Weak enforcement was always regarded as a major factor affecting the effectiveness of environmental policy efforts in China. One part of this problem is that most works are carried out by local governments under China’s decentralised system of environmental governance. However, local governments are often more concerned with the economic growth rather than environmental protection. Therefore, there is a strong incentive for the local governments who are the environmental policy executors to ignore the environmental violations for the sake of getting more fiscal revenue. Although the new Environmental Protection Law is called “the strictest environmental law” in recent Chinese history after it was issued in 2015, its validity is still being questioned to some extent. The working arrangement of Central Environmental Inspection was set up by China’s central government for the purpose of monitoring the local implementation of environmental laws and policies from 2016 which is an

The Stock Market’s Reaction to Strict Environmental Inspection: Evidence from Heavily Polluting Listed important institutional innovation in the field of ecological civilisation system reform. In the last two years, the first round of inspection saw investigations conducted in 31 provinces and that had successfully solved more than 80,000 environmental problems. Specifically, the first round of inspection received more than 135,000 complaints from the public in which it handled 29,000 cases and issued fines totaling 1.43 billion Yuan. Approximately, 18,448 officials were investigated and 18,199 were found responsible for some cases. In the name of the CPC Central Committee and the State Council, central inspection was endowed with the higher authority and rigidity, emphasising the same responsibilities of CPC committees and governments at all levels in environmental protection. Through the implementation of inspection, the targets of environmental supervision are shifted from sole enterprises to both Local party committees and governments, making environmental supervision more deterrent than before. In the process of inspection, environmental supervision mainly targets provincial party committees, governments and their relevant departments and can directly sink into municipal party committees and governments or conduct in-depth investigation on some related enterprises if necessary. As shown in table 1, the specific situation of the first round of Central Environmental Inspection was sorted out. On January 4, 2016, the first inspection group was stationed in Hebei Province to carry out the pilot work of supervision. After the pilot project, the Central Environmental Inspection carried out four batches of supervision work in 2016 and 2017, covering the remaining 30 provinces. According to the work plan of environmental inspection, each batch of inspection lasts for about one month and involves seven or eight provinces in general. The whole inspection process was divided into three stages, each of which lasts for about 10 days. The first stage was to talk with the leaders of provincial CPC committees, governments and related departments to consult them on the relevant materials on local environmental protection work, to visit relevant departments and at the same time to receive the public complaints. The second stage was to conduct an investigation and verification aiming at the problems and clues sorted out during the first period. The purpose of this stage was to assess the seriousness of the problems and implement the allocation of responsibilities. The third stage was to summarize basic conclusions of inspection, present reporting framework and carry out targeted supplementary supervision. And then, about three months later, each environmental inspection group would provide feedback on the inspection results and corrective suggestions on supervised provinces which mainly include problems discovered and penalty opinions.

To implement an event study, we first needed to identify a clear event date. According to the supervision process, each batch of the Central Environmental Inspection involved two important dates; one was the date when the supervision groups were stationed, the other was the date when the supervision group provided feedback on specific opinions to the local goverment and also to the public synchronously. The latter reflects the inspected problems and corresponding processing results, such as the officials held accountable and the amount of the fine, which may have a direct deterrent effect. Before the first batch of Central Environmental Inspection, a pilot project was carried out in Hebei province. But from the perspective of media exposure, public attention is relatively lower. In addition, as the central environmental inspection was a newly proposed mechanism, the specific impact on economy and enterprises may still be unknown to investors. After the pilot study period, investors may develop a certain understanding of the central environmental inspection. Therefore, the first large-scale central environmental inspection may have a more significant impact on the stock market. For the second one, during the period of the third and fourth batch of central environmental inspection,

The Stock Market’s Reaction to Strict Environmental Inspection: Evidence from Heavily Polluting Listed investors may already have obvious expectations after experiencing the first inspection. And also, the days of the inspection teams’ arrival are close to the feedback days of the previous batch which make it difficult to identify the effective impact of the event. Because the first batch of central environmental inspection involved seven provinces, the dates of entry of each inspection group were different, though the dates were close. The date range of stationing was from July 12 to July 19, 2016 and the date range of feedback was from November 14 to November 23, 2016. The keyword search trend and information exposure were provided by the 360 search index (Fig.1) The 360 search is a search engine which is commonly used in China as it can draw the public attention to the central environmental supervision events. In terms of trend index, the focus of the keywords of central environmental inspection and environmental inspection increased sharply on July 14, 2016 and July 15, 2016, respectively. In terms of exposure, the two keywords rose sharply on July 15. In the following month, there was still a continuous concern trend of central environmental supervision. But in September and October, after the end of the first batch of central environmental inspection, the public concern on the incident declined dramatically. Until November 14, the inspection groups started to send feedback opinions to the inspected provinces and the public attention increased sharply again. Thus, we used both the supervision groups station (July 12) and the supervision group feedback (November 14) in the first batch of central environmental supervision as the event dates.

Event study requires the occurrence of events to be unpredictable. However, sometimes it is difficult to ensure that information is not disclosed in advance. And also, the station and feedback process of the inspection was gradual which would cause the attention of media and the public be lagged behind. From figure 1, it can be seen that the attention on environmental

EventEvent study requires the occurrence of events to be unpredictable. However, sometim 3.1.2 window difficult to ensure that information is not disclosed in advance. And also, the statio

Event studyprocess requiresofthetheoccurrence events to be which unpredictable. However, feedback inspectionofwas gradual would cause the attention of med sometimes it is difficult to ensure that information is not disclosed in advance. the also, public lagged figure 1, inspection it can be seen that thewhich attention on environ And thebe station andbehind. feedbackFrom process of the was gradual would cause the of media and theTherefore, public be lagged behind. From figure relevant researc inspection wasattention gradually increasing. according to the existing 1, it can be seen that the attention on environmental inspection was gradually time windows [-5,5], [-5,10], [-10,10], [-10,20] are set to test the reaction of the stock m increasing. Therefore, according to the existing relevant research, four time windows [-5,5], [-5,10], [-10,10], [-10,20] are set to test the reaction of the stock market. 3.1.3 Estimation window and calculation of CAR

The data of listed companies used in this paper are from the China Stock Market and Accounting Research Database (CSMAR). According to the Guidelines for Environmental Information Disclosure of Listed Companies issued by the Ministry of Environmental Protection, 19 industries are defined as heavily polluting industries. At last, corresponding to launch event and feedback event of Central Environmental Inspection, there are 790 and 814 listed companies in our research, which cover all polluting corporations listed on China’s stock exchanges. Table 2 is the descriptive statistics of relevant variables.

The Stock Market’s Reaction to Strict Environmental Inspection: Evidence from Heavily Polluting Listed

To examine the market responses to the event, we first estimated a market model over a 120 estimation window ending 11days before the event date. A valueweighted average return of all stocks in our sample is adopted as the market return. We then calculated cumulative abnormal returns (CARs) over an 11 day (-5, 5) event window centred on the start date and feedback date of the inspection and tested their statistical significance respectively. To establish the robustness of our results, we also considered two longer event windows, a 21-day window and a 31-day window. As shown in Table 3, the average of the CARs is negative and significantly different from zero. And compared with the start-up day event, we can see a greater negative response of the stock market of the feedback event. According to the CARs in the event window of 11 days, the central environmental supervision results in a 0.35 and 0.47 percentage point decline of the market returns during the start-up and feedback event window respectively, but the results are not so significant. In the event window of 21 days, we can see a 1.95 and a 2.87 percentage point significant decline of market returns after the start date and feedback date of the supervision. The results show that with the increasing attention and exposure of the central environmental supervision, the impact on the market value of heavily polluted enterprises become more significant and persistent. Using a natural experiment generated by the National Specially Monitored Firms (NSMF) in China, Zhang et al. (2018) found that central supervision significantly reduced the emission intensity of industrial enterprises. These results highlight the substantial room for improvement in Chinese environmental regulations via central supervision.

Note: ***, ** and * indicate significant levels of 1%, 5% and 10% respectively.

In Figure 2, taking 21 days as the event window, we reported the change trend of the CAR of heavily polluting listed companies. The results show that CARs

CAR（-10, 20）

Note: ***, ** and * indicate significant levels of 1%, 5% and 10% respectively.

In Figure 2, taking 21 days as the event window, we reported the change trend of the CAR of listed companies in heavy polluting industries dropped dramatically and the of heavily polluting listed companies. The results show that CARs of listed companies in heavy shock of the start-up day event was more obvious. However, the shocking effect industries dropped dramatically and the shock of the start-up day event was more ofpolluting the feedback event became more and more significant only on the fifth day obvious. However, the shocking effect of the feedback event became more and more significant after the launch of the Central Environmental Inspection. One possible reason is only the fifth day the launch of the Central Inspection. One possible that theonfeedback is aafter sequential process. With Environmental the mass reporting of the feedback reason is that the feedback is a sequential process. With the mass reporting of the feedback results in various news media, public attention increased quickly which induced results in various news media, public attention increased quickly which induced investment investment behaviors to react quickly and lead to the decrease of returns of behaviors to react quickly and leadpolluting to the decrease of returnsIn of the Listed Companies heavyly Listed Companies in heavyly industries. whole eventinwindow polluting the industries. In thefeedback whole event the start and feedback made of central (-10,10), start and of window central (-10,10), environmental inspection the environmental inspection the listed companies in the heavyly industries - 2% listed companies in themade heavyly polluting industries getpolluting - 2% and - 3%get CARs respectively. and - 3% CARs respectively.

Note: The event day on the left is the start date. The event day on the right is the feedback day.

Note: The event day on the left is the start date. The event day on the right is the feedback day.

The key of the event study is to identify the causal effect of central environmental inspection on the stock market. It is necessary to judge whether the change trend of CARs after event day was really caused by the central environmental inspection. Firstly, we tested whether the CARs day by day before the event day was significant. If the CARs was only significant after the event day, it could be proved that the result above was valid. Secondly, we examined the result again excluding the listed companies that were likely to be affected by other related events in the event window. Specifically, we deleted the sample with more than five days trading interval after the event day and the sample with dividend payment, changes in equity and major trading events happened during the event window.

The Stock Market’s Reaction to Strict Environmental Inspection: Evidence from Heavily Polluting Listed

In Table 4, we presented the estimated timing results with the sample deleting the listed companies that may be affected by other important events during the event window. Firstly, the shocked effects of start-up event and feedback event are still significantly negative, though the effect has become smaller. As shown in Table 4, the CARs are – 2.3% and - 2.5% respectively in the event window (-10,10). We can also find that the CARs are all insignificant within five days before the event date. Thus, our estimated results are not disturbed obviously by other events, such as information disclosed in advance. The result in Table 4 means that the central environmental inspection brings a significant negative market response.

Note: ***, ** and * indicate significant levels of 1%, 5% and 10% respectively.

We calculated the average CARs of different industries over the 21 days event window. In Table 5, most of the industries made significant negative reactions to the first batch of central environmental inspection. And more bigger shock effects were observed in coal mining, ferrous metal mining, chemical industry and rubber industry.

Note: ***, ** and * indicate significant levels of 1%, 5% and 10% respectively.

The Stock Market’s Reaction to Strict Environmental Inspection: Evidence from Heavily Polluting Listed

In order to investigate the heterogeneous impact of central environmental inspection on the market value of heavily polluted listed enterprises, we selected enterprise ownership, political connection and enterprise scale totally three comparable variables and tested the different shock effects. 5.1

Table 6 shows that t the central environmental inspection has a significant negative impact on the market value of private companies in different event windows. However, the start-up and the feedback event of the supervision both have no significant impact on state-owned enterprises in the 11 days and 21 days event window. Only in the 31 days event window, the negative impact starts to be obvious. Through comparing the results between state-owned and private companies, the negative impact of central environmental inspection on private companies is significantly higher than that of state-owned companies, with significant difference of 2.81 percent. On the one hand, the environmental performance of state-owned listed companies, such as environmental information disclosure, is relatively better than private companies. The study by Cheng et al. (2017) showed that the corporate political connection can influence companies to more actively disclose environmental information but it can also mask political rent-seeking in the guise of protecting the environment. On the other hand, the ability for the state-owned companies to withstand policy risks is higher than that of the private companies which will bring different anticipation to the investors.

Note: Double-tailed test and t-value in parentheses are shown in the table. ***, ** and * indicate significant levels of 1 percent, 5 percent and 10 percent respectively under t test.

Table 7 shows that environmental supervision has a significant negative impact on companies with and without political connection. That means that political affiliation is no longer the main factor affecting the environmental behaviours of polluting companies under the deterrence of central environmental inspection. Even though the impact of central environmental inspection on political affiliated companies is relatively higher, the difference is not so obvious between the companies with and that without the political connections. The results are to a great extent related to the ousting of some local officials brought by the Central Environmental Inspection. Just as the study conducted by Wang et al. (2018), in general, corporate investment expenditures of listed firms decline significantly after the ouster of the politicians, especially for non-SOEs relative to SOEs.

Note: Double-tailed test and t-value in parentheses are shown in the table. ***, ** and * indicate significant levels of 1%, 5% and 10% respectively under t test.

According to the total assets, companies are divided into large-scale and small-scale groups. Specifically, companies below the median are identified as small-scale companies and those above the median are identified as large-scale companies. By comparison, the impact of central environmental inspection on the market value of small-scale companies is significantly greater than that of large-scale companies. As shown in table 8, the impact of the startup and the feedback event on the market value of large-scale companies is 6 and 5 percentage points smaller than that of the small-scale companies in the event window of 31 days. Many studies have proved that environmental regulation brought a more adverse effect on small-scale enterprises, such as large firms can successfully lobby government and are less likely to exit a heavily regulated industry than small firms.

The Stock Market’s Reaction to Strict Environmental Inspection: Evidence from Heavily Polluting Listed

Large-scale Difference

Feedback Event

Note: Double-tailed test and t-value in parentheses are shown in the table. ***, ** and * indicate significant levels of 1 percent, 5 percent and 10 percent respectively under t test.

According to target provinces of the first batch of central environmental inspection, we divided the research sample into inspected and non-inspected provinces. The result shows that although the start and the feedback event of the central environmental inspection both exert a significant negative impact on the companies in inspected and non-inspected provinces, there is no significant difference between the two categories which illustrates that the supervision has a diffusive impact on other provinces (Table 9).

Note: Double-tailed test and t-value in parentheses are shown in the table. ***, ** and * indicate significant levels of 1%, 5% and 10% respectively under t test.