The impact of Corporate Environmental Performance disclosure on Corporate Financial Performance: Evidence from China

The impact of Corporate Environmental Performance

disclosure on Corporate Financial Performance:

Evidence from China

Master thesis Finance

Author: E.C.A. Steenge Student number: S2759896 Focus area: Sustainable Society

Date: June 12, 2019 Supervisor: Dr. N. Selmane

Word count: 11.509

Abstract

This research examines the relationship between Corporate Environmental Responsibility (CER) and Corporate Financial Performance (CFP) using the CSR disclosure mandate enacted in China in 2008. Specifically, the study examines the change in financial performance among mandatory CER disclosing firms with the change among voluntary CER disclosing firms, using a sample of 250 Chinese companies between 2009 and 2017. The study does find a statistically negative relationship between CER disclosing firms on CFP. When the sample is distinguished between mandatory and voluntary disclosing firms, the same negative relationship holds, but the relationship is only statistically significant for voluntary disclosing firms.

Keywords: Corporate environmental performance, corporate environmental responsibility,

1. Introduction

In recent years, the topic of corporate social responsibility (CSR) has been heavily debated by academics. Milton Friedman (1970) proposed his shareholder view where he argued: "The social responsibility of business is to increase its profits." According to Friedman's shareholder theory, a firm should only act in the interest of its shareholders. Nowadays, this view is minimal, and people are more concerned about the environment and the contribution of society and companies. There has been a growing global focus on economic and environmental sustainability. The growing global focus has triggered a trend toward increasing the pressure for firms to disclose their corporate social responsibility (CSR) activities. CSR activities encompass corporate social and environmental behavior that goes beyond the legal or regulatory requirement of the relevant market and economy (Kitzmueller and Shimshack, 2012). CSR suggests that companies should do more than they are obligated under applicable laws. It also suggests that companies should consider not only the interests of shareholders but also those of stakeholders, such as employees, suppliers, consumers, and local communities.

In this study, the impact of mandatory CSR disclosure on firm’s financial performance will be examined. More specifically, the impact of the CSR disclosure mandate enacted in China in 2008 will be examined. The CSR disclosure mandate enacted in 2008 is the first regulation in China where the Shanghai Stock Exchange and the Shenzhen Stock Exchange mandated a subset of listed firms to issue CSR reports along with their annual reports starting from fiscal year 2008.

To assess the impact of the CSR disclosure mandate, this study focuses on environmental performance. As one of the three pillars of CSR, the emphasis on environmental, social, and governance is shifting more towards environmental focus due to an increase in environmental awareness. At the same time, environmental responsibility is one of the key targets of the CSR disclosure mandate. The economic growth of China over the last decades has been incredibly rapid – China has become the world's second largest economy and has become a critical region for companies from around the world to manufacture in, source from and sell to. However, the rapid economic expansion has brought several social and environmental pollution problems, such as considerable energy consumption, and water and air pollution, which in return, delivered a significant contribution to climate change. Since 2007 China has been the largest greenhouse gas emitter, and responsible for 27% of the global emissions in 20171_{. However, in} recent years, things are changing. Currently, China is leading with the highest engagement in green practices. The Chinese government has undertaken several corporate environmental responsibility (CER) activities to make its economy "greener" by lowering the acceptable level of carbonate emissions and reducing pollution. The rationality behind the initiatives reflects the view that CER can contribute to "building a harmonious society" which is a key goal outlined by the Chinese Communist Party at the 2006 National People's Congress (Lin, 2010). At the same time, China needs to find a balance between development and growth on the one hand and the protection of the environment on the other hand. Opportunities have to be investigated to improve the environment, and at the same time, it should be considered to provide financial incentives to the industry by offering "green practices." The opportunities might have a positive impact on industries' environmental performance and the industries' financial position in the long term. Part of the challenge of the government lies in increasing awareness among companies and investors about the importance of timely environmental action by disclosing CER activities besides their annual reports. Nevertheless, despite the considerable attention, the answer to the question remains contradictory: does it pay to be environmentally responsible?

By taking advantage of the setting, the study aims to provide insights into the differences between voluntary and mandatory CSR disclosure and the effect of the CSR disclosure on environmental performance. Therefore, the study states the following research question: "Do Chinese companies that disclose their environmental performance voluntarily differ in their financial performance, compared to companies that disclose their environmental performance mandatory?"

Using a difference-in-difference (DiD) approach, the change in financial performance among mandatory CSR disclosing firms (also called treatment firms) with the change among voluntary CSR disclosing firms (also called controlling firms) will be compared. The study focuses on A-share listed firms on the Shanghai Stock Exchange (SSE) and the Shenzhen Stock Exchange (SZSE) between 2009 and 2017. The study starts in 2009 as this is the first year that the CSR disclosure mandate enacted in China is put into effect. By starting in 2009, the (direct) change of mandatory disclosing firms with the change among voluntary disclosing firms on financial performance can be measured.

The remainder of the paper is organized as follows. Section 2 provides an overview of the relevant literature in this field and construct the hypothesis that will be examined. Section 3 describes the research design, the regressions, and the variables used in this study. Section 4 presents the sample and data used in the paper. Section 5 presents and describes the empirical results and robustness tests. Section 6 concludes and discusses the limitations of this study.

2. Literature

This section provides an overview of the relevant literature in the field of the study. The first part provides background information of the CSR disclosure mandate enacted in China. The second part describes the existing literature of CER on CFP. Lastly, the third part focuses on the effect of CSR disclosure on CFP.

2.1 Background on China’s CSR activities

"building a harmonious society," the long-term goal outlined by the Chinese Communist Party (Lin, 2010).

In December 2008, both the SSE and SZSE began mandating a subset of firms listed on their exchanges to issue CSR reports along with their annual reports beginning with the reports for the fiscal year 2008, which were released in 2009 (Noronha et al., 2013). This is the first time the two stock exchanges required a subset of firms to release CSR reports to ensure public transparency about their CSR activities. Since the SSE and SZSE are (entirely) owned by the government and directly supervised by the China Securities Regulatory Commission (CSRC), the CSR disclosure mandate of 2008 is a government requirement, virtually. As a consequence, the regulation can be seen as an exogenous shock. Specifically, on December 30, 2008, the SSE required three types of listed companies: firms listed in the SSE "Corporate Governance Sector," financial firms and firms with shares listed overseas. On December 31, 2008, the SZSE realized a similar announcement that required all firms included in its "SZSE 100 Index". Appendix A presents each news announcement. The news announcements indicate that firms that failed to file their CSR reports by May 1, 2009, will be subjected to delisting, and that both the firms and related persons will be subject to public condemnation.

2.2 The relation between CER and firm performance

(CEP). Examples of environmental actions by businesses are reducing the emission of greenhouse gases or recycling of waste.

Within the existing literature, two major theories can be found related to the relation between CEP and CFP. The first theory states that firms with higher environmental performances face less idiosyncratic risks in comparison to firms with lower environmental performances. Firms that improves its CEP reduces its current and future hazards, thus reducing the risk of future financial claims concerning e.g., emission reductions (Sharfman and Fernando, 2008). Moreover, firms with higher environmental ratings are considered better investment opportunities in the eyes of investors, since high environmental ratings are associated with more exceptional stakeholder management making the company more attractive for long-term investments (Cheng, Ioannou and Serafeim, 2014). At the same time, firms with higher corporate environmental performance disclose higher amounts of corporate information by being more transparent and accountable for their actions. The higher amount of disclose leads to making investing in the company more attractive and less risky. The second theory is related to increased efficiency resulting from the shift towards more innovative production methods within firms. Investment in environmental performance, especially investments which are done in the R&D department, could lead to additional sources of income and environmental innovation technology can minimize the costs generated by inefficient production processes. As a result, the firm can be able to license new technologies, lower production costs and improve the competitiveness of the firm in the long-run (Porter, 1990; Porter and Van der Linde, 1995). In this context, the higher the degree of environmental performance, the higher the environmental-friendliness.

companies and is harmful to competitiveness. On the other hand, reducing emissions increases and saves money, which may give firms a cost advantage. The most recent empirical evidence points toward a positive relationship between environmental performance and the financial performance and market valuation of the firm. Kong, Liu, and Dai (2014) and Lee, Cin, and Lee (2016) found that environmental responsibility enhances firms' market value and financial performance in China and South Korea, respectively. Taken together, most of the studies found a positive relationship between CER and CFP. Therefore, the paper proposes the following hypothesis:

Hypothesis 1: Corporate environmental responsibility increases corporate financial performance.

2.3 Mandatory CSR disclosure, voluntary CSR disclosure and financial performance

In the field of CSR activities and reporting, whereas Western companies focus on its customers and community as its most important stakeholders, the government is the essential stakeholder in business in China (ChinaCSR, 2009). According to the stakeholder theory, pressure from stakeholders is the leading force in driving firms to issue CSR reports beside their annual reports. Previous theoretical studies suggest that mandatory disclosure and voluntary disclosure differ in several ways (Dye, 1990; Fishman and Hagerty, 2003). To start with, voluntary disclosure is an endogenous choice of the firm, whereas mandatory disclosure is an exogenous shock to the firm. Secondly, voluntary disclosing firms self-select the disclosure of their CSR activities based on their firm-specific factors in contrast to mandatory disclosing firms who are forced to disclose under a regulation. As a result, the predicted relations can be different. Lastly, firms that are mandated to disclose CSR reports face additional regulatory pressure in comparison to firms who disclose voluntarily. When firms are mandated to disclose their CSR activities, the firms feel pressure to increase their commitment to CSR. Due to the mandatory disclosure requirement, stakeholders (e.g., suppliers, consumers, employees, and communities) can more closely watch the firms and react upon their CSR activities. The rationale behind this is that stakeholders benefit from increased transparency since increased transparency makes it easier for governments and interest groups to pressure firms to engage in more CSR activities.

environment-related information. They found several factors which may influence a firm's voluntary disclosing patterns, namely litigation concerns, regulatory concerns, and the effect the disclosure may have on access to capital. Zeng et al. (2012) found that firms in China that are state-owned, those that operate in environmentally sensitive industries, those having more industrial peers engaged in environmental information disclosure and those with better reputation are more likely to disclose environmental information.

Environmental disclosure is a useful tool for reducing exposure to potential political costs (Patten and Trompeter, 2003). However, concerning the relationship between financial performance and disclosure, prior literature documents a weak association between CSR disclosure and profitability (Belkaoui & Karpik, 1989). Chen, Hung, and Wang (2018) argue that firms would have undertaken CSR activities before the mandate if it were beneficial for the firm's (financial) performance. They further argue that the change in CSR behavior due to regulation may lead to a decrease in firm performance since the increase in CSR activities comes at a cost to performance. Consequently, the following hypotheses are proposed for mandatory and voluntary disclosing firms:

Hypothesis 2a: Firms who need to disclose their CER mandatory experience a decrease in financial performance, measured by ROA and ROE.

Hypothesis 2b: Firms who need to disclose their CER voluntary experience an increase in financial performance, measured by ROA and ROE.

3. Methodology

This section outlines the methodology used in this study. The section starts by discussing the regression model that the study uses. The second part elaborates on the assumptions regarding the research method and the regression model. The third part explains the research design the study uses. Lastly, the fourth part discusses the dependent financial performance variables, the independent explanatory variables, and control variables used in the study.

3.1 Estimation model

𝐹𝑃#,% = 𝛽)+ 𝛽+ 𝐸𝑛𝑣𝑆𝑐𝑜𝑟𝑒#,% + 𝛽4 𝑀𝑎𝑛𝑑𝑎𝑡𝑒#,% + 𝛽9 𝑆𝑖𝑧𝑒#,%+ 𝛽< 𝐿𝑒𝑣#,% +

𝛽_> 𝑂𝑤𝑛𝑒𝑟𝑠ℎ𝑖𝑝_#,%+ 𝛽_D 𝑅&𝐷_#,% + 𝛽_H 𝐶𝑎𝑝𝐸𝑥_#,% + 𝑢_#,% (1)

where 𝐹𝑃_#,% is the estimate of the financial performance (FP) for firm i at time t, measured by Return On Assets and Return On Equity. 𝐸𝑛𝑣𝑆𝑐𝑜𝑟𝑒#,% measures the environmental score of firm i at time t. 𝑀𝑎𝑛𝑑𝑎𝑡𝑒#,% is a dummy variable that equals one for firms who need to disclose mandatory and equals zero if firms are disclosing voluntary. The control variables display the logarithm of size, leverage, state ownership, R&D intensity and investment intensity. As referred to by Horváthová (2010), industry fixed effects are controlled for since environmental performance can be very sector specific. Additionally, the study controls for entity and year fixed effects Lastly, the error term, 𝑢_#,%, is the disturbance term which measures the accuracy of the estimated values.

3.2 Diagnostic tests

Similar to Barnett and Salomon (2012), Hu, Wang and Xie (2018) and among other researchers, this study uses a regression model with Ordinary Least Squares (OLS) to investigate the effect of CEP on CFP. OLS is the best linear unbiased estimator (BLUE) if all five assumptions related to the disturbance term of the sample data hold. First, the mean of the error term is zero. Secondly, the variance of the error is constant and finite over all values of x. Third, all errors are linearly independent of one another. Fourth, no relationship exists between the error term and the corresponding x. Last, the error term is normally distributed (Brooks, 2014).

Various diagnostic tests will be performed to make sure that the sample data is BLUE. Additionally, the diagnostic tests are conducted to ensure that the results from the different estimations are not biased. By including a constant term in the estimation model, the first assumption will never be violated.

Thirdly, the study tests for autocorrelation in the estimation models by performing the Durbin-Watson test. The value of the Durbin-Durbin-Watson test always lies between 0 and 4, depending on the sample. There are three essential values: 0, 2, and 4. If the value is two, there is no autocorrelation in the residuals. If the values are 0 or 4, there is positive autocorrelation and negative autocorrelation in the residuals, respectively. For the first model, with ROA as the dependent variable, the value of the Durbin-Watson test is 0.63. The second model, with ROE as the dependent variable, the value of the Durbin Watson test is 1.00. Both models indicate positive autocorrelation in the residuals, indicating that the OLS standard errors are biased downwards relative to the true standard errors (Brooks, 2014). As a result, there exists the possibility that the wrong inferences may be made about whether a variable is or is not an essential determinant of variations in the dependent variable. To control for heteroskedasticity and autocorrelation, the standard errors are clustered at the firm-level.

Lastly, the Gauss-Markov assumptions of OLS are checked. This study uses the Jarque-Bera test to test if the sample is normally distributed by analyzing if there are extreme outliers and fat tails that cause skewness and kurtosis in the data. The null hypothesis for a normally distributed sample for all variables is rejected at the 1% level. It is known that if the variables violate the assumption of a normal distribution, outliers exist in the sample that may affect the estimated regression coefficient. Following Cai, Cui and Jo (2016) and Jo, Park, and Kim (2015) and many other studies, all the variables are winsorized at the 1st percentile and 99th percentile to reduce the impact of outliers.

3.3 Research design

The procedure is implemented by first estimating a probit regression to model the probability of being a treatment firm. Next, each treatment firm is matched to the control firms using the nearest neighbor matching technique. Panel A of Appendix B presents the estimation results of the logit regression. Panel B of Appendix B presents the test results of the propensity score matches. The PSM procedure resulted in a PSM sample of 240 firm-year observations, 137 of which are treatment firm-years and 103 of which are control firm-years.

3.4 Variables

3.4.1 Financial performance

The dependent variables of the model are related to the financial performance of the firms. In line with previous literature, accounting-based measures will be taken into account, since they are most used in defining the financial performance of the firm related to ESG performance (Lee et al., 2016; Qui et al., 2016).

Accounting-based measures take accounting profitability of companies into account. Return on assets (ROA) and return on equity (ROE) are widely used accounting-based measurements of firm performance related to ESG-metrics. They both measure the efficiency of a firm's operation. ROA reflects the ability of a firm to generate profits by using its resources. The proxy is a ratio and expressed as a percentage: 𝑅𝑒𝑡𝑢𝑟𝑛 𝑜𝑛 𝐴𝑠𝑠𝑒𝑡𝑠 = _{TR%UV UWWN%W}MN% OPQRSN. ROE relates to the return made by a firm for its shareholders with the amount of money that the shareholders have invested in the firm. The proxy is a ratio and expressed as a percentage: 𝑅𝑒𝑡𝑢𝑟𝑛 𝑜𝑛 𝐸𝑞𝑢𝑖𝑡𝑦 = MN% OPQRSN

Z[U\N[RV]N\W^_{_`a#%b}.

3.4.2 Environmental performance

ENVSCORE is composed of 61 different environmental indicators, which are divided into three environmental performance sub-pillars: emissions reduction, resource reduction, and product innovation. The sub-pillar emission reduction measures the commitment of a company towards reducing environmental emission in the production and operational processes. Resource reduction focusses on the ability of the firm to reduce the use of materials, energy, and water, and improvements in the supply chain to find more eco-efficient solutions. Lastly, R&D efforts towards inventing and producing environmentally friendly technologies, processes, and products are part of the product innovation sub-pillar (Thomson Reuters, 2017).

The Environmental Pillar Score takes all three subcategories into account. The score is between 0 and 100, with 0 indicating inferior environmental performance and 100 excellent environmental performance. A higher overall score corresponds to better environmental performance overall.

Table 1: The table reports the summary statistics (mean, median, minimum, maximum, standard deviation,

and the number of data points) of the Environmental Pillar Score for each firm in year t.

EnvScore _{data points} Number of Mean Mean Median Minimum Maximum Std. Dev.

2009 45 34.99 34.99 31.72 8.97 80.24 19.03 2010 63 35.81 35.81 30.98 9.06 85.98 19.64 2011 65 37.17 37.17 34.56 10.29 82.58 19.03 2012 65 36.06 36.06 32.39 6.85 89.60 22.06 2013 69 36.71 36.71 34.01 7.64 90.38 22.23 2014 75 37.73 37.73 35.50 6.64 87.73 21.72 2015 76 43.76 43.76 41.16 7.64 95.34 21.99 2016 77 49.85 49.85 47.63 6.71 93.61 20.63 2017 246 41.51 41.51 39.47 6.73 92.53 20.51 2009-2017 781 40.11 40.11 36.97 6.64 95.34 21.13

average of 40.11. Moreover, over the years, the minimum environmental score decreased, and the maximum environmental score increased. This evidence shows that firms are putting more effort into CER activities in the last couple of years.

3.4.3 Control variables

Control variables are factors that can systematically impact the independent variable and dependent variables (Barnett & Salomon, 2012). Numerous comparable CER and CSR are consulted, concluding that controlling for Size, Leverage, Industry, Year, State ownership, Research & Development (R&D) expenditures and Capital Expenditures (CAPEX) are relevant. The usage of the different control variables has generated mixed results in prior researches. The predicted effects of the control variables on financial performance are positive and negative, depending on the sample.

Size: Substantial studies have found that the size of the firm has a significant impact on CFP

(Habimana, 2014; Karagiorgos, 2010). Fama and French (1992) argued that larger firms are expected to earn higher ex-post returns. Large companies have an abundance of resources available to invest in society and the environment. Furthermore, larger companies attract more attention from the media, and they face more pressure from shareholders and stakeholders. Also, larger companies give out more information to the market, leading to lower information asymmetry. The size of the company will be measured by the natural logarithm of the firm's total assets (Karagiorgos, 2010).

Leverage: Leverage refers to the ratio of a firm's total liability to the total value of assets. It is

Ownership: Due to the nature of Chinese companies, state ownership has a significant impact

on the value of the company. If a firm has a higher level of state ownership, it is likely to generate more profits than other public or private listed companies in its industry. A firm with a higher level of state ownership has more funds allocated by the government and more political connections (Yu, 2013). Firms that have close ties to the government can enjoy preferential treatment (Claessens et al., 2000). Additionally, prior researches suggest a positive relationship between ownership concentration and ex-post firm performance measures (Demsetz & Lehn, 1985; Claessens et al., 2000; Hovey et al., 2003). Ownership is measured through a dummy variable, where the value is one when the largest shareholder is the state, and the value of zero is otherwise.

Research & Development expenditures: R&D expenditures are used to control for differences

in the level of development within firms related to environmental practices. In this study, R&D is defined as the allocation of firm's resources and capabilities for new products or services, processes and technological developments which increases firm's operational efficiency and reduces environmental adversities (Manrique & Martí-Ballester, 2017). Firm's technological capabilities are an essential determinant of long-term economic performance (Lin, Yang, and Liou, 2009). Firms with high investments in (environmental) research and development projects improve their long-term corporate financial performance (Luo and Bhattacharya, 2009; McWilliams and Siegel, 2000), but decrease their short-term corporate financial performance (Graham et al., 2005). A possible explanation could be that (environmental) research, and development projects require initially high investments, which increases the costs for firms, reducing their profits and in return, their return on assets. Following prior studies (Lin, Yang, and Liou, 2009; McWilliams and Siegel, 2000), R&D is controlled by dividing the research and development expenses of the firm by its total sales revenue.

Capital expenditures: Another variable to control for differences in the level of development

4. Data

This section describes the sample used in this study. The first part outlines how the data of the study is collected. The second part gives the descriptive tables to give an overview of the data.

4.1 Sample selection

The sample size depends on the amount of data available from Thomson Reuters' Eikon database. The initial sample consists of all the public listed firms trading A shares on the Stock Exchange (SSE) and Shenzhen Stock Exchange (SZSE) from 2009 to 2017. The companies are gathered from the constituents lists "Shanghai SE A Shares" and "Shenzhen A-Shares," which consists of 3,589 firms. The sample excludes all the public listed firms with B-shares, also called foreign shares, as they are subjected to different regulations and market trading mechanisms. Additionally, the sample is narrowed down to include only listed firms who have been rated once by ASSET4 in the period 2009-2017. After merging all the listed firms with the corporate environmental responsibility scores, the sample is left with 250 firms and 781 firm-year observations. The dataset yields an unbalanced panel since a portion of the (financial) data is unavailable for several firms from 2009 to 2017. The sample period starts in 2009 as this is the first year that the CSR disclosure mandate enacted in China is put into effect. The firms in the final sample are all public listed companies and represent a wide variety of industries. The sample covers companies that are predominantly presented in China.

Several databases are used to research the relation between corporate environmental performance and corporate financial performance. Annual corporate financial data are extracted from Thomson Reuters' Eikon database. Data relating to the state ownership of the firms are collected from the CSI State-owned Enterprise Composite Index from www.csindex.com.cn. Furthermore, environmental performance scores are obtained from Thomson Reuters' ASSET4 database. Lastly, the news announcements relating to the mandate come from the Shanghai Stock Exchange and Shenzhen Stock Exchange website. Besides, the two stock exchanges also provide the list of indices to define which firms are mandatory to disclose their CSR scores.

4.2 Descriptive statistics

and the PSM sample by industry sector. The companies in the sample are mainly doing business in the financial, industrial, and material industries, as shown as in the full sample. Surprisingly, for the PSM sample, all the financial companies dropped out. The reason behind may be that financial companies could not be matched due to their higher environmental scores.

Table 2: The table shows the sample distribution of the full sample and the propensity score matched (PSM)

sample. Panel A reports the sample distribution by year. Panel B reports the distribution of the sample by industry sector based on the Global Industry Classification Standard (GICS).

Panel A: Distribution of treatment and control firms by year

Full sample PSM sample

Total Treatment Control Total Treatment Control

Year N % N % N % N % N % N % 2009 45 5.76 34 6.72 11 4.00 1 0.42 1 0.73 0 0.00 2010 63 8.07 45 8.89 18 6.55 2 0.83 2 1.46 0 0.00 2011 65 8.32 46 9.09 19 6.91 3 1.25 2 1.46 1 0.97 2012 65 8.32 46 9.09 19 6.91 7 2.92 5 3.65 2 1.94 2013 69 8.83 49 9.68 20 7.27 20 8.33 13 9.49 7 6.80 2014 75 9.60 54 10.67 21 7.64 27 11.25 20 14.60 7 6.80 2015 76 9.73 55 10.87 21 7.64 32 13.33 24 17.52 8 7.77 2016 77 9.86 56 11.07 21 7.64 32 13.33 24 17.52 8 7.77 2017 246 31.50 121 23.91 125 45.45 116 48.33 46 33.58 70 67.96 Total 781 100.00 506 100.00 275 100.00 240 100.00 137 100.00 103 100.00

Panel B: Distribution of treatment and control firms by industry sector

Full sample PSM sample

Total Treatment Control Total Treatment Control

Industry sector N % N % N % N % N % N % Communication Services 10 1.28 1 0.20 9 3.27 5 2.08 0 0.00 5 4.85 Consumer Discretionary 64 8.19 53 10.47 11 4.00 28 11.67 22 16.06 6 5.83 Consumer Staples 31 3.97 13 2.57 18 6.55 10 4.17 2 1.46 8 7.77 Energy 66 8.45 27 5.34 39 14.18 34 14.17 14 10.22 20 19.42 Financials 185 23.69 169 33.40 16 5.82 0 0.00 0 0.00 0 0.00 Health Care 27 3.46 14 2.77 13 4.73 16 6.67 6 4.38 10 9.71 Industrials 193 24.71 117 23.12 76 27.64 75 31.25 50 36.50 25 24.27 Information Technology 29 3.71 17 3.36 12 4.36 25 10.42 15 10.95 10 9.71 Materials 108 13.83 76 15.02 32 11.64 37 15.42 26 18.98 11 10.68 Real Estate 30 3.84 5 0.99 25 9.09 1 0.42 1 0.73 0 0.00 Utilities 38 4.87 14 2.77 24 8.73 9 3.75 1 0.73 8 7.77 Total 781 100.00 506 100.00 275 100.00 240 100.00 137 100.00 103 100.00

non-normality in the distribution by winsorizing at the top and bottom 1% of their distributions, as discussed in section 3.2. For conciseness, the paper only presents the descriptive statistics for the PSM-sample2_.

Table 3: The table shows the summary statistics (mean, median, minimum, maximum, the standard deviation, and

number of observations) for all of the firm-level variables employed in this study. The scores define the average scores over the period 2009-2017.

Table 4: The table shows the Pearson correlation coefficients between all the variables used in this study. ROA

represents the Return On Assets, ROE represents to the Return On Equity, EnvScore represents the Environmental Pillar Score provided by the ASSET4 database and Ownership represents the state ownership. The stars (*, **, ***) denote the statistical significance at the 10%, 5%, and 1% levels, respectively.

Variable (1) (2) (3) (4) (5) (6) (7) (8) (1) ROA 1.000 (2) ROE 0.683*** 1.000 (3) EnvScore -0.238*** -0.114*** 1.000 (4) Size -0.484*** 0.010 0.505*** 1.000 (5) Leverage -0.295*** -0.370*** -0.071** -0.203*** 1.000 (6) Ownership -0.317*** -0.213*** 0.095*** 0.253*** 0.056 1.000 (7) R&D 0.099 0.083 0.132** -0.296*** -0.223*** -0.435*** 1.000 (8) CapEx -0.046 -0.088** 0.062 0.022 0.359*** 0.026 -0.048 1.000

Table 4 presents the Pearson correlation coefficients between all the variables used in this study. From the correlation matrix, there are a few points that stand out. First of all, the CER score is significantly and positively related to size (0.505), indicating that larger firms are more likely to have higher environmental scores. The same counts for the level of capital expenditures

2 _{The descriptive statistics of the full sample are not show but are available upon request.}

Descriptive statistics on firm-level variables

PSM Treatment firms PSM Control firms

(0.132) showing that firms who are using their capital to invest in sustainable resources are more likely to have better environmental performances. Noteworthy, the correlation between environmental performance and ROA and ROE are both going in the negative direction. Environmental performance has a significant, negative correlation with ROA (-0.238) and ROE (-0.114). A possible explanation for the negative relationship may be that the environmental performance scores of Chinese companies are quite low, as shown in Table 1, which results in a decrease in financial performance.

The relationship between the variables is significant if the correlation coefficient is larger than 0.7. If that is the case, multicollinearity exists, and one of the variables must drop out of the regression to ensure the accuracy of the results. As Table 4 shows, the highest correlation is 0.683. Nevertheless, none of the variables are highly correlated with another, which means that the estimation model with the variables can be used to analyze the relationship between CEP and CFP.

5. Results

This section examines the relationship between CER and CFP. In the first part, a univariate test is performed that compares the corporate financial measures of mandatory and disclosing firms with above and below median CER scores. In the second part, a multivariate analysis is performed to examine the effect of CER on ROA and ROE, while controlling for other factors previously shown that may affect the firm's financial performance. In the third part, robustness tests are performed to ensure that the results presented in section 5.2 hold.

5.1 Univariate model

To provide initial evidence on the CER-CFP relationship, a univariate analysis is used to answer the question if CER affects the ROA and ROE. Table 5 compares the mean of the dependent variables across firms with high CER scores and firms with low CER scores, where high and low are those firms with above- and below median CER scores, respectively. The differences in means help to identify if the mean of the ROA and the ROE measures are significantly different from each other between the treatment and control group.

for the control group are 1.5146 and 0.8863. As a result, a t-test assuming equal variances is performed for both ROE's, while a t-test assuming unequal variances is performed for both ROA's.

Table 5: The table shows the results of the mean comparison test. ROA represents the Return On Assets, and ROE

represents the Return On Equity. The sample consists out of 642 year observations for ROA and 704 firm-year observations for ROE between 2009-2017 due to missing data points. The stars (*, **, ***) denote the statistical significance at the 10%, 5% and 1% levels, respectively.

Mandatory disclosing firms - Voluntary disclosing firms ROA (Obs. = 431) ROE (Obs. = 465) ROA (Obs. = 211) ROE (Obs. = 239) CER Score > median (1) 2.8494 11.9370 4.4311 10.4896 CER Score < median (2) 4.8395 14.5850 7.2348 14.9260 Difference (1)-(2) -1.9900 -2.6480 -2.8037 -4.4364

t-test statistic 4.8473*** 2.9960*** 3.4365*** 2.9473***

5.2 Multivariate analysis

In the multivariate analysis, the results of the regression analyses are presented. The study uses a panel dataset (also called longitudinal data) to examine the relationship between CER and CFP further. Panel data contain observations of multiple variables over multiple periods for the same firms. To further examine the relationship between CER and CFP, the CEP is regressed on CFP and a diverse set of control variables. All the regression models presented use the PSM sample to control for selection bias in the results, as discussed in section 3.3.

Table 6: The table reports the overall effect of the CEP on CFP. The model regresses the ROA (in panel A) and

ROE (in panel B) on CER scores and different control variables. Robust standard errors are reported in the parentheses. The stars (*, **, ***) denote the statistical significance at the 10%, 5% and 1% levels, respectively.

Panel A: ROA

ENV (1) ENV (2) ENV (3) ENV (4) ENV (5)

EnvScore -0.048** -0.049** -0.008 -0.015 -0.003 (0.023) (0.024) (0.020) (0.019) (0.023) Mandate 0.133 0.038 0.549 1.559 (1.264) (1.049) (1.204) (1.063) Size -0.954* -0.936* -1.091** (0.542) (0.506) (0.532) Leverage -18.269*** -17.152*** -18.006*** (3.790) (3.736) (2.938) Ownership -5.281*** -4.633*** -3.821*** (1.139) (1.030) (1.161) R&D-ratio -56.734*** -56.560*** -19.344 (20.633) (20.146) (21.256) CAPEX-ratio 2.976 3.238 4.508 (4.102) (3.993) (3.733) Intercept 8.020*** 7.965*** 38.308*** 37.122*** 38.487*** (1.140) (1.127) (12.315) (11.353) (12.118)

Entity fixed effects No No No Yes No

Industry fixed effects No No No No Yes

Year fixed effects No No No Yes Yes

Observations 240 240 240 240 240

Adjusted R-squared 0.018 0.014 0.430 0.431 0.547

Panel B: ROE

ENV (1) ENV (2) ENV (3) ENV (4) ENV (5)

EnvScore -0.097** -0.096** -0.063 -0.078* -0.028 (0.046) (0.048) (0.045) (0.043) (0.050) Mandate -0.562 -0.847 0.742 0.594 (2.383) (2.071) (2.175) (2.186) Size 0.010 0.028 -0.237 (1.158) (1.034) (1.102) Leverage -25.353*** -21.773*** -24.082*** (7.568) (7.528) (7.518) Ownership -11.245*** -9.313*** -8.153*** (2.430) (2.101) (2.440) R&D-ratio -78.731* -80.286** -38.937 (40.326) (38.420) (34.364) CAPEX-ratio 3.732 4.568 5.230 (8.696) (8.270) (9.273) Intercept 16.986*** 17.220*** 32.210 29.168 32.226 (2.129) (2.005) (26.117) (23.157) (24.866)

Entity fixed effects No No No Yes No

Industry fixed effects No No No No Yes

Year fixed effects No No No Yes Yes

Observations 239 239 239 239 239

to make some inferences about the relationship between environmental performance and corporate financial performance. It is important to note that much of the explanatory power of the model comes from including control variables and adding in industry fixed effects; the adjusted R2 rises from 0.018 to 0.547. This indicates that much of the change in ROA is due to differences in industries since environmental performance can be very sector specific, as referred to by Horváthová (2010). Column (1), in panel B, tests the effect of environmental performance on the ROE without the usage of control variables and fixed effects. The environmental performance is still negative, but statistically significant at the 5% significance level. However, an increase in environmental performance leads to a decrease in ROE with 9,7%, which is almost twice a decrease as the decrease in ROA. Furthermore, in column (2), the variable ‘mandate' has switched sign and became negative, but still statistically insignificant. The negative coefficient of the ‘mandate' variable indicates that investors anticipate a decrease in firm performance after the disclosure mandate since firms will spend money on CER activities (Chen, Hung, and Wang, 2018). In line with previous results, almost all control variables are negative but statistically significant. However, now only the sign of the variable ‘ownership' is not as expected. The variable ‘size' has switched sign and became positive, following the findings of previous literature. To summarize, the findings lead to conclude that environmental performance does influence corporate financial performance. When looking at column (1) and (2) of both panels, it is possible to say with certainty that there is a significant, negative relationship between environmental performance and corporate financial performance, as expected by previous literature (Lioui and Sharma, 2012; García-Sánchez and Prado-Lorenzo, 2012). The effect of an increase in CEP is related to more costs for the firm, leading to more reduced financial performance. As a consequence, hypothesis 1 is rejected. Nevertheless, it is essential to note that the negative relationship may not hold when looking at columns (3) to (5) since the variables are statistically insignificant.

behavior might lead to a decrease in firm performance since the increase in CSR activities comes at a cost to performance. However, the sign of the environmental performance switches when the control variables are included in the model, and the model controls for fixed effects. For the ROA, the signs of environmental performance become

Table 7: The table reports the effect of the CEP on CFP when companies are mandated to disclose their

environmental performance. The model regresses the ROA (in panel A) and ROE (in panel B) on CER scores and different control variables. Robust standard errors are reported in the parentheses. The stars (*, **, ***) denote the statistical significance at the 10%, 5% and 1% levels, respectively.

Panel A: ROA

MAND (1) MAND (2) MAND (3) MAND (4) MAND (5) MAND (6)

EnvScore -0.016 -0.041 -0.027 0.019 0.004 0.010 (0.026) (0.029) (0.037) (0.023) (0.020) (0.028) Size -1.934** -1.812** -1.994** (0.751) (0.695) (0.858) Leverage -21.932*** -21.489*** -22.956*** (3.058) (3.065) (3.977) Ownership -3.924** -3.553** -2.275 (1.780) (1.698) (1.728) R&D-ratio -69.090*** -70.450*** -36.996* (23.016) (23.151) (19.859) CAPEX-ratio 4.578 5.634 5.842 (6.684) (5.886) (9.430) Intercept 6.714*** 7.783*** 7.202*** 60.252*** 57.617*** 60.327*** (1.618) (1.842) (1.660) (18.274) (17.005) (20.430)

Entity fixed effects No Yes No No Yes No

Industry fixed effects No No Yes No No Yes

Year fixed effects No Yes No No Yes Yes

Observations 137 137 137 137 137 137

Adjusted R-squared -0.005 0.091 0.138 0.556 0.556 0.597 Panel B: ROE

MAND (1) MAND (2) MAND (3) MAND (4) MAND (5) MAND (6) EnvScore -0.039 -0.082 -0.033 -0.022 -0.054 0.016 (0.052) (0.055) (0.079) (0.053) (0.048) (0.065) Size -1.347 -0.972 -1.082 (1.608) (1.520) (1.931) Leverage -38.091*** -37.089*** -44.291*** (8.123) (7.936) (10.063) Ownership -9.535** -8.643** -6.063 (3.827) (3.569) (3.920) R&D-ratio -109.619** -113.195** -53.192 (49.551) (49.904) (43.098) CAPEX-ratio 16.671 19.087 19.040 (16.973) (15.111) (17.059) Intercept 14.251*** 16.089*** 13.991*** 63.149 54.748 53.087 (3.107) (3.318) (3.297) (39.031) (36.903) (45.552)

Entity fixed effects No Yes No No Yes No

Industry fixed effects No No Yes No No Yes

Year fixed effects No Yes No No Yes Yes

Observations 137 137 137 137 137 137

Table 8: The table reports the effect of the CEP on CFP when companies are voluntarily disclosing their

environmental performance. The model regresses the ROA (in panel A) and ROE (in panel B) on CER scores and different control variables. Robust standard errors are reported in the parentheses. The stars (*, **, ***) denote the statistical significance at the 10%, 5% and 1% levels, respectively.

Panel A: ROA

VOLUN (1) VOLUN (2) VOLUN (3) VOLUN (4) VOLUN (5) VOLUN (6) EnvScore -0.106*** -0.086** -0.055 -0.053 -0.050 0.004 (0.038) (0.035) (0.034) (0.034) (0.038) (0.036) Size 0.229 0.141 -0.609 (0.413) (0.509) (0.735) Leverage -13.941** -13.019** -13.546*** (5.969) (6.025) (4.681) Ownership -6.095*** -5.646*** -4.797*** (1.373) (1.331) (1.584) R&D-ratio -10.968 -11.827 18.233 (29.740) (31.664) (27.623) CAPEX-ratio 2.113 1.876 2.807 (4.971) (4.937) (4.344) Intercept 10.216*** 9.448*** 8.228*** 11.160 12.552 26.678* (1.691) (1.683) (1.502) (8.909) (10.467) (15.540)

Entity fixed effects No Yes No No Yes No

Industry fixed effects No No Yes No No Yes

Year fixed effects No Yes No No Yes Yes

Observations 103 103 103 103 103 103

Adjusted R-squared 0.079 0.121 0.361 0.349 0.312 0.548 Panel B: ROE

VOLUN (1) VOLUN (2) VOLUN (3) VOLUN (4) VOLUN (5) VOLUN (6) EnvScore -0.196** -0.131** -0.116 -0.151** -0.115* -0.033 (0.075) (0.064) (0.069) (0.067) (0.066) (0.077) Size 1.692* 1.211 -0.603 (0.947) (0.799) (1.293) Leverage -10.178 -2.085 2.010 (9.831) (9.411) (10.077) Ownership -11.363*** -8.850*** -6.607** (2.589) (2.509) (3.165) R&D-ratio 21.414 21.004 40.283 (49.742) (48.801) (46.232) CAPEX-ratio -4.223 -7.631 -6.437 (7.913) (8.248) (7.919) Intercept 21.146*** 18.604*** 18.004*** -8.795 -2.822 32.839 (2.942) (2.808) (2.900) (20.392) (16.746) (28.096)

Entity fixed effects No Yes No No Yes No

Industry fixed effects No No Yes No No Yes

Year fixed effects No Yes No No Yes Yes

Observations 102 102 102 102 102 102

Adjusted R-squared 0.085 0.246 0.254 0.240 0.310 0.430

there is an inconclusive relationship between mandatory disclosure of environmental performance and corporate financial performance. Since all the models show statistically insignificant results, the treatment group does not experience any change in ROA and ROE due to the disclosure mandate. As a result, hypothesis 2a is rejected.

On the other hand, Table 8 reports the regression analysis of the control group, measured by the ROA (panel A) and the ROA (panel B). The signs of environmental performance are negative but statistically significant associated with corporate financial performance, indicating that an increase in environmental performance leads to a decrease in ROA and ROE. However, the signs are not as expected, since previous literature argues that a firm will disclose their environmental performance voluntarily when the firm expects that it will benefit the firm. A possible explanation for the decrease in both financial measures may be that a company is pressured in spending money on CSR activities due to the companies' shareholders and stakeholders. This may be since investors believe that the company is doing too little in their green practices, as measured by the environmental scores shown in Table 1. As a result, the company creates more costs, which leads to a decrease in the companies' net income.

5.3 Robustness checks

To determine the accuracy of the previous results, a robustness test is conducted. In Appendix C, the same regressions as for section 5.2 are ran, but this time solely for the full sample in order to check if the data fit the results of the PSM sample. Table 11 reports the overall effect of disclosing CER on CFP without making a distinction between mandatory and voluntary disclosing firms. As shown in Table 11, the relationship between environmental performance and corporate financial performance is negative but statistically significant for both financial measures. Previous results found the same statistically significant, negative relationship when using the PSM sample. However, when adding in control variables, panel A of Table 11 improves and all control variables are significant at the 5% and 1% significance level. For panel B, the significance of the model stays the same. Again, all the signs are the same as the results in Table 6.

variables and fixed effects are included, the significance of the environmental performance turns insignificant. For panel B, the opposite holds, meaning that the environmental performance measure turns significant when control variables are included and when entity and year fixed effects are active. When the model controls for industry and year fixed effects, the sign of the environmental performance measure turns insignificant again, indicating that industry fixed effects do not play an essential role if firms need to disclose mandatory. To summarize, for the full sample, the environmental performance measure becomes significant for several models. This indicates that it is possible to say that there exists a negative relationship between environmental performance and corporate financial performance for mandatory disclosing firms, compared to the conclusion of Table 7.

Lastly, a robustness check for the voluntarily disclosing firms has been carried out. As shown in Table 10, the negative but statistically significant relationship between environmental performance and corporate financial performance holds for both measures. The results are again in contrast to previous literature. When adding control variables and controlling for fixed effects, the environmental performance measure turns insignificant in five out of six models. At the same time, almost all control variables lose their significance, except for leverage and ownership. Concluding, it is possible to say that there exists a negative relationship between environmental performance and corporate financial performance for voluntarily disclosing firms.

6. Conclusion

This thesis examines the impact of the mandatory CSR disclosure mandate enacted in China in 2008. The CSR disclosure mandate enacted in 2008 is the first regulation in China where the Shanghai Stock Exchange and the Shenzhen Stock Exchange mandated a subset of listed firms to issue CSR reports along with their annual reports starting from the fiscal year 2008. On December 30, 2008, SSE required three types of listed companies: firms listed in the SSE "Corporate Governance Sector", financial firms and firms with shares listed overseas. On December 31, 2008, the SZSE realized a similar announcement that required all firms included in its "SZSE 100 Index". Since the SSE and SZSE are (entirely) owned by the government and directly supervised by the China Securities Regulatory Commission, the CSR disclosure mandate of 2008 is government regulation, essentially.

performance. More specifically, using a difference-in-difference (DiD) approach, we compare the change in financial performance among mandatory disclosing firms (also called treatment firms) with the change among voluntary disclosing firms (also called controlling firms). The study focuses on A-share listed firms on the Shanghai Stock Exchange (SSE) and the Shenzhen Stock Exchange (SZSE) between 2009 and 2017. The study uses a propensity-score-matching (PSM) sample of 240 firm-year observations, divided into 137 firm-years observations for the treatment group and 103 firm-year observations for the control group.

To examine the relationship between the disclosure of environmental performance and corporate financial performance, a regression analysis is used to measure the change in financial performance. The hypothesis is that there may be a difference between financial performance, measured by ROA and ROE, between mandatory and voluntary disclosing firms. The findings of the study show a negative relation between environmental performance disclosure and corporate financial performance after the (mandatory) CSR disclosure. When distinguishing the sample in mandatory disclosing firms and voluntary disclosing firms, the same negative relationship holds. However, the negative relationship for mandatory disclosing firms is inconclusive due to the insignificant result between environmental performance and corporate financial performance, whereas voluntarily disclosing firms experience a statistically significant decrease in ROA and ROE. The negative results hold up to a variety of robustness checks but for both groups statistically significant this time. All in all, firms included in the CSR disclosure mandate experience a decrease in profitability after the mandate, consistent with Chen, Hung, and Wang (2018). The findings suggest that investors anticipate a decrease in corporate financial performance and an increase in CSR spending after the mandate, although the mandate does not require firms to spend on CSR. For voluntarily disclosing firms, the decrease in profitability is since investors perceive environmental practices as potential sources of cost for the firm, leading to more mediocre financial results in the short run (Lioui and Sharma, 2012; García-Sánchez and Prado-Lorenzo, 2012).

are not assigned relative to each other. Moreover, using an aggregated score to measure environmental performance indicates that environmental strengths can substitute environmental weaknesses. As a result, the score is not entirely objective. Secondly, to measure the financial performance of the firm, only the ROA and ROE has been used. Even though they are the most frequently used ones, they are not the only variables which could be used to measure financial performance. Other measures such as return on sales (ROS), return on investment (ROI) or earnings per share (EPS) may influence the relationship as the predicted effects may be different. Thirdly, the sample of the study is quite small as the sample size consists of 250 Chinese companies over nine years. The small sample size is due to the limited amount of ESG data of the ASSET4 database for Chinese companies. Due to missing many CER scores, multiple companies had to be removed from the original sample. It is possible that due to the small sample size, the disclosure does not result in portraying an accurate picture of the situation. Fourthly, to measure the relationship between environmental performance and financial performance, it would have been wise to lag the environmental performance by one or more years. The rationality behind this is that it may take time to affect financial performance. Previous literature argues that firm's with higher environmental ratings shows better financial performance levels in the future. By lagging the environmental performance behind the financial performance, the lag measures the ‘future' and the confidence in the direction of the relationship increases. For this study, it is not possible to lag the environmental performance since lagging the environmental performance leads to a decrease in the sample size of almost one third due to multiple one-year observations of the environmental score. Since the sample of the study is already small, the study decides to measure the effect on financial performance for the short run. Lastly, global environmental concerns are hard to cope with since it is a relatively new topic for businesses. Since the recent appearance of the environmental aspect within CSR, the data around the environmental aspect is still relatively minimum, and the results are still inconclusive. In the next years, the focus on collecting environmental data will improve, which makes it possible to enlarge research and come to more conclusive conclusions.

References

Al Farooque, O., Van Zijl, T., Dunstan, K., & Karim, W. 2005. A simultaneous equations approach to analyzing the relationship between ownership structure and performance in

Bangladesh. Unpublished working paper. University of Auckland Business School, Auckland.

Barnett, M.L., & Salomon, R.M. 2012. Does it pay to be really good? Addressing the shape of the relationship between social and financial performance. Strategic Management Journal, 33(11), p. 1304-1320.

Barth, M.E., & McNichols, M.F. 1994. Estimation and market valuation of environmental liabilities relating to superfund sites. Journal of Accounting Research, 32, p. 177-209.

Belkaoui, A.R., & Karpik, P.G. 1989. Determinants of corporate decision to disclose social information. Accounting Auditing & Accountability Journal, 2(1), p. 36-51.

Brooks, C. 2014. Introductory econometrics for finance. Cambridge: Cambridge University Press.

Cai, L., Cui, J., & Jo, H. 2016. Corporate environmental responsibility and firm risk. Journal of Business Ethics, 139(3), p. 563-594.

Chen, Y-C., Hung, M., & Wang, Y. 2018. The effect of mandatory CSR disclosure on firm profitability and social externalities: Evidence from China. Journal of Accounting and Economics, 65(1), p. 169-190.

Cheng, B., Ioannou, I., & Serafeim, G. 2014. Corporate social responsibility and access to finance. Strategic Management Journal, 35(1), p. 1-23.

Claessens, S., Djankov, S., & Lang, L.H. 2000. The separation of ownership and control in East Asian corporations. Journal of financial Economics, 58(1), p. 81-112.

Demsetz, H., & Lehn, K. 1985. The structure of corporate ownership: causes and consequences. Journal of Political Economy, 93(6), p. 1155-1177.

DesJardins, J. 1998. Corporate environmental responsibility. Journal of Business Ethics, 17(8), p. 825-838.

Dowell, G., Hart, S., & Yeung, B. 2000. Do corporate global environmental standards create or destroy market value? Management Science, 46(8), p. 1059-1074.

Dye, R.A. 1990. Mandatory versus voluntary disclosures: The case of financial and real externalities. The Accounting Review, 65(1), p. 1-24.

Fama, E.F., & French, K.R. 1992. The cross-section of expected stock returns. Journal of Finance, 47(2), p. 427-465.

Fishman, M.J., & Hagerty, K. 2003. Mandatory versus voluntary disclosure in markets with informed and uninformed customers. Journal of Law, Economics & Organization, 19(1), p. 45-63.

Friedman, M. 1970. A Friedman doctrine – the social responsibility of business is to increase its profits. New York Times Magazine, 6(33), p. 122-126.

García-Sánchez, I.M., & Prado-Lorenzo, J.M. Greenhouse gas emission practices and financial performance. International Journal of Climate Change Strategies and Management, 4(3), p. 260-276.

Habimana, O. 2014. Capital structure and financial performance: Evidence from firms

operating in emerging markets. International Journal of Academic Research in Economics and Management Sciences, 3(6), p. 159-166.

Hart, S., & Ahuja, G. 1996. Does it pay to be green? An empirical examination of the relationship between emission reduction and firm performance. Business Strategy and the Environment, 5(1), p. 30-37.

Hong, H., & Kacperczyk, M. 2009. The price of sin: the effect of social norms on markets. Journal of Financial Economics, 93(1), p. 15-36.

Horváthová, E. 2010. Does environmental performance affect financial performance? A meta-analysis. Ecological Economics, 70(1), p. 52-59.

Horváthová, E. 2012. The impact of environmental performance on firm performance: Short-term costs and long-Short-term benefits. Ecological Economics, 84(C), p. 91-97.

Hovey, M., Li, L., & Naughton, T. 2003. The relationship between valuation and ownership of listed firms in China. Corporate Governance: An International Review, 11(2), p. 112-122.

Hu, J., Wang, S., & Xie, F. 2018. Environmental responsibility, market valuation and firm characteristics: Evidence from China. Corporate Social Responsibility and Environmental Management, 25(6), p. 1376-1387.

Jo, H., Kim. H., & Park, K. 2015. Corporate environmental responsibility and firm

performance in the financial services sector. Journal of Business Ethics, 131(2), p. 257-284.

Jones, D.A., Willness, C.R., & Madey, S. 2014. Why are job seekers attracted by corporate social performance? Experimental and field test of three signal-based mechanisms. Academy of Management Journal, 57(2), p. 383-404.

Kitzmueller, M., & Shimshack, J. 2012. Economic perspectives on corporate social responsibility. Journal of Economic Literature, 50(1), p. 51-84.

Kong, D., Liu, S., & Dai, Y. 2014. Environmental policy, company environment protection, and stock market performance: Evidence from China. Corporate Social Responsibility and Environmental Management, 21(2), p. 100-112.

Lee, K.H., Cin, B.C., & Lee, E.Y. 2016. Environmental responsibility and firm performance: The application of an environmental, social and governance model. Business Strategy and the Environment, 25(1), p. 40-53.

Lin, L.W. 2010. Corporate social responsibility in China: window dressing or structural change. Berkeley Journal of International Law, 28(1), p. 64-100.

Lin, C.H., Yang, H.L., & Liou, D.Y. 2009. The impact of corporate social responsibility on financial performance: Evidence from business in Taiwan. Technology in Society, 31(1), p. 56-63.

Lioui, A., & Sharma, Z. 2012. Environmental corporate social responsibility and financial performance: Disentangling direct and indirect effects. Ecological Economics, 78(C), p. 100-111.

Luo, X., & Bhattacharya, C.B. 2009. The debate over doing good: Corporate social

performance, strategic marketing levers, and firm-idiosyncratic risk. Journal of Marketing, 73(6), p. 198-213.

Manrique, S., & Martí-Ballester, C.P. 2017. Analyzing the effect of corporate environmental performance on corporate financial performance in developed and developing countries. Sustainability, 9(11), p. 1-30.

McWilliams, A., & Siegel, D. 2000. Corporate social responsibility and financial

performance: Correlation or misspecification? Strategic Management Journal, 21(5), p. 603-609.

Meyers, S.C. 1984. The capital structure puzzle. Journal of Finance, 39(3), p. 575-592.

Moneva, J., & Ortas, E. 2010. Corporate environmental and financial performance: A multivariate approach. Industrial Management & Data Systems, 110(2), p. 193-210.

Nakao, Y., Amano, A., Matsumara, K., Genba, K., & Nakano, M. 2006. Relationship between environmental performance and financial performance: An empirical analysis of Japanese corporations. Business Strategy and the Environment, 16(2), p. 106-118.

Noronha, C., Tou, S., Cynthia, M.I., & Guan, J. 2013. Corporate social responsibility reporting in China: an overview and comparison with major trends. Corporate Social Responsibility and Environmental Management, 20(1), p. 29-42.

Patten, D., & Trompeter, G. 2003. Corporate responses to political costs: an examination of the relation between environmental disclosure and earnings management, 22(1), p. 83-94.

Porter, M.E. 1990. The competitive advantage of nations. Harvard Business Review, 1(1), p. 73-91.

Porter, M., & Kramer, M. 2006. Strategy and society: the link between competitive advantage and corporate social responsibility. Harvard Business Review, 84(12), p. 78-92.

Porter, M.E., & Van der Linde, C. 1995. Toward a new conception of the environment-competitiveness relationship. Journal of Economic Perspectives, 9(4), p. 97-118.

Qui, Y., Shaukat, A., & Tharyan, R. 2016. Environmental and social disclosures: Link with corporate financial performance. The British Accounting Review, 48(1), p. 102-116.

Sharfman, M.P., & Fernando, C.S. 2008. Environmental risk management and the cost of capital. Strategic Management Journal, 29(6), p. 569-592.

Taipi, E., & Ballkoci, V. 2017. Capital expenditure and firm performance: Evidence from Albanian construction sector. European Scientific Journal, 13(28), p. 231-238.

Trinks, A., Ibikunle, G., Mulder, M., & Scholtens, B. 2017. Greenhouse gas emissions intensity and the cost of capital. Unpublished working paper. University of Groningen, Groningen.

Qi, G.Y., Zeng, S.X., Jonathan J. Shi, J.J., Meng, X.H., Lin, H., & Yang, Q.X. 2014. Revisiting the relationship between environmental and financial performance in Chinese industry. Journal of Environmental Management, Volume 145, 1 December 2014, p. 349-356.

Yu, M. 2013. State ownership and firm performance: Empirical evidence from Chinese listed companies. China Journal of Accounting Research, 6(2), p. 75-87.

Appendix A

Mandatory CSR disclosure announcements

Shanghai Stock Exchange (SSE) Shenzhen Stock Exchange (SZSE)

Date December 30, 2008 December 31, 2008 Heading “Notice on listed companies’ 2008 annual

report”

“Notice on listed companies’ preparation for 2008 annual reports”

Disclosure deadline and consequences

2. All companies listed before December 31, 2007 should finish the preparation, submission, and disclosure of the 2008 annual reports by April 30, 2009… If a firm fails to file the report by May 1, 2009, the SSE will delist the firm’s stock and publicly condemn the persons in charge.

2. Listed companies should disclose the 2008 annual reports by April 30, 2009… Firms that do not file reports by May 1, 2009 will be subject to delisting, and the firms and their related personnel will be subject to public condemnation.

Scope of the CSR reporting

11. Firms listed on the SSE “Corporate Governance Sector”, firms with shares listed overseas, and financial companies should disclose the CSR report when releasing 2008 annual report… The CSR report should be separately approved by the board of directors… The CSR report should be attached to the annual report.

11. Firms included in the “Shenzhen 100 index” should follow “corporate social responsibility guidelines” in reference to Annex 3 – CSR disclosure requirements to disclose CSR report, while encouraging other companies to disclose CSR reports. CSR reports should be approved by the board of directors and reported separately in the form of annual disclosure.

Appendix B

Probit regression used for Propensity Score Matching (PSM) sample

The table describes the propensity score matching (PSM) approach. The procedure is implemented by first estimating a probit regression to model the probability of being a treatment firm. Secondly, each treatment firm is matched to the control firms using the nearest neighbor matching technique. Panel A presents the estimation result of the probit regression. Panel B reports the outcomes of the nearest neighbor matching technique by showing the size of the PSM-sample.

Panel A: Probit model used to find propensity scores Dep. var. = Mandate

EnvScore -0.003 (0.005) Size 0.745*** (0.120) Leverage -0.425 (0.718) Ownership -0.568** (0.250) R&D-ratio 7.587** (3.821) CAPEX-ratio -1.803* 1.008 Intercept -16.737*** (2.593) Entity fixed effects No Year fixed effects No Observations 240 Pseudo r-squared 0.229

Panel B: Sample outcomes of propensity score matching

Treatment assignment Common support Total

Treatment 137 137

Control 103 103

Appendix C

Robustness test: Full sample

Table 9: The table reports the effect of the CEP on CFP when companies are mandated to disclose their

environmental performance. The model regresses the ROA (in panel A) and ROE (in panel B) on CER scores and different control variables. Robust standard errors are reported in the parentheses. The stars (*, **, ***) denote the statistical significance at the 10%, 5% and 1% levels, respectively.

Panel A: ROA

MAND (1) MAND (2) MAND (3) MAND (4) MAND (5) MAND (6) EnvScore -0.039** -0.040** -0.034*** 0.003 -0.003 -0.001 (0.015) (0.017) (0.012) (0.014) (0.013) (0.021) Size -1.300*** -1.326*** -1.261** (0.362) (0.339) (0.516) Leverage -22.831*** -22.573*** -22.560*** (2.332) (2.370) (3.439) Ownership -2.170* -2.010 -0.977 (1.200) (1.229) (1.516) R&D-ratio -45.866*** -47.161*** -33.456 (14.777) (15.427) (20.552) CAPEX-ratio 8.845** 9.470** 9.314 (4.112) (4.294) (5.720) Intercept 5.526*** 5.580*** 5.303*** 43.920*** 44.623*** 41.903*** (0.861) (0.909) (0.652) (8.725) (8.278) (12.563)

Entity fixed effects No Yes No No Yes No

Industry fixed effects No No Yes No No Yes

Year fixed effects No Yes No No Yes Yes

Observations 431 431 431 137 137 137

Adjusted R-squared 0.041 0.054 0.314 0.534 0.531 0.538 Panel B: ROE

MAND (1) MAND (2) MAND (3) MAND (4) MAND (5) MAND (6) EnvScore -0.037 -0.025 -0.038 -0.070** -0.084** -0.030 (0.033) (0.035) (0.029) (0.034) (0.033) (0.050) Size 0.173 0.236 0.604 (0.802) (0.859) (1.307) Leverage -43.083*** -42.596*** -47.687*** (5.872) (5.769) (7.867) Ownership -5.284** -4.831** -2.861 (2.307) (2.364) (3.606) R&D-ratio -47.420 -49.150 -52.256 (33.021) (35.566) (48.711) CAPEX-ratio 16.605 17.414 19.844 (10.513) (10.432) (15.059) Intercept 14.843*** 14.304*** 14.859*** 25.752 24.388 13.106 -0.037 -0.025 -0.038 -0.070** -0.084** -0.030

Entity fixed effects No Yes No No Yes No

Industry fixed effects No No Yes No No Yes

Year fixed effects No Yes No No Yes Yes

Observations 465 465 465 146 146 146