Bankruptcy courts and the marketization of bond issuance

Abstract

Purpose

This paper aims to study the effect of the establishment of bankruptcy courts on bond issuance market. This paper helps to predict that the introduction of bankruptcy courts in China can mitigate price distortions caused by the implicit government guarantees and promote the development of the high-risk bond market.

Design/methodology/approach

This paper exploits the staggered introduction of bankruptcy courts across cities to implement a differences-in-differences strategy on bond issuance data. Using bonds issued in China between 2018 and 2020, the impact of bankruptcy courts on the bond issuance market can be analyzed.

Findings

This paper reveals that bond issuance credit spreads increase and is more sensitive to firm size, profitability and downside risk of issuance entity after the introduction of bankruptcy courts. It also reveals a substantive increase in bond issuance quantity and a decrease in issuer credit ratings following the establishment of bankruptcy courts. In addition, the increase of credit spreads is more prominent for publicly traded bonds, those whose issuers located in provinces with lower judicial confidence, bonds issued by SOEs and bonds with stronger government guarantees. Finally, the role of bankruptcy courts is more pronounced in regions with higher marketization.

Originality/value

This paper relates to previous studies that investigate the impact of laws and institutions on external financing. It helps provide new evidence to this literature on how improvements of efficiency and quality in bankruptcy enforcements relate to the marketization of bond issuance. The results provide further evidence on legal institutions and bond financing.

Citation

Huang, X., Teng, F., Xin, Y. and Xu, L. (2022), "Bankruptcy courts and the marketization of bond issuance", China Accounting and Finance Review, Vol. 24 No. 3, pp. 359-390. https://doi.org/10.1108/CAFR-05-2022-0056

1. Introduction

The law and its enforcement are important determinants of financial development (La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 1997, 1998; Shleifer & Wolfenzon, 2002; Djankov, McLiesh, & Shleifer, 2007). Recent years have witnessed the rapid development of the stock market and the banking system in China, thanks in part to improvements in the country's law and institutions (Zheng & Deng, 2010; Luo, He, & Xue, 2016). The bond market has become one of the crucial components of China's multi-level capital markets. China's bond market size ranks second in the world, serving as one of the most important financing channels for Chinese firms, second only to the bank financing. However, weaknesses in default resolution, especially the inefficient bankruptcy regimes [1], have prevented small- and medium-sized enterprises from gaining access to bond issuance and thus hindered the development of a high-yield bond market (Becker & Josephson, 2016). Increasing direct financing, enlarging bond market size and marketizing the bond market are among the main goals to be achieved during China's 14th Five-Year Plan period. The marketization reform of the bond market relies heavily on the efficiency of the judicial system. Current literature on “law and finance” focuses mainly on the stock market and the banking sector, but much less on the bond market. Our analysis of the impact of bankruptcy law enforcement on bond issuances adds to the current literature. This paper not only contributes to the literature on law and bond market, but also has practical implications for promoting the marketization of bond issuances and the establishment of China's multi-level capital markets.

The rule of law serves as the institutional foundation of the market economy. It plays important roles in defining property rights and is regarded as the breeding ground of the pricing mechanism in free and fair trades (Liu, 2015). Higher quality of law enforcement, as an essential element of the rule of law, is associated with stronger investor protection (Ergungor, 2004), and helps to pave the way towards the multi-layer securities market (Gu, Zhuang, & Cao, 2016). The 14th Five-Year Plan proposes the blueprint for the development of China's bond market, and the improvements of default resolution serves as an important channel. Defaults are inevitable in a well-functioning bond market and bankruptcy is one of the most important resolutions for bond defaults (Dou, 2016). However, in the past years, government intervention to bail out ailing enterprises has enabled a large number of inefficient firms to survive, leading to the absence of an efficient bankruptcy system (Shen, 2016; Wang & Liu, 2018; Zhang & Pu, 2018). Default risk is the most important determinant for pricing bonds (Merton, 1974). The implicit guarantee and the lack of bankruptcy regimes can distort risk-pricing in bond issuances, leading to systematic overvaluation of bonds (Leland & Toft, 1996; Fan & Sundaresan, 2000; Francois & Morellec, 2004). Without an efficient bankruptcy system, bond prices could hardly reflect fundamental values and the establishment of the market-based bond issuance mechanism has extreme difficulties.

In July of 2020, the Supreme People's Court of China issued a notice on the “Minutes of the National Courts Symposium on the Trial of Bond Disputes”, in an effort to emphasize the importance of bankruptcy as a tool for default resolution (Supreme People's Court, 2020). The Enterprise Bankruptcy Law was approved and enacted in 2006. However, the outcome of the legislation of 2006 Bankruptcy Law turned out to be far from expected because of the weak legal environment and the public's misconceptions about bankruptcy (Li & Wang, 2011). The quality of law enforcement could compensate for the incompleteness of laws. In China, the poorly functioning bankruptcy procedures and the shortage of professionals for handling bankruptcy cases resulted in disproportional bankruptcy rates of potentially inviable firms, a large number of pending untried cases and many difficulties in enforcing bankruptcies. To correct this situation, bankruptcy courts have been introduced across cities since 2019. With the appointment of specialized personnel, bankruptcy courts are able to accelerate the disposal speed of suits and provide distressed firms with higher quality of in-court resolution. The establishment of bankruptcy courts could help improve the default resolution system through more cost-efficient bankruptcy regimes.

Improvements in the bankruptcy system facilitate corporate bonds issuance, especially for high-risk firms, and thus promote the development of high-yield bond market (Becker & Josephson, 2016). The introduction of bankruptcy courts is an important move towards the marketization of the bond sector. In a market-based financial system, price distortions caused by implicit government guarantees would be mitigated and credit spreads would become more risk sensitive (Zhang, 2021). Exploiting the staggered introduction of bankruptcy courts across cities in China, we use a differences-in-differences approach to estimate the effect of bankruptcy courts on bond issuance spreads. We find that the spread is significantly higher and more sensitive to firm fundamentals, including firm size, profitability and downside risk, after the introduction of bankruptcy courts. The results imply that the establishment of bankruptcy courts enables bondholders to price risk. Moreover, our findings show a significantly larger number of bond issuances and lower issuer credit ratings after bankruptcy courts were established. It indicates that more high-risk firms are able to access bond financing. The results conform with the predictions of Becker and Josephson (2016), showing that improvements in bankruptcy enforcement are associated with increases of bond financing by high-risk issuers.

In addition, this study finds that the effect of bankruptcy courts on bond issuance spreads is more profound for publicly traded bonds and for those whose issuers located in regions with lower level of Judicial Confidence Index. These empirical results support the argument that bankruptcy courts could strengthen creditors' protection and have a positive impact on lenders' confidence in regional judiciary. We also find evidence that the introduction of bankruptcy courts helps mitigate price distortions caused by state ownership and the implicit government guarantees. The rise of bond issuance spreads following bankruptcy court establishment is more profound for SOEs and issuers located in regions with higher government solvency and higher unemployment rate. Finally, we show that the functioning of bankruptcy courts relies on the development of the local financial market for the effect of bankruptcy courts is more pronounced in regions with higher Marketization Index.

This paper contributes to the “law and finance” literature. First, with an emphasis on emerging economies, our findings support the law–finance–growth nexus from the perspective of the bond market. Current literature focuses more on the impact of legal environment on the stock market and the banking sector, with limited evidence related to the bond market. Using cross-country data, La Porta et al. (1997, 2000) and La Porta, Lopez-de-Silanes, and Shleifer (2008) document that countries with better judicial institutions have larger financial markets. The theoretical model of Shleifer and Wolfenzon (2002) also suggests that stronger investor protection is associated with faster growth of the stock market. Based on firm-level analysis of China's financial market, several studies show the improvement of legal institutions helps the development of China's stock market and banking sector (Zheng & Deng, 2010; Zhang & Wang, 2012; Luo et al., 2016). There are also studies investigate how creditors' protection affects bank loan contracting or corporate debt financing (Djankov, Hart, McLiesh, & Shleifer, 2008; Visaria, 2009; Haselmann, Pistor, & Vig, 2010; Chemin, 2012; Vig, 2013; Ponticelli & Alencar, 2016; Qian & Fang, 2017). This paper brings new evidence to law and finance literature by examining the effect of court enforcement on the bond market.

Second, most studies on bankruptcy law reform focus on the “law on the books” and the effects of law enactment but largely ignore the effect of law enforcement. We depart from the existing literature and shift our attention to judicial enforcement. There has been rich evidence related to bankruptcy law enactment. The 1978 US Bankruptcy Act weakened creditors' bargaining power against ailing enterprises and resulted in higher credit spreads for weak firms (Hackbarth, Hennessy, & Leland, 2007). In 2005, the Brazilian Congress approved a new bankruptcy law aimed at increasing creditor protection, followed by a significant increase in the total amount of firms' debt (Araujo, Ferreira, & Funchal, 2012). Moreover, evidence from Ponticelli and Alencar (2016) shows that differences in Brazilian court enforcement affected the impact of the bankruptcy reform on firm access to finance, investment and outputs. Several studies investigate the impact of China's 2006 Bankruptcy Law. Wei and Xue (2010) show that stock market investors believed the law would enhance creditor protection and they upgraded the value of major banks. Jiang, Shen, Jiang, and Yin (2017) document that improved creditor protection promoted innovation at the firm level after the passage of the 2006 Bankruptcy Law. Instead of focusing on bankruptcy codes written on books, this study explores the impact of bankruptcy enforcement.

Existing literature related to bankruptcy law enforcement has largely focused on the increase in market efficiency following court reforms. Visaria (2009), Lilienfeld-Toal, Mookherjee, and Visaria (2012) and Vig (2013) document that the introduction of Debt Recovery Tribunals in India had a positive impact on firms' debt financing. Chemin (2012) shows that speedy disposal of civil suits encouraged investments and facilitated access to finance. Using a database provided by the World Bank's Doing Business Survey, several studies show how the quality of bankruptcy law enforcement affects innovation and firms' choice of debt financing exploiting cross-country differences (Lu & Dang, 2015; Zheng, Zhou, Chen, & Yang, 2019; Becker & Josephson, 2016). These studies have looked for cross-country correlations between creditor protection and financial outcomes from the law and finance perspective, which probably subjects to endogeneity problems due to the existence of several omitted factors at the country level in cross-country studies. Following a quasi-experiment approach, we conduct a within-country analysis based on China's judicial practice, in an effort to provide evidence for the effect of bankruptcy law enforcement on the bond market, correcting for endogeneity.

Third, this paper also provides implication for judicial practice. Different from the existing evidence that bankruptcy enforcement helps to reduce cost of debt (Visaria, 2009; Araujo et al., 2012; Gross, Kluender, Liu, Notowidigdo, & Wang, 2021), we find that the bankruptcy courts increase bond issuance spreads in China. However, it's inappropriate to directly draw the conclusion that China's introduction of bankruptcy courts increases the cost of debt. Different from the institutions in developed economy, China has maintained a state-dominated financial system. The implicit government guarantee and soft budget constraints have resulted in the ill-functioned bankruptcy system and obstacles for market exits. The introduction of bankruptcy courts improves the efficiency of the bankruptcy system and strengthens investor protection. It enables firms to borrow at costs that reflect their default risks and facilitates the development of high-risk bond market.

The remainder of the article is organized as follows: Section 2 discusses the development of China's bond market and the institutional background of China's bankruptcy law reform. Section 3 presents the theoretical analysis and Section 4 describes our empirical strategy. Section 5 reports the results and Section 6 concludes.

2. Institutional background

2.1 China's bond market

China's bond market comprises a significant segment of the multi-layer securities market. Bond financing has become an increasingly important financing channel, with an aggregate size second only to the banking sector. According to statistics from the People's Bank of China, the total value of China's bond market reached RMB 131.7 trillion as at the end of November in 2021, ranking second in the world [2]. Despite its growing size, the weakness of the law and institutions has hampered the development of the marketization process of the bond sector, which, in turn, impaired its function of supporting the direct financing of small- and medium-sized enterprises (Cai & Tang, 2011).

In early days, bond issuance was regulated and examined by the government authorities and the approval criteria were strict. In a sense, the strict approval criteria constituted a quality guarantee or insurance of the bond issued. The new Securities Law was enacted in 2019 and went into effect on March 1st, 2020, which signaled the shift towards a registration bond issuance system (Standing Committee of the National People’s Congress, 2019). Under the registration issuance system, the regulatory authorities no longer provide endorsement for bond value. As the criteria for bond issuance are lowered, the market opens up to high-risk issuers.

Inefficient exit mechanism and lagged clearing system increase financial risks, which, in turn, have a negative impact on bond issuance. A well-functioning exit system serves as a prerequisite for the marketization of bond issuance. Defaults release credit risks and sound exit systems help avoid risk accumulation. Both are indispensable in preventing the outbreak of financial crisis (Yin & Xie, 2021). China's bond market had maintained a zero-default rate in the past because of the implicit government guarantees and bailouts (Peng & Cheng, 2018). It was not until 2014 that China experienced its first domestic bond default when Shanghai Chaori Solar Co. failed to pay out interest on bonds sold two years ago. Since then, both the number and the amount of bond defaults have been increasing. At the same time, the market-oriented debt resolution mechanisms have been developing. More firms have gained access to the bond market and bond financing cost has reduced. Figure 1 shows the monthly bond issuance spread during the period from 2018 to 2020, indicating a decreasing trend [3].

The number of bankruptcy cases has been gradually increasing since China experienced its first bond default in 2014. However, the enforcement difficulty is still a prominent problem in bankruptcy regimes. Low efficiency in bankruptcy system and slow exits for defaulting issuers result in risk accumulation in the bond market. The absence of an enforcement judiciary invites government interference in bankruptcy, leading to distortions of bond pricing, mismatch of risk and return and inefficient market resource allocation. Improving exit mechanisms has become especially important under the increasing trend of bond defaults. In April 2020, the CPC Central Committee and the State Council issued the “Opinions on Building a More Perfect Factor Market-based Allocation System”, proposing to accelerate bond market development through improvements in default resolution (Central Committee of the Communist Party of China and State Council, 2020). Later in July 2020, the People's Bank of China, the National Development and Reform Commission and the China Securities Regulatory Commission issued the “Notice on Matters Concerning the Disposal of Corporate Credit Bond Defaults” (People's Bank of China, National Development and Reform Commission, & China Securities Regulatory Commission, 2020). The Notice emphasizes to promote a unified law enforcement system for the bond market and calls for the establishment of an efficient mechanism for default resolution based on the market-oriented and rule-based principles. In the meantime, the Supreme People's Court promulgated the “Minutes of the National Courts Symposium on the Trial of Bond Disputes” to address the importance of bankruptcy mechanism in default resolution and further clarify the obligations of the bankruptcy administrator in promptly confirming claims and continuing information disclosure during the bankruptcy proceedings (Supreme People's Court, 2020).

2.2 China's bankruptcy law and bankruptcy courts

2.2.1 Bankruptcy law reform and the dilemma in legal practice

The trial version of the Enterprise Bankruptcy Law was promulgated in 1986 and came into effect in 1988. In early days, bureaucratic approval was required before the filing of a bankruptcy case. China experienced a low rate of bankruptcy because it was overregulated by bureaucrats. In August 2006, the Enterprise Bankruptcy Law was effectively approved for enactment by the Standing Committee of the 10th National People's Congress. The new law gave the creditor or the debtor in possession the right to voluntarily decide whether to accept a reorganization plan, to adopt a reconciliation procedure or to apply for bankrupt liquidation. The new law also authorized the court to designate a bankruptcy administrator. Accounting firms, law firms and any certified intermediaries might serve as an administrator. The market-based bankruptcy proceedings could help improve default resolution efficiency. The 2006 Bankruptcy Law intended to put an end to the old practice of bankruptcy procedures with government intervention and set up a market-based bankruptcy system (Li & Wang, 2011).

However, the number of bankruptcy filings had been declining over the years because of the weakness in law enforcement and the insufficient supporting institutions (Wang & Xu, 2014). First, the acceptance rate of bankruptcy suits remained low. The enactment of the 2006 Bankruptcy Law set up the infrastructure for market-oriented default resolutions. Yet, social stability and government performance were still issues to be considered before a bankruptcy suit was accepted by the court (Lu, 2020). Bureaucrats used their discretion to exert pressure on the judicial branch and set up high barriers for the acceptance of bankruptcy applications [4]. Inadequate court performance evaluation and lack of professionals were also reasons that led to low incentives to accept bankruptcy suits in civil courts. Compared to other civil suits, the bankruptcy procedures are more complex and time-consuming. Despite its greater complexity, bankruptcy suits were subject to exactly the same performance evaluation standards as other civil suits, which had resulted in courts' lower willingness to process bankruptcy cases. Moreover, there were not enough professionals to handle bankruptcy cases.

Second, the old practice for bankruptcy procedures was to designate a default resolution group led by the local governments. Judicial candidates in charge of the bankruptcy cases, as group members, were usually subject to the jurisdiction of bureaucrats. Such arrangements facilitated successive government interventions in bankruptcy suits and would lead to erosion of judicial independence (Lu, 2020).

Third, although the new law authorized the court to appoint a bankruptcy administrator who might be a member of recognized legal or accounting specialist who possessed relevant professional expertise and qualifications, the old practice of designating a government-appointed liquidation group have remained for a long time. The government still acted as the bankruptcy administrator, which created opportunities for it to exert influence on civil courts to serve political interests. Weak judicial independence in bankruptcy proceedings remained a problem to be solved (Li & Wang, 2011).

2.2.2 Establishment of bankruptcy courts

As discussed above, distorted performance evaluation and lack of professionals have resulted in perceived unwillingness to accept bankruptcy applications in civil courts. The absence of judiciary institutions invited government interference in bankruptcy proceedings and eroded judicial independence. In an effort to promote the development of a rule-based mechanism for market-oriented debt resolutions, an effective judiciary branch may need to insulate interference from the government. Therefore, it has been argued that specialized courts for bankruptcy cases be introduced to facilitate speedy disposal of bankruptcy suits (Xu, 2016).

The motive of the introduction of bankruptcy courts is to improve the legal channels for bankruptcy resolution by promoting the appointment of specialized personnel and improving the quality of law enforcement. In December 2018, Shenzhen Intermediate People's Court became the first to announce the establishment of a bankruptcy court. Beginning in 2019, bankruptcy courts were set up by the intermediate people's courts across cities. Table 1 presents information about the introduction of bankruptcy courts by the end of 2020.

According to Article Three of the Enterprise Bankruptcy Law (2006), bankruptcy cases shall be assigned to territorial jurisdictions where the creditor resides. In 2020, the Supreme People's Court issued the “Minutes of the National Courts Symposium on the Trial of Bond Disputes”, which further specified the territorial jurisdictions of bankruptcy suits (Supreme People's Court, 2020). As stated in the document, “for a bankruptcy case in which a trustee petition or bondholders petition for reorganization or liquidation of the issuer, or in which the issuer petitions for reorganization, reconciliation, or liquidation, the jurisdiction shall belong to the intermediate people's court of the place where the issuer is located.” Therefore, the scope of jurisdiction by a bankruptcy court covers bankruptcy of firms registered at the municipal level and above in the city where bankruptcy court locates. All issuers in the city are thus considered to be exposed to bankruptcy courts once a bankruptcy court is established in a city.

2.2.3 Bankruptcy courts and improvements in bankruptcy resolution

The establishment of bankruptcy courts improves the efficiency of the bankruptcy system and strengthens the independence of judicial branch in bankruptcy proceedings. It promotes the development of rule-based infrastructures and market-oriented default resolution procedures as explained below.

1. Speedy disposal of bankruptcy suits and higher quality of law enforcement

Bankruptcy courts have separate personnel, trial place and office space. The separate organizational design enjoys independent performance evaluation for bankruptcy suits, which may provide stronger incentives for judges to process bankruptcy cases. Specialized courts are better able to assign cases to judges in line with their expertise, speeding up the execution of the verdict and having better judgment for bankruptcy cases.

The quality and efficiency of bankruptcy case resolution improve due to more reliable performance evaluation and the appointment of specialized personnel. After the setup of the bankruptcy court, processing time is shortened, and bankruptcy costs are lowered. For example, Guangzhou Bankruptcy Court follows a streamlined “summary procedure” for bankruptcy suits in order to facilitate faster processing. By the end of November 2020, the proportion of bankruptcy cases processed through the “summary procedure” was 52.7%, and the average time taken to process such cases was 0.35 year, about 51% shorter than that of the common bankruptcy processing [5]. At the same time, the recovery rate of debt resolution improves. Evidence shows that more viable debtors under financial distress have been rescued through reorganization, leading to higher recovery rate for debtors in possession.

Speedy disposal of bankruptcy suits and higher quality of law enforcement increase the willingness of debtors and creditors to file for bankruptcy and impose a positive feedback effect on the exit through bankruptcy. Since the establishment of Shenzhen Bankruptcy Court, the number of bankruptcy cases accepted has increased by 50.3% and the number of suits resolved rose by 23% in 2019 [6]. Similarly, after the introduction of the Wenzhou Bankruptcy Court, the number of bankruptcy cases accepted and resolved by courts in the region increased by 78.6% and 54.2% respectively in 2020 [7].

1. Greater judicial independence in bankruptcy resolution

The bankruptcy court is authorized to designate a bankruptcy administrator to manage the affairs of the debtor, distinguishing from the previous practice of a government-appointed liquidation group. For example, the Shenzhen Bankruptcy Court was actively involved in setting up the Shenzhen Bankruptcy Administrator Association to encourage self-regulation for qualified administrators [8]. As for the Shanghai Bankruptcy Court, an administrator was appointed through random selection from the pool of certified intermediaries or through competition [9]. Such appointments allow for less bureaucratic discretion in the designation of administrators.

In the meantime, the bankruptcy court is able to strengthen its supervision over the administrator's subsequent performance of duties. For example, in cases of “Bestway Marine and Energy Technology” and “China Electric Equipment Group” processed by the Shanghai Bankruptcy Court during 2020, administrators were instructed to hold creditors' meetings, select investors for reorganization and propose reorganization plans under the supervision of the bankruptcy court. This signals the dominant role of the judicial branch in bankruptcy proceedings after the establishment of bankruptcy courts [10].

In sum, the introduction of bankruptcy courts greatly enhances the independence of judiciary in bankruptcy resolution. Greater judicial authority and stronger market discipline allow for little room for government intervention and promote the development of rule-based mechanisms for market-oriented bankruptcy resolution.

2.3 Bond issuance characteristics after the establishment of bankruptcy courts

The establishment of bankruptcy courts provides a market-based exit mechanism for default resolution and has positive impacts on the bond market. Figure 2 shows the trends of bond issuance characteristics before and after the bankruptcy courts were introduced. The sample covers corporate bonds, enterprise bonds and medium-term notes issued by firms located in provinces with at least one bankruptcy court established. The Figure presents the bond issuance spread, the number of bonds issued and credit ratings of bond issuers around the establishment of a bankruptcy court from month −10 to month 10, counting the month of bankruptcy court's establishment as zero [11].

As shown in Figure 2 Panel A, the issuance spreads for bonds issued by firms located in cities with bankruptcy courts are lower than those issued by firms located in cities without bankruptcy courts during the 10-month period before the establishment of bankruptcy courts, and a parallel trend is observed. However, after the setup of the bankruptcy counts, the issuance spreads for bonds issued by firms located in cities with bankruptcy courts increase compared to those issued by firms located in cities without bankruptcy courts. Figure 2 Panel B shows that the number of monthly bond issuance for cities with bankruptcy courts rises in comparison to those without bankruptcy courts. In Figure 2 Panel C, the difference in credit ratings for issuers located in cities with bankruptcy courts and for those in cities without bankruptcy courts narrows after the setup of the bankruptcy courts. It indicates that after the introduction of bankruptcy courts, the newly issued bonds by firms located in cities with bankruptcy courts are perceived to have higher default risk.

The changes of bond issuance characteristics indicate that after the establishment of the bankruptcy courts, the market-based pricing mechanisms come into play and bond issuance spreads increase, consistent with the Zhang's (2021) prediction. The number of bond issuance has also increased after the introduction of bankruptcy courts, with more high-risk firms being able to issue bonds, conforming with the theoretical prediction by Becker and Josephson (2016).

3. Theoretical analysis

The law and finance literature argues that the extent of legal protection differs across countries and such differences affect the development of the countries' financial systems (La Porta et al., 1997, 1998; Shleifer & Wolfenzon, 2002; Djankov et al., 2007). There are two categories of law enforcement mechanisms, private enforcement and public enforcement and they substitute for one another. When the presence of various obstacles precludes private enforcement from playing a significant role in market regulation, strong public enforcement serves as an alternative to substitute for the weakness of private enforcement (Layton, 2008). However, public enforcement requires devotion of public resources and can lead to the government's ex ante interference in the market, which may have a negative impact on financial market order (Liu, 2015). Strong law enforcement could reduce the reliance on public enforcement through legal discipline and market participants' self-regulation (Coffee, 2007; La Porta et al., 2008). Higher degree of self-regulation in securities market facilitates the development of the market-based financial system. La Porta, Lopez-de-Silanes, and Shleifer (2006) provide evidence that private enforcement of laws is superior to strong regulation by the government in promoting securities market development.

High-quality law enforcement protects investors against expropriation and raises their willingness to provide finance (La Porta et al., 1997). Coffee (2007) argues that the intensity of law enforcement may be the factor that best distinguishes the US securities markets from others. The effective legal enforcement mechanism makes investors in the US securities market willing to pay a high premium, which increases the valuation of US listed companies relative to other capital markets. In countries with weak legal environments, however, the size of the stock market and the bond market is relatively small (La Porta et al., 1997, 2000, 2008). Therefore, legal institutions have important influences on the width and depth of financial markets.

The rule of law constrains the power of the government to intervene and ensures the order of market economy (Qian, 2000). Excessive government intervention weakens the effectiveness of securities market pricing efficiency and thus the security price fails to reflect fundamental values of listed firms (Allen, Qian, & Qian, 2005). Strengthened judicial independence and better legal institutions restraint government interventions in economic activities and therefore impose stronger market discipline. Improved legal protection improves securities pricing, facilitates access to external finance, reduces financial constraints and encourages investors' participation (La Porta, Lopez-de-Silanes, Pop-Eleches, & Shleifer, 2004; Berkowitz, Lin, & Ma, 2015; Mclean, Zhang, & Zhao, 2012). In the case of China's bond market, credit spreads might not accurately reflect credit risk factors due to the presence of implicit government guarantees. Excessive bureaucratic interference in the bankruptcy regime has resulted in price distortions and hampered bond market growth. Therefore, higher judicial quality in the bankruptcy system matters for the establishment of a market-based bond pricing system and for enlarging the width and depth of the bond market in China.

3.1 Bankruptcy courts and bond issuance pricing

The value of bonds depends essentially on the risk of bond default (Merton, 1974). Default risk is the uncertainty surrounding a firm's ability to service its debts and obligations. Investors evaluate the likelihood of default and firms have to pay a spread premium that is proportional to their default probability to compensate lenders for this uncertainty (Crosbie & Bohn, 2003). Weber, Duffy, and Schram (2016) show that the default probability affects the issuing price in the bond market. The financial condition of a firm serves as an important indicator of its default probability (Madan & Unal, 2000). According to the bond pricing model of Merton (1974), credit spreads are expected to increase with leverage. Default is triggered when the leverage ratio approaches 1 (Collin-Dufresn, Goldstein, & Martin, 2001).

There are two ways of default resolution including negotiation out of court and bankruptcy, the in-court resolution. Negotiation out of court is more flexible and incurs lower costs. In spite of this, decentralized bondholders with weak bargaining power in negotiation find the in-court bankruptcy resolution more favorable (Becker & Josephson, 2016).

However, the implicit government guarantee has prevailed in China's bond market for a long time (Dou, 2016; Yuan & Wang, 2021). One the one hand, the government would provide bailouts for the distressed issuers when they encounter financial difficulties to maintain credit rating of local firms and stability of the local economy. On the other hand, in absence of professional judges and sound performance evaluation system, court judges have little incentive to handle the complex and troublesome bankruptcy cases (Xu, 2016). The lack of sufficient judicial support in bankruptcy proceedings invites government interference in bankruptcy.

Credit spreads represent a premium required by investors to compensate for the expected loss from defaults. The government's involvement in default resolution reinforces investors' expectations that the government would guarantee the obligations of distressed issuers when they are not able to satisfy some or all of the indenture requirements (Wang & Chen, 2015; Wang, Lv, & Ye, 2016). Therefore, investors tend to weigh less on the financial condition of firms and attach more importance to state influence characteristics, such as state-ownership and political connections in bond pricing (Wei, Lai, & Zhang, 2017; Wang, Guo, Xue, & Liu, 2015). Firms would thus borrow at costs that do not reflect their default risks (Wang, Shen, & Zhong, 2019). The default events may alter investors' belief in future government support and bondholders will perceive a higher likelihood of default. Wang et al. (2015) support this notion by providing evidence that credit spreads increased dramatically after the Shanghai Chaori Solar event. However, the Shanghai Chaori Solar bond default event and those followed mitigate but fail to eliminate bailout expectations.

The establishment of bankruptcy courts unblocks the market-based exit mechanism for bond default resolution. First, bankruptcy courts appoint professional judges and adopt performance evaluation systems that incentivize judges to devote to bankruptcy proceedings, thus leading to higher acceptance of bankruptcy petitions. At the same time, higher expertise gives rise to higher quality of case resolution and market participants would be more willing to file bankruptcy cases. Evidence from the Shenzhen Bankruptcy Court shows that the number of cases accepted and resolved has been increasing since its establishment. Second, specialized courts with greater independence would be able to resist the intervention of local governments and thus strengthen the dominant role of judiciary in bankruptcy proceedings.

As the number of bankruptcy cases rises, China's bond market risk also increases. The introduction of market-based bankruptcy regimes has altered investors' perception of default probability. When bankruptcy turns out to be a formal mechanism for default resolution, investors would evaluate an issuer's fundamentals and incorporate those assessments into securities' prices, demanding higher yields for greater risks assumed (Wang, Shen, & Zhong, 2019).

3.2 Bankruptcy courts and bond market entry

The development of market-based and rule-based mechanisms for default resolution serves as the premise of the market-oriented reform of bond issuance. First, an effective bond risk pricing mechanism is necessary to guide the efficient allocation of bond market resources. In case of implicit government guarantee, credit risk and bond price do not match properly, resulting in resource misallocation and obstacles for the bond market entry. Second, inefficient default resolution and slow market exits increase the probability and severity of future crises, and in turn hamper the bond issuance reform (Yin & Xie, 2021). In addition, for less efficient bankruptcy regimes, market participants are more likely to negotiate out of court, in which distribution of liquidation value depends on lenders' bargaining power. Because decentralized bondholders are generally less favored compared to concentrated lenders (banks) in this situation, they are less willing to fund risky firms, leading to the underdeveloped high yield bond market (Becker & Josephson, 2016).

The introduction of bankruptcy courts promotes the development of a market-based bankruptcy system. On the one hand, the market-based mechanism could mitigate distortions in risk pricing and improve resource allocation efficiency. Local governments would no longer bear the responsibility to provide guarantees for the value of bonds issued and investors tend to be more willing to monitor and price the risks of issuers (Shen & Cao, 2014). On the other hand, speedy disposal of bankruptcy suits could facilitate faster exits for distressed firms to avoid the outbreak of financial crisis. Meanwhile, the appropriate distribution of value and higher quality of default resolution have provided bondholders with higher incentives to fund risky firms, and thus encouraged high-risk firms to issue bonds. Therefore, bankruptcy, as a risk release instrument, has contributed to the growing size of bond issuance and the introduction of high-yield bonds in the market.

Lower threshold to bond financing could enable more risky firms to issue bonds and thus lead to higher credit risk in the bond market. Default risk is an essential determinant of credit ratings. As the risk of default increases, issuers would receive lower credit ratings accordingly (Zhao & Long, 2007; Livingston, Poon, & Zhou, 2018). Moreover, confronted with the increasing number of default events, credit rating agencies would become more sensitive to default risk and more cautious about credit ratings for they have higher pressure to provide high quality products, informative credit ratings (Wang, Shen, & Zhong, 2019).

In sum, higher quality of law enforcement enables the judicial branch to play a significant role in bankruptcy proceedings and thus fosters the development of a market-based mechanism for default resolution, which in turn promotes the reform of bond issuance and encourages the entry of high-yield bonds. With the change in the perceived implicit government guarantees, the bond risk is more accurately reflected in bond prices and credit spreads thus increase (Zhang, 2021). The introduction of bankruptcy courts not only allows for lower entry threshold for bond issuance, but also enables the development of a market-based bond pricing system.

4. Identification strategy

4.1 Data and sample

As shown in Table 1, the first bankruptcy court (Shenzhen Bankruptcy Court) was established in China on January 14th 2020 and the latest bankruptcy court established during our sample period was the Xiamen Bankruptcy Court, on August 18th 2020. The original sample contains 9,831 issuances of corporate bonds, enterprise bonds and medium-term notes from 2018 to 2020. We exclude bonds issued by financial institutions and observations with missing values and get 8,603 bond issuances. To further control for differences across regions, we restrict our sample to issuers located in provinces with at least one bankruptcy court established. The final sample contains 5,451 bond-date observations. The bond-level data and issuer (firm)-level data are obtained from the Wind database. We winsorize all continuous variables at the 1 and 99% percentiles.

4.2 Empirical model

To investigate the impact of bankruptcy courts on bond issuance, we examine changes of bond issuance spreads before and after bankruptcy court establishments. Table 1 shows the staggered introduction of bankruptcy court. Accordingly, we implement the time-varying DID (differences-in-differences) strategy as shown in Equation (1):

(1)IssueSpreadi,t=α+β1Posti,t+β2Bond+β3Firm+β4Issuenumj,t+β5GDPg+Date+City+Ind+εi,t

In Equation (1), IssueSpreadi,t is the credit spread of bond i at issuance date t. Following Jiang (2008) and Wang and Gao (2017), IssueSpreadi,t is computed as the yield to maturity on the bond issuing date minus the yield to maturity on duration-equivalent treasury bond on that date. Post is a dummy variable coded 1 for periods after bankruptcy court establishments and 0 otherwise. For bonds whose issuers are located in cities with no bankruptcy court, Post equals 0 all the time. We expect β1 to be positive if the establishment of bankruptcy courts leads to higher credit spreads.

Following Wang and Gao (2017) and Wang and Xu (2019), we control for a range of bond-level characteristics (Bond) and firm-level financial variables (Firm) in the model. The bond-level characteristics include issuance size (Bondsize), duration (Bondterm), credit rating (Bondrate), whether the bond is puttable (Bondput) or callable (Bondcall), and whether the bond is secured (Bondsecu). The firm-level variables are firm size (Firmsize), leverage (Lev), the growth rate of operating income (Growth), return on asset (ROA), coverage ratio (Cover), tangibility (Tangibility), current ratio (Current), auditing quality (Big4), state ownership (SOE) and shareholdings of the largest stockholder (Top1). We also control for the number of bond issuance (Issuenum) and the growth of regional GDP (GDPg). In addition, we include issuing date, city and industry fixed effects in our regression. The detailed definition of variables is listed in Table 2.

5. Empirical results

5.1 Descriptive statistics

Table 3 presents sample descriptive statistics. The table shows that the mean value of bond issuance spreads is 2.162 in percentile, and the minimum and maximum value equals 0.522 and 5.839 respectively. The average value of Post is 0.375, indicating that over one third of the total observations are exposed to bankruptcy courts. Other variables fall within a range of reasonable values.

Table 4 presents correlation coefficients for main variables. The correlation between IssueSpread and Post is −0.162. However, we cannot draw the conclusion based on the unconditional correlation that the bankruptcy court leads to lower bond issuance spreads. First, the causal relationship cannot be established without taking into account the impacts of bond characteristics, firm-level financial variables and fixed effects. Second, the negative correlation may reflect a downward trend in bond issuance spreads during the 2018–2020 period. The correlation of IssueSpread and SOE is −0.225, indicating that state-owned enterprises can borrow at lower costs. Our findings also show a strong correlation between Bondrate and Firmsize (ρ= 0.583), which implies that credit ratings are positively associated with firm size. Other correlations fall within a range of reasonable values, suggesting that multicollinearity is not a concern in the regressions.

5.2 Univariate analysis

The bonds in the sample are classified into four groups according to whether their issuers are located in cities with bankruptcy courts and whether they are issued before or after the bankruptcy courts were established. Table 5 presents the differences of the mean bond issuance spreads between groups. For cities without bankruptcy courts (Group2), we use the date of first bankruptcy court's establishment in the province to divide the sample period into pre- and post-establishment periods.

As shown in Table 5, before bankruptcy courts were introduced, the mean value of IssueSpread for bond issuances in cities with bankruptcy courts was 2.136, significantly lower than that, 2.733, in cities without bankruptcy courts. The average IssueSpread for bonds issued in cities with bankruptcy courts falls to 1.889 after the establishment of bankruptcy courts. But for cities with no bankruptcy courts, the value reduces to 1.987, signaling a narrower gap between the two groups. The t test for differences in changes of means shows that the decline of IssueSpread for cities with bankruptcy courts (0.247) following the establishment is significantly lower than that in cities without bankruptcy courts (0.746). Results in Table 5 indicate that the bond issuance spread experiences a downward trend within the sample period. However, the fall of spreads is less significant for bonds issued in cities with bankruptcy courts. The findings conform with our predictions in that the bond issuance spreads increase for cities with bankruptcy courts relative to those without following the introduction of these specialized courts.

5.3 Bankruptcy courts and marketization of bond issuance

5.3.1 Bankruptcy courts and bond issuance spreads

We employ regression (1) to examine the impact of bankruptcy courts on bond issuance spreads and results are shown in Table 6. In column (1), only date, city and industry fixed effects are included. The coefficient of Post is 0.320 and positively significant at 1% level. Then we add bond-level characteristics in column (2) and finds that the coefficient of Post is 0.132 and statistically significant at 1% level. In column (3), we further include firm-level financial variables and control for regional characteristics, and the coefficient of Post remains significantly positive. The value of r-squared increases as we include control variables in our regression, indicating that the increase of bond issuance spreads could not be fully explained by bond-level characteristics, firm-level financial variables and regional differences. The introduction of bankruptcy courts did give rise to higher issuance spreads. The coefficient of Post in column (3) is 0.084, which implies that the bond issuance spread increases by 0.084 following the introduction of bankruptcy courts, about 3.88% of the mean issuance spreads. Our findings of increased issuance spreads after the introduction of bankruptcy courts are consistent with our predictions.

5.3.2 Bankruptcy courts and bond market entry

We predict that the establishment of bankruptcy courts lowers the threshold for bond issuance and allow more risky firms to access the bond market. We first examine the unconditional differences in the means between groups to investigate changes in the number of bonds issued and issuer credit ratings around the introduction of bankruptcy courts. Univariate tests results are reported in Table 7. Panel A shows that the average number of bonds issued (Issuenum) in cities with bankruptcy courts increases significantly after the establishment of the specialized courts while the mean of Issuenum in cities without bankruptcy courts remains almost the same as before. Panel B presents the results for univariate test of issuer credit ratings. It can be seen that credit ratings are significantly higher after bankruptcy courts were established and rating upgrades tend to be less profound for issuers located in cities with bankruptcy courts. These results indicate that the issuers' credit ratings are relatively lower after the introduction of bankruptcy courts in comparison to those in cities without.

We then implement a differences-in-differences strategy to further investigate changes in the number of bonds issued (Issuenum) and issuers' ratings (CorpRate) around the introduction of the bankruptcy courts. The results are shown in Table 8. The coefficients of Post are negative and statistically significant in both regressions, with Issuenum and CorpRate used as the dependent variable in column (1) and (2) respectively. The results indicate that the number of bonds issued increases significantly and the issuers receive lower credit ratings after bankruptcy courts were established.

Overall, our findings in Tables 7 and 8 show that the bankruptcy court promotes the marketization of bond issuance in China and lowers the bond issuing threshold. After the introduction of bankruptcy courts, the number of new bonds issued increases and more risky firms are able to access the bond market.

5.3.3 Bankruptcy courts and bond pricing

By improving the quality of default resolution, bankruptcy courts facilitate the transition towards a market-based pricing mechanism and enable firms to borrow at costs that better reflect their fundamental values. After bankruptcy courts were established, bond pricing would be more informative to firm fundamentals, indicating higher sensitivity of credit spreads to risk factors. We empirically investigate these predictions by testing whether the introduction of bankruptcy courts increases the sensitivity of bond issuance spreads to risk factors, including firm size, profitability and downside risk.

The regression results are presented in Table 9. Column (1) reports the sensitivity of bond issuance spread to firm size. We sort firms according to their size and divide them into two groups based on the industry-year median. The dummy variable HFirmsize equals 1 for the group of firms with firm size above the industry-year median, and 0 otherwise. The coefficient of the interaction term HFirmsize×Post in column (1) is negatively significant at 1% level, indicating that bond issuance spreads become more sensitive to firm size following the introduction of bankruptcy courts.

Column (2) of Table 9 reports the sensitivity of bond issuance spread to issuer's profitability. We create a dummy variable HROA, which is coded 1 for firms with ROA higher than the industry-year median, and 0 otherwise. In column (2), the coefficient of the interaction term HROA×Post is negative and statistically significant at 1% level. The result implies that bond issuance spreads turn out to be more sensitive to profitability after bankruptcy court establishments.

Column (3) of Table 9 presents the sensitivity of bond issuance spread to downside risks. Previous studies employ organizational downside risk to measure financial distress (He, Zhu, Yang, & Wang, 2017). Based on the research of Miller and Leiblein (1996), we use the second-order root central moment of ROA for the past five years to calculate the downside risk: RERi,t=15∑t=15(ROAi,t−1−iROAi,t−1)2. ROAi,t−1 represents firm i's ROA in year t−1. iROAi,t−1 is the previous year industry average for ROA, serving as proxies for organizational targets for firm i. The underlying premise of the calculation is that firm i's ROA in year t−1 is lower than the target value. When firm i's ROA in year t−1 is higher than its target value, the term (ROAi,t−1−iROAi,t−1) is coded 0. Therefore, higher RER is associated with higher downside risk. Having calculated organizational downside risk, we sort firms according to RER and divide them into two groups based on the industry-year median. The dummy variable HRER equals 1 for the group of firms with downside risk higher than the industry-year median and 0 otherwise. In column (3) of Table 9, the coefficient of the interaction term HRER×Post is positive and statistically significant at 10% level, indicating that bond issuance spreads become more sensitive to downside risks after bankruptcy courts were established.

5.4 Additional analysis

5.4.1 Bankruptcy courts and investor protection

Bankruptcy courts are introduced in the wake of China's legal reform, and we have seen stronger investor protection following the establishment of bankruptcy courts. We empirically investigate whether bankruptcy courts enhance investor protection in China from the perspective of investor types and legal environments. The results are listed in Table 10.

Individual investors, especially the public investors will be more willing to invest in the bond market as creditor protection is strengthened. We therefore expect that more public investors would enter the bond market following the introduction of bankruptcy courts. We use the dummy variable PInvestor to identify bonds available to different types of creditors. PInvestor is coded 1 for publicly traded bonds and 0 when the bond is only accessible to qualified investors or qualified institutional investors. In column (1), the coefficient of PInvestor×Post is positive and statistically significant at 1% level, indicating that public investors have more incentives to fund for risky firms after bankruptcy courts were established.

Bankruptcy courts also serve to substitute for the weakness in law enforcement with higher professionalism and greater independence. The introduction of bankruptcy courts has given investors stronger incentives to invest in bonds because of their greater confidence in regional legal regimes. We use the Justice Confidence Index constructed by the World Bank's Investment Climate Survey to measure the quality of regional judicial system. The index reveals how much confidence the respondents have that their commercial disputes would be settled with justice by the local legal system. The sample is separated into two groups according to the median provincial Justice Confidence Index. The dummy variable HJudicial_Confidence is coded 1 if the province has a Justice Confidence Index above the median and 0 otherwise. The coefficient of HJudicial_Confidence×Post in column (2) is negative and statistically significant at 5% level. It implies that bankruptcy courts are strong complements to legal protection and have given investors greater incentives to fund for risky firms, especially in regions with severe weaknesses in legal institutions.

5.4.2 Bankruptcy courts and the implicit government guarantees

The deviation of bond pricing from the fundamentals can be attributed to the implicit government guarantees. Governments are likely to provide bailouts when state-owned enterprises fail to service obligations due to financial distress, which enables state-owned enterprises to borrow at lower costs and thus leads to lower issuance spreads for SOEs (Fang, Shi, & Zhang, 2013). We predict that the introduction of bankruptcy courts would mitigate the implicit government guarantee and therefore test the changes in bond issuance spreads caused by state ownership. The results are reported in Table 11. As shown in column (1) of Table 11, the coefficient of SOE×Post is positively significant at 1% level, indicating a narrower gap of credit spreads between SOEs and non-SOEs after bankruptcy court establishments.

Further, we explore the role of bankruptcy courts in eliminating bondholders' expectations of government support from the aspect of local governments' incentives and capabilities to provide bailouts. Unemployment rate is an important issue to be considered for local governments. Facing high unemployment rate, local governments are more likely to provide support to firms in an effort to maintain sufficient employment to meet the government's employment objective (Lu, Sun, Zhou, & Yang, 2014). Therefore, we employ the regional unemployment rate to measure the government's incentive to provide support and use it as a moderating variable in the regression. The results are reported in column (2) of Table 11. Here, the coefficient of Unemployment×Post is positively significant at the 5% level. It reveals that the introduction of bankruptcy courts could mitigate the negative correlation between incentives of government support and bond issuance spreads.

Local governments with stronger solvency will be more likely to provide bailouts to defaulting issuers. Government solvency therefore serves as an important indicator of the capability of local governments to provide support. Following Lei and Chen (2021), we use GDP per capita to estimate the capability of local governments to provide support and results are reported in column (3) of Table 11. The coefficient of GDP×Post in column (3) is positively significant at the 10% level. It reveals that the introduction of bankruptcy courts could eliminate the negative association between capabilities of providing government support and bond issuance spreads.

5.4.3 Bankruptcy courts and regional financial development

The functioning of bankruptcy courts depends on the extent of marketization of local financial markets. The well-established financial infrastructures along with well-developed capital markets serve as the foundation of well-functioning bankruptcy courts. Therefore, we predict that the impact of bankruptcy courts is more profound in regions with higher extent of marketization. We use the “Factor market development index” constructed by Wang, Fan, and Hu (2019) to measure the development of regional financial markets. The dummy variable HFin_Mar_index equals 1 for the observations with factor market development index higher than provincial median and 0 otherwise. The results are reported in Table 12. In column (1) of Table 12, the coefficient of HFin_Mar_index×Post is positively significant. We also employ the “financial resources allocation index” constructed by Wang, Shen, and Zhong (2019). The dummy variable HFin_Credit_index is coded 1 for firms with “financial resources allocation index” higher than the provincial median and 0 otherwise. As shown in column (2) of Table 12, the coefficient of HFin_Credit_index×Post is positively significant. Both regressions show that the increase of bond issuance spreads following the introduction of bankruptcy courts is more pronounced in regions with higher level of marketization. The results thus indicate that bankruptcy courts contribute more significantly to the development of high-yield bond market by giving investors greater incentives to finance risky firms in regions with higher extent of marketization.

5.5 Robustness tests

5.5.1 Testing for parallel trends

To test the parallel trends assumption of the time-varying DID design, Beck, Levine, and Levkov (2010) exclude the year of treatment counting as year 0 from the sample period in concern of the potential noise during the treatment year. Our analysis is based on day-level observations. The credit spreads shortly before and after the establishment of bankruptcy courts might be noisy for the following reasons. First, bond trading is less frequent compared to the stock market, which in turn leads to less timely bond pricing. Second, information about the introduction of bankruptcy courts might be publicly known ahead through the media news. Expectations of bankruptcy court establishment might have already had an impact on the bond market before bankruptcy courts were established. Lastly, it takes time for the functions of bankruptcy courts to take effect. Thus, we split the sample into 30-day intervals and drop the (−30,30) event window following Beck et al. (2010).

We employ Equation (2) to test the parallel trend assumption. Time intervals and are described in Table 13. Dt (t = −2, …, 6) is a set of time dummies signaling the time periods before and after the introduction of bankruptcy courts. D−2 is coded 1 if the bond was issued during (−90, −60) trading days relative to the date of bankruptcy court establishment, and 0 otherwise. For bonds whose issuers located in cities without bankruptcy courts, D−2 is coded 0 all the time.

(2)IssueSpreadi,t=α+β1D−2+β2D−1+β3D1+…+β8D6+δControls+Date+City+Ind+εi,t

Parallel trend test results are shown in Table 14. No matter whether bond- and firm-level characteristic variables are included or not, estimates in column (1) and (2) both show that the coefficients of D−2 and D−1 are not significant while the coefficients of D2, D3, and D6  are positively significant at 5% level. D4 is positively significant when bond- and firm-level variables are not controlled, and D5 is positively significant when those variables are controlled. The results in Table 14 therefore indicates an increase in bond issuance spreads after the introduction of bankruptcy courts. Figure 3 depicts the dynamic impact of bankruptcy courts on bond issuance spreads. Notice that the effect of bankruptcy courts on issuance spreads is close to 0 during the period (−90,−30), and as time progresses the issuance spreads display a significant increase during the period (60,90), a trend that continues thereafter. The results show a lagged impact of bankruptcy courts on bond issuance spreads and an increasing trend of issuance spreads for cities with bankruptcy courts and the trend persists over time.

5.5.2 Alternative samples

Separate regressions using alternative samples are estimated to assess the robustness of our findings. The results are reported in Table 15. First, in column (1) of Table 15, bond issuances in Beijing, Shanghai, Chongqing and Tianjin are excluded since we could hardly define a control group for comparison for these municipalities. The coefficient of Post remains positively significant in column (1). Second, we enlarge our sample to include bond issuances in provinces without bankruptcy courts. The results are similar as reported in column (2) of Table 15.

5.5.3 PSM approach

We employ a propensity score matching procedure to correct for the probable selection bias. Specifically, we pair each treated observation of bond issuances (Post = 1) with a similar control observation (Post = 0), selected from the same industry-year and matched on the firm- and bond-level characteristics included in Equation (1) using the kernel algorithm. The PSM estimators are listed in column 3 of Table 15. The coefficient of Post is positively significant at 1% level, which conforms with the results in the main regressions.

5.5.4 Placebo test

We also run placebo tests as an additional robustness check. We generate the treatment group and control group by random selection and replicate estimation of Equation (1) as if the bankruptcy courts were established in the generated pattern. Bootstrapping is performed up to 500 times to obtain the distribution of the coefficients on Post. As shown in Figure 4, the coefficients of Post are normally distributed, and the mean value is close to 0. The placebo tests find no significant effects of bankruptcy courts on issuance spreads using the randomly selected treatment group. Therefore, the increase of bond issuance spreads is very likely to be linked to the introduction of bankruptcy courts.

6. Conclusion

The number of bond defaults has been increasing with the growing size of China's bond market. Defaults are inevitable in a well-functioning bond market, and bankruptcy serves as an important legal channel for default resolution. However, weak bankruptcy institutions in the past resulted in low quality of law enforcement. The implicit government guarantees led to high barriers for bond exits through bankruptcy, which hampered the development of the bond market. The law and finance literature shows that the differences in the width and depth of capital markets can be partially attributed to disparities in legal rules and the quality of their enforcement. However, the existing studies on law and finance give relatively little attention to the bond market. Exploiting the staggered introduction of bankruptcy courts across cities in China, this paper discusses the effect of improvement in law enforcement on the bond market, providing micro level evidence on law and public debt market.

The establishment of bankruptcy courts enables judges to be insulated from interference by governments in bankruptcy proceedings. Greater judicial independence along with the appointment of professional administrators shortens the processing time of bankruptcy cases and lowers bankruptcy costs. Moreover, higher recovery rate encourages market participants to fund for high-risk firms in expectation of being well compensated when defaults triggered. Our analysis finds increase in bond issuance spreads, increase in the number of bonds issued and a trend of lower issuer credit ratings after bankruptcy courts were established. Further, following the establishment of bankruptcy courts, the bond market is more efficient in that prices better reflects fundamentals including firm size, profitability and downside risk. The increase of bond issuance spreads after the establishment of bankruptcy courts have been more pronounced for publicly traded bonds and bonds issued in regions with lower level of Justice Confidence Index. In addition, the introduction of bankruptcy courts serves to mitigate pricing distortions caused by the implicit government guarantees. Bonds issued by state-owned enterprises and those issued in regions with more intentions or rooms for government interventions experienced a significantly greater increase of issuance spread. Finally, the effects of bankruptcy courts are more prominent in provinces with higher extent of marketization.

Our analysis has significant implications for the construction of rule-based infrastructures and the development of Socialist Market Economy with Chinese Characteristics. Legal rules on books are only one element of the legal system. The enforcement of these rules could be equally, or even more important. Confronted with the rapidly changing social environment and conflicts between different stakeholders, it's almost impossible even for the most elaborate legal rules to cover every aspect of the emerging problems in economic activities. In the absence of such “perfect” laws, a well-functioned enforcement system is needed. The effects of legal rules can be realized only through a strong system of legal enforcement. This is also the underlying premise of legal development in China. A well-functioning legal system puts formal restraints on local governments to facilitate the development of a “free, open, and fair” market. Higher quality of legal institutions for Socialist Market Economy is required to achieve the goal of “allowing the market to play a decisive role in allocating resources and improving the government's role” [12]. The introduction of bankruptcy courts is regarded as an attempt to develop a rule-based mechanism for market exits. By allowing ailing enterprises to default and enter into bankruptcy without subsequent interferences, government could signal its commitment to not to engage in future bailouts. Market participants therefore are motivated to monitor and bear the credit risk of corporate debt. After bankruptcy courts were established, bond issuing spreads more accurately reflect credit risk, signaling higher efficiency in the bond pricing system and inviting more firms to enter the bond market. The findings in this study support the view that higher quality of legal institutions can promote the development of high yield bond market and the growth of China's multi-layered securities market. In conclusion, the rule of law serves as the foundation of the marketization reform and helps disentangle the relationship between government and the market.

Figures

Figure 1

Bond issuance spreads for the 2018–2020 period

Figure 2

The trends of bond issuance characteristics around the establishment of bankruptcy courts

Figure 2

Figure 3

The dynamic impact of bankruptcy courts on bond issuance spreads

Figure 4

The distribution of coefficients in placebo tests

Table 1

Establishment of bankruptcy courts for the 2019–2020 period

No	Bankruptcy court	Date of establishment	Subordination of courts
1	Shenzhen Bankruptcy Court	2019.01.14	Shenzhen Intermediate People's Court
2	Beijing Bankruptcy Court	2019.01.30	Beijing First Intermediate People's Court
3	Shanghai Bankruptcy Court	2019.02.01	Shanghai Third Intermediate People's Court
4	Tianjin Bankruptcy Court	2019.12.19	Tianjin Second Intermediate People's Court
5	Guangzhou Bankruptcy Court	2019.12.20	Guangzhou Intermediate People's Court
6	Wenzhou Bankruptcy Court	2019.12.28	Wenzhou Intermediate People's Court
7	Chongqing Bankruptcy Court	2019.12.31	Chongqing Fifth Intermediate People's Court
8	Hangzhou Bankruptcy Court	2019.12.31	Hangzhou Intermediate People's Court
9	Jinan Bankruptcy Court	2020.04.15	Jinan Intermediate People's Court
10	Qingdao Bankruptcy Court	2020.04.22	Qingdao Intermediate People's Court
11	Nanjing Bankruptcy Court	2020.06.12	Nanjing Intermediate People's Court
12	Xiamen Bankruptcy Court	2020.08.18	Xiamen Intermediate People's Court

Note(s): The table lists the dates when bankruptcy courts were established across cities by the end of 2020

Table 2

Definition of variables

Variable	Definition
IssueSpread	Yield to maturity on the bond issuance date minus the yield to maturity on a treasury bond of similar maturity
Post	A dummy variable that equals 1 on the dates after bankruptcy courts were established and 0 otherwise. For bonds whose issuers are located in cities with no bankruptcy court, Post equals 0 all the time
Bondsize	The natural logarithm of bond offering amount
Bondterm	The natural logarithm of bond duration in years
Bondrate	The ordered value of bond's credit rating (A−1 = 1, AA− = 2, AA = 3, AA+ = 4 and AAA = 5)
Bondput	A dummy variable that equals 1 if there's put provision, and 0 otherwise
Bondcall	A dummy variable that equals 1 if there's call provision, and 0 otherwise
Bondsecu	A dummy variable that equals 1 if the bond is secured, and 0 otherwise
Firmsize	The natural logarithm of the issuer's total assets
Lev	The ratio of total liabilities to total assets of the issuer
Growth	The annual percentage change in the issuer's operating income from year t to year t−1
ROA	The ratio of net income to total assets
Cover	The ratio of net operating cash flow to interest expense
Tangibility	The ratio of tangible assets to total assets
Current	The ratio of current assets to current liabilities
Big4	A dummy variable that equals 1 for Big 4 clients, and 0 otherwise
SOE	A dummy variable that equals 1 if the controlling shareholder is a government agency, and 0 otherwise
Top1	The proportion of shares owned by the largest stockholder of the firm
Issuenum	The number of bonds issued at the issuance date at city level
CorpRate	The ordered value of firm's rating. (BBB = 1, BBB+ = 2, A− = 3, A = 4, A+ = 5, AA− = 6, AA = 7, AA+ = 8, AAA− = 9, AAA = 10)
GDPg	The growth rate of regional GDP

Note(s): This table describes definitions of variables in model 1

Table 3

Descriptive statistics

	N	Mean	Sd	Min	Max
IssueSpread	5,451	2.162	1.312	0.522	5.839
Post	5,451	0.375	0.484	0	1
Bondsize	5,451	20.667	0.681	18.951	22.227
Bondterm	5,451	1.401	0.371	0.693	2.303
Bondrate	5,451	4.544	0.672	1	5
Bondput	5,451	0.281	0.450	0	1
Bondcall	5,451	0.203	0.402	0	1
Bondsecu	5,451	0.103	0.304	0	1
Firmsize	5,451	25.316	1.401	22.480	28.191
Lev	5,451	0.629	0.13	0.246	0.858
Growth	5,451	0.185	0.333	−0.411	2.032
ROA	5,451	0.02	0.018	−0.001	0.089
Cover	5,451	−7.387	105.996	−804.788	386.645
Tangibility	5,451	0.236	0.182	−0.170	0.731
Current	5,451	1.898	1.636	0.227	13.500
Big4	5,451	0.127	0.333	0	1
SOE	5,451	0.874	0.332	0	1
Top1	5,451	82.399	25.462	16.910	100
Issuenum	5,451	2.128	1.599	1.000	10
GDPg	5,451	0.061	0.009	0.033	0.083

Note(s): This table reports descriptive statistics of main variables. Variable definitions are reported in Table 2

Table 4

Univariate correlations

		(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
(1)	IssueSpread	1
(2)	Post	−0.162***	1
(3)	Bondsize	−0.194***	0.238***	1
(4)	Bondterm	−0.179***	−0.184***	−0.086***	1
(5)	Bondrate	−0.352***	0.297***	0.461***	−0.064***	1
(6)	Bondput	−0.096***	−0.061***	−0.078***	0.346***	−0.133***	1
(7)	Bondcall	0.475***	0.200***	0.161***	−0.306***	0.221***	−0.202***	1
(8)	Bondsecu	0.113***	−0.112***	−0.205***	0.204***	0.083***	−0.007	−0.070***	1
(9)	Firmsize	−0.220***	0.359***	0.579***	−0.127***	0.583***	−0.126***	0.261***	−0.369***	1
(10)	Lev	0.129***	0.200***	0.174***	−0.210***	0.215***	−0.056***	0.301***	−0.129***	0.449***	1
(11)	Growth	0.022	−0.049***	−0.048***	0.006	−0.048***	0.095***	−0.024*	0	−0.045***	0.013
(12)	ROA	−0.073***	0.074***	0.032**	−0.133***	0.128***	0.081***	−0.036***	−0.033**	−0.024*	−0.149***
(13)	Cover	−0.031**	0.040***	0.049***	−0.070***	0.075***	−0.01	0.021	−0.055***	0.089***	0.014
(14)	Tangibility	−0.045***	−0.243***	−0.217***	0.258***	−0.346***	0.105***	−0.304***	0.195***	−0.506***	−0.755***
(15)	Current	0.119***	−0.268***	−0.236***	0.298***	−0.400***	0.067***	−0.207***	0.227***	−0.432***	−0.424***
(16)	Big4	−0.063***	0.173***	0.214***	−0.083***	0.220***	0.005	0.097***	−0.091***	0.234***	0.174***
(17)	SOE	−0.225***	−0.001	0.088***	0.116***	0.048***	−0.271***	0.132***	−0.003	0.022	−0.073***
(18)	Top1	−0.070***	−0.043***	−0.011	0.169***	−0.092***	−0.072***	−0.039***	0.004	−0.003	−0.133***
(19)	Issuenum	−0.096***	0.378***	0.197***	0.051***	0.264***	−0.095***	0.185***	−0.043***	0.354***	0.145***
(20)	GDPg	0.117***	−0.042***	0.062***	−0.016	−0.013	−0.082***	0.157***	−0.01	0.024*	0.076***

		(11)	(12)	(13)	(14)	(15)	(16)	(17)	(18)	(19)	(20)
(11)	Growth	1
(12)	ROA	0.083***	1
(13)	Cover	−0.028**	0.102***	1
(14)	Tangibility	0.011	−0.016	−0.093***	1
(15)	Current	−0.002	−0.188***	−0.179***	0.551***	1
(16)	Big4	−0.044***	0.191***	0.014	−0.128***	−0.204***	1
(17)	SOE	−0.085***	−0.399***	−0.070***	0.076***	0.130***	−0.158***	1
(18)	Top1	0.017	−0.312***	−0.037***	0.135***	0.210***	−0.314***	0.350***	1
(19)	Issuenum	−0.082***	−0.034**	0.005	−0.196***	−0.173***	0.101***	0.102***	0.023*	1
(20)	GDPg	−0.069***	0.011	−0.023*	−0.131***	−0.032**	0.016	0.077***	0.001	0.246***	1

Note(s): This table reports correlation matrix for main variables. The sample consists of 5,451 bonds issued between 2018 and 2020. Variable definitions are reported in Table 2. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 5

Univariate test of bond issuance spread

IssueSpread	Before		After		MeanDiff
	N	Mean	N	Mean
Group1	1,613	2.136	2047	1.889	0.247***
Group2	1,228	2.733	563	1.987	0.746***
MeanDiff		−0.597***		−0.098*	−0.499***

Note(s): The table presents univariate analysis of bond issuance spreads (IssueSpread). Bonds are classified into Group1(Group2) if their issuers are located in cities with (without) bankruptcy courts. Mean values of IssueSpread are reported for periods before and after the introduction of the bankruptcy courts and for the two groups. We conduct t-test for differences in mean. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 6

Bankruptcy courts and bond issuance spreads

Dependent: IssueSpread	(1)	(2)	(3)
Post	0.320*** (4.943)	0.132*** (3.021)	0.084** (2.047)
Bondsize		−0.032 (−1.400)	0.053** (2.533)
Bondterm		−0.237*** (−6.034)	−0.046 (−1.250)
Bondrate		−0.774*** (−21.992)	−0.698*** (−18.357)
Bondput		−0.154*** (−4.263)	−0.352*** (−10.843)
Bondcall		2.041*** (50.998)	2.048*** (51.869)
Bondsecu		0.632*** (12.281)	0.430*** (7.393)
Firmsize			−0.103*** (−6.136)
Lev			0.620*** (3.935)
Growth			−0.018 (−0.515)
ROA			−5.836*** (−7.307)
Cover			−0.0001 (−1.004)
Tangibility			−0.410*** (−3.834)
Current			0.033*** (3.414)
Big4			−0.050 (−1.279)
SOE			−1.363*** (−22.865)
Top1			−0.001 (−1.387)
Issuenum			0.017* (1.669)
GDPg			12.512*** (3.730)
_cons	2.042*** (71.785)	6.176*** (14.144)	6.745*** (13.323)
Date	Yes	Yes	Yes
City	Yes	Yes	Yes
Industry	Yes	Yes	Yes
N	5,451	5,451	5,451
r2_a	0.261	0.659	0.729

Note(s): The table shows regressions of bond issuance spreads on the introduction of bankruptcy courts. The dependent variable is the bond issuance spreads (IssueSpread). In column (1), only date, city and industry dummies are included. In column (2), bond-level characteristics are included. In column (3), firm-level controls and regional characteristics are included. Variable definitions are reported in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 7

Univariate analysis on bond issuance and issuer's rating

	Before		After
	N1	Mean1	N2	Mean2	MeanDiff
Panel A: Number of bonds issued (Issuenum)
Group1	1,613	2.003	2047	2.906	−0.903***
Group2	1,228	1.352	563	1.352	0.000
		0.651***	2610	1.554***	−0.903***
Panel B: Issuer credit ratings (CorpRate)
Group1	1,613	9.234	2,047	9.510	−0.276***
Group2	1,228	7.981	563	8.393	−0.411***
		1.252***	2610	1.117***	0.135**

Note(s): The table presents univariate analysis of the number of bonds issued (Issuenum)and credit ratings for issuers (CorpRate). Bonds are classified into Group1(Group2) if their issuers are located in cities with (without) bankruptcy courts. We apply t-test for differences in means. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 8

Bankruptcy courts and bond market entry

	(1)	(2)
Dependents	Issuenum	CorpRate
Post	0.131** (2.309)	−0.072*** (−2.850)
IssueSpread	0.036* (1.670)	−0.051*** (−3.167)
Bondsize	−0.111*** (−3.635)	0.037*** (3.055)
Bondterm	0.608*** (11.317)	−0.017 (−0.784)
Bondrate	0.018 (0.447)	1.189*** (23.278)
Bondput	−0.087** (−2.134)	−0.037* (−1.949)
Bondcall	0.031 (0.449)	0.144*** (3.605)
Bondsecu	0.203*** (3.260)	−1.601*** (−27.059)
Firmsize	0.138*** (6.346)	0.150*** (9.991)
Lev	−0.222 (−0.974)	−0.445*** (−4.566)
Growth	0.070 (1.476)	0.051** (2.268)
ROA	−0.837 (−0.726)	2.716*** (5.135)
Cover	−0.000 (−1.462)	0.00002 (0.372)
Tangibility	0.405** (2.467)	−0.202*** (−2.869)
Current	0.014 (0.983)	−0.016*** (−2.599)
Big4	−0.059 (−1.044)	0.094*** (4.433)
SOE	−0.077 (−1.079)	0.076** (2.081)
Top1	−0.00003 (−0.042)	0.001*** (3.587)
Issuenum		0.001 (0.172)
GDPg	11.516*** (2.770)	−2.211 (−1.139)
_cons	−0.728 (−1.032)	−0.440 (−1.536)
Date	Yes	Yes
City	Yes	Yes
Industry	Yes	Yes
N	5,451	5,451
r2_a	0.618	0.897

Note(s): The table shows regressions of the number of bonds issued and issuer credit ratings on the introduction of the bankruptcy courts. In column (1), the dependent variable is the city-level measure: the number of bonds issued (Issuenum). In column (2), the dependent variable is the issuers' credit ratings (CorpRate). In both regressions, bond-level and firm-level controls and regional characteristics are included. Variable definitions are reported in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 9

Bankruptcy courts and bond pricing

Dependent: IssueSpread	(1)	(2)	(3)
Post	0.170*** (3.344)	0.168*** (3.475)	0.083* (1.923)
HFirmsize	−0.079*** (−2.031)
HFirmsize×Post	−0.161*** (−3.293)
HROA		−0.024 (−0.765)
HROA×Post		−0.192*** (−4.269)
HRER			−0.008 (−0.227)
HRER×Post			0.081* (1.763)
Controls	Yes	Yes	Yes
Date	Yes	Yes	Yes
City	Yes	Yes	Yes
Industry	Yes	Yes	Yes
N	5,451	5,451	5,050
r2_a	0.728	0.728	0.733

Note(s): The table shows regressions of bond issuance spreads on firm level risk factors. In columns (1)–(3), the dependent variable is bond issuance spreads (IssueSpread). HFirmsize is a dummy variable that equals 1 for firms with firm size above the industry-year median, and 0 otherwise. HROA is a dummy variable that equals 1 for firms with ROA above the industry-year median and 0 otherwise. HRER is a dummy variable that equals 1 for firms whose downside risk is greater than the industry-year median, and 0 otherwise. Control variables are the same as in Table 6. Other variables are defined as in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represents statistical significance at the 10, 5 and 1% level, respectively

Table 10

Additional analysis of bankruptcy courts and investor protection

Dependent: IssueSpread	(1)	(2)
Post	−0.033 (−0.629)	0.339** (2.542)
PInvestor	−0.058** (−2.062)
PInvestor×Post	0.198*** (4.315)
HJudicial_Confidence×Post		−0.262** (−2.048)
Controls	Yes	Yes
Date	Yes	Yes
City	Yes	Yes
Industry	Yes	Yes
N	5,451	5,451
r2_a	0.730	0.729

Note(s): The table shows regressions of bond issuance spreads on the introduction of bankruptcy courts and the effect of creditor type and justice confidence. In columns (1) and (2), the dependent variable is bond issuance spreads (IssueSpread). PInvestor is a dummy variable that equals 1 for publicly tradable bonds and 0 for bonds only available to qualified investors or qualified institutional investors. HJudicial_Confidece is a dummy variable that equals 1 for bond issuances in provinces with Judicial Confidence Index above median and 0 otherwise. Control variables are the same as in Table 6. Other variables are defined as in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 11

Additional analysis of the effects of implicit government guarantees

Dependent: IssueSpread	(1)	(2)	(3)
Post	−0.172 (−1.624)	−0.018 (−0.315)	−0.039 (−0.691)
SOE	−1.495*** (−21.559)
SOE×Post	0.295*** (2.876)
Unemployment		9.370 (0.836)
Unemployment×Post		8.611** (2.524)
GDP			−0.306 (−1.343)
GDP×Post			0.284* (1.667)
Controls	Yes	Yes	Yes
Date	Yes	Yes	Yes
City	Yes	Yes	Yes
Industry	Yes	Yes	Yes
N	5,451	4,925	4,976
r2_a	0.730	0.725	0.724

Note(s): The table shows regressions of bond issuance spreads on the introduction of bankruptcy courts and its interactions with the implicit government guarantees. The dependent variable is bond issuance spreads (IssueSpread). SOE is a dummy variable that equals 1 if the controlling shareholder is a government agency, and 0 otherwise. Unemployment is the unemployment rate at city level. GDP refers to GDP per capita at city level. Control variables are the same as in Table 6. Other variables are defined as in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 12

Additional analysis of the effects of regional financial development

Dependent: IssueSpread	(1)	(2)
Post	−0.038 (−0.660)	−0.015 (−0.265)
HFin_Mar_index×Post	0.162*** (2.704)
HFin_Credit_index×Post		0.133** (2.221)
Controls	Yes	Yes
Date	Yes	Yes
City	Yes	Yes
Industry	Yes	Yes
N	5,451	5,451
r2_a	0.729	0.729

Note(s): The table reports regressions of bond issuance spreads on the introduction of bankruptcy courts, considering the effect of the extent of financial market development. The dependent variable is bond issuance spreads (IssueSpread). HFin_Mar_index is a dummy variable that equals 1 for firms in provinces with factor market development index above the provincial median and 0 otherwise. HFin_Credit_index is a dummy variable that equals 1 for firms in provinces with financial resources allocation index above the provincial median and 0 otherwise. Control variables are the same as in Table 6. Other variables are defined as in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 13

Time intervals for parallel trend testing

	t−2	t−1	t1	t2	t3	t4	t5	t6
Variables	D−2	D−1	D1	D2	D3	D4	D5	D6
Day	(−90, −60)	(−60, −30)	(30,60)	(60,90)	(90,120)	(120,150)	(150,180)	>180

Note(s): The table presents the time dummies and their corresponding time periods for parallel trend testing

Table 14

The dynamic effect of bankruptcy courts on bond issuance spreads

Dependent: IssueSpread	(1)	(2)
D−2	0.179 (1.184)	−0.037 (−0.341)
D−1	0.018 (0.134)	−0.015 (−0.215)
D1	0.108 (0.652)	0.012 (0.101)
D2	0.337*** (2.792)	0.193** (2.363)
D3	0.395*** (2.715)	0.194** (2.129)
D4	0.283* (1.896)	0.039 (0.469)
D5	0.163 (1.030)	0.165* (1.746)
D6	0.487*** (5.447)	0.128** (2.298)
Controls	No	Yes
Date	Yes	Yes
City	Yes	Yes
Industry	Yes	Yes
N	5,126	5,126
r2_a	0.265	0.729

Note(s): The table reports estimated coefficients from Equation (2). The dependent variable is bond issuance spreads (IssueSpread). D^t is a set of time dummies representing different time periods before and after the establishment of bankruptcy courts (t = −2, −1, 1, 2, 3, 4, 5, 6). The corresponding time periods are listed in Table 13. In column (1), only date, city and industry fixed effects are included. In column (2), bond-level and firm-level controls are included. Controls are the same as in Table 6. Other variables are defined as in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Table 15

Robustness checks using alternative samples and the PSM approach

Dependent: IssueSpread	(1)	(2)	(3)
Post	0.158** (2.553)	0.086** (2.511)	1.059*** (3.144)
Controls	Yes	Yes	Yes
Date	Yes	Yes	Yes
City	Yes	Yes	Yes
Industry	Yes	Yes	Yes
N	3,215	8,603	1,743
r2_a	0.743	0.736	0.693

Note(s): The table shows regressions of bond issuance spreads on the introduction of bankruptcy courts for alternative samples. The dependent variable is bond issuance spreads (IssueSpread). In column (1), bond issuances in Beijing, Shanghai, Chongqing and Tianjin are excluded. In column (2), bond issuances in provinces without bankruptcy courts are included. In column (3), the PSM approach is used to pair the treatment group with control group and the matched sample is employed. Control variables are the same as in Table 6. Variable definitions are given in Table 2. Robust t-statistics are reported in parentheses. *, ** and *** represent statistical significance at the 10, 5 and 1% level, respectively

Notes

1. According to China's 2006 Bankruptcy Law, a debtor “may file an application with the people's court for reorganization, reconciliation, or bankrupt liquidation” when it “fails to pay off its due debts” or “the relevant assets are not enough to pay off the debts”.

2. See “Operation of the Financial Market in November, 2021” by the People's Bank of China, Dec 21, 2021, available at http://www.pbc.gov.cn/jinrongshichangsi/147160/147171/147173/4424924/index.html

3. There are two possible reasons for the decreasing trend of bond issuance spread: first, greater tolerance for defaults allowed more financially distressed firm to exit and the financial performance of issuers improved on average, which resulted in lower average cost of debt for the bond market. Second, more “structured bonds” were issued under the “deleveraging” policy environments, in which issuers attempted to lower its financing cost by purchasing subordinated class of the bonds sold.

4. See “The bankruptcy law enforcement over a 10-year period in Zhuangji” by Zhang Heng, July 1, 2017, The Economic Observer, available at http://www.eeo.com.cn/2017/0701/307615.shtml

5. See “The first anniversary of the Guangzhou Bankruptcy Court: Improved efficiency in bankruptcy procedures in support of the Greater Bay Area”, December 12, 2020, People's Court Daily, available at http://rmfyb.chinacourt.org/paper/html/2020-12/19/content_174829.htm?div=-1

6. See “The role of the Shenzhen Bankrupt Court in promoting the development of the Greater Bay Area and the Pilot Demonstration Zone for Socialism with Chinese Characteristics”, Shenzhen Intermediate People's Court, January 17, 2020, available at https://www.thepaper.cn/newsDetail_forward_5551354

7. See “What we have done by the time of the first anniversary of the introduction of Wenzhou Bankruptcy Court”, Wenzhou Intermediate People's Court, December 28, 2020, available at https://mp.weixin.qq.com/s?__biz=MzU5MTgwMDE2NA==&mid=2247500788&idx=1&sn=67ace4edfb55ff36ef525e32f16dc6d4&chksm=fe2bf1ffc95c78e973d2b4ff73e1bfc02e4ab1e1004b9854b6d68b0fa1f1bf50e944bf1ca562#rd

8. See “Ten issues to be mentioned by the time of the second anniversary of the introduction of Shenzhen Bankruptcy Court”, Shenzhen Intermediate People's Court, January 14, 2021, available at https://mp.weixin.qq.com/s/lpXe6S4AFD7HyLSfWwDWDQ

9. See “Typical cases of the Shanghai Bankruptcy Court during 2020”, Higher People's Court of Shanghai Municipality, February 3, 2021, available at https://www.pkulaw.com/lar/4a46ad9550da6835e2b7a4e5ac03c0d2bdfb.html

10. See “The annual report of typical cases of the Shanghai Bankruptcy Court (Cases of Reorganization)”, Shanghai Bankruptcy Court, January 30, 2021, available at https://mp.weixin.qq.com/s/gj0E5jygsU5Uu9f_6uPEVQ

11. For cities with bankruptcy courts established, we count the month of bankruptcy court's establishment as 0 (t_month = 0). As for cities without bankruptcy courts, we count the month of first bankruptcy court's introduction at province level as 0 (t_month = 0).

12. See “Decision of the Central Committee of the Communist Party of China on Some Major Issues Concerning Comprehensively Deepening the Reform” adopted at the Third Plenary Session of the 18th Central Committee of the Communist Party of China on November 12, 2013, available at http://www.gov.cn/jrzg/2013-11/15/content_2528179.htm