Regional Anti-Corruption and CSR Disclosure in a Transition Economy: The Contingent Effects of Ownership and Political Connection

Based on the empirical data of China’s Shanghai and Shenzhen A-share market, this paper examined the impact of regional anti-corruption intensity on corporate social responsibility (CSR) disclosure. The results indicate that (1) regional anti-corruption intensity has a significant positive effect on firms’ CSR disclosure; (2) through the grouping test based on the ownership of firms, it was found that the positive effect of anti-corruption intensity on CSR disclosure in the sample of non-state-owned enterprises was more significant and positive than that of state-owned enterprises (SOEs); and (3) through the grouping test of whether or not the enterprises had political connections, the positive effect of regional anti-corruption intensity on CSR disclosure was stronger and more significant in firms with political connections (relative to those with no or weak political connections). This paper sheds light on the research into anti-corruption policies by linking government macro policy and enterprises’ micro social economic behaviors, and it provides empirical evidence for this linkage. This paper also contributes to organizational legitimacy theory and CSR theory by probing the impact of anti-corruption policies on firms’ CSR disclosure. At the same time, the effects of two contingency factors (ownership and political connection) also provide some practical implications to the relevant government departments by: (1) speeding up the market-oriented reform of state-owned enterprises including clarifying the boundaries of authority and responsibility between SOEs and the government, and conducting the de-administration of managers of SOEs; and (2) encouraging firms to focus on market competition and be more socially responsible, instead of speculating with political rents.


Introduction
Global sustainable development cannot be separated from the participation of enterprises. As the largest developing country, the issue of corporate social responsibility (CSR) has attracted much attention in China. Based on the specific background of the current economic transition in China, the CSR disclosure of listed firms in mainland China adopts the method of a combination of compliance and voluntary disclosure. Except for some index components, there is no mandatory disclosure requirement according to the regulation. On the one hand, paying attention to CSR can help corporations to develop sustainably, and strengthen their long-term survivability [1]. Particularly in the information era, any information about corporations such as their attitudes towards CSR can change stakeholders' decisions [2]. For example, customers may arrange their purchase volume while they have learned of a firm's CSR performance [3]. Meanwhile, firms that have made great efforts in CSR will enjoy good reputation [4]. Hence, the number of domestic enterprises disclosing CSR is rising. On the other hand, scholars have carried out exploratory and complementary studies on the influencing factors of firms' (voluntary) disclosure of CSR information. These studies have mainly focused on firm size, debt, market value, CEO characteristics, board structure, and so on [5][6][7], and covered various factors [8], however, there are controversial conclusions on these factors [6].
Recently, some scholars have begun to pay attention to the impact of external corporate governance factors. For example, Reverte (2009) [9] believes that media exposure has a driving effect on CSR disclosure. When Wang et al. (2016) [10] discussed the contingency effect of the internal characteristics of firms on the disclosure of CSR and the cost of equity capital, they extended the research framework to include external factors such as the marketization process, industrial and environmental protection policies, and so on. Unfortunately, they did not conduct further empirical analysis of the direct impact of the external institutional environment on firms' CSR disclosure.
As the largest emerging economy, China's 31 provinces and municipalities are heterogeneous in the degree of development and marketization process. Although all regions must implement the laws and regulations promulgated by the state council, the different degrees of marketization might lead to the different implementation of national laws and regulations in different regions [11,12]. Since the 18th National Congress of the Communist Party of China, China has vigorously launched an anti-corruption campaign and has normalized and institutionalized anti-corruption efforts. During the transition period in emerging economies, there are different effects of the implementation of anti-corruption policies in various regions, which provides good context for exploring the relationship between regional anti-corruption intensity and the CSR disclosure of firms.
During the economic transition period, the Chinese government controls many key resources. If a top manager of a firm establishes a personal relationship with a government official, it means that access to the key resources controlled by the government could be used to generate advantages in market competition by the firm. The implementation of anti-corruption policies may suppress the above rent-seeking activities and urge firms to attach importance to strategic options such as manifesting social responsibility. From the perspective of motivation, the disclosure of CSR is to obtain the recognition and support of stakeholders through the transmission of information to them [13,14], thus being considered by society as "legitimate" and "suitable", that is, improving organizational legitimacy [15]. According to this view, the higher the organizational legitimacy of the enterprise, the easier it is to receive positive feedback from stakeholders and to gain access to the resources needed for their development and long-term survival [16,17].
To a large extent, the "anti-corruption" campaign has cut off any illegitimate connections between government officials and firms, thus increasing the cost and difficulty of seeking political asylum by the firm. Therefore, as a "rational" micro-individual operating in the market economy, the firm will evaluate the gains and costs of pandering to officials and other development strategies (such as the active implementation of CSR, the disclosure of CSR information, etc.), and select the optimal option that could generate more benefits. From this point of view, continuous and in-depth anti-corruption actions may become an institutional factor of the external governance of enterprises, which helps promote firms obtain the "legitimacy" basis of their existence by improving their governance through measures such as the disclosure of CSR information.
It is worth indicating that the discussion on the causal relationship between regional anti-corruption intensity and firms' CSR disclosure only represents the sample as a whole. If there are different effects of anti-corruption intensity on CSR disclosure across different types of firms with heterogeneous factors, the ignorance of the above issue will lead to a conclusion that is incomplete and discrepant from reality. Thus, this paper identified two contingent factors that may lead to the main effect diverging (change) from the previous literature.
(1) Ownership. At present, China's special economic system determines the inextricable connections between SOEs and the government. SOEs assume the responsibility of resettling employment and maintaining social stability, which belongs to the government, and thus are often faced with soft budget constraints [18]. In addition, the existing political system in China often means that senior executives of state-owned enterprises have the legitimate dual identities of "official" and "businessman", that is, they hold positions as both government officials and enterprise managers at the same time. Thus, SOEs are naturally endowed with abundant resources from the government compared with their non-state-owned counterparts. To obtain access to resources controlled by the government, non-SOEs might put "bets" on individual officials seeking to benefit from corruption. By bribing these officials, the enterprises can gain key resources for development and create a favorable external environment. To these non-SOEs, the impact of anti-corruption efforts may be even stronger, and consequently allow them to improve relationships with their stakeholders by acting in a more socially responsible way.
(2) Political connection. During the process of economic transition, some firms will establish connections with government officials, and try to obtain protection from officials and the key resources for firm development through bribery. However, the strong anti-corruption efforts of the government at all levels has forced these firms to switch to market-oriented behaviors. The CSR strategy provides a feasible path for the market-oriented transformation of such firms. By actively fulfilling the social responsibility and disclosing CSR reports, enterprises can obtain support from their stakeholders to alleviate the impact of anti-corruption policies.
Based on the above analysis, this paper employed a sample of listed companies on the Shanghai and Shenzhen stock markets to test the impact of regional anti-corruption intensity on firms' disclosure of CSR information. At the same time, we introduce two contingent factors, ownership and the political connection of firms, to examine the different effects across different groups.
This paper makes the following contributions to the literature: (1) Our research revealed that anti-corruption actions are an important influencing factor on firms' disclosure of CSR in a transition economy, which might enrich the knowledge in the field of CSR. Most of the existing literature has focused on the impact of the internal characteristics of firms on the disclosure of CSR. However, as a transition and developing economy, there are considerable regional differences in the legal environment in China. The strong anti-corruption efforts of the national government also work as an external governance factor for firms, and could impel them to be more aware of the importance of CSR. Furthermore, we also studied the contingent effects of ownership and political connection on the linkage between anti-corruption intensity and the disclosure of CSR. These context specific antecedents and their contingent factors of CSR activities may enrich the CSR literature.
(2) Our research focused on the external governance effect of the government's anti-corruption actions on the firms' disclosure of CSR, and provided empirical evidence on the positive economic consequence of anti-corruption activities. The existing research on the economic consequences of anti-corruption policies is still controversial: one view holds that corruption can help firms circumvent inefficient regulation by creating close ties with officials [19], consequently, anti-corruption actions would reduce the efficiency of business operations [20]. Another perspective is that anti-corruption actions weakens the collusion effect of political connections, which can guide firms to regulate their operations (fulfilling social responsibilities, attaching importance to R & D and innovation, etc.) and is good for the long-term development of firms [21,22]. This paper found a positive effect of anti-corruption intensity on firms' CSR disclosure, which provides empirical evidence for the positive economic consequences of China's current anti-corruption efforts, and enriches the knowledge on anti-corruption policy in transition economies.

Literature Review and Hypotheses
CSR refers to the moral obligations of firms, which means that firms should consider internal and external relevant needs [23]. Ansoff (1965) [24] put forward the "stakeholder theory" in the definition of corporate objectives, where firms should balance the claims of different stakeholders that might be in conflict with each other. Just like shareholders, all other stakeholders such as investors, activist groups, and government bodies may influence an organization's outcome [25]. With the integration of the two theories, Roberts (1992) [26] pointed out that in addition to shareholders, firms are also responsible for a series of internal and external interest groups such as employees, creditors, consumers, governments, media, and so on. To summarize, CSR is how a firm gives consideration to claims from its stakeholders. A company that engages in excellent CSR performance could establish a closer relationship with its stakeholders [3]. Therefore, the disclosure of the CSR of a firm means that the firm is conveying its attitude, actions, and results toward the claims of its stakeholders to society, and indicates that they are socially responsible.
Why do some firms choose to disclose their CSR? In economic development, the financial information itself cannot depict corporate behavior very well, so the non-financial information, for example, CSR information, becomes a beneficial complement [27]. Mamun et al. (2014) [28] found that banks have increasingly disclosed their CSR information in recent years. Theoretically, corporations disclose their CSR information voluntarily, mainly to meet their stakeholder's demands, releasing the information asymmetry between shareholders and managers, and reducing financing costs [2]. Most of the existing empirical studies have focused on the internal characteristics of firms and industrial factors [29]. For example, Gamerschlag et al. (2011) [30] conducted an empirical study based on 130 listed companies in Germany. Their results showed that high visibility, decentralized ownership structure, and American shareholder ownership contributed to the disclosure of CSR. However, profitability only had a significant impact on information disclosure of the environmental dimension of CSR.
Studies using data from developing countries have also drawn similar conclusions. Lu et al. (2017) [31] selected Chinese forestry listed firms as their sample. The results indicated that the sizes and the degrees of ownership concentration of firms had significant positive effects on the disclosure of CSR. However, ROE and financial leverage did not have significant effects on CSR disclosure. Muttakin et al. (2018) [7] used a sample of listed firms in Bangladesh, and found that the board capital had a significant positive impact on the firms' disclosure of CSR, but the power of the CEO had the opposite negative effect on the disclosure of CSR and a weakening effect on the relationship between board capital and CSR disclosure.
Recently, external drivers of CSR disclosure have been of concern. International buyers, foreign investors, international media, and the World Bank have a profound impact on the CSR disclosure of firms in developing countries [32]. Baldini et al. (2016) [33] conducted an empirical study based on 14,174 firms in many countries. The results showed that the political system, labor system, and cultural system have significant impacts on the disclosure of environmental, societal, governance, and other CSR information.
Corruption is a thorny problem in a transition economy [34]. The central government of China is aware of the harm caused by corruption and has vigorously implemented a high-pressure anti-corruption policy. There is evidence of the relationships between firm level corruption and firm social reputation [35] and CSR performance [36] as well as a relationship between country level corruption and a firm's CSR information disclosure [33]. Will anti-corruption actions in transition economies impact on firms' disclosure of CSR? However, there is still a lack of knowledge on this area. In order to fill the gap in the literature, we focused on anti-corruption policy as an important external driver of CSR disclosure, and looked for evidence in China, which is the largest transition economy.

Regional Anti-Corruption and CSR Disclosure
There is a lot of evidence that suggests that external environmental factors have an impact on corporate governance [37] as well as evidence of the impact of external environmental factors on CSR information disclosure [32,33]. There is also evidence from China that the external environment has an influence on the information disclosure of firms. Some studies have pointed out that differences in the external environment, institutions, and policies of firms will lead to heterogeneity in the motivation of information disclosure, consequently affecting information disclosure [27]. Specific external factors such as regional marketization processes [10] and industry competition [38] were found to have a significant impact on the CSR information disclosure of firms.
During the economic transition process, the relationship between the government and the firms is a government dominant non-equivalent game relationship [39], as the former has the right to allocate most of the key social resources. By paying "economic rent" to government officials, the enterprises obtain special "care" from the latter (such as environmental protection, tax subsidies, government procurement, media propaganda, and so on), which makes it easier for the firms to survive and develop. However, under the pressure of strong anti-corruption policies, rent-seeking through "personal relations" with government officials is less feasible, which would force firms to consider other solutions.
One of these solutions is to increase the legitimacy of the firm, and gain a positive response from society [15]. CSR information is a defensive or compliance statement made by a firm that considers the reality of its impact on society [40,41]. Disclosing CSR information could help firms mitigate the threat of legitimacy, fill the gap in legitimacy [40], and reduce management challenges and operational risks [42]. Therefore, anti-corruption policies might force firms to consider CSR information disclosure and convey their efforts to cater to stakeholder requirements, therefore enhancing their legitimacy. Meanwhile, stakeholder theory has revealed that companies must pay close attention to CSR disclosure as it could meet and satisfy their stakeholders' demand for information if they want to gain their continuous support [2].
However, as an emerging market country with a vast territory, there are differences in the social-economic development and execution of government policy across various regions in China [12], and the anti-corruption policy is no exception. In regions where the intensity of anti-corruption actions is higher, firms may be more likely to disclose their CSR information, and vice versa. Accordingly, we propose the following hypothesis: H1. The intensity of regional anti-corruption has a positive impact on a firm's voluntary disclosure of CSR information.

Ownership, Regional Anti-Corruption and CSR Disclosure
Existing studies have mainly discussed the direct impact of ownership on a firm's CSR disclosure. Ghazali (2007) [43] carried out empirical research based on listed firms in Malaysia and found that the government was commonly the controlling shareholder of firms in Malaysia, and that government ownership was helpful for firms to disclose more CSR information. Furthermore, a cross-country study based on 203 firms in China, Malaysia, India, and the United Kingdom indicated that government ownership is associated with a higher quality of CSR reports [44].
Studies in transition economies and emerging economies have received mixed results in the relationship between state ownership and CSR disclosure. An empirical study of Bangladeshi enterprises found that public ownership had a positive influence on CSR disclosure [45]. An empirical study on Vietnamese listed firms found that state ownership had a significant negative impact on CSR information disclosure [46]. Similarly, empirical studies on Chinese listed firms have found that there was a positive but insignificant effect of private ownership on CSR reporting [47] and a negative impact on CSR disclosure by state ownership [48]. An empirical study based on Indian listed firms indicated that multinational ownership and family ownership were associated with higher levels of CSR engagement, while public ownership had a negative impact on CSR engagement [49].
In a transition economy such as China, SOEs and private firms are allowed to coexist. The ownership (controlling stake) of SOEs belongs to the state, and this nature determines that the SOEs have an indivisible internal relationship with the government. On one hand, the SOEs have assumed some social functions of the government such as political responsibility, charity donation, environmental protection, etc. [50]. At the same time, the SOEs also have special rights authorized by the government such as soft budget restrictions, market entry, and so on [18]. On the other hand, many SOE executives are government officials, or can be transferred to work in government departments. Thus, SOEs have a close fundamental relationship with the government, and this relationship does not change with the change in individual government officials.
Private firms have an inferior position when compared with SOEs in terms of their relationships with the government. If they want to address their "innate weaknesses" in terms of access to crucial resources such as industry access, land tenure, bank credit, and so on, they must try their best to improve their relationships with the government (or with government officials). The institutionalization and regularization of anti-corruption policies will cause the opportunity cost of the illegal operation of officials and firms to rise sharply [51], and its external governance effect would benefit performance enterprises in the long-term [21,22]. However, the impact of anti-corruption actions tends to be weak for firms in regulated sectors and low-marketization areas, but strong in non-regulated sectors or high-marketization areas [52]. Anti-corruption events are more likely to have a positive impact on the market value of private firms, but negative to that of SOEs [52], thus forcing private firms to release more negative information, while SOEs still suppress the release of negative information [53].
Therefore, faced with more intensive anti-corruption actions, private firms who operate mostly in non-regulated industries are more likely to legitimize themselves through disclosing CSR information. Thus, we propose the following hypothesis: H2. Compared with SOEs, the intensity of anti-corruption has more impact on CSR disclosure in private firms.

Political Connection, Regional Anti-Corruption, and CSR Disclosure
A great deal of evidence has indicated that political connections could generate certain economic benefits for firms [54,55]. Political connections allow firms to obtain government subsidies and policy support much easier than those without [56]. Khwaja and Mian (2008) [57] argued that the establishment of political connections by a firm was beneficial to obtaining debt financing. Besides obtaining higher debt financing, a study across nearly 50 countries indicated that the market shares of firms with political connections were also higher than firms with less political connections [58]. Private firms connected to a politician's networks could enjoy a larger increase in public procurement when compared with non-connected firms [59].
While enjoying the many benefits, enterprises with political connections also bear risks, especially when faced with anti-corruption campaigns. Fisman (2001) [60] carried out a study of Indonesian firms and found that the share price of politically dependent firms were more negatively impacted by rumors regarding the politician. A study of Chinese private list firms whose managers had bribed or had political connections with corrupt bureaucrats experienced the loss of competitive advantages in the M&A market [61].
Since the promotion of anti-corruption institutions, incentives to establish political connections among firms have weakened, and the potential costs of officials providing hidden shelter to businesses have risen sharply [62]. However, only more stable connections could have a positive impact on firm performance [63]. Under anti-corruption pressure, firms with political connections would lose their rent-seeking opportunities, and therefore try to reallocate their resources in competition-oriented areas. By increasing investment in CSR and disclosing information on CSR, firms could meet their stakeholder expectations and become more legitimized, which can attract further support from stakeholders to enhance market competition [16,17].
Therefore, we propose the following hypothesis: H3. Compared with non (weak) politically connected firms, the influence of regional anti-corruption intensity on the disclosure of CSR information is stronger in firms with political connections.

Data Selection
The Chinese government has established a "high-pressure anti-corruption" course of action since the end of 2012. Against this policy background, we constructed non-equilibrium short panel data of Shanghai and Shenzhen A-share listed firms from 2013 to 2015. Firms were eliminated from the sample for the following reasons: (1) Due to the lack of anti-corruption data in Tibet, samples of firms registered in Tibet were eliminated; (2) The financial indicators of financial firms are different from that of other firms, so we eliminated these financial firms in order to ensure the consistency of the research; (3) The financial indicators of special treatment (ST) firms are usually abnormal and thus were excluded in our sample; (4) We eliminated observations of new IPO firms in each of the studied years as their data may have been incomplete in that year; (5) According to regulatory policy, listed firms in index stocks are compulsorily required to disclose CSR information, thus, we eliminated these firms as their disclosure of CSR information was not voluntary; and (6) Observations with missing data of key indicators and that were difficult to verify were eliminated. Finally, we obtained a sample with 5925 valid observations.

Dependent Variable
The dependent variable in this research was the disclosure CSR information (Disc). We searched the data of the CSR disclosure of listed firms from websites of the Shanghai Stock Exchange and Shenzhen Stock Exchange. To ensure accuracy, we verified the data with the CSR ranking data released by Rankins CSR Ratings (RKS). The disclosure of CSR information (Disc) was defined as a dummy variable; if the firm voluntarily disclosed social responsibility information, it was assigned a value of 1, otherwise it was assigned a value of 0.

Independent Variable
The independent variable in this research was regional anti-corruption intensity (Anti-corr). In China, the number of registered crimes of public officials [64] and the ratio of criminal records to the number of public officials [65] are good measures of anti-corruption intensity. Related foreign studies have also used similar measures. For example, Glaeser and Saks (2006) [66] used the ratio of the numbers of corruption crimes of state civil servants to the total state population (or the total number of civil servants) to measure the extent of regional corruption. Following the above studies, we employed the ratio of the number of registered duty crimes to the number of public officials (ten thousand people) in each province as the proxy of the regional anti-corruption intensity (Anti-corr).

Control Variables
Referring to similar studies [3,6,31], we controlled the factors that may affect the disclosure of CSR information: (1) firm size (Lnemp), which was measured by the natural logarithm of the number of employees; (2) firm age (Firmage), which was measured by the number of years from firm establishment to the observation; (3) referring to related research [38], we controlled the degree of the monopoly of the industry (Concen), which was measured by the sales of the top four firms as a percentage of the total sales of the industry [67]; (4) firm financial leverage (Lever), which was measured by the asset-liability ratio of each firm; and (5) the growth of the firm (Grow), which was measured as the growth rate of the main business revenue. We also tried to control the year and industry effects by including year dummies and industry dummies, however, the statistical tests indicated that there were no significant year and industry effects, thus we excluded them in regressions.

Grouping Variables
To test Hypotheses 2 and 3, we defined two grouping variables. (1) Ownership (State). If the firm was owned by the state, it was assigned a value of 1, otherwise it was 0. (2) Political connection (Policonn). If the top managers of the listed firm were currently or had served as officials of any government organization at the state level, provincial level, municipal level, county level, town level and below (including party representatives, deputies to the people's Congress, and members of the Chinese People's Political Consultative Conference, etc.), then the variable was assigned a value of 1, otherwise it was assigned as 0. Table 1 reports the descriptive statistics and correlation coefficients of the main variables. In terms of data distribution, the absolute values of the standard deviation of variables such as CSR information disclosure (Disc) and firm growth (Grow) were greater than their respective mean values, which shows that the distribution was relatively discrete. The correlation coefficient matrix showed that there was a significant positive correlation between regional anti-corruption intensity (Anti-corr) and the firms' CSR information disclosure (Disc) (r = 0.071, p < 0.001), which suggests that both of them changed in the same trend. Among the control variables, firm size (Lnemp) and financial leverage (Lever) were positively correlated with dependent variables at a significant level of 1% . In addition, the correlation coefficients of each variable were relatively small, thus the regression model should not have a serious collinearity problem.

Regressions
The data of this research were non-equilibrium short panel data and the dependent variable was a dummy variable. Therefore, we adopted the logit regression model for panel data. Considering that a large number of firms showed no changes in the value of dependent variables (Disc) from 2013 to 2015, if the fixed effect model was used, the observations of these firms would have been eliminated in regressions, which may lead to potential bias. In light of this, we selected the random effect model and the mixed effect model instead of the fixed effect model. The results of the LR test of the main model in the random effect test indicated that the random effect model was more suitable than the mixed effect model for regression (Prob ≥ chibar2 = 0.000). Table 2 reports the regression results of the total sample and each sub-sample divided by the grouping variables. It is necessary to note that we tried to control the year effect and the industrial effect, and the regression results showed that we could not reject the original hypothesis of "non-time effect" (Chi 2 = 0.28, p = 0.871) and "non-industrial effect" (Chi 2 = 13.61, p = 0.479). Thus, we did not include year dummies and industry dummies in the regression models, as the statistical test indicated that there were no significant time and industry effects. Model (1) was the base model, which only included the control variables and showed that Lnemp had significant positive effects on the dependent variable.
Hypothesis 1 predicted a significant positive relationship between regional anti-corruption intensity and CSR information disclosure. Model (2) included the independent variable Anti-corr, and the results showed that the coefficient of Anti-corr was significant and had a positive effect on Disc (r = 0.059, p < 0.05). The results support Hypothesis 1. Furthermore, among the control variables, firm size (Lnemp) played a significant and positive impact on Disc (r = 1.838, p < 0.001), which suggests that the larger the scale, the more pressure there was to impel to company to disclose CSR. Meanwhile, Concen exerted a negative influence on Disc (r = −0.222, p < 0.05), revealing that the higher the degree of monopoly of the industry, the easier it was for a company to refuse to publish its CSR information.
Hypothesis 2 predicted that a more significant effect existed between regional anti-corruption intensity and CSR information disclosure in private firms. Models (3) and (4) examined the relationship between Anti-corr and Disc in the two groups of private firms and SOEs. The results indicated that Anti-corr had a significant positive effect on Disc (r = 0.098, p < 0.01) in the group of private firms, while the relationship between them in the group of SOEs was negative and not significant (r = −0.081, p > 0.1). The results confirmed Hypothesis 2. Simultaneously, regardless of whether it was Model (3) or Model (4), firm size (Lnemp) had a significant and positive influence on Disc.
Hypothesis 3 predicted that there was a more significant impact of regional anti-corruption intensity on CSR information disclosure in firms with political connections. Models (5) and (6) examined the relationships between Anti-corr and Disc in the two groups of firms with political connections and firms without political connections. The results indicated that Anti-corr had a significant positive effect on Disc (r = 0.206, p < 0.001) in the group of firms with political connections, while the relationship between them in the group of firms without political connections was positive but not significant (r = 0.049, p > 0.1). Therefore, Hypothesis 3 was supported. At the same time, like the models above-mentioned, in both Models (5) and (6) the coefficient of Lnemp was significant and had a positive effect on Disc. Furthermore, Firmage has a significant and positive influence on Disc (r = 0.310, p < 0.001) only in Model (5), meaning that the greater the age, the stronger the company's disclosure of CSR. Standard error in parentheses; * p < 0.05, ** p < 0.01, *** p < 0.001.

Impact of Regional Anti-Corruption Intensity on Firms' Disclosure of CSR
The regression results showed that regional anti-corruption intensity had a positive effect on firms' disclosure of CSR information, which reveals that the governance effects of anti-corruption policies were an exogenous institutional variable to firms. The finding is consistent with some previous studies [9,32,33], confirming the significant impacts of external factors on firms' CSR information disclosure. Existing research indicates that the intensive industry competition encourages firms to disclose CSR information [38]. This finding is also consistent with cross-country studies that indicated that firm level corruption was negative associated with firm social reputation [35] and CSR performance [36], and that the level of corruption in a country was negatively related to firms' CSR information disclosure [33]. Our explanation of this finding is that when firms are less likely to seek "rents" from corruption under the pressure of intensive anti-corruption actions, they are more likely to disclose CSR information. By doing so, firms could increase their legitimacy and gain support from their stakeholders. This logic is also consistent with literature that suggests that being socially responsible could help firms to legitimize themselves in front of their stakeholders [40] and mitigate potential risks [42]. Compared to those studies, our research revealed that in a transition economy, anti-corruption actions work as an important external governance factor, and could impel firms to disclose CSR information.

Effect of Ownership on Regional Anti-Corruption Intensity and CSR Disclosure
We introduced ownership as a contingent factor, as SOEs are naturally endowed with abundant resources from government when compared with private firms. The results indicated a significant positive impact of regional anti-corruption intensity on CSR disclosure in the sub-sample of private firms. In the sub-sample of SOEs, there was a negative and insignificant effect. The existing studies have mainly focused on the direct effects of state ownership or private ownership on firms' CSR disclosure [45][46][47][48][49] and the quality of CSR information [43,44], or the economic consequences of anti-corruption policies [21,22,52]. Our research further explored the contingent effect of state ownership on the link between anti-corruption intensity and firms' CSR information disclosure, and revealed that ownership also had an indirect effect on firms' CSR disclosure. A possible explanation for this finding is that, to those private firms with an innate deficiency in resource endowment, anti-corruption actions could be more effective in encouraging them to convey their CSR information to the public, and thus gaining support from their stakeholders.
From this point of view, private firms may be more likely to adopt a CSR strategy as a substitute for "rent-seeking" to enhance their economic performance. As private firms are more likely to be under public supervision after disclosing CSR information, they would become more market competition oriented and more socially responsible. Thus, our research also indicates that positive economic consequences [21,22] and positive social consequences of anti-corruption policies might appear in the long-run in private firms.

Effect of Political Connection on Regional Anti-Corruption Intensity and CSR Disclosure
Our study explored the contingent effect of political connection on the relationship between regional anti-corruption intensity and firms' CSR disclosure. The results indicate that regional anti-corruption intensity exerts a significant positive influence on firms' CSR disclosure in the sub-sample of firms with political connections, but had an insignificant influence in the sub-sample of firms without political connections. This finding indicates that firms with political connections have to find solutions while their "rent-seeking" behaviors are at risk due to intensive anti-corruption actions. This is consistent with previous studies that have indicated the risks and negative impact of political connections [60,61]. This finding also reflects the positive consequence of anti-corruption actions in a transition economy. While "rent seeking" through political connections is less feasible, firms would have to change to become more socially responsible and solicit support from their stakeholders instead of government officials. Thus, constant anti-corruption actions might push forward the process of marketization in a transition economy.

Implications
The findings of this paper may have some implications for policy makers. First, we provided evidence on the effect of regional anti-corruption intensity upon the firms' CSR disclosure, which indicates that in a transition economy, institutionalized and constant anti-corruption efforts could improve the market environment, and regulate firms to be more socially responsible and change to market-oriented solutions. Second, the different effects of regional anti-corruption intensity between private firms and SOEs reflect the problems of SOEs' "mixed functions of government and business" and "ambiguity of power and duties". Hence, it is necessary to push forward the reform of SOEs through de-administration and mixed ownership to improve corporate governance. Finally, the different effects of regional anti-corruption intensity between firms with political connections and without political connections reflect the sharp increase of the risks of "rent seeking" before the anti-corruption campaign. Even intensive anti-corruption actions might negatively impact on a firm's economic performance in a short period, but would be beneficial to the firm's performance and marketization process in the long-run. Thus, the government should institutionalize anti-corruption to generate a constant governance effect on individual firms and the economic environment.

Conclusions
The influencing factors of CSR information disclosure have been widely discussed in the literature, but most of them have focused on the internal factors of firms. This study focused on the effect of anti-corruption intensity on CSR disclosure, and introduced two contingent factors as grouping variables, namely, ownership and political connection. Based on the panel data of China's Shanghai and Shenzhen A-share listed firms from 2013 to 2015, we tested our hypotheses and drew the following conclusions: Regional anti-corruption intensity has a positive impact on firms' CSR information disclosure. Our finding was consistent with those studies that suggest that external environmental factors have an impact on firms' CSR information disclosure [10,32,33,38], and the logic implicated by organizational legitimacy theories [15,40]. Compared with the existing literature, this paper provides evidence that anti-corruption policy works as an external impact factor on firms' CSR information disclosure in a transition economy. This finding indicates that in a transition economy, anti-corruption actions increased the illegal costs of "rent seeking" behaviors, worked as an external governance factor, and created an external environment of fair competition for firms. It also provided implications for the Sustainable Development Goals (SDGs) of the United Nations. Intensive anti-corruption actions provide an institutional environment that is favorable toward market competition and leads firms to be more socially responsible. Therefore, three core elements of SDGs-economic growth, social inclusion, and environmental protection-are more likely to be harmonized as the needs of the different stakeholders in society are better balanced.
The effect of regional anti-corruption intensity on firms' CSR disclosure is different among private firms and SOEs. Namely, the positive effect of anti-corruption intensity in the sub-sample of private firms on CSR disclosure was more positive and significant than that of the SOEs. Our finding was consistent with studies that found different effects of anti-corruption policies on the behaviors [53] and market value [52] of private firms and SOEs. Compared to related studies that have focused on the direct impact of ownership on CSR information disclosure [43][44][45][46][47][48][49], we further explored the contingent effect of state ownership on the linkage between anti-corruption policies and firms' CSR information disclosure. This finding indicates that the problems such as "mixed functions of government and business" and "ambiguity of power and duties" would lead to the lack of market-oriented operation of the SOEs, and thus provides new evidence that SOEs need further reform to improve their governance.
Compared with firms without political connections, the effect of regional anti-corruption intensity on CSR disclosure was more significant and positive in firms with political connections. Our finding was consistent with studies that revealed the risky aspect of political connections [60,61]. Compared to studies that have focused on the economic consequences of political connections [54][55][56][57][58][59][60][61], our study further explored the contingent effect of political connections on the linkage between anti-corruption policies and firms' CSR information disclosure. This finding indicates that intensive anti-corruption actions could force firms to transform from a "relation economy" to a "market economy", and consequently change to market-oriented solutions such as adopting a CSR strategy.
Our research also has its limitations, which are as follows. First, the sample used in this research was only compiled of China's listed firms, which means that our conclusions might not be generalized to other transition economies due to differences in national cultures and political systems. Second, the quantitative research method was also limited in the interpretation of our conclusions. Therefore, future research could expand this study by using a cross-country sample of firms to enhance the generalizability of conclusions and adopting qualitative research methods to provide in-depth and richness evidence, which would be helpful in explaining how and why firms tend to disclose their CSR information before intensive anti-corruption policies.