. At the same time, existing literature does not introduce entrepreneurship and its configuration into the analytical framework of institutions and zombie enterprise disposal; accordingly, no adequate theoretical foundation yet exists.

In China, high entry-exit barriers resulting from administrative monopolies under the government’s investment approval system destroy the market mechanism, breed unproductive entrepreneurship, and distort the allocation of economic resources. Although China’s market-oriented reforms continue to advance, state-led resource allocation remains essential to China’s economic development. Private enterprises have always faced institutional barriers, such as market access and approval permits, and have been unable to operate freely and compete fairly. Thus, fundamentally, market entry barriers under the administrative approval system have raised transaction costs and institutional uncertainty for firms; this has led to a lack of attention to innovation and the identification of market opportunities, severely inhibiting entrepreneurship (Bennett, 2021) and causing the formation of zombie enterprises. Based on the institutional roots of zombie enterprises, reducing government intervention and eliminating administrative monopolies through administrative system reform are fundamental ways to stimulate productive and destructive entrepreneurship. China’s investment approval system reform involved a dramatic simplification and unprecedented decentralization of the corporate investment approval process, reducing the government’s power to allocate operating resources and removing unfair business competition. Can the administrative system reform solve the zombie dilemma and stimulate an entrepreneurial spirit?

Two types of productive entrepreneurship facilitate the creation of profits and revive zombie enterprises, namely, Schumpeterian “creative destruction” (Schumpeter, 1934) and Kirznerian “arbitrage” opportunities (Kirzner, 1973). Schumpeterian entrepreneurship implies that entrepreneurship is an innovative process in which entrepreneurs seek profits by creating new products with better value to replace old ones (Lafuente et al., 2020). Administrative system reforms cut business registration procedures, time, and costs, resulting in lower entry barriers and institutional transaction costs that stimulate the creation of new businesses (Amici et al., 2016). As entrepreneurial behavior increases, market competition and incumbent firms’ exit risks increase. This encourages firms to allocate cost and resource savings to product development and innovation capital investment, thereby facilitating incentives for Schumpeterian “creative destruction” (Arrow, 1962). Given that political affiliation and innovation are alternate methods of firm development, rent-seeking zombie enterprises may shift their resources to innovation due to the smaller rent-seeking space created by the improved business environment, thereby easing their rigidity and gradually reviving them into healthy firms.

Beyond Schumpeter’s “creative destruction,” Kirznerian entrepreneurship also creates value through market discovery. Entrepreneurs profit by discovering new untapped markets for existing products or identifying arbitrage opportunities that have not yet been exploited (Kirzner, 1973). The administrative approval system allows the government to decide whether enterprises can carry out investment activities in relevant areas, hindering Kirznerian entrepreneurship. The reform of the administrative system has eliminated or decentralized enterprise investment approval projects and abolished administrative licenses in some industries. Notably, it allows enterprises to become more independent in production and management, promotes the efficient mobilization and acquisition of resources given market opportunities, and stimulates productive Kirznerian entrepreneurship. Firms can combine their strengths and market needs to make rapid strategic adjustments, leading to scale expansion and efficiency gains. This process will eventually manifest in increased corporate efficiency and profitability, transforming zombie enterprises into healthy and sustainable operations.

3 Investment approval system reform and identification methods

3.1 Reform of theNational Development and Reform Commission (NDRC) investment approval system reform

Following the reforms and opening-up, China’s market-oriented policies continued to advance. However, as of 2004, the issue of the optimal allocation of investment resources had not yet been resolved, and planning-era practices, such as the strict examination and approval of investment activities, became an essential means of macroeconomic regulation and control. Investment approval system reform also followed: projects—regardless of their investment subject, source of capital, or nature—had to accord with the government at various levels and the relevant departments of hierarchical examination and approval according to the investment size. Project proposals, feasibility studies, and construction reports had to be examined and approved by government departments before implementation.

Although the pattern of the state as a single investment entity has gradually weakened, this examination and approval system has not substantially changed. Along with establishing the socialist market economy system, it was imperative to reform the anachronistic investment system. The state’s relevant departments began to study proposals for widening investment system reform in the early 1990s.

After more than a decade of proof and dozens of revisions, on July 16, 2004, “the State Council on the investment system reform decision” was promulgated. The decision mandated that “enterprises that do not use government funds in the construction of their projects, could no longer implement the project ‘examination and approval’ (审批 shenpi) system and distinguish between different situations to implement the project ‘registry’(核准 hezhun) or project ‘record’ (备案 beian) system.” The State Council also formulated and promulgated the “Catalog of Government Approved Investment Projects (2004)” (the Catalog), which restricts the approval authority of different levels of government in various industries.Footnote 1 The Catalog specifies the investment program, whether “registry” or “record” system, in the areas of agriculture, forestry and water conservancy, energy, transportation, information industry, raw materials, machinery manufacturing, light industry, tobacco, high-tech, urban construction, social services, and finance.

There are significant differences between the “registry” system and the “record” system. The government’s investment decision-making power for the “record” system projects is more decentralized than the “examination and approval” system. The government no longer interferes with the project’s investment decisions. The process of enterprise investment is straightforward. After completing legal procedures related to environmental protection, land use, resource utilization, production safety, urban planning, and submitting basic business information, the company can carry out investments related to business operations.

The project “registry” system is highly similar to the past “examination and approval” system. The materials offered are about meeting the market access conditions, and government departments will rigorously review the access conditions of the project. It is necessary to go through a project proposal, feasibility study report, and construction report approval before implementation for any investment projects. The approval period is generally between 6 months and 2 years, and a few project lengths even reach 10 years. The government departments will strictly approve the access conditions of the project, and the enterprise will incur public law penalties and even criminal liability if it engages in production without passing the administrative approval liability.

The approval behavior of the NDRC is a direct intervention in micro-subjects. Taking the feasibility of investment projects as an example, the government must review the content involving consumer preferences, producer costs, potential producers’ willingness to enter, and other business operation information such as market, resources, industrial technology, and investment efficiency. Hence, the NDRC still has a great deal of discretionary power in approving investment projects. The project “registry” system essentially becomes an “examination and approval” system in disguise.

3.2 Identifying zombie enterprises

Caballero et al. (2008) first proposed that firms in Japan that pay a lower interest rate than the risk-free rate are likely to be zombie enterprises. However, we may also label healthy enterprises with risk-free interest rates on loans as zombie enterprises. Additionally, zombie enterprises that borrow new debt to pay off old debt can be incorrectly classified as healthy companies. Fukuda and Nakamura (2011) added two new criteria to address these issues. The “profitability criterion” identifies firms with an annual earnings before interest and taxes (EBIT) higher than the minimum interest payable in Caballero et al. (2008) as healthy firms, aiming to reduce the possibility of misclassifying non-zombie enterprises as zombie enterprises. The “evergreen borrowing criterion” identifies firms with debt-to-asset ratios of more than 50% in year t − 1 and whose borrowing still increases in year t as zombie enterprises, thus reducing the possibility of misclassifying zombie healthy firms.

Short-term business cycle shocks may resemble a firm’s poor operating conditions in a given year. Coupled with the fact that some young firms still have unstable revenues and are prone to operating fluctuations, these firms are highly likely to be misclassified as zombie enterprises. McGowan et al. (2017a) argue that a firm in an OECD country can be defined as a zombie enterprise when it is older than 10 years old, and its operating income is lower than its interest expense for 3 consecutive years. Notably, Carreira et al. (2022) suggest that it is essential to focus on poor profitability and high debt risk rather than interest coverage ratios for Portuguese firms. A company is classified as a zombie enterprise if its return on assets has been below the low-risk rate for at least 3 consecutive years and its leverage ratio has been above the industry median for the exit group with a low return on assets for more than 5 years. They exclude “one-shot” zombie enterprises (Nakamura, 2017), which are enterprises that are only identified as zombie enterprises once during the research sample period.

In Japan and Europe, where the market dominates the allocation of resources, zombie enterprises exist due to the help of banks. However, most of the zombie enterprises in China are attributed to excessive government intervention under China’s unique economic and political system. Driven by the goals of developing the economy, maintaining social stability, and political promotion, local governments usually support local enterprises and avoid bankruptcy through bank credit and direct government subsidies. Improper government intervention has led to high barriers to exit for zombie enterprises. Therefore, it is necessary to consider government subsidies when identifying zombie enterprises in China. Shen and Chen (2017) argue that only firms with leverage ratios above 50%, annual liabilities exceeding those reported for the previous year, and actual negative profits (less fiscal subsidies) for 3 consecutive years are considered zombie enterprises; these criteria are more in line with the Chinese situation.

However, temporary fluctuations in business operations due to economic cycles and the instability of young firms’ profit models can cause the above approach to overestimate the number of zombie enterprises in the economy. Two or 3-year zombie status is vital to ensure that a zombie enterprise is not just a profitable firm with transitory financial problems. It is also essential to distinguish between actual zombie enterprises and young innovative start-ups (Carreira et al., 2022).

In this paper, after considering the Chinese context and the existing literature, we propose identification criteria that are more suitable for identifying Chinese zombie enterprises. Specifically, there are four points: (i) net profits (EBIT deducting government subsidy) are below the minimum level of interest payments for at least 2 consecutive yearsFootnote 2; (ii) liability in period t − 1 is higher than 50%, and period t liabilities are greater than period t-1 liabilities; (iii) age is more than 5 years old; and (iv) not a “one-shot” zombie enterprise.

In addition, the official concept of zombie enterprises in China first emerged from an executive meeting of the State Council in 2015. Premier Li Keqiang suggested that zombie enterprises refer to enterprises with excess capacity, have sustained losses for more than 3 years, and do not meet structural adjustment requirements. Therefore, Chang et al. (2021) consider companies with negative net profit for 3 consecutive years as zombie enterprises. As a robustness check, we re-identified zombie enterprises using the methods in Carreira et al. (2022), McGowan et al. (2017a), Shen and Chen (2017), and Chang et al. (2021). Tables 11, 12, and 13 in the Appendix show that the correlation coefficients of zombie enterprises’ proportions with different identified methods almost exceed 0.5, with a large proportion exceeding 0.7. It indicates that this paper’s zombie enterprise identification method is reasonable.

3.3 Methodology

Based on the “quasi-natural experiment” of the 2004 NDRC investment approval system reform, we assess the influence of the administrative reform on eradicating zombie enterprises using a DID approach. The model is:

$${\mathrm{Zombie}}_{\mathrm{ipt}}=\alpha+\beta{\mathrm{Treat}}_i{\;\times\;\mathrm{Post}}_t+\gamma X_{\mathrm{ipt}-1}+\delta_i+\delta_p+\delta_t+\varepsilon_{\mathrm{ipt}}$$
(1)

In model (1), \({\mathrm{Zombie}}_{\mathrm{ipt}}\) denotes the proportion of zombie enterprises in year t in industry i of province p.

The DID method requires a treatment and a control group. Below, we further explain how the treatment and control groups were computed based on the reform. The approval system reform in 2004 significantly reduced the declaration procedures for firms with the “record” system. Firms in these industries can be the treatment group in the DID method. The approval power of governments on the project “registry” system has not changed significantly. Firms in these industries can be the control group. We take the provisions of the light industry tobacco industry as an example. The Catalog mentioned that “pulp, denatured fuel ethanol, polyester, salt production, sugar processing and cigarette, tobacco cellulose diacetate projects should be strict ‘registry’ system, other light tobacco industry projects suit for the ‘record’ system.” So we consider enterprises in the other light industry tobacco industry as the treatment group and enterprises in pulp, denatured fuel ethanol, polyester, salt production, sugar processing, cigarette, and tobacco cellulose diacetate industry as the control group.

\({\mathrm{Treat}}_{i}\) is used to identify the industry in which the treatment group of companies is located. The industry corresponding to the Catalog used in the project “registry” system is assigned a value of 0, and the others are assigned a value of 1.Footnote 3\({\mathrm{Post}}_{t}\) is used to identify the timing of the policy shock, with a value of 1 for years after 2004 and 0 for others. Treat_Post is the core explanatory variable of the model, and its coefficient measures the change in the proportion of zombie enterprises with a “registry” system and a project “record” system after the reform by the NDRC.\({X}_{\mathrm{ipt}-1}\) is a one-period-lagged control variable. \({\delta }_{i}\), \({\delta }_{p}\), and \({\delta }_{t}\) denote fixed effects in the four-digit industry, province, and year, respectively. \({\varepsilon }_{\mathrm{ipt}}\) is a random disturbance term. The standard errors were clustered by the four-digit industry codes for possible heteroskedasticity and autocorrelation problems. Tables 7 and 8 in the Appendix provide the measurements of core variables and descriptive statistics.

China’s economic system reform is a typical incremental reform. When the reform of the investment approval system is enacted, several reforms are usually carried out in parallel or by way of crossover. To control for the impact of China’s World Trade Organization (WTO) accession, SOE reform, and FDI liberalization on zombie enterprises, we further incorporate the share of SOEs in each industry, the import tariff rate, and the share of wholly foreign-owned enterprises among foreign enterprises in 2000 with the year dummy interaction term to control for potential policy disturbances. The main reason is that the characteristic variables for each year may result from market development and related policy reforms. Using the proportion of a pre-policy year multiplied by the year dummy ensures the exogeneity of the policy control variables and better controls for the effect of relevant measures on the share of zombie enterprises over time (Lu et al., 2017).

3.4 Data

Data was collected from the ASIF, spanning 1998 through 2007.Footnote 4 The model setting of this study also involves tariff data and provincial economic and administrative statistics, which were taken from the World Bank and China Statistical Yearbook, respectively. Before conducting the econometric analysis, we processed the data as follows:

First, the same enterprise is first identified based on the same enterprise code, then based on the same enterprise name, and finally further matched by basic information such as enterprise address, legal representative’s name, telephone number, and year of opening. Second, we delete the observations from the original data set if the values of key variables are either larger than the 99 percentile or smaller than the first percentile. Again, we unify the four-digit national economic industry classification before and after 2002. Finally, we match the industry codes with GB/T4754-2002 according to the HS1996 and HS2002 conversion tables provided by the United Nations Statistics Division and the HS2002 to ISIC (Rev3) conversion table provided by the Purdue University to obtain the industry-level tariff data. In addition, we removed Tibet Autonomous Region information due to a severe lack of data.

4 Empirical results

4.1 The impact of NDRC approval system reform on the proportion of zombie enterprises

Column (1) of Table 1 shows the regression results of model (1) estimated with the fraction of zombie enterprises based on asset weights as the explanatory variable. Treat_Post is significantly negative at the 5% level with a coefficient size of − 0.030. This indicates that after the investment approval system reform, the fraction of zombie enterprises in registry-based industries measured by asset weight declined. The economic implications can be derived: in comparison to the average percentage of zombie enterprises (22.5%), the reform resulted in a 13.3% (0.03/0.225) decrease in the share of zombie enterprises. The proportion of zombie enterprises measured by debt weights and quantity weights, respectively, is employed as explanatory variables in model (1). According to columns (2) and (3), the reform also reduces the proportion of zombie enterprises by 2.8% and 3.4% in absolute terms and 10.6% (0.028/0.265) and 17.3% (0.034/ 0.196) in relative terms. The above results imply that reforming the approval system to reduce the government’s control over resources has effectively crowded out the “ossified economy.” The market mechanisms are essential for the disposal of zombie enterprises.

Table 1 Effects of approval system reform on the percentage of zombie enterprises
Full size table

4.2 Event study

A parallel trend test must be passed when using the DID approach. The policy effect estimated may be caused by the differences between the two samples rather than the actual policy effect if it does not satisfy the parallel trend. The model for the parallel trend test is as follows:

$${\mathrm{Zombie}}_{\mathrm{ipt}}=\alpha+\beta_t{\mathrm{Treat}}_i{\;\times\;\mathrm{Year}}_t+\gamma X_{\mathrm{ipt}-1}+\delta_i+\delta_p+\delta_t+\epsilon_{\mathrm{ipt}}$$
(2)

\({\mathrm{Year}}_{t}\) is the year dummy variable, and the other variables have the same meaning as in Eq. (1). \({\beta }_{t}\), the coefficient of \({\mathrm{Treat}}_i{\;\times\;\mathrm{Year}}_t\), indicates whether there is a substantial difference between the treatment and control groups in the percentage of zombie enterprises in year t. The dependent variables are the proportion of zombie enterprises with asset weight. In order to present the findings of the parallel trend test more visually, Fig. 1 shows the estimated coefficient of \({\mathrm{Treat}}_i{\;\times\;\mathrm{Year}}_t\) in model (2). Before 2004, there was no notable change between the treatment and control groups, indicating that our sample satisfied the parallel trend test. However, the estimated coefficients have changed from positive to negative since 2004. The economic implications and statistical significance grow progressively larger after policy implementation, indicating that the approval system reform significantly reduces the proportion of zombie enterprises. Due to their similarity and space limits, the graphs illustrating the proportion of zombie enterprises based on number weights or debt weights as explanatory variables have been omitted.

Fig. 1
figure 1

Effect of approval system reform on the share of zombie enterprises. Note: the dashed line indicates the 95% confidence interval of the estimated effect. The dots marking the estimated beta in model 2 represent the difference between the control and treatment groups for each year. The solid line reflects the dynamic process resulting from the reforms by demonstrating the difference, over time, between the proportion of zombie enterprises in the control groups and those in the treatment groups

Full size image

4.3 Excluding other relevant policies

First, since 2000, administrative system reform has been seen as an essential step to stimulate economic vitality and boost resource allocation efficiency. The central government has strongly advocated for the establishment of administrative approval centers at all levels of government to bring together various independent and scattered approval departments in the same office hall. The fundamental purpose is to establish one-stop public services by integrating different administrative departments, reduce the time required for enterprises to travel between multiple functional departments, and encourage firms to build plants quickly in streamlined ways. To investigate whether this affects causal identification, we control the proportion of administrative approval centers established in prefecture-level municipalities in each province to re-estimate model (1). Table 2 shows that the coefficient of approval center is notably negative, implying that establishing administrative approval centers reduces the proportion of zombie enterprises. However, after controlling for the above effects, the coefficient of interest is still considerably negative at the 5% level with a coefficient of − 0.030, which means that the policy effect of the investment approval system reform is not affected by the establishment of administrative approval centers.

Table 2 Excluding other contemporaneous policies from interfering with causal identification
Full size table

Second, industrial policy is also an institutional cause of zombie enterprises. China’s policymakers have been concerned about blind investment and overcapacity in steel, coal, power, and other industries since 2002 and have implemented policies accordingly. We analyze the effect of industrial policy by combining the cross-product term of the industrial dummy variable and the year dummy variable into model (1) to analyze whether it would mix the policy effects. The coefficient of industrial policies in Table 2 is not significant, but our interest variable remains robust.

Subsequently, this paper excludes other contemporaneous policies.Footnote 5 In 2003, the State Council issued “Several Opinions on the Implementation of the Northeast Region’s Old Industrial Bases Revitalization Strategy,” which vigorously promoted the reform of the corporate tax system in the northeast region and may have greatly reduced the proportion of zombie enterprises. For this reason, this study introduces a northeast region dummy variable with the policy year dummy variable cross multiplication term to control for possible bias brought by the revitalization strategy of the northeast region. Column (3) of Table 2 shows that the coefficient of northeast revitalization is negative but not significant. The core explanatory variables in this study are still significant at the 5% level after controlling for the policy, and the policy effect of the approval system reform is not reduced. Therefore, the causal relationship in this study was not driven by the northeast revitalization strategy.

China’s economic development, corporate investment, and financing are cyclical. Under the effect of the policy uncertainty and the incentive for official promotion, businesses are conservative in the year before the Party Congress. In contrast, since the Party Congress, market agents have more optimistic expectations, affecting the proportion of zombie enterprises. For this reason, this paper introduces the year before the Party Congress (PPCt−1), the year of the Party Congress (PPCt), and the year after the Party Congress (PPCt+1) in model (1) to investigate whether the Party Congress cycle affects the causal identification in this study.Footnote 6 Table 2 shows that the estimated coefficients of the PPCt−1, PPCt, and PPCt+1 are not significant. The coefficients and significance of the core explanatory variables remain unchanged. This indicates that the Party Congress cycle may affect the zombie enterprises but has no bearing on the conclusions of this paper.

Finally, there was a new round of market-oriented reforms in China’s banking system in 2003, resulting in a major increase in domestic banking competition. The optimization of the financing environment may be more conducive to the survival of zombie enterprises, thus reducing the proportion of domestic zombie enterprises. This study takes one minus the market share of the top five state-owned commercial banks as the bank’s degree of competition and re-estimates it by adding it to model (1) to exclude policy interference.Footnote 7 Column (5) of Table 2 shows that the percentage of zombie enterprises is unaffected by banking sector rivalry, while the policy effect remains significant. In conclusion, reforming the investment approval system aimed at transforming government functions, reducing the government’s power to allocate resources, and breaking administrative monopolies have destroyed the survival space of existing zombie enterprises, squeezed out unproductive entrepreneurship, and significantly alleviated China’s “ossified economy.”

5 Further analysis and mechanism test

5.1 Competitive effects

The typical feature of the NDRC investment approval system is that the government has set up an invisible threshold for market entry. The reform’s greatest direct effect is a decrease in governmental transaction costs, which leads to the ongoing opening of new market actors into the original industry. The market competition effect brought by the reform makes firms change their previous rent-seeking business model, and innovative development strategies become an inevitable choice. This study retains only a sample of zombie enterprises; we exclude any consistently identified as non-zombie enterprises during the sample period to analyze whether the policy reduces their rent-seeking behavior. The model of the relevant mechanism test was as follows:

$${\mathrm{Zfirm}}_{\mathrm{kipt}}=\alpha +\beta {\mathrm{Treat}}_{i}{\times \mathrm{Post}}_{t}+\gamma {\mathrm{X}}_{\mathrm{ipt}-1}+\tau {F}_{\mathrm{kipt}-1}+{\delta }_{i}+{\delta }_{p}+{\delta }_{t}+{\vartheta }_{\mathrm{kipt}}$$
(3)

\({\mathrm{Zfirm}}_{\mathrm{kipt}}\) is firm transaction cost, innovation capacity, size, and TFP. \({F}_{\mathrm{kipt}-1}\) is a control variable at the firm level,Footnote 8\({\vartheta }_{\mathrm{kipt}}\) is the random disturbance term, and the other variables have the same meanings as in model (1).

The management fee reflects the institutional transaction costs imposed on firms by excessive government intervention. It can be considered a proxy variable for unproductive firm expenditures, such as rent-seeking. We define institutional transaction cost as the share of the management fee in product sales revenue. Column (1) of Table 3 shows a 2% decline in zombie enterprise rent-seeking expenditures in the record-based system industry compared to zombie enterprises in the registry-based industry, which corresponds to 12.8% (2/15.6) of the average. Subsequently, we use the new firm entry rate at the province-industry level as the explanatory variable to test whether the reform of the approval system significantly affects the entry of new firms. A firm is considered to enter the market in period t if it does not exist in period t − 1 but appears in period t. The new firm entry rate in period t is defined as the number of firms entering the market in period t divided by the total number of firms in the market in period t − 1. Column (2) of Table 3 shows that the reform increases the new firm entry rate by 6.1% (1.1/18) relative to the average new firm entry rate of 18%. Next, we define innovation as the share of new product output in total output. Column (3) of Table 3 shows that firm innovation capacity increases 16.6% (0.53/3.19) relative to the average innovation capacity of 3.19% after the reform. These findings validate the mechanism of the market competition effect. This implies that breaking administrative monopolies reduces enterprises’ rent-seeking scope and allows enterprises with high entry barriers to compete in the market. This facilitates the natural selection of market mechanisms and stimulates productive and destructive entrepreneurship.

Table 3 The mechanism of approval system reform affecting zombie enterprises
Full size table

5.2 Scale effects

The investment approval system reform established enterprise status as the main investment entity, breaking the barrier that restricts enterprise growth. At the same time, market mechanisms regulating the scale and capital flow in the industry will increase the efficiency of resource allocation in the industry and expand market size. Zombie enterprises may use this opportunity to improve their total factor productivity (TFP) and financial position, thus enhancing their viability. Based on the above analysis, we hypothesize that the reform may increase the TFP of zombie enterprises through the scale expansion effect, which eventually results in a decrease in the proportion of zombie enterprises.

To test whether the administrative reform can affect the expansion of zombie enterprises, we define the number of employees as firm size to re-estimate model (1). Similar to the above economic implications of the empirical results, column (4) of Table 3 indicates that the size of zombie enterprises significantly increases by 7.1% (0.28/3.94), and that the employment opportunities furnished by zombie enterprises significantly increase. Taking the share of net fixed assets at the industry level in the total fixed assets of all industrial industries as a proxy variable for industry market size, column (5) indicates that the policy significantly increases industry market size by 13.2% (0.08/0.61).

TFP is total factor productivity at the firm level. We obtain unbiased TFP based on the estimation method proposed by Olley and Pakes (1996), which is used as the explanatory variable for the analysis. Column (6) shows that zombie enterprises can gradually achieve economies of scale in the process of firm and market size expansion, and the reform can significantly improve TFP by 3.4% (3.1/90.1). This implies that reducing the government’s control over resources enables firms to face the changing market demand head-on, develop differentiated competitive advantages, and promote their operational efficiency and productivity.

5.3 Viability of zombie enterprises

As the previous analysis suggests that market competition and scale expansion effects increase the viability of zombie enterprises, we examine firms’ short-term business risk and profitability, zombie enterprises’ resurrection and exit rates, and the efficiency of industry resource allocation. Firm debt is defined as the ratio of firm debt to total assets. Firm profit is the logarithm of the ratio of firm profit to main business revenue. First, using firm debt and firm profit as explanatory variables, Table 4, columns (1) and (2), shows that the reform decreases the liability-to-asset ratios of zombie enterprises (suggesting they may also revamp their business models), which thereby reduces their business risks. Their average profitability changes from − 1.7% in the sample period to 1% under the policy, indicating significant recovery.

Table 4 The approval system reform’s impact on the viability of zombie enterprises
Full size table

Second, to explore the mechanism underlying the reduction of the share of zombie enterprises by the administrative reform, this section further analyzes only the sample of zombie enterprises; that is, it excludes those that have been non-zombie enterprises since 1998 to 2007. We examine the changes in zombie enterprises’ recovery and exit rates. Following Nakamura (2017), if a zombie enterprise in period t − 1 becomes a non-zombie enterprise in period t, which means that the firm did not satisfy the zombie criterion, the zombie enterprise will be considered the “revived” status in period t. If the zombie enterprise does not become healthy until it exits the market in period t, which means that the firm is no longer in the database, then it is considered to be in the “exit” status in period t. The zombie enterprise recovery rate in period t is expressed as the proportion of the number of resurrected zombie enterprises in period t to the total number of zombie enterprises in period t − 1. The zombie enterprise exit rate in period t is expressed as the proportion of exited zombie enterprises in period t to the total number of zombie enterprises in period t − 1. Columns (3) and (4) suggest that the investment approval system reform significantly increases the recovery rate and decreases the exit rate of zombie enterprises. Breaking administrative monopolies has facilitated the transformation of entrepreneurship from unproductive to productive. The market mechanism has optimized the business strategies of zombie enterprises, which have significantly increased their viability.

Finally, Caballero et al. (2008) have shown that zombie enterprises lead to a misallocation of resources in their industries. If the reform improves the recovery rate of zombie enterprises, it will translate into higher resource allocation efficiency. In this paper, we define resource allocation efficiency as the standard deviation of TFP estimated, based on Olley and Pakes’ (1996) approach within the same four-digit industry, where a larger standard deviation of productivity implies less efficient resource allocation. Column (5) of Table 4 shows that the productivity dispersion of firms significantly decreases after the reform, and the efficiency of industry resource allocation significantly improves. This suggests that lowering barriers to entry and granting economic freedom to firms are imperative for optimizing the allocation of economic resources and promoting dynamic creative destruction.

The findings show that the administrative reform improved the business environment, created several new market players, and substantially stimulated market dynamics and entrepreneurship through the competition mechanism. Furthermore, the scale effects (enterprise expansion, significantly enhanced market capacity) had significantly positive impacts on zombie enterprises’ TFP, which, in turn, reduced their short-term business risks and revived their profitability. Hence, market equilibrium improved, but not through bankruptcy, liquidation, and other market elimination mechanisms. Instead, the reform encouraged competitive mechanisms and scale effects that alleviated China’s “ossified economy.”

5.4 Heterogeneity mechanisms based on firm ownership and industry competitiveness

5.4.1 Firm ownership

Chinese SOEs’ nature and close ties to the government may make them behave differently from private firms, especially under the pre-2004 administrative approval system, under which SOEs had more advantages than private firms. Does this also suggest that investment approval system reform has a heterogeneous impact? We re-estimated model (1) for a sub-sample of SOEs and private firms.Footnote 9 Columns (1) and (2) in Table 5 show that the reform significantly reduces the percentage of private zombie enterprises, while the policy does not affect state-owned zombie enterprises. We then further examine the sources of the differences in institutional transaction costs, innovation capacity, firm size, and TFP. Table 5 shows that the reform resulted in a 0.5 percentage point reduction in institutional transaction costs, a 0.9 percentage point rise in technical innovation, a 0.404 increase in size, and a 0.019 improvement in TFP for private businesses. Additionally, it significantly reduced institutional transaction costs for SOEs. However, the effects on firm innovation capacity and TFP are insignificant.

Table 5 Mechanism analysis: a perspective based on firm ownership
Full size table

For the above conclusion, we speculate that SOEs are often an important source of tax for local governments; even after the reform, SOEs have continued to enjoy more invisible policy tilts—the unique status makes it difficult for external competitive effects to drive its entrepreneurship. The existence of equity structures and corporate governance problems also makes it more challenging to improve TFP. Before the reform of the investment approval system, private enterprises were excluded from the market, investment opportunities were limited, and it was complicated to grow bigger and stronger. The reform brought rare development opportunities for small- and medium-sized private zombie enterprises to revive their entrepreneurship under the role of market mechanisms.

5.4.2 Industrycompetitiveness

We then further examine the heterogeneity of the policy between non-competitive and competitive industries from the perspective of industry competitiveness, utilizing the industry’s median number of enterprises as a dividing factor. The results in columns (1)–(10) of Table 6 show that the reform significantly reduces zombie enterprises in competitive industries but has little effect in non-competitive industries. The policy still leads to a significant reduction in institutional transaction costs for firms in competitive industries. However, the innovation capacity of competitive firms is not significantly affected, possibly because the competition effect already exists in competitive industries, and policy strengthening does not significantly incentivize firms to innovate. In contrast, non-competitive industries may not have a competition effect. At the same time, the scale effect also exists only in competitive industries, and the reform significantly increases the size of firms in competitive industries and increases TFP. Finally, there are notable policy effect differences in industries with different levels of competition.

Table 6 Mechanism analysis: an industry-competitiveness-based perspective
Full size table

Therefore, the market competition mechanism plays a vital role in alleviating the “ossified economy.” When enterprises are in competitive industries, they face more pressure to survive, motivating productive and disruptive entrepreneurship. They will try to exploit their comparative advantages, thus enhancing the viability or speeding up the resurrection of zombie enterprises. In non-competitive industries, enterprises have insufficient incentives to innovate and improve efficiency. The administrative reform policies still fail to motivate zombie enterprises to enhance their viability through innovation and efficiency.

6 Concluding remarks, discussion, and policy implications

6.1 Concluding remarks

Global economies are shifting from “entrepreneurial economies” to “rigid economies.” In this context, inefficient zombie enterprises are not exiting, and creatively destructive SMEs are struggling to enter (Naudé, 2022). A critical question for countries worldwide is what policies governments can adopt to avoid “ossified economies?” Administrative monopoly systems create barriers to entry that lead to unproductive entrepreneurship. However, it is unclear whether investment approval reforms to reduce government control over resources can improve the allocation of entrepreneurship and transform the growing “ossified economy” into an “entrepreneurial economy.” This study is the first to analyze how the shift in the role of government affects the allocation of entrepreneurship and the disposal of zombie enterprises from the perspective of breaking administrative monopolies.

This study investigates the impact of administrative reform on the proportion of zombie enterprises in China and its mechanism based on investment approval system reform, using data from ASIF from 1998 to 2007. The study shows that the investment approval system reform significantly reduced the proportion of zombie enterprises and crowded out the “ossified economy.” According to the mechanism analysis, the reform resulted in the entry of new enterprises. Market competition promoted the configuration of entrepreneurship and pushed zombie enterprises to change their previous “rent-seeking” business practices, significantly improving their innovation capacity. Simultaneously, the improved business environment and increased investment autonomy broke up the barrier limiting enterprise growth, expanded market boundaries, and contributed to TFP growth. Eventually, zombie enterprises’ profitability and recovery and survival rates increased significantly, and they survived in a healthier, more dynamic and sustainable way. Heterogeneity mechanism analysis reveals that the administrative reform differs between state-owned and private enterprises and non-competitive and competitive industries. It can only reduce the proportion of private zombie enterprises and competitive industries’ zombie enterprises; the proportion of state-owned zombie enterprises and non-competitive industries’ zombie enterprises did not change significantly.

6.2 Discussion

This study contributes to current knowledge in three ways. First, this paper contributes to the emerging literature on zombie enterprise disposal. The fundamental assertion in zombie enterprise disposal literature is that firm restructuring (Carreira et al., 2022; Fukuda & Nakamura, 2011), government funding (Yang et al., 2021), and bankruptcy system reform (Nieto-Carrillo et al., 2022) are critical. However, existing studies ignore the root cause of zombie enterprises: government intervention, which has led to a big gap in research on how to efficiently dispose of zombie enterprises. This study is the first to reveal the impact of lowering entry barriers on the behavior of zombie enterprises based on the reform of China’s administrative system. Administrative approval reforms aimed at reducing the government’s power to allocate resources have not led to the exit of zombie enterprises but rather to the enhancement of their viability by reducing institutional uncertainty, which in turn has reduced the share of zombie enterprises in the economy. After the reform, zombie enterprises’ profitability and debt patterns became significantly more viable. These findings challenge the conventional wisdom in zombie enterprise disposal research and supplement existing literature.

Second, this paper provides fertilizer for the debate on what mechanisms can contribute to reducing the rate of zombie enterprises in the economy. McGowan and Andrews (2018) argue that efficient bankruptcy regimes can enhance market selection and promote zombie enterprises’ bankruptcy. Yang et al. (2021) find that government innovation funding revives zombie enterprises by improving innovation through sharing innovation costs and encouraging firm downsizing. This paper contends that administrative reforms stimulate a virtuous circle between competition, innovation, and productivity growth, which resurrects zombie enterprises. We agree that market mechanisms and firm innovation are essential for reviving zombie enterprises. However, we find that the market competition forces zombie enterprises to shift from rent-seeking to innovation. In addition, we reveal a new mechanism: administrative reform revives zombie enterprises by granting them the right to invest autonomously and scale their expansion. Lowering barriers to entry helps them regain their lost investment opportunities and earn profits based on market demand and opportunities, thus increasing their TFP. Therefore, to reduce the proportion of zombie enterprises through resurrection, it is necessary to enhance market functioning mechanisms, inhibit unproductive behaviors such as rent-seeking, incentivize creative destruction, and enhance the efficiency of productive operations.

Third, this paper enriches the literature on entrepreneurship. On the one hand, entrepreneurship has been central to the analysis of economic growth (Schumpeter, 1934), yet it is unclear what policies improve the allocation of entrepreneurship. Djankov et al. (2002) found that countries with strict entry regulations have higher levels of corruption and larger unofficial economies. Additionally, economic freedom reduces entrepreneurs’ exposure to uncertainty and has become an essential driver of entrepreneurship (Bennett, 2021). This study explored whether the barriers to entry created by the administrative monopoly system led to unproductive entrepreneurship. When the exogenous administrative approval system reform occurred, the market acted as an invisible hand to shift zombie enterprises away from rent-seeking and toward innovation-driven behavior. This study extends research on institutions and the configuration of entrepreneurship to the field of administrative institutions for the first time. On the other hand, this paper unifies Schumpeterian and Kirznerian entrepreneurship into Baumol’s (1990) framework for the first time. Notably, it provides micro-level evidence from a transition country for the theory of entrepreneurship allocation. Moreover, Kirzner’s theory of entrepreneurship suggests that information asymmetry can be exploited by entrepreneurs who understand the market better than others. By investigating this in the context of transition economies, this study found that it works with the institutional condition of well-functioning markets, which ensures that firms can arbitrage. This also extends the application of Kirznerian entrepreneurship to transition economies.

6.3 Policy implications

Given the contribution of this study, we can draw several policy implications. First, in general, this study found that administrative system reform can be a cure for zombie enterprises. Since scholars and governments have not acknowledged that the administrative system may have caused the economy to ossify, policies have mainly focused on restructuring (Fukuda & Nakamura, 2011) and bankruptcy system reform (Nieto-Carrillo et al., 2022). In practice, however, it is complicated and politically unpopular due to employment and the enforcement of bankruptcy. This study found that the administrative system reform revived zombie enterprises rather than prompting them to go bankrupt and exit the market. Therefore, the zombie enterprise puzzle can be solved by reforming the administrative system, lowering access barriers, improving market-based allocation, and minimizing government intervention. These solutions are relatively easy to implement and, more importantly, can help reduce bureaucracy and create a more efficient market environment. Accordingly, this study’s insights can provide critical guidance to governments worldwide on how to face an increasingly ossified economic environment.

Second, according to the mechanism analysis, entrepreneurship will be the key to the recovery of zombie enterprises. This study found that administrative reform can both promote firm innovation through competitive effects and inspire alert entrepreneurs to identify market opportunities through market demand–based scale expansion. Notably, it stimulates both Schumpeterian and Kirznerian entrepreneurship and inhibits unproductive entrepreneurship, improving zombie enterprises’ competitiveness and thereby reviving them. From a policy perspective, countries can take measures to optimize the allocation of entrepreneurship to prevent economic ossification in the post-pandemic era. For example, the government should shift its focus from administrative interventions to serving business development by reducing bureaucratic discretion. It is also important to note that neglecting the vital role of Kirznerian entrepreneurship can considerably damage national productivity. Countries in transition, which have much room for improvement in efficiency gains based on market discovery functions, should not focus solely on promoting creative destruction, even though it is essential for a country’s long-term growth (Baumol & Strom, 2007). Instead, governments should also strive to enhance the market discovery function by breaking down the institutional barriers preventing Kirznerian entrepreneurship. Specifically, governments should simplify and speed up the approval process and enhance the efficiency of administrative services to improve the autonomy, convenience, and flexibility of business operations.

Third, the priority for accelerating economic recovery is creating a business environment conducive to developing small- and medium-sized private enterprises, which are the main drivers of national growth. The heterogeneity mechanism analysis in this paper suggests that reducing entry regulations and allowing firms to compete can greatly encourage small- and medium-sized private enterprises to innovate, expand, and increase productivity; however, it does not work for state-owned monopolies. Small- and medium-sized private enterprises are inherently more dynamic and competitive in the market and more subject to the ravages of excessive government intervention. Successful productivity development should be built on creating a fair, transparent, and predictable business environment that removes the institutional barriers that inhibit the growth of small- and medium-sized private enterprises, such as anti-competitive, unfair entry controls on small firms, and excessive licensing requirements and restrictions. Policymakers would do well to consider limiting discriminatory and restrictive provisions against private capital and small- and medium-sized enterprises to reduce policy distortions, thereby stimulating innovation and job creation for small- and medium-sized private enterprises. In practice, the government could further liberalize market access, implement a negative list system for market access, cut down the number of items restricted or prohibited among SMEs, and break all unreasonable restrictions.