CSR法規對企業股價表現之影響 : 以台灣股市為例 - 政大學術集成

全文

(1)國立政治大學社會科學學院應用經濟與社會發展英語碩士學位學程 International Master’s Program of Applied Economics and Social Development College of Social Sciences National Chengchi University. 治碩士論文政大. 立 Master’s Thesis. ‧. ‧ 國. 學. CSR 法規對企業股價表現之影響:以臺灣股市為例. y. Nat. er. io. sit. Does CSR Regulation Affect Company's Stock Performance?. n. from Taiwan aEvidence iv l C n hengchi U. Student: 嚴蔓舸 Man-Ko Yen Advisor: 陳敦源 Don-Yun Chen. 中華民國 106 年 12 月 December 2017.

(2) Abstract This study examines the effectiveness of the new regulations regarding CSR report on stock returns in Taiwan. The sample of corporations is divided into five groups depending on “compulsory or noncompulsory” and “voluntary or non-voluntary before the rules.” In the empirical study, “the enforcement date of the regulations (Part 1)” and “the filing date of CSR report (Part 2)” are two basis dates; we observe the difference of 180 transaction days before and after the basis date in both parts. The results are as follows. 1. With controlled relevant. 政治大. variables, the whole market increased 26.5% on stock return after the release of the rules. 2.. 立. Whether the new regulations are released or not and with controlled relevant variables, the. ‧ 國. 學. compulsory firms (which are with higher customer awareness) have better stock performance than others; while the firms which are noncompulsory but volunteered to generate a CSR. ‧. report did not show significant difference on stock performance; the firms that are. y. Nat. io. sit. noncompulsory and do not submit CSR report, however, perform significantly worse than. n. al. er. others. 3. After the enforcement of the new CSR regulations, with controlled relevant. i n U. v. variables, the companies which are compulsory but voluntary to file CSR report before the. Ch. engchi. rules performs worse than before; the companies that are noncompulsory and do not file CSR report perform significantly better than before; while the new rules do not have significant influence on the other corporations. 4. Companies which have issued CSR report increase 11.11% on stock return after the issuance of CSR report of each company. 5. Companies that are compulsory and voluntary to submit CSR report before the rules shows better stock performance; while the issuance of CSR report does not make other corporations perform better.. Key words: CSR report, regulations, disclosure, stock return, voluntary I.

(3) A Table of Contents Abstract………………………………………………………………………………….........I A Table of Contents………………………………………………………………………….II A Table of Contents – Tables……………………………………………………………….III A Table of Contents – Figures……………………………………………………………...IV. II.. 政治大 Literature Review and Hypotheses Development………………………………….4 立. 學. III.. Introduction…………………………………………………………………………..1. ‧ 國. I.. ‧. Data and Methodology……………………………………………………………...17 1. Data Description…………………………………………………………………...17 1.1 Sample Selection………………………………………………………………17 1.2 Sample Period Selection……………………………………………………….17 1.3 Population Classification………………………………………………………18. y. Nat. sit. n. al. er. io. 2. Measurement of Variables…………………………………………………………21 3. Empirical Model Setting…………………………………………………………..22. Ch. i n U. v. IV.. Empirical Results…………………………………………………………………...24 1. Part 1—the enforcement date of new CSR regulations……………………………24 2. Part 2—the filing date of CSR report of each company…………………………...34. V.. Conclusion…………………………………………………………………………...45. engchi. References…………………………………………………………………………………49 Appendix…………………………………………………………………………………….52. II.

(4) A Table of Contents – Tables Table 1 Questions and relative hypotheses…………………………………………………..16 Table 2 Categories of population and numbers of each category……………………………21 Table 3 Descriptive statistics - the enforcement date of new CSR regulations……………...24 Table 4 T test - the enforcement date of new CSR regulations………………………………27 Table 5 Regressions of the enforcement of CSR regulations on stock return - 180 transaction days before and after the enforcement date of new CSR regulations………………31. 立. 政治大. Table 6 Descriptive statistics - the filing date of CSR report of each company……………..35. ‧ 國. 學. Table 7 T test - filing date of CSR report of each company…………………………………38. ‧. sit. y. Nat. Table 8 Regressions of CSR disclosure on stock return - 180 days before and after the filing date of CSR report of each company……………………………………………….41. n. al. er. io. Table 9 Results of the hypotheses under corresponding questions…......................................44. Ch. i n U. v. Table 10 The control variables and the corresponding codes and calculations and sources...52. engchi. III.

(5) A Table of Contents – Figures Figure 1 Classification of listed companies…………………………………………………...9 Figure 2 Part 1 database selected period – 180 transaction days before and after the enforcement date of the new regulations of CSR report…………………………..11 Figure 3 Part 2 database selected period – 180 transaction days before and after the filing date of CSR report of each company……………………………………………...11. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. IV. i n U. v.

(6) I. Introduction This study focuses on the effect of the enforcement of new CSR (corporate social responsibility) report regulations in Taiwan on stock returns of listed companies. On November 26, 2014, Taiwan’s government released the new regulations requiring certain corporations to submit annual CSR report. Since CSR report discloses nonfinancial information of a corporation, the new CSR regulations may therefore affect investors’ evaluation of corporations, which then affects their decision and directly reflects on stock return, the first indicator of market.. 立. 政治大. ‧ 國. 學. In La Porta, Lopez-de-Silanes, Shleifer, and Vishny’s article (1997), it mentions that a good legal environment will protect potential financiers, hence their willingness in investing. ‧. securities will increase, and eventually the capital markets will expand. If the new CSR. sit. y. Nat. regulations in Taiwan can improve the legal environment of market, the regulations may also. io. al. er. have direct impact on stock return. In Jizi, Nehme, and Salama’s research (2016), they mention that voluntary disclosure of CSR is appreciated by stock participants; also, according. n. v i n Ch to Wang, Chuang, and Xu’s finding (2016), voluntary information disclosure is positively engchi U related to firm’s equity. There are few studies regarding mandatory nonfinancial disclosure in Taiwan. Therefore, the main purpose of this research is to shed light on the effect of CSR disclosure on corporate stock performances under legal enforcement in Taiwan.. The legal system of Taiwan’s market was not compulsory before. Financial Supervisory Commission R.O.C. (Taiwan) (FSC) issued a press release on September 18, 2014 announcing that an enterprise which is either operating livelihood-related business or have direct interaction with people or above certain scale of capital shall prepare CSR report.

(7) annually as a means of communication with its stakeholders. FSC intensified the disclosure of corporations’ non-financial information to the publics in order to boost the confidence of consumers and supply chain manufacturers and strengthen the internal governance of corporations at the same time. The detailed practices of the new regulation are released by Taiwan Stock Exchange (TWSE) on November 26 in the same year. The name of the regulation is called “Taiwan Stock Exchange Corporation Rules Governing the Preparation and Filing of Corporate Social. 政治大 OTC (over the counter) company under one of the following circumstances shall prepare and 立. Responsibility Reports by TWSE Listed Companies1.” According to the rule, a listed or an. ‧ 國. 學. file a corporate social responsibility report in Chinese: 1. company that falls into food industry. ‧. 2. company whose revenue no less than 50% is derived from food and beverage. y. Nat. n. al. Ch. er. io. 4. company that falls into financial and insurance industry. sit. 3. company that falls into chemical industry. i n U. v. 5. company whose capital stock achieves no less than NT$10 billion. engchi. The rule also mentions that the listed company must disclose its corporate social responsibility report and link it to the online system of TWSE by June 30. However, with some conditions, the listed company may complete the filing by December 312. In account of the new regulations, the number of the disclosure of CSR reports of year 2014 surged in 2015. The quantity of CSR reports of listed companies is 267, comparing to 1. Data Source: Taiwan Stock Exchange - Rules & Regulations Directory If the listed company does not prepare a corporate social responsibility report in the most recent year or does not prepare the report by referring to the GRI Guidelines, or the corporate social responsibility report has obtained a CPA's letter of opinion according to the rules mentioned in the preceding paragraph, the filing may be completed by December 31.. 2. 2.

(8) 171 reports of year 2013 is commendable. The quantity of CSR reports of OTC companies also increased from 41 to 75. The total amount of listed and OTC companies hiked from 212 to 342. With the surge of CSR report, TWSE amended the rules to an extent the requirement of the listed companies on October 19, 2015. From 2017 onwards, besides the company that falls into food industry, chemical industry, financial and insurance industry, the company whose revenue no less than 50% is derived from food and beverage and the company whose. 政治大 report of 2016, the company whose capital stock has achieved no less than NT$5 billion but 立. capital stock is no less than NT$10 billion shall disclose a corporate social responsibility. ‧ 國. 學. no more than NT$10 billion shall comply with the rule as well. In the following study, the previous rules released on November 26, 2014 are adopted, which means that the company. ‧. whose capital stock has achieved no less than NT$5 billion but no more than NT$10 billion is. sit. y. Nat. disregarded as the observed object.. n. al. er. io. However, it is noteworthy that since the new CSR regulations only restrict certain. i n U. v. corporations, the effectiveness of the regulations on the noncompulsory corporations may be. Ch. engchi. different. Furthermore, before the new regulations released, there are already some companies file CSR report spontaneously. According to Dhaliwal, Radhakrishnan, Tsang, and Yang’s study (2012), the issuance of stand-alone CSR reports is related with lower analyst forecast error. With lower analyst forecast error, corporations have lower cost of equity, which implies higher stock price. Since the companies which have issued CSR report before the regulations may have already benefited from the disclosure behavior, the new CSR regulations may have different influences on these companies and other companies.. 3.

(9) This study contributes to the influences of the rules of CSR disclosure on Taiwan’s stock market. Since CSR used to be the activities practiced voluntarily by companies, it is the first time to make CSR into enforcement in Taiwan. Therefore, there are few, if any, research related CSR regulations. The results of this study can examine the effectiveness of these regulations and be a reference basis for Taiwanese government to improve CSR related laws. Moreover, most of CSR associated studies concentrated on the effects on risk, forecast error, equity financing cost and volatility. This study, however, analyzes the angles directly from investors. Stock performance is the straightforward factor that investors consider; meanwhile,. 政治大. it is also the clear-cut indicator which shows investors’ preference. The study contributes to. 立. the influences of the rules of CSR disclosure on Taiwan’s stock market.. ‧ 國. 學. Since La Porta, Lopez-de-Silanes, Shleifer, and Vishny’s article (1997) has proved that. ‧. sound legal environment improves stock market, the new CSR regulations, which enforce. sit. y. Nat. corporations to disclose information, are supposed to benefit stock performance as well.. io. n. al. er. Therefore, the research questions of the study are as follows:. i n U. v. 1. Whether the new CSR regulations can bring positive effectiveness on companies’ stock performance?. Ch. engchi. 2. Whether the issuance (disclosure) of CSR report can bring positive effectiveness on companies’ stock performance?. II. Literature Review and Hypotheses Development Stock return reveals the valuation of investors; while the valuation of investors is affected by legal investor protection, which is necessary (La Porta, Lopez-De-Silanes, and Shleifer 2006). Demirguc-Kunt and Maksimovic (1998) find that an active stock market plus. 4.

(10) sound legal system improve corporation growth because these companies are available to obtain external funds. In Mclean, Zhang, and Zhao’s research (2012), it shares similar result that investment protection laws are not only positive related to access to external fund, they also improve efficient investment, and accurate share prices. For the countries which have poor investor protection, the valuation is also low (La Porta, Lopez-De-Silanes, Shleifer, and Vishny 2002); although mandatory dividend can work as alternative (La Porta, Lopez-De-Silanes, and Shleifer 1998), yet some studies (La Porta, Lopez-De-Silanes, Shleifer, and Vishny 2000; Shleifer and Wolfenzon 2002) indicate that higher dividend is. 政治大. followed by higher investor protection. In Grossman and Hart’s study (1980), it indicates that. 立. voluntary disclosure is favored by good companies since they want to show difference from. ‧ 國. 學. other competitors. This implies shareholders can obtain more information, which may increase the valuation. The function of law, moreover, can decrease the social transaction cost. ‧. caused by non-disclosure (Grossman and Hart 1980). But mandatory adherents of rules seem. y. Nat. io. sit. to attain better forecast accuracy than voluntary adherents (Horton, G. Serafeim, and I.. n. al. er. Serafeim 2010). Overall, to conclude the viewpoints above, the new regulations protecting. i n U. v. shareholders are supposed to have positive effectiveness on stock market.. Ch. engchi. The studies above demonstrate the definite influences of laws on security market and the necessity of legal system; Hope (2003) even points out that strong enforcement enhances higher forecast accuracy, which will raise shareholders’ valuation on firms. Hermalin and Weisbach (2012), however, have the opposite findings. They argue that disclosure will increase agency problem and related costs, including CEO compensation and CEO turnover rates; besides, to follow stricter disclosure rules, shareholders should pay higher managerial compensation, which will then decrease the valuation of stock. Nevertheless, there are still others support the positive relationship between disclosure and the valuation of stock. The 5.

(11) study conducted by Lundholm and Myers (2002) found that increasing disclosure will enhance the influence of future earnings on current returns. Further discussion pointed out that firms with greater external financing demands have higher voluntary disclosure levels, which will later lead to a lower financing cost (Francis, Khurana, and Pereira, 2005). The most surprising discovery in this study is that firm-level voluntary disclosure incentives seem to be applicable to all places regardless a country’s legal and financial system. Moreover, the result also implies the importance of voluntary disclosure in improving financial performance through lowering cost of capital.. 政治大 Above references discuss about the relationship between financial disclosure and 立. ‧ 國. 學. financial performance. Yet, for the nonfinancial disclosure, Dhaliwal, Radhakrishnan, Tsang, and Yang (2012) indicate that issuing CSR reports is positively associated with lower analyst. ‧. forecast error, which will increase valuation of shareholders; the relationship is particularly. sit. y. Nat. stronger in the countries that are more stakeholder-oriented and for the firms with more. io. al. er. ambiguous financial disclosure. Thus, based on all the arguments above, the new CSR report. n. regulations in Taiwan are assumed to enhance valuation of investors.. Ch. engchi. i n U. v. When focusing exclusively on the influences of CSR on CFP, there are still numerous studies upholding the positive relationship between CSR and CFP. It is originated earliest from the book of Freeman (1984); wherein he pointed out that a firm that practices CSR activities is able to manage good relationships with its stakeholders, which is beneficial for its financial performance. El Ghoul, Guedhami, Kwok, and Mishra (2011) found that firms with higher CSR scores will display cheaper equity financing cost, especially in improving such CSR activities as employees’ relationship, environment policies, and product strategies.. 6.

(12) Furthermore, the evidence found in bank industry by Simpson and Kohers (2002) verifies that social and financial performances are positively connected. CSR has significant risks, which are relevant to valuation of shareholders as well. Kim, H. Li, and S. Li (2014) observed that CSR performance is negatively associated with crash risk, especially in the firms with less effective corporate governance or a lower level of institutional ownership. Additionally, firms with higher CSP have lower firm-idiosyncratic risk as well (Luo and Bhattacharya, 2009). CSR is clearly connected to idiosyncratic risk, but. 政治大 idiosyncratic risk of firms while negative CSR increase so. It is noticeable that when firms 立 not all of them reduce the risk. Mishra and Modi (2012) found that positive CSR will reduce. ‧ 國. 學. with higher financial leverage, positive CSR may not reduce idiosyncratic risk. The other findings by Lee and Faff (2009) in contrast show that leading CSP results in lower. ‧. idiosyncratic risk while lagging CSP results in higher idiosyncratic risk.. y. Nat. sit. However, CSR may have different effects on different types of companies. For instance,. n. al. er. io. Hull and Rothenberg (2008) discovered that CSP has greater influence on the companies with. i n U. v. lower innovation and in the industries with little differentiation. Another interesting study. Ch. engchi. (Servaes and Tamayo 2013) revealed that CSR can enhance firm value for those with higher customer awareness; for those with low customer awareness, on the contrary, the effect of CSR is either negative or nonrelative. Therefore, CSR activities have different degree of effect in different time. For example, the theory developed by A. Mackey, T. B. Mackey, and Barney (2007) showed that managers in publicly traded firms might fund CSR activities that do not maximize the present value of their firm’s future cash flows but maximize the market value instead. The inferences above could explain the different effects of new CSR report regulations on different groups in this study.. 7.

(13) The effects of CSR on firms have diverse outcomes probably because of different sorts of CSR. Hillman and Keim (2001) suggested that CSR activities related stakeholder management improves shareholder value; while activities regarding social issue participation are negatively related to the shareholder value. One research corroborates the opposite i.e. participation in institutional CSR activities which focused on stakeholders and society provides an “insurance-like” benefit; while participation in technical CSRs which aimed at firm’s trading partners generates no such advantages (Godfrey, Merrill, and Hansen, 2009). Another research focuses on the dissimilar results of announcement of positive and negative. 政治大. CSR toward abnormal stock returns. Huang, Wang, and Chang (2013) revealed empirical. 立. evidence that positive (negative) CSR announcement has significantly associated with. ‧ 國. 學. positively (negatively) abnormal stock returns; the impact of negative CSR events is stronger than that of positive one. The discovery of the research conducted by Khanna, Quimio, and. ‧. Bojilova (1998) is of consistency that the disclosure of negative CSR information like Toxic. y. Nat. io. sit. Release Inventory has negative link to stock market returns. Barnett and Salonmon (2012) on. n. al. er. the other hand hypothesize that the corporate social performance to corporate financial. i n U. v. performance (CSP-CFP) is U-shaped; the consequence turned out to meet the hypothesis that. Ch. engchi. firms with low CSP have higher CFP than moderate CSP, but firms with high CSP have highest CFP instead. Martin (2002) introduced a tool called the virtue matrix, to categorize CSR activities into four types, which are Strategic, Structure, Choice, and Compliance3; each type of CSR activities is based on different intentions and may contribute to different financial performances. The studies above imply that the contents of CSR report may cause different results on corporations, which will influence the valuation of investors. 3. A CSR activity counted in Strategic means that it is beneficial to society and environment, but corporations should try to make it beneficial to shareholders as well; when an activity belongs to Structure, it benefits society but damages shareholders’ advantages; if a corporation practice CSR for increasing its self-interest, the activity belongs to Choice; while if a corporation fulfills CSR due to law or regulations, the activity is regarded as Compliance. 8.

(14) Group 1: Forerunners Voluntary before the rules. Compulsory Group 2: RuleFollowers Non-voluntary before the rules. Listed Companies (2014). Group 3A: Saints Voluntary before and after the rules. Group 3: Voluntary before the rules. 立. Group 3B: Fake Saints Voluntary only before the rules. 政治大. Noncompulsory. ‧ 國. 學. Group 4A: RuleSurfers Voluntary only after the rules. ‧. Group 4: Non-voluntary before the rules. sit. y. Nat. Group 4B: Rebels Non-voluntary before and after the rules. n. al. er. io. Figure 1. Classification of listed companies. The listed companies are classified into six groups, which are “Group 1: Forerunners,” who are compulsory and voluntary before the rules, “Group 2: Rule-Followers,” who are compulsory but non-voluntary before the rules, “Group 3A: Saints,” who are noncompulsory but voluntary before and after the rules, “Group 3B: Fake Saints,” who are noncompulsory and voluntary only before the rules, “Group 4A: Rule-Surfers,” who are noncompulsory but voluntary only after the rules, and “Group 4B: Rebels,” who are noncompulsory and non-voluntary before and after the rules.. Ch. engchi. i n U. v. Source: This study and the website of Taiwan Stock Exchange. In view of the circumstances above, all observed listed companies are divided into six groups. The company which is under one of the conditions of the regulations mentioned above is regarded as the sample of “compulsory group.” The rest of the listed companies are control group called “noncompulsory group.” Next, in each group there are “voluntary group”. 9.

(15) and non-voluntary group” shown in Figure 1. The companies which are in the voluntary group under compulsory group, namely voluntary before rules, are counted in Group 1: Forerunners. On the contrary, the companies which are in the non-voluntary group under compulsory group belong to Group 2: Rule-Followers. As for the voluntary group under noncompulsory group, however, there are two kinds of companies: one is spontaneously filing CSR report before the release of the new regulations; they are regarded as Group 3A: Saints, which are voluntary before and after the rules, and Group 3B: Fake Saints, which are only voluntary before the rules, namely as they noticed that they are not under obligation of. 政治大. filing CSR report, they stopped issuing CSR report. 4 The other kind of companies is. 立. voluntary filing CSR report after the release of the new rules; they are called Group 4A:. ‧ 國. 學. Rule-Surfers. They file CSR report since they find that it becomes a trend of issuing CSR report. Last, the companies which belong to non-voluntary group under noncompulsory. ‧. group are named Group 4B: Rebels.. sit. y. Nat. io. er. It is also notable that the filing date of each company may be different, which means the. al. content of CSR report is open to the public on different days. Therefore, two different. n. v i n C hdates are developed: assumptions based on different event e n g c h i U For the first one, say Part 1, the. base date is the release date of the regulation. Part 2, on the other hand, the base date is the filing date of CSR report of each company. Under Part 1, investors only care about whether a company files its CSR report regardless the content and the quality. Therefore, the stock market will reflect as soon as the new rules are announced. As for part 2, it emphasizes the actual content in CSR reports. Thus, a company’s stock price will increase after the CSR report is available to the public.. 4. Since the numbers of Group 3B: Fake Saints are quite few, this group is just a figurehead. Therefore, it will not be discussed in the following. 10.

(16) Part 1. -180 transaction days. 2014.03.11. +180 transaction days. 2015.08.21. 2014.11.26 The release date of new regulations. Figure 2. Part 1 database selected period – 180 transaction days before and after the enforcement date of the new regulations of CSR report Source: This study. 立. -180 transaction days. +180 transaction days. ‧. ‧ 國. 學. io. sit. y. Nat The filing date of CSR. er. Part 2. 政治大. n. a l report of each company v i n Ch e–n180 Figure 3. Part 2 database selected period h i U days before and after the filing g ctransaction date of CSR report of each company Source: This study In Figure 2, it shows the selected period of Part 1, which is 180 transaction days before and after the enforcement date of the new regulations of CSR report. The purpose is to examine whether the enforcement of CSR regulations has effectiveness on corporations’ stock performance. In Figure 3, on the other hand, it depicts the observed period of Part 2, which is 180 transaction days before and after the filing date of CSR report of each company. Noted that since the sample firms of Group 4B did not issue CSR reports, they are excluded. 11.

(17) in Part 2. The intention of this part is to explore the effect of the content of disclosure of CSR report on firms’ stock performance. Briefly, the given research examines the comprehensive influences combined with regulations, information disclosure, and corporate social responsibilities on financial performances of companies; with the knowledge base of the references above, we establish a relationship between regulation, between disclosure, and between CSR and CFP one-to-one; with the classification of listed companies and the observed period selection, the following. 政治大 effectiveness of new CSR report regulations on stock return, the hypotheses are developed 立. hypotheses are developed. Recalling that the core purpose of the study is to examine the. ‧ 國. 學. under two assumptions. In Part 1, since the shareholders only care about whether companies issue CSR report, the base date falls on the enforcement date of the new regulations, which is. ‧. November 26, 2014. Hypotheses from 1 to 6 are based on the enforcement date of the new. sit. y. Nat. regulations. For Part 2, since the assumption is that shareholders care more about the content. io. al. er. of CSR report, the base date is the filing date of CSR report of each company. Below are the. n. hypotheses; the hypotheses from 1 to 6 belong to Part 1, and for Part 2 the hypotheses are from 7 to 9.. Ch. engchi. i n U. v. Hypothesis 1: After the release of the rules, all companies will have better stock performance.5. The new CSR regulations will decrease information asymmetry and therefore enhance the efficiency of the market and provide better legal environment for investors. When the market becomes much sound, potential investors will emerge and the stock market will expand.. 5. This means the whole market will have better stock performance after the release of the rules than before the release of the rules. 12.

(18) Hypothesis 2: After the release of the rules, among forced companies, those voluntarily filed CSR reports do not have better stock performance.6. After the release of the rules, the behavior of filing CSR reports of forced companies shifts from “Strategic” to “Compliance,” which makes the companies who voluntarily practiced gained nothing while the companies which never intended to volunteer benefited from it because of the positive side-effect of the new regulations.. 政治大 CSR reports before but are now compulsory to do so have better stock performance. 立. Hypothesis 3: After the release of the rules, companies which did not voluntarily file 7. ‧ 國. 學. According to the research above, regulations have positive influences on stock market. In this situation, investors will gain much more information of these companies due to the. ‧. new CSR regulations. When the information asymmetry decreases, the stock will perform. n. al. er. io. sit. y. Nat. better.. i n U. v. Hypothesis 4: After the release of the rules, noncompulsory companies that voluntarily. Ch. engchi. filed CSR reports before and after the rules may have better or irrelevant stock performance.8. Since these companies are not compulsory, they are affected simply by the new rules, which are thought to be beneficial to the whole market. According to the research result of Horton, G. Serafeim, and I. Serafeim (2010), however, mandatory regulation adopters will improve their financial performances, which will then improve the valuation of shareholders;. 6 7 8. This means Group 1 does not have better performance after the release of the rules than before. This indicates Group 2 has better stock performance after the release of the rules than before. This indicates the release of the rules may have positive or irrelevant relationship with Group 3A. 13.

(19) while for the noncompulsory adopters, the effects are not robust. So the effect of new CSR regulations on noncompulsory but voluntary corporations may be positive or irrelevant.. Hypothesis 5: After the release of the rules, noncompulsory companies that did not voluntarily file CSR reports before but are voluntary after the rules may have better or irrelevant stock performance.9. The situation is similar as the former hypothesis. Since these companies are not. 政治大 the whole market. According to the research result of Horton, G. Serafeim, and I. Serafeim 立. compulsory, they are affected simply by the new rules, which are thought to be beneficial to. (2010), mandatory regulation adopters will improve their financial performances, which will. ‧ 國. 學. then improve the valuation of shareholders; while for the noncompulsory adopters, the effects. ‧. are not robust. So the effect of new CSR regulations on noncompulsory but voluntary. io. sit. y. Nat. corporations may be positive or irrelevant.. n. al. er. Hypothesis 6: After the release of the rules, noncompulsory companies that do not. i n U. v. voluntarily file CSR reports before and after the rules have better stock performance.10. Ch. engchi. The companies which did not submit a CSR report are affected by the new rules, which are thought to be beneficial to the whole market. Moreover, according to Barnett and Salomon’s study (2012), CSP-CFP is U-shaped, which means that firms with low CSP have higher CFP than moderate CSP, so among noncompulsory corporations, those who do not issue CSR report will have better stock performance than those who issue a CSR report.. Hypothesis 7: All companies will have better stock performances after filing CSR 9 10. This indicates the release of the rules may have positive or irrelevant relationship with Group 4A. This indicates Group 4B has better stock performance after the release of the rules than before. 14.

(20) reports.11. Dhaliwal, Radhakrishnan, Tsang, and Yang (2012) indicate that issuing CSR reports is positively associated with lower analyst forecast error, which will increase valuation of shareholders. Thus, companies that issue CSR reports are thought to have better financial performance.. Hypothesis 8: The mandatory companies have better stock performance after filing CSR reports.12. 立. 政治大. Mandatory companies are regarded as ones with higher customer awareness. According. ‧ 國. 學. to the study conducted by Servaes and Tamayo (2013), CSR can enhance firm value especially for these companies. As a result, filing CSR report is inferenced to improve. ‧. corporate financial performance. sit. y. Nat. al. er. io. Hypothesis 9: The noncompulsory companies may have worse or irrelevant stock. n. performance after filing CSR reports.13. Ch. engchi. i n U. v. On the contrary, the noncompulsory companies are regarded as ones with lower customer awareness. CSR, however, will be irrelevant or even negative for these firms (Servaes and Tamayo, 2013). Furthermore, disclosure will increase agency problem and related costs (Hermalin and Weisbach, 2012), which may also be the reason why CSR has negative influences on these corporations’ stock performance. Hence, the effect of issuance of. 11. This means the whole market except Group 4B will have better stock performance after the issuance of CSR report. 12 This means Group 1 and Group 2 have better performance after filing CSR reports. 13 This indicates Group 3A and Group 4A may or may not have worse stock performance after filing CSR reports. 15.

(21) CSR report on these firms may be irrelevant or negative. The hypotheses and the corresponding research questions are as below. Table 1 Questions and relative hypotheses Questions. Hypotheses. 1. Whether the new CSR regulations can bring positive effectiveness on. Hypothesis 1: After the release of the rules, all companies will have better stock performance.. companies’ stock performance?. companies, those voluntarily filed CSR reports do not have better stock performance.. Hypothesis 2: After the release of the rules, among forced. 政治大. Hypothesis 3: After the release of the rules, companies which did not voluntarily file CSR reports before but are now compulsory to do so have better stock performance.. 立. ‧. ‧ 國. 學. Hypothesis 4: After the release of the rules, noncompulsory companies that voluntarily filed CSR reports before and after the rules may have better or irrelevant stock performance. Hypothesis 5: After the release of the rules, noncompulsory. y. sit. io. er. Nat. companies that did not voluntarily file CSR reports before but are voluntary after the rules may have better or irrelevant stock performance.. n. aHypothesis l C 6: After the releasenofi vthe rules, companies that do not voluntarily h e n g cfilehCSR i Ureports before and are now not voluntary and noncompulsory to do so has better stock performance.. 2. Whether the issuance (disclosure) of CSR report. Hypothesis 7: All companies will have better stock performances after filing CSR reports.. can bring positive effectiveness on companies’ stock performance?. Hypothesis 8: The mandatory companies have better stock performance after filing CSR reports. Hypothesis 9: The noncompulsory companies may have worse or irrelevant stock performance after filing CSR reports.. Source: This study. 16.

(22) III. Data and Methodology The Research method is divided into three parts: 1. DATA Description; 2. Measurement of Variables; and 3. Empirical Model Setting.. 1. Data Description In the research, the database used is “TSE/OTC Adjusted Price (Daily)- Excluding Right and Dividend” from Taiwan Economic Journal (TEJ). This database only incorporates the. 政治大. data of the companies that are now listing, namely the data of the companies that have already delisted is excluded.. 立. ‧ 國. 學. The list of companies that shall submit CSR report Year 2014 and the previous CSR. n. al. er. io. sit. y. Nat. 1.1 Sample Selection. ‧. reports of corporations are collected at Market Observation Post System (MOPS) of TWSE.. i n U. v. According to the database of TEJ, the population of part 1 includes 787 observations of. Ch. engchi. all listed companies in Taiwan Stock Exchange from March 11, 2014 to August 21, 2015. 14 The population of part 2 is almost the same as that of part 1, except that since the companies of Group 4B do not file their CSR reports, there is no base date of them. Therefore, Group 4B is unqualified for the research of part 2. In part 2, the number of samples is 259.. 1.2 Sample Period Selection. There are two parts of observed period adopted in this study. The first part of the period 14. TDR and F-shares are not included. The company that has terminated listing before April 20, 2017 is not included as well. 17.

(23) of observation is determined by 180 transaction days before and after the release of the new regulations, which falls on November 26, 2014.15 The hypothesis of this setting of the period is that investors will care only about whether a company files its CSR report instead of the content inside. On the contrary, the other part of the period of observation is calculated depending on each company’s filing day of the CSR reports of year 2014. The total observation period will be 180 days before and after the filing day.16 In the study of Waddock and Graves (1997), corporates social performance (CSP) is. 政治大 future financial performance. Hence both before and after 180 days of the reference dates are 立. proved to be positively related to prior financial performance and positively associated with. ‧ 國. 學. selected in this study.. ‧. 1.3 Population Classification. Nat. sit. y. These corporations are divided into four groups. As mentioned in the introduction, all. n. al. er. io. the listed companies are divided into “compulsory group” and “noncompulsory group” at the. i n U. v. beginning. Afterwards, each group is split into “voluntary group” and “non-voluntary group.”. Ch. engchi. The “voluntary group” under “compulsory group” is Group 1: Forerunners, namely “voluntary before compulsory.” The “non-voluntary group” under “compulsory group” is Group 2: Rule-Followers, namely “non-voluntary before compulsory.” The “voluntary group” under “noncompulsory group” is Group 3, namely “voluntary before the rules.” The “non-voluntary group” under “noncompulsory group” is Group 4, namely “non-voluntary before the rules.” In Group 3, it can be split into Group 3A: Saints, the company voluntary before and after the release of new regulations, and Group 3B: Fake Saints, the company. 15 16. November 26, 2014 is counted in the later period. The filing day of each company is counted in the later period. 18.

(24) only voluntary before the rules. It is noteworthy that the numbers of this group are quite few; from a statistical point of view, insufficient sample size can lead to an increase in statistical inference uncertainty; therefore it will not be discussed in the following. Group 4 can also be divided into Group 4A: Rule-Surfers, the company voluntary only after the release of new regulations, and Group 4B: Rebels, namely non-voluntary before and after the rules. The conditions of each group are mentioned as follows.. Group 1: Forerunners—Voluntary before compulsory. 政治大 This group includes the company voluntary to disclose CSR reports already before the 立. law and is obligated after the law. In the sample, the company has filed CSR report of year. ‧ 國. 學. 2013.17 There are 163 companies are obligated to prepare and file CSR reports of 2014.. ‧. Among them, there are 67 companies that have already filed CSR reports of year 2013. As a. n. al. Ch. engchi. er. io. Group 2: Rule-Followers—Non-voluntary before compulsory. sit. y. Nat. result, there are 67 companies in Group 1.. i n U. v. This group includes the company which does not disclose CSR report before the law and is obligated after the law. The number will be 163 of the total compelled listed companies deducts 67 of Group 1. Therefore, there are 96 companies in Group 2. In Group 2, it is noteworthy that under the rules, few companies perform even better as predicted. Some companies file not only CSR reports of year 2014, but also prepare the previous year(s) of CSR reports. Some even provide English version of CSR reports although it is not necessary. This is consistent with the results of the research which Kagan and 17. Even the CSR report is filed after the release date of new regulations, the company that filed CSR report of year 2013 is counted as voluntary. 19.

(25) Thornton (2003) conducted that companies have the tendency to go “beyond compliance” with tightening regulatory requirements.. Group 3A: Saints—Voluntary before and after the release of new regulations. This group includes the one voluntary to disclose CSR report already before the law and is not obligated after the law. The total companies filing CSR report of year 2013 are 125. The total companies of Group 1 are 67. Still, two firms are excluded since they do not have. 政治大. enough transaction days to observe. Thus, the companies of firms in Group 3A are 57.18. 立. Group 4A: Rule-Surfers—Voluntary only after the release of new regulations. ‧ 國. 學. Group 4A incorporates the one voluntary to disclose CSR report after the release of the. ‧. rules but is not obligated after the law. The total companies filing CSR report of year 2014. sit. y. Nat. are 252. After excluding the compulsory companies, namely Group 1 and Group 2, and the. al. n. companies in Group 3A, the companies of Group 4A are 39.19. Ch. engchi. er. io. companies that have already been voluntary before the release of new regulations, namely the. i n U. v. Group 4B: Rebels—Non-voluntary before and after the rules. The group is inclusive of the one which does not disclose CSR report before or after the law nor is obligated after the law. The total number of listed companies excluding the firms which have filed CSR reports of year 2013 or 2014 will be the total number of this group. That is, all listed companies deduct those of Group 1, Group 2, Group 3A, and Group 4A. 18. Firm 2029, 2062, 2369, 3266, 3669, 4119, 4919, 4960, 6166, 8271, and 9911 filed CSR report of year 2013 but did not file CSR report of year 2014. They are still counted in Group 3A. 19 The companies which have already suspended or have not listed in the observed period are not included. In Group 3A, there are 11 companies did not file CSR report of year 2014. 20.

(26) Nevertheless, few companies who suspend listing during this period are removed from the observed objects. Therefore, there are 528 firms in Group 4B.. Table 2 Categories of population and numbers of each category Categories. Numbers of firms. All listed companies. 787. Group1. 67. Group2. 96. Group3A. 57. Group4A. 39. 528 政治大 Source: This study and Taiwan Economic Journal (TEJ) database 立 Group4B. ‧ 國. 學. 2.. Measurement of Variables. ‧. Since the preference of investors will have a direct and immediate impact on the stock. sit. y. Nat. market, the dependent variable used in this study is “stock return.” In case of control. io. er. variables, Atiase (1985) indicates that capitalized value should be control when requiring. al. v i n C as one of the control value, volume (Volume) is adopted h e n g c h i U variables as well (Kyle, 1985; n. control for disclosure information; and the study adopts market value (MV). Besides market. Glosten, 1988). According to Capital Asset Pricing Model (CAPM)20, beta value measures the risk of stock, which may influence investors’ valuation; thus, beta value counted by the return of last year (BETA_1Y) is used. Furthermore, referring to the results of Rajgopal, S., and M. Venkatachalam (2011), the study also employs ratio of leverage ratio (LeverageR), return on equity (ROE), cash flow ratio (CashFlowR), and market return (MKT_R) as control variables. Additionally, turnover (Turnover) and current asset to total asset (CATA) are also used since they are the ratios that investors will consider. 20. CAPM is a model used to estimate the appropriate rate of return and the corresponding risk of a security. 21.

(27) 3.. Empirical Model Setting. Analytically, the model established to estimate the return on investment of each firm is as follow:. 0. 1 | | ∑ 𝑅𝑖,𝑗 , 𝑑 𝑗=𝑑 𝑑−1. 1 ∑ 𝑅𝑖,𝑗 , 𝑑. ,. (1). 學. 𝐴𝑅𝑖,𝑗 =. d<0. (2). d≥0. { 政治大 𝑗=0. 𝐶𝐿𝑂𝑆𝐸𝑖,𝑗 − 𝐶𝐿𝑂𝑆𝐸𝑖,𝑗−1 , 𝐶𝐿𝑂𝑆𝐸𝑖,𝑗−1. th. transaction day of stock i,. sit. Nat. where 𝐴𝑅𝑖,𝑗 represents the average return on investment for j. ‧. ‧ 國. 𝑅𝑖,𝑗 =. y. 立. where. io. er. 𝑅𝑖,𝑗 represents the return on investment for j th transaction day of stock i, d represents the day. al. before or after the enforcement date of new CSR regulations, and i represents the stock of. n. v i n C h of each group isUthe average of ARs in each group. listed company. The return on investment engchi. For instance, in part 1, each company of those 67 companies in Group 1 will yield an average rate of 180 days before the release day of the new rules; Group 1 will then yield a mean by these 67 average rates as its average rate of 180 days before. Group 2, Group 3A, Group 4A, and Group 4B are calculated by the same approach. For the average rate of 180 days after the release day of the new rules are the same. According to the references above, the empirical model in part 1 is constructed as follows:. 22.

(28) 𝑅𝑖,𝑗 = 𝛽0 + 𝛽1 𝑅𝐸𝐺𝑈𝑖,𝑗 + 𝛽2 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 + 𝛽3 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 + 𝛽4 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 + 𝛽5 𝐺𝑟𝑜𝑢𝑝4𝐴𝑖,𝑗 + 𝛽6 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 + 𝛽7 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 + 𝛽8 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 + 𝛽9 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝4𝐴𝑖,𝑗 + 𝛿𝑋𝑖,𝑡 + 𝜀𝑖,𝑡 (3) In formula (3), 𝑅𝑖,𝑗 represents the return on investment for j th transaction day of stock i, while 𝑅𝐸𝐺𝑈𝑖,𝑗 represents the dummy of the enforcement of new CSR regulations for j. th. transaction day of i th stock. 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 is the dummy of Group 1, if 𝑅𝑖,𝑗 is from the stock in Group 1, then 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 = 1 , else 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 = 0 ; 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 , 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 , and 𝐺𝑟𝑜𝑢𝑝4𝐴𝑖,𝑗 are the dummies of Group 2, Group 3A, and Group 4A consequently.. 政治大 In part 2, the method is the same when calculating the average rates, but the base date 立. ‧ 國. 學. will be the filing day of CSR report of each company. That is, formula (1) and formula (2) are the same in part 2, but d here represents the day before or after the filing date of CSR report. ‧. of each company.. y. Nat. io. sit. Since Group 4B is not included in part 2, the empirical model is modified as follow:. n. al. er. 𝑅𝑖,𝑗 = 𝛽0 + 𝛽1 𝑅𝐸𝑃𝑂𝑖,𝑗 + 𝛽2 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 + 𝛽3 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 + 𝛽4 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 + 𝛽5 𝑅𝐸𝑃𝑂𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 + 𝛽6 𝑅𝐸𝑃𝑂𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 + 𝛽7 𝑅𝐸𝑃𝑂𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 + 𝛿𝑋𝑖,𝑡 + 𝜀𝑖,𝑡. Ch. engchi. i n U. v. (4). In formula (4), 𝑅𝑖,𝑗 represents the return on investment for j th transaction day of stock i, while 𝑅𝐸𝑃𝑂𝑖,𝑗 represents the dummy of the filing date of CSR report for j th transaction day of i th stock. 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 is the dummy of Group 1, if 𝑅𝑖,𝑗 is from the stock in Group 1, then 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 = 1 , else 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 = 0 ; 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 , and 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 are the dummies of Group 2 and Group 3A subsequently.. 23.

(29) IV. Empirical Results Due to the different time point division of this study, the empirical results are composed of two parts: Part 1 is on the subject of the enforcement date of new CSR regulations; while part 2 is regarding the filing date of CSR report of each company.. 1. Part 1—the enforcement date of new CSR regulations. Table 3 presents descriptive statistics of 787 listed companies with the time period of. 政治大. 180 days before and after the enforcement date of new CSR regulations. In panel A we can. 立. see that the average rate of return (Return) of all observed samples is -0.00053; the average. ‧ 國. 學. trading volume (Volume) is 2583.17; the average turnover rate (Turnover) is 0.0049; the average market value in million (MVmillion) is 29698.5; the average market value (MV) is. ‧. 8.8703; the average ratio of current asset to total asset (CATA) is 0.5717; the average. y. Nat. al. er. io. the average cash flow ratio (CashFlowR) is 32.8861.. sit. leverage ratio (LeverageR) is 0.1119; the average return on equity (ROE) is 9.3736; and that. n. v i n In panel B, C, D, E, and F, it C shows Group 1 decreases the least (-0.00018) on its h ethat ngchi U average rate of return (Return) while Group 3A decreases the most (-0.00062). For average trading volume (Volume), Group 1 has the highest (13635.31) figure while Group 4B has the lowest (1137.99) one. When it comes to turnover ratio (Turnover), Group 4A has the highest ratio (0.0056) whereas Group 2 has only 0.0027.. Table 3 Descriptive statistics - the enforcement date of new CSR regulations This table includes descriptive statistics of variables applied in this paper. The sample is composed of 787 listed companies from March 11, 2014 to August 21, 2015. Panel A displays the data of overall sample; panel B, panel C, panel D, panel E, and panel F display the data of 67 companies in Group 1, 96 companies in Group 2,. 24.

(30) 57 companies in Group 3A, 39 companies in Group 4A; and 528 companies in Group 4B respectively. The variables include rate of return (Return), trading volume (Volume), turnover rate (Turnover), market value in million (MVmillion), market value (MV), ratio of current asset to total asset (CATA), leverage ratio (LeverageR), return on equity (ROE), and cash flow ratio (CashFlowR). The calculation basis of Return, Volume, and Turnover is the daily value of each stock in the sample period. The calculation basis of the other control variables is the value of each stock in the end of 2013.. Panel A All listed companies Q1. -0.00053 2583.17. 0.0196 7972.07. -0.1875 0. -0.0088 174. 0 563. 0.0049 29698.5 8.8703 0.5717. 0.0095 125452.8 1.4325 0.2123. 0 121 4.7958 0.0531. 0.00081 2679 7.8932 0.4247. 0.0019 6062 8.7098 0.5776. 0.0049 0.2972 14529 2735469 9.5839 14.8218 0.7388 0.9960. 3.11 5.53. 0.0781 8.96 21.63. 0.1651 0.9575 15.12 250.76 44.76 1558.96. 政治大 0 0.0284 立0.1126. 0.1119 9.3736 32.8861. 17.8333 112.0297. -85.55 -1062.9. Panel B Group 1 Min. Nat. io. -0.00018 0.0161 Return 13635.31 21499.03 Volume 0.0030 0.0043 Turnover MVmillion 187371.24 361117.13 11.322616 1.2578 MV 0.4344 0.2136 CATA 0.1897 0.1275 LeverageR 10.1412 18.7321 ROE 42.5152 54.5889 CashFlowR. n. al. Ch. Q1. -0.1 0 0 8096 8.9991 0.1372 0.0054 -27.99 -7.61. Median. -0.0078 2413 0.00086 31551 10.3594 0.2690 0.1008 3.73 10.48. engchi. 0 6889.5 0.0016 89132 11.3979 0.3816 0.1772 9.71 27.055. y. Std.. sit. Mean. Median. ‧. Variable. Min. 學. Turnover MVmillion MV CATA LeverageR ROE CashFlowR. Std.. er. Return Volume. Mean. ‧ 國. Variable. i n U. v. Q3. Max. 0.0067 1922.5. 0.1 367103. Q3. Max. 0.0066 15962 0.0035 213431 12.2711 0.5270 0.2597 14.38 42.37. 0.0993 367103 0.1076 2735469 14.8218 0.9237 0.5521 137.36 208.05. Q3. Max. Panel C Group 2 Variable Return Volume Turnover MVmillion MV. Mean. Std.. -0.00044 0.0153 3391.98 7617.4 0.0027 0.0057 50696.72 134057.93 9.5785 1.5210. Min. Q1. -0.1 0 0 541 6.2934 25. -0.0071 203 0.00061 4918.5 8.5001. Median 0 1005 0.0013 15333 9.6375. 0.0054 3459.5 0.0027 34150.5 10.4385. 0.1 253299 0.2808 1051609 13.8658.

(31) CATA LeverageR ROE CashFlowR. 0.4982 0.1394 8.6535 33.2993. 0.1986 0.1056 10.5806 38.7700. 0.1130 0.00038 -23.26 -42.44. 0.3497 0.0472 2.77 8.3. 0.4697 0.1167 8.165 22.15. 0.6490 0.2190 14.015 56.38. 0.9194 0.4282 47.96 232.43. Panel D Group 3A Mean. Std.. Return Volume. -0.00062 1896.34. 0.0188 3261.25. -0.1 0. -0.0088 279. 0 728. 0.0071 2129. 0.1 54642. Turnover MVmillion MV CATA. 0.0052 0.0085 17691.89 30574.08 9.1933 0.9533 0.5828 0.1636 0.0884 0.0819 9.8551 10.0217 33.1853 33.9800. 0 2328 7.7528 0.1755 0.0012 -29.52 -24.96. 0.00097 5238 8.5637 0.4937 0.0203 3.57 8.8. 0.0024 7210 8.8832 0.5564 0.0609 10.37 26.83. 0.0058 15907 9.6745 0.7272 0.1395 16.04 50.57. 0.1851 199059 12.2014 0.8515 0.3078 33.85 125.69. Q3. Max. Panel E Group 4A Q1. io. -0.00039 0.0192 2174.66 3790.92 0.0056 0.0085 21728.03 27771.78 9.4169 1.0826 0.5742 0.2203 0.1042 0.0889 12.3841 11.9583 34.3944 47.3697. -0.1 0 0 1302 7.1717 0.0550 0.0028 -22.94 -106.11. -0.0093 362 0.0013 5357 8.5862 0.4149 0.0350 5.94 13.23. n. al. Ch. engchi. Median. 0 1057 0.0029 11688 9.3663 0.5723 0.0782 12.29 24.63. y. Min. sit. Std.. er. Mean. Median. 政治大. Nat. Return Volume Turnover MVmillion MV CATA LeverageR ROE CashFlowR. 立. Q1. ‧. Variable. Min. 學. LeverageR ROE CashFlowR. ‧ 國. Variable. i n U. v. Q3. 0.0077 2419.5 0.0065 24101 10.0900 0.7545 0.1442 21.02 50.34. Max. 0.1 111479 0.2570 126600 11.7488 0.8873 0.3888 36.86 210.9. Panel F Group 4B Variable Return Volume Turnover MVmillion MV. Mean. Std.. -0.00060 0.0208 1137.99 2355.79 0.0055 0.0106 8060.7 15751.14 8.3591 1.0601. Min. Q1. -0.1875 0 0 121 4.7958 26. -0.0092 139 0.00082 2080.5 7.6404. Median 0 388 0.0020 4437 8.3977. Q3. Max. 0.0069 1128 0.0054 8462.5 9.0434. 0.1 104408 0.2972 200993 12.2110.

(32) CATA LeverageR ROE CashFlowR. 0.5934 0.2117 0.1041 0.1137 8.9310 19.6190 31.4716 131.7328. 0.0531 0 -85.55 -1062.9. 0.4565 0.0244 2.02 1.845. 0.6062 0.0711 8.29 19.8. 0.7530 0.1544 14.965 44.105. 0.9960 0.9575 250.76 1558.96. Source: This study. Table 4 depicts the t test of the differences of the rate of return before and after the enforcement date of new CSR regulations. Panel A reports the difference between 5 days before and after the enforcement date of new CSR regulations in Group 1, Group 2, Group 3A, Group 4A, and Group 4B. Panel B, C, D, E, F, and G reports the difference between 10,. 政治大. 15, 30, 60, 90, and 180 transaction days before and after the enforcement date of new CSR. 立. regulations respectively. In panel A, all groups have significantly negative difference of. ‧ 國. 學. positive 5 days minus negative 5 days of the enforcement date of new CSR regulations. In panel B, on the contrary, they all have positive difference and except for Group 1 they have. ‧. all significant difference. In panel C and D, aside from the negative difference of Group 1, all. Nat. sit. y. groups have positive difference and most of them are significant. When it comes to panel E. n. al. er. io. and F, all groups have significantly positive difference. For panel G, it portrays a huge. i n U. v. opposite change that all groups have negative and significant difference.. Ch. engchi. Table 4 T test - the enforcement date of new CSR regulations This table includes rate of return in Group 1, Group 2, Group 3A, Group 4A, and Group 4B with 67, 96, 57, 39, and 528 companies respectively. The observed period of Panel A is 5 transaction days before and after the enforcement date of new CSR regulations in Group 1, Group 2, Group 3A, Group 4A, and Group 4B. The observed periods of Panel B, Panel C, Panel D, Panel E, Panel F, and Panel G, are 10, 15, 30, 60, 90, and 180 transaction days correspondingly. The differences in Table 1 are the average returns of positive days minus those of negative days.. Panel A The difference between 5 transaction days before and after the enforcement date of new CSR regulations -5 days. +5 days 27. Difference. t-value.

(33) 0.004461 0.002761 0.003091 0.006575 0.004635. Group 1 Group 2 Group 3A Group 4A Group 4B. -0.001745 -0.000463 0.001123 -0.000188 0.001660. -0.00621 -0.00322 -0.00197 -0.00676 -0.00297. -5.20*** -3.99*** -1.97* -4.64*** -5.67***. Panel B The difference between 10 transaction days before and after the enforcement date of new CSR regulations -10 days Group 1 Group 2 Group 3A Group 4A Group 4B. +10 days. Difference. t-value. 0.000225. 0.001292. 0.00107. 1.37. -0.000588 -0.002237 0.000123 -0.000636. 0.001170 0.003379 0.002409 0.003169. 0.00176 0.00562 0.00229 0.00381. 2.68*** 6.66*** 2.43** 9.88***. 立. 政治大. -15 days. -0.000120. -0.00033. 0.000047 -0.001054 0.000053 -0.000298. 0.000542 0.002326 0.001265 0.002404. 0.00050 0.00338 0.00121 0.00270. n. Ch. Panel D. sit. io. al. y. 0.000213. Nat. Group2 Group3A Group3B Group4. Difference. er. Group1. +15 days. ‧. ‧ 國. 學. Panel C The difference between 15 transaction days before and after the enforcement date of new CSR regulations. i n U. v. t-value -0.52 0.97 4.70 *** 1.53 9.12 ***. e n gbefore c h i and after the enforcement date of The difference between 30 transaction days new CSR regulations -30 days Group 1 Group 2 Group 3A Group 4A Group 4B. +30 days. Difference. t-value. 0.001168 0.000266. 0.001090 0.001180. -0.00008 0.00091. -0.16 2.17 **. -0.000765 0.000268 -0.000441. 0.002078 0.001605 0.001983. 0.00284 0.00134 0.00242. 5.03 *** 2.57 ** 11.45 ***. Panel E The difference between 60 transaction days before and after the enforcement date of new CSR regulations -60 days. +60 days 28. Difference. t-value.

(34) -0.000739 -0.001222 -0.001869 -0.001434 -0.001576. Group 1 Group 2 Group 3A Group 4A Group 4B. 0.001195 0.000840 0.001496 0.001174 0.001356. 0.00193 0.00206 0.00336 0.00261 0.00293. 6.80 *** 8.85 *** 8.32 *** 6.65 *** 21.46 ***. Panel F The difference between 90 transaction days before and after the enforcement date of new CSR regulations -90 days Group 1 Group 2 Group 3A Group 4A Group 4B. +90 days. Difference. t-value. -0.000422. 0.000880. 0.00130. 5.30 ***. -0.000859 -0.001178 -0.001048 -0.001051. 0.000602 0.001110 0.001005 0.001215. 0.00146 0.00229 0.00205 0.00227. 7.80 *** 9.07 *** 6.29 *** 18.69 ***. 立. 政治大. -180 days. -0.00122. -5.46 ***. -0.000083 -0.000158 -0.000042 -0.000246. -0.000791 -0.001084 -0.000741 -0.000947. -0.00071 -0.00093 -0.00070 -0.00070. -4.70 *** -4.29 *** -2.72 ** -7.51 ***. n. Ch. engchi. sit. io. al. y. -0.000787. *p < 0.10, **p < 0.05, ***p < 0.001. Source: This study. t-value. 0.000433. Nat. Group 2 Group 3A Group 4A Group 4B. Difference. er. Group 1. +180 days. ‧. ‧ 國. 學. Panel G The difference between 180 transaction days before and after the enforcement date of new CSR regulations. i n U. v. Table 5 depicts the regression of the enforcement of CSR regulations on stock return. The period employs 180 transaction days before and after the enforcement date of new CSR regulations. It includes sample of Group 1, Group 2, Group 3A, Group 4A, and Group 4B. In Panel A, Model 1 to Model 7 display the results without controlling other variables whereas in Panel B, Model 8 to Model 14 present the results with control variables such as volume (Volume), turnover (Turnover), beta value calculated by the return of the previous year of the enforcement date (BETA_1Y), market value (MV), ratio of current asset to total asset. 29.

(35) (CATA), leverage ratio (LeverageR), return on equity (ROE), cash flow ratio (CashFlowR), and market return (MKT_R). Model 1 shows the impact of the enforcement of new CSR regulations on stock return with a dummy variable “regulation (REGU).” Model 2 to Model 6 display the results of the enforcement of new CSR regulations with adding Group 1, Group 2, Group 3A, Group 4A, and Group 4B as dummy variables respectively. Model 7 reveals the influence of the enforcement of new CSR regulations by employing all variables (Group 4B is regarded as the constant term).. 立. 政治大. In Panel A, without controlling other variables, the coefficient of regulation (REGU) is. ‧ 國. 學. significantly negative. Under this condition, the enforcement of new CSR regulations makes the whole stock return of samples of listed companies decrease approximately 5 times. ‧. (0.00076/0.00015). In Model 2 to Model 6, the coefficients of regulation (REGU) are all. y. Nat. sit. significantly negative, which means when excluding Group 1, Group 2, Group 3A, Group 4A,. n. al. er. io. and Group 4B respectively, the stock returns of whole market decline due to the enforcement. i n U. v. of new CSR regulations as well. When only considering Group 1, overall, it has the highest. Ch. engchi. positive and significant performance (0.00064) comparing to other groups regardless of before or after the enforcement of new CSR regulations; but the interactive dummy (REGUGroup 1) in Model 2 is significantly negative (-0.00050). Group 1 decreases 143.53% ((-0.00072-0.0005)/(0.00021+0.00064)) after the enforcement of new CSR regulations. In Model 3 and Model 5, Group 2 and Group 4A have positive but not significant returns before and after the enforcement of new CSR regulations. In Model 4, Group 3A, on the other hand, has negative but insignificant figures. When it comes to Model 6, Group 4B has significantly negative performance (-0.000286) comparing to other groups no matter. 30.

(36) before or after the enforcement of new CSR regulations. The interactive dummy (REGUGroup 4B) is positive but insignificant (0.000186). In Model 7, nevertheless, considering all groups at the same time, the enforcement of new CSR regulations decreases the return of Group 4B by 0.0007007, and its total impact on Group 1 is -0.0012201 (REGU+REGUGroup 1). Model 8, as Model 1 does, shows the impact of the enforcement of new CSR regulations on stock return with dummy variable “regulation (REGU),” but it employs control variables. 政治大 new CSR regulations with adding Group 1, Group 2, Group 3A Group 4A, and Group 4B as 立 mentioned above. Model 9 to Model 13, similarly, depict the results of the enforcement of. ‧ 國. 學. dummy variables respectively and also with control variables. Model 14, on the other hand, reports the influence of the enforcement of new CSR regulations by employing all variables. ‧. (Group 4B is regarded as the constant term).. sit. y. Nat. n. al. er. io. Table 5 Regressions of the enforcement of CSR regulations on stock return - 180 transaction days before and after the enforcement date of new CSR regulations. Ch. i n U. v. This table reports the estimate of the following regression: 𝑅𝑖,𝑗 = 𝛽0 + 𝛽1 𝑅𝐸𝐺𝑈𝑖,𝑗 + 𝛽2 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 + 𝛽3 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 + 𝛽4 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 + 𝛽5 𝐺𝑟𝑜𝑢𝑝3𝐵𝑖,𝑗 + 𝛽6 𝐺𝑟𝑜𝑢𝑝4𝑖,𝑗 + 𝛽7 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 + 𝛽8 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 + 𝛽9 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 + 𝛽10 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝3𝐵𝑖,𝑗 + 𝛽11 𝑅𝐸𝐺𝑈𝑖,𝑗 × 𝐺𝑟𝑜𝑢𝑝4𝑖,𝑗 + 𝛿𝑋𝑖,𝑡 + 𝜀𝑖,𝑡 The sample is composed of 787 listed companies from March 11, 2014 to August 21, 2015. Panel A displays stock return under model 1 to 7 without control variables. 𝑅𝑖,𝑗 represents the return on investment for j th transaction day of stock i, while 𝑅𝐸𝐺𝑈𝑖,𝑗 represents the dummy of the enforcement of new CSR regulations for j th transaction day of i th stock, which implies if the stock return of a company falls on the day after the enforcement date of new CSR regulations, 𝑅𝐸𝐺𝑈𝑖,𝑗 will be 1; otherwise it will be 0. 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 is the dummy of Group 1, if 𝑅𝑖,𝑗 is from the stock in Group 1, then 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 = 1, else 𝐺𝑟𝑜𝑢𝑝1𝑖,𝑗 = 0; 𝐺𝑟𝑜𝑢𝑝2𝑖,𝑗 , 𝐺𝑟𝑜𝑢𝑝3𝐴𝑖,𝑗 , 𝐺𝑟𝑜𝑢𝑝3𝐵𝑖,𝑗 , and 𝐺𝑟𝑜𝑢𝑝4𝑖,𝑗 are the dummies of Group 2, Group 3A, Group 4A, and Group4 consequently. Panel B displays stock return under model 8 to 14 with control variables such as volume (Volume), turnover (Turnover), beta value counted by the return of last year (BETA_1Y), market value (MV), ratio of current asset to total asset (CATA), leverage ratio (LeverageR), return on equity (ROE), cash flow ratio (CashFlowR), and market return (MKT_R).. Constant. engchi. Panel A Without control variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 *** *** *** *** *** -0.00015 -0.00021 -0.000161 -0.0001513 -0.000158 0.000040 31. Model 7 -0.0002458***.

(37) (-3.00) (-3.84) (-2.91) (-2.87) (-3.04) (0.53) (-3.75) -0.00076***-0.00072***-0.000770***-0.0007492***-0.000765***-0.000887***-0.0007007*** (-10.35) (-9.21) (-9.54) (-9.77) (-10.12) (-7.99) (-7.35) 0.00064*** 0.0006787*** Group 1 (4.26) (4.39) -0.00050** -0.0005194** REGUGroup 1 (-2.26) (-2.28) 0.000078 0.0001624 Group 2 (0.62) (1.24) 0.000062 -0.0000069 REGUGroup 2 (0.34) (-0.04) -0.0000067 0.0000878 Group 3A (-0.04) (0.47) -0.0001764 -0.0002249 REGUGroup 3A (-0.65) (-0.81) 0.000115 0.0002037 Group 4A (0.50) (0.86) 0.000067 0.0000020 REGUGroup 4A (0.20) (0.01) -0.000286*** Group 4B (-2.85) 0.000186 REGUGroup 4B (1.27) 0.0004 0.0004 0.0004 0.0004 0.0004 0.0004 0.0004 R2 283320 283320 283320 283320 283320 283320 283320 N Panel B With control variables REGU. 立. 政治大. Group 2 REGUGroup 2. ‧. ‧ 國. 學. REGUGroup 1. y. Ch. Model 11 -0.0020022*** (-6.06) 0.0005320*** (7.10). sit. Model 10 -0.00193*** (-5.84) 0.00055*** (7.05). Model 12 -0.00201*** (-6.08) 0.00052*** (7.06). er. al. n. Group 1. Model 9 -0.00188*** (-5.42) 0.00059*** (7.92) 0.00076*** (3.85) -0.00097*** (-3.91). io. REGU. Model 8 -0.00200*** (-6.05) 0.00053*** (7.38). Nat. Constant. engchi. i n U. v. Model 13 -0.00142*** (-3.79) 0.00030*** (2.64). 0.00045*** (3.42) -0.00016 (-0.86) -0.0000075 (-0.04) -0.0000363 (-0.14). Group 3A REGUGroup 3A. -0.00023 (-1.06) 0.00016 (0.53). Group 4A REGUGroup 4A. -0.00040*** (-3.72). Group 4B. 32. Model 14 -0.001718*** (-4.83) 0.000631*** (6.99) 0.000952*** (4.53) -0.001004*** (-3.98) 0.000574*** (4.06) -0.000245 (-1.26) 0.000142 (0.82) -0.000134 (-0.52) -0.000053 (-0.24) 0.000052 (0.17).

(38) REGUGroup 4B. CATA LeverageR ROE CashFlowR MKT_R R2 N. 立. 政治大. 學. *p < 0.10, **p < 0.05, ***p < 0.001. Source: This Study. ‧. Nat. In Panel B, addition of other control variables into the regression, the results change. y. MV. dramatically. The coefficient of regulation (REGU) becomes significantly positive in all. io. sit. BETA_1Y. er. Turnover. ‧ 國. Volume. 0.00033** (2.30) 0.000000049*** 0.000000047*** 0.000000049*** 0.000000049*** 0.000000049*** 0.000000048*** 0.000000045*** (5.02) (4.56) (5.01) (5.01) (4.99) (4.86) (4.39) 0.275382529*** 0.276222730*** 0.275891421*** 0.275393370*** 0.275532697*** 0.276005874*** 0.277259816*** (32.04) (31.94) (32.06) (32.02) (32.01) (32.07) (31.95) -0.002520268***-0.002528584***-0.002475352***-0.002519798*** -0.00252578***-0.002493358***-0.002480638*** (-17.60) (-17.65) (-17.19) (-17.60) (-17.62) (-17.34) (-17.20) 0.000225909*** 0.000206933*** 0.000205974*** 0.000226472*** 0.00022949*** 0.000186095*** 0.000170017*** (6.43) (5.52) (5.77) (6.44) (6.50) (4.83) (4.24) -0.000976190***-0.000958436*** -0.00093266***-0.000976523***-0.000976899*** -0.00092887***-0.000893585*** (-5.01) (-4.92) (-4.77) (-5.01) (-5.01) (-4.74) (-4.56) 0.000206409 0.000162530 0.000197716 0.000201618 0.000196846 0.000222937 0.000137188 (0.55) (0.43) (0.52) (0.53) (0.52) (0.59) (0.36) 0.000007882*** 0.000008150*** 0.000008357*** 0.000007870*** 0.000007876*** 0.000008535*** 0.000008924*** (2.91) (2.99) (3.07) (2.90) (2.90) (3.11) (3.24) 0.000000340 0.000000351 0.000000361 0.000000340 0.000000338 0.000000366 0.000000383 (0.99) (1.02) (1.05) (0.99) (0.99) (1.07) (1.12) 0.826886917*** 0.826889971*** 0.826888238*** 0.826886965*** 0.82688737*** 0.826889026*** 0.826893176*** (151.12) (151.12) (151.13) (151.12) (151.12) (151.12) (151.13) 0.1202 0.1203 0.1203 0.1202 0.1202 0.1203 0.1203 272520 272520 272520 272520 272520 272520 272520. Models. The enforcement of new CSR regulations makes the whole stock return of samples. n. al. of listed companies increase. v i n 26.5% C h(0.00053/0.002). UFor Model engchi. 9 to Model 13, the. coefficients of regulation (REGU) also turn significantly positive. This implies when excluding Group 1, Group 2, Group 3A, Group 4A, and Group 4B respectively, the stock returns of whole market improve due to the enforcement of new CSR regulations. When only considering Group 1 with control variables, in general, it still has the highest positive and significant performance (0.00076) comparing to other groups regardless before or after the enforcement of new CSR regulations; but the interactive dummy (REGUGroup 1) in Model 9 is significantly negative (-0.00097), even higher than that without control variables. Group 1 decreases 14.39% ((0.00059-0.00097)/(0.00188+0.00076)) after the enforcement of new. 33.

(39) CSR regulations. In Model 10 Group 2 has positive and significant return before and after the enforcement of new CSR regulations (0.00045); but the interactive dummy (REGUGroup 2) is negative and insignificant. In Model 11 and Model 12, Group 3A and Group 4A have insignificant figures. When it comes to Model 13, Group 4B has significantly negative performance (-0.00040), comparing to other groups no matter before or after the enforcement of new CSR regulations. The interactive dummy (REGUGroup 4B) is positive and significant (0.00033). Group 4B increases 61.76% ((0.00030+0.00033)/(0.00142-0.0004)) after the enforcement of new CSR regulations. To consider all groups at a same time, in. 政治大. Model 14, it reveals that the enforcement of new CSR regulations significantly improves. 立. Group 4B by 0.000631. For Group 1 and Group 2, they have significantly positive. ‧ 國. 學. performances (0.000952 and 0.000574) which are higher than other groups before the enforcement of new CSR regulations. However, for Group1, it significantly decreases after. ‧. the enforcement of new CSR regulations by a total of 0.000373 (REGU+REGUGroup 1).. y. Nat. er. io. sit. The changes in other groups are not significant.. n. a CSR report of each company 2. Part 2—the filing date of v i l C n hengchi U. In table 6, it presents descriptive statistics of 259 listed companies (Group 1 + Group 2 + Group 3A + Group 4A) with the time period of 180 days before and after the filing date of CSR report of each company. In panel A we can see that the average rate of return (Return) of all observed samples is -0.000067; the average trading volume (Volume) of them is 5309.43; the average turnover rate (Turnover) is 0.0034; the average of market value in million (MVmillion) is 74458.18; the average market value (MV) is 9.9153; the average ratio of current asset to total asset (CATA) is 0.5157; the average leverage ratio (LeverageR) is. 34.