Does religion matter to corruption? Evidence from China

    Does religion matter to corruption? Evidence from China Xixiong Xu, Yaoqin Li, Xing Liu, Weiyu Gan PII: DOI: Reference:

S1043-951X(16)30159-6 doi:10.1016/j.chieco.2016.11.005 CHIECO 999

To appear in:

China Economic Review

Received date: Revised date: Accepted date:

29 July 2015 21 November 2016 21 November 2016

Please cite this article as: Xu, X., Li, Y., Liu, X. & Gan, W., Does religion matter to corruption? Evidence from China, China Economic Review (2016), doi:10.1016/j.chieco.2016.11.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Does Religion Matter to Corruption? Evidence from China1

IP

T

Xixiong Xu (Corresponding author) School of Economics & Business Administration, Chongqing University, Chongqing, PR China 400030 [email protected] Tel: 86-13996250466

SC R

Yaoqin Li School of Economics & Business Administration, Chongqing University, Chongqing, PR China 400030

NU

Xing Liu School of Economics & Business Administration, Chongqing University, Chongqing, PR China 400030

MA

Weiyu Gan School of Management Southwest University of Political Science and Law, Chongqing, PR China 401120

AC

CE P

TE

D

Abstract There is a growing interest in understanding how religion affects corruption. Using provincial-level panel data from 1998 to 2009, this paper investigates the effect of religious beliefs on bureaucratic corruption in China. The empirical results show that, bureaucratic corruption is negatively associated with local religious heritage, implying that religious culture plays a positive role in restraining official’s corruption since religion has influence on political preference and work ethic. We also find that the negative association between religion and corruption is weaker in provinces with stronger law enforcement, which identifies the substitution effect between religious ethic and legal supervision in curbing corruption. Our findings also reveal that, among the different religions, the anti-corruption effects of China’s native religions (i.e., Taoism and Buddhism) are more significant than those of foreign religions (i.e., Christianity and Islam). These conclusions are consistent and robust to various measures of main variables and a variety of robustness checks. Given the very few studies and limited data resources in the context of China, this paper as a tentative study provides new evidences of the relationship between religion and corruption. Keywords Religion; Corruption; Legal Institutions; Buddhism; Taoism JEL Classification: Z120, D730 1. Introduction 1

We acknowledge the National Natural Science Foundation of China (approval number 71102063;

71232004; 71572019) and the Fundamental Research Funds for the Central Universities (approval number 106112015CDJSK02XK12; 106112016CDJXY020003) for their financial support. We are also especially grateful to editors and anonymous referees for their helpful comments and suggestions. Any remaining errors are our responsibility. 1

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Since the establishment of the People's Republic of China in 1949, corruption has vexed the national leadership. Especially with the launch of economic reforms, corruption has become even more widespread and exists at every level of China's political system. Despite the introduction of several anti-corruption campaigns, corruption has managed to flourish and seemingly become more virulent over time (Wedeman, 2012; Dong & Torgler, 2013). Even the Chinese government has admitted that corruption “is now worse than during any other period since New China was founded in 1949. It has spread into the Party, into Government administration and into every part of society, including politics, economy, ideology and culture” (Liang, 1994, p.122). The seriousness of this problem is exemplified by the new wave of ferocious fighting corruption and recent charges against five national-level officials, Yongkang Zhou, Caihou Xu, Boxiong Guo, Jihua Ling and Rong Su .2 Corruption in contemporary China has generated much literature in sociology, political science and economics (Yao, 2002; Gong, 2006; Cai, Fang, & Xu, 2011; Dong & Torgler, 2013; Ramirez, 2014). These studies have identified several possible causes of bureaucratic corruption, including political institutions, the judicial system and the level of economic development. However, to date, religious culture has received scant attention in the analysis of government corruption. It is a surprising omission given the extensive literatures on how religion influences individual-level political behavior (Flevin & Ledet, 2013). One of the oldest criticisms of corruption points to morality, which closely relates to religion. From the point of morality and religious principles, corruption is wrong because it involves theft, dishonesty, abuse of others, and illegality (Douglas, 2007). Therefore, religious culture may play a unique role in curbing public officials’ rent-seeking activities, and thus depressing bureaucratic corruption. In some previous studies, religious beliefs are viewed as a contributory factor to corruption (Treisman, 2000; Mensah, 2014). However, these studies are mostly based on western contexts and focused on the Protestant Christian. The traditional eastern religions such as Buddhism and Taoism and their effects on corruption have been ignored. The majority of corruption literatures, on the other hand, are cross-national investigations that use subjective survey data. As Treisman (2007) admits, perception-based data reflect impressions of corruption intensity rather than corruption itself, meaning that the data are correlated with survey respondents' beliefs and other social and economic conditions. The main purpose of this paper is to use province-level objective data from China to investigate whether and how religion affects the prevalence of bureaucratic corruption. China presents an interesting case for this analysis not only because it is the largest transitional and developing country, but also because corruption has become 2 After the 18th CPC National Congress in late 2012, China launched a new wave of anti-graft campaign, which targeted both “tigers and flies”, referring to high and low ranking corrupt officials. According to public reports, more than 130 provincial and ministerial-level officials and over 200,000 officials at all levels have been charged in the past three years, among which even include Yongkang Zhou who was former Politburo Standing Committee Member, Caihou Xu who was former members of the Politburo and vice chairman of the Central Military Commission, Boxiong Guo who was former members of the Politburo and vice chairman of the Central Military Commission, and Rong Su and Ling Jihua who both are incumbent vice-chairman of the CPPCC.

2

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

more rampant since China launched its economic reforms. Although serious corruption in China has generated much literature (e.g., Gong, 2006; Dong & Torgler, 2013; Ramirez, 2014), few studies have examined the impact of religion on corruption in China, where the Communist government advocates atheism. Second, different from western developed countries, China lacks independent and efficient judicial system, so existing laws, regulations, and rules are performed poorly. When formal systems and ethical codes do not work effectively, we may rely more on informal power, such as religious culture, to curb corrupt behavior. North (1990) emphasizes the importance of informal institutions and suggests researchers to pay attention to economic consequence of informal systems. Williamson (2000) also argues that informal institutional arrangements such as culture and religion have an important impact on formal systems. In this sense, we predict that substitution effect between informal religious constraint and formal legal monitoring exists in reducing officials’ proclivity to corrupt crimes, and thus religious traditions can play much more positive role in curbing corruption in transitional China. One may presume religion does not work in China because Chinese Communist Party (CCP) members are atheism. China has 86 million CCP members by the end of 2013 and they are inclined to atheism. However, Chinese government advocates the freedom of religious beliefs, and more than 1.3 billion people do have the right to choose religious beliefs. Especially after the termination of political suppression in the modern Chinese society, long repressed religious beliefs were released (Du, 2013). The World Values Survey (2007) shows that 11% of Chinese people have religious beliefs. According to Chinese General Social Survey (CGSS), the population of religious believers in China has increased 120% from 2003 to 2010, and the annual growth rate reaches 18%. Some scholars even mentioned that, “a wave of religious climate is sprouting in China” (Yang, 2010). This paper shines light on the impact of religion on corruption in China and goes further to explore whether the substitution effect between informal religious constraint and formal legal supervision exists in curbing officials’ inclination to corruption. We construct a province-level index of “religious density” based on the number of religious sites, and examine whether religious adherence in different provinces affects regional corruption. Our empirical results show that provinces with stronger religious density are less likely to suffer from bureaucratic corruption, indicating that religion does matter and can reduce corruption. We also find that the negative association between religion and corruption is weaker in provinces with stronger law enforcement. This implies that, to some extent, religious constraint and legal supervision are two substitutive mechanisms in curbing corrupt crimes. In addition, after separating China’s native religions (i.e., Taoism and Buddhism) from foreign religions (i.e., Christianity and Islam), our findings indicate that native religions have stronger impact in reducing corruption than foreign religions. The anti-corruption effect of Buddhism is much more significant than Taoism. Our study contributes to the extant literature in several ways: First, to our knowledge, this paper is the first to examine the influence of religion on corruption in the context of China. Prior studies have identified several antecedents of corruption in 3

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

China (Yao, 2002; Gong, 2006; Cai, Fang, & Xu, 2011; Dong & Torgler, 2013; Ramirez, 2014), but the religious traditions have been ignored. A study on the role of religion playing in fighting corruption is meaningful since religion has influence on political preference and work ethic; it is also timely for China as the country is faced with severe official corruption. Second, our paper also adds to existing literature on the value of religion. Although religion has long been part of economic thought, more recent stream of the literature has focused on the relationship between religion and macro-economic growth. For example, Barro and McCleary (2003, 2006) indicate that macroeconomic development has a negative correlation with church attendance across countries (but a positive one with beliefs in heaven and hell). Guiso, Sapienza, and Zingales (2003) find that religious beliefs are associated with “good” economic attitudes. Wang and Lin (2014) reveal that, among the different religions, Christianity has the most significant effect on economic growth. These researches provide valuable insights, but they are mainly focused on economic growth and generally used country-level cross-sectional samples. Moreover, they focus on the US context in which country the majority of the residents are Christians, but provide little evidence on the influence of other religions. Different with these studies, we analyze the role of religion in China, using provincial- level panel data. The study also identifies itself from focusing on China, whose government advocates atheism. The results can reveal how religions affect peoples’ mind and unethical behavior in an invisible way. Finally, we provide strong and robust evidence that religious culture and legal supervision have reciprocal substitution effects in reducing corruption. Although the impact of the two factors on corruption has been examined independently, little is known about their interactions and the corresponding effects. Our study indicates that the negative association between religion and corruption is weaker for provinces with strong legal institutions. This finding is consistent with the view that religious constraint is an important alternative mechanism to reduce unethical behavior or corruption crimes in emerging markets like China, where formal legal systems are incomplete and many external monitoring mechanisms are still under construction. The rest of this paper is structured as follows. Section 2 presents a description of Chinese religious tradition and summary of related literature, and then develops our research hypotheses. Section 3 presents the methodology, including the variable measurement, data sources and empirical models specification. Section 4 reports empirical results and the robust check is presented in Section 5. We conclude in Section 6. 2. Institutional Background, Literature Review, and Hypotheses Development 2.1. Religion in Contemporary China China is a country with a long history of religious practice. The history of the five major religions being practiced can be traced back for centuries. Buddhism and Taoism are the two most influential religions. Taoism, which originated in China, was established around 25–220 AD in the late Eastern Han Dynasty. Buddhism was introduced from India as early as the 1st century and gained continuing popularity in China. In comparison, Christianity did not reach China until the 7th century AD. It 4

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

disappeared for hundreds of years until it was re-introduced at the end of the Ming dynasty in the 16th century. Islam can be dated back to a mission in 651 AD, and the first mosque is the Great Mosque of Xi’an. Some also consider that Confucianism should be seen as a kind of religion. Upon the establishment of the People’s Republic of China, the Communist Party of China, which controlled the government and was the authoritative source of acceptable policy positions on all matters, adopted the Marxist-Leninist view towards religion. According to Kolodner, a learned author on religion in China, the Marxist-Leninist perspective views religion as a product of history and further claims that religion will disappear “only when socioeconomic and cultural conditions have improved to the extent that people no longer require this ‘opiate.” Thus, religious activities were largely reduced from the founding of the People’s Republic of China in 1949 to the termination of the Cultural Revolution in 1976, because atheism is the fundamental doctrine of Chinese Communist Party by that time. Since 1978, China began an unprecedented open-up reform. During the process of reform and opening, CCP realized that people have more diversified spiritual demands which called on religious beliefs. In fact, religion can never be eradicated in any society. Since the adoption of the open door policy in 1978, China has become a more open society, tolerating more civil differences and disputes and lessening compulsory collective orthodoxy. From the 1980s, a number of monasteries, temples, mosques, churches, and some other religious sites are repaired and reopened for religious activities. An example was the rise of the “house church” movement, where Chinese Christians who prefer to worship outside the state-controlled religious movements have begun to meet in unregistered house churches which have mushroomed everywhere in China. This phenomenon is tolerated to a certain degree but not encouraged (Xiong, 2014). In the past 30 years, religious activities in China have been flourishing beyond expectation. Some Buddhist monasteries and Taoist temples have been crammed during holidays and festivals.3 The World Values Survey (2007) notes that 11% of people in China believe in a religion. Yang (2010) argues that this amount may be under-estimated. Based on the aforementioned, it cannot be denied that religion is indispensable in China. Especially, in modern Chinese society, polarization of rich and poor become rampant. People turn to traditional religions in search of comfort and transfer of resentment. Therefore, the number of religious activities and adherents are increasing rapidly. Ashiwa and Wank (2006) and Yang (2010) notice this phenomenon and urge religion-based research in China. 2.2. Literature Review Corruption has been defined as “the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as ‘capture’ of the state by elites and private interests”. Corruption frequently takes the form of bribery and extortion, but it also includes the abuse of insider information, 3 For example, the peak number of daily adherents crowded into a religious site (i.e., Famen Si, a famous Buddhist temple) is 216 thousand in order to worship the gods in 2012.

5

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

procurement fraud, embezzlement of public money, and money laundering. Empirical studies have demonstrated that corruption is associated with undesirable social outcomes such as lower economic growth, less foreign investment, more social unrest and so on (Mauro, 1995; Aidt, Javasri, & Vania, 2008; Aidt, 2009; Javorcik & Wei, 2009). Therefore, from a social welfare perspective, it is important to contain, and even reduce, the level of corruption. Reducing corruption requires a thorough understanding of its causes. To explain corruption, scholars take different approaches. The traditional approach, adopted by many political scientists and economists, focuses on economic and administrative factors. For example, Treisman (2000), in a comprehensive cross-country study, finds that long exposure to democracy, higher average income, and high levels of imports lead to a decrease in corruption, whereas decentralization encourages it. Glaeser & Saks’ (2006) analysis of within-country data for the U.S. indicates that economic development and education can decrease corruption, whereas income inequality and racial fractionalization may increase it. According to Treisman (2007), the negative relationship between the incidence of corruption and income level is the strongest. Goel and Nelson (2010) show that, historical and geographic factors, as well as the size and scope of government, significantly influence corruption rates. Recently, some researchers have criticized the traditional approach for not taking the moral and cultural dimensions of corruption into account. They considered religion as a starting point for understanding these dimensions, suggesting that religion influences corrupt actions by shaping cultural attitudes toward social hierarchy and malfeasance in government (Lipset & Lenz, 2000; Dreher, Kotsogiannis, & McCorriston, 2007). In fact, religion has long been part of economic thought. For example, Adam Smith (1776) suggests that participation in religion could be viewed as a rational action by which individuals enhanced the value of their human capital (Anderson, 1988). Later, Weber (1905) suggests that the Protestant ethic was at the core of the economic development of capitalism. Modern economic theory has revisited the analysis of religions. At the micro-economic level, religion has been linked to a large range of social decisions4. Nevertheless, religion-based research is relatively lacking for a long time. Iannaccone (1998) concludes that religious factor has been neglected by social science scholars and suggests that researchers should pay their attention to religious influence on human behavior, business behavior, and even economic development. There has been a growing interest in understanding how religion affects corruption. Many empirical studies have suggested that countries with strong hierarchical religions are more likely to suffer from corruption. For example, Treisman (2000) have examined the historical, economic, political and other socio-cultural settings as the antecedents of corruption. In a society where hierarchical religions dominate, it is costly for people to challenge corrupt public officials because compliance with social hierarchy and loyalty to family are the norm. In addition, state-sponsored hierarchical religions may not monitor and denounce abuse of public office as actively as Protestantism. Moreover, La Porta et al. (1997) argued that strong 4

see Iannaccone (1998) for a review. 6

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

hierarchical religions–Catholicism, Orthodox Christianity and Islam–discourage the formation of trust and in turn promote corrupt acts. Mensah (2014) finds that both cultural and religious differences are incrementally related to perceived corruption, even after controlling for other economic and political factors. Paldam (2001) considers religion as a proxy of culture and analyzes the corruption-religion relationship. He divides Christianity into two categories, and finds that there is a lower level of corruption in countries practicing reform Christianity than in non-Christianity as well as in pre-reform Christianity countries. He also warns that although questions such as why religions affect corruption are interesting, it is difficult to propose a single answer because ‘religions differ in many and subtle ways’. For example, Sanholtz and Gray (2003) document that the Protestant religion has a reducing effect on the level of perceived corruption and that the Catholic religion has a positive effect on the level of perceived corruption within nations. Overall, extant literature has examined the relationship between religion and corruption and suggested that there is a significant statistical relationship between them. However, paradoxes exist in these studies. Some of them show that religion as a cultural index has a positive effect on corruption, while others comes to the opposite conclusion (Shadabi, 2013). Especially, most studies in this field are based on developed countries and mainly focused on western religions (Protestantism and Catholicism). Little evidence was provided for the questions of whether and how eastern religions can play a role in resisting corruptions. In this paper, we address the aforementioned gap and use a sample of province-level data for the period of 1998-2009 to explore whether and how religion matters to bureaucratic corruption in transition China. The reasons why we choose the Chinese background are as follows. First, corruption in contemporary China, the largest transitional and developing country, is a longstanding problem. However, there have been no empirical studies comprehensively analyzing the impact of religious factors on corruption in China. Second, different from western developed countries, the formal anti-corruption systems (i.e. legal institutions and the rule of law) in China are still not complete, so informal power, such as culture and religious belief, may play much more positive role in curbing corruptive crimes. We expect that related studies based on the Chinese background will bring new insight into the determinants of corruption and also complement to the existing literature which was dominated by the U.S. samples. 2.3. Hypotheses Development Religion can influence human behavior and actions. Many scholars asserted that religion can explain why corruption is more rooted in one society than another (Treisman, 2000; Shadabi, 2013). Why to expect religious factors to influence corruptive activities? Beliefs associated with religions are known to influence a wide range of human behaviors. An individual’s ideas about what behavior is correct, as well as what consequences the deviant behavior will cause are to a great extent shaped by religious beliefs (Feavin & Ledet, 2013). Religious doctrines inform believers that “do-gooders” in this life will receive in the heaven, and the price “evildoers” will pay 7

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

in the hell. Moral behavior, and the resulting moral order of society, is enforced by religion. Hull and Bold (1994) agree that by shaping perceptions of the afterlife, religion encourages people to conduct in socially desirable ways. They go further by suggesting that in some cultural contexts the rule-enforcement power of religion may be stronger than other factors, such as government and community powers and rules. Guiso, Sapienza, and Zingales (2003) use international survey data and find that religious people had higher levels of trust in others and in the government, were less willing to break the law, and were more likely to believe that market outcomes are fair—attitudes that are generally associated with decreased corruption. Treisman’s (2000) cross-national research comes to a similar conclusion—religion reduces corruption because it assists in the organization of a civil society and makes citizens more likely to monitor elites. Prior literature also addressed the concerns about the ethical mechanisms through which religion can temper illicit behavior in the public sphere. Religion provides a language of ethics, as it serves as a constant reminder of what is considered good and evil. As such, religion may be translated into political virtuousness and integrity. Indeed, studies show that individual-level religiosity is usually connected to ethical political behavior (Weaver & Agle, 2002). For example, Longenecker, McKinney, and Moore (2004) find that respondents who indicated that religious interests were of high or moderate importance to them demonstrate a higher level of ethical judgment (less accepting of unethical decisions) than others. Using survey data, Conroy and Emerson (2004) indicate that religiosity (e.g., church attendance) is a significantly negative predictor of responses in unethical scenarios. All these researches indicate that persons with higher level of religiosity will be less likely to engage in illegal (corrupt) behavior. Note that any corrupt activities have great risks although it can bring enormously illegitimate gains. Because once these corruption crimes and illegal practices were detected and caught, severe punishment will be implemented. Culture matters to individual decision since the decision-makers’ perceptions of the risks and the rewards are, to some extent, affected by their values, beliefs, preferences, and constraints. Prior research has also shown a positive correlation between risk aversion and individual religiosity. For example, Miller and Hoffmann (1995) report a negative correlation between religiosity and self-reported attitudes towards risk and danger. Osoba (2003) uses individual level panel regressions to show that risk-averse individuals attend church more often than risk-seeking individuals, risk avoidance and church attendance are consistently positively correlated at the 1% level. Therefore, from the risk-averse viewpoint, religious factor may contribute to a public official’s willingness to participate in risky corrupt transaction. Therefore we hypothesize: Hypothesis 1: Religion is negatively associated with bureaucratic corruption. Corruption is derived from the abuse of public power. In order to prevent rent-seeking behavior, especially that of public officials, governments needs strong legal institutions. Thus, an effective legal system has been considered as a key component in resisting corruption (Herzfeld and Weiss, 2003; Dreher, Kotsogiannis, and McCorriston, 2007). A well-designed and enforced regulatory system can increase the costs of corruptive crimes and therefore resist corruption. China lacks 8

ACCEPTED MANUSCRIPT

MA

NU

SC R

IP

T

independent and efficient judicial system, and the existing laws, regulations, and rules are poorly enforced. When formal system and ethical codes cannot perform effectively, we may rely more on the informal power, such as religious culture, to curb corruptive behavior. North (1990) emphasizes the importance of informal institution and suggests researchers to pay attention to economic consequence of informal system. Williamson (2000) also argues that informal institutional arrangements have an important impact on formal systems. Therefore, we predict substitution effect between religious constraint and legal monitoring in influencing public officials’ inclination to corruption, and furthermore, religious tradition may play a greater role in anti-corruption activities in China as its legal system is weak. We use the interactive terms between religion and legal environment, rather than control variables, to address our concerns. The aim is to examine whether religion and law affect corruption jointly. This leads to our second hypothesis: Hypothesis 2: The negative association between religion and corruption is attenuated for regions with strong legal institutions. 3. Research Design

D

3.1. Model Specification

TE

To test Hypothesis 1, we estimate Eq. (1) to link corruption with religion and province-specific variables:

(1)

CE P

Corruptionit  1 Re ligionit   2GDPit  3GDPit 2   4Wageit  5 HCit  i  t   it

AC

In Eq. (1), the dependent variable (Corruption) refers to one of the two province-level corruption index: Corruption1 and Corruption2. Religion, the main independent variable, is measured as the number of religious sites within a given province (see “measures of religion” in detail). In Eq. (1), the coefficient on Religion (i.e., 1 ) captures the influence of regional religion intensity on bureaucratic corruption after controlling other determinants. Hypothesis 1 predicted that religion is negatively associated with corruption. Therefore, the coefficient 1 should be negative. Following Herzfeld and Weiss (2003), Glaeser and Saks (2006), Treisman (2007), and Dong and Torgler (2013), we include a set of control variables into Eq. (1). This specification measures GDP as the provincial GDP divided by the regional population and human capital (HC) as the proportion of the regional population that has completed a college degree. We also control for the role of the salary of civil servants (Wage).5 Finally, we control for province effects and time effects using province fixed effects (  i ) and year fixed effects (  t ), respectively. 5 There is a viewpoint called “nourish honesty with high salary”，which attend to achieve anti-corruption by considerably increasing income of civil servants.

9

ACCEPTED MANUSCRIPT To test Hypothesis 2, we estimate Eq. (2) including regional law environment (i.e., Law) and the interaction items between Religion and Law (i.e., Religion*Law): Corruptionit  1 Re ligionit   2 Lawit   3 Re ligionit  Lawit   4GDPit   5GDPit 2

T

 6Wageit   7 HCit   i   t   it

SC R

IP

(2) where Law denotes proxies for the external monitoring mechanisms. Hypothesis 2 predicted that the negative association between religion and corruption is attenuated for regions with strong legal institutions. Therefore, the coefficients for Religion*Law should be negative. 3.2. Measures of Corruption

AC

CE P

TE

D

MA

NU

We explore the impact of religion on corruption using a province-level dataset composed of information from all provincial areas in mainland China during the period 1998 to 2009. The dataset covers 22 provinces, 4 autonomous regions and 4 municipalities but excludes data from Tibet, Hong Kong, Macao and Taiwan.6 To ensure comparability of the corruption data, we use only corruption data from 1998 to 2009 because until the Fifth Session of the Eighth National People's Congress of China passed the 1997 Criminal Law (which includes Chapter VIII: Crimes of Embezzlement and Bribery), the definition and measurements of corruption-related crimes has changed. To measure the overall extent of bureaucratic corruption at the provincial level, we used two categories. The first category, which we call Corruption Frequency, is defined as Corruption1.We derive the number of registered cases of corruption per 10,000 people in one province and in a given year, which is listed in the China Procuratorial Yearbooks. Similar measures have been adopted in previous US-related studies (e.g. Glaeser & Saks, 2006). The second category, Corruption Seriousness (Fan, Lin, & Treisman, 2009) was marked as Corruption2. We collect the total amount of money involved in all registered corruption cases in one province in a given year which is listed in China Audit Yearbook, and then divide this number by the regional GDP. The two measures of corruption are complementary, capturing different dimensions that may not always coincide (official’s corruption could be frequent but tiny, or rare but large). We adopt conviction data7 because they offer a less subjective measure of corruption, and allow us to work with longer time spans (Glaeser & Saks, 2006). Fig.1. displays the trends of corruption in China from 1998 to 2009, and the resulting average regional corruption data are listed in Table 1. 6

As a part of China, the political system and law environment in Hong Kong, Macao and Taiwan are quietly different from mainland of China. Therefore, we exclude data from Hong Kong, Macao and Taiwan. In addition, we also remove Tibet from our sample because Tibetan Buddhism is very popular in Tibet and its religion environment seems utterly different from other regions in China mainland. 7 Theoretically, the conviction rate and the number of registered cases of corruption are different. However, in China they are likely to be highly correlated, if not identical. In most cases in China, suspect officials are first investigated by the discipline inspection commission of the Chinese Communist Party and its local branches. Only after they have obtained enough evidence, the discipline inspection commission and its local branches will refer corrupt cases to the procuratorates, and then the procuratorates will register the cases. Furthermore, the courts and the procuratorates are both controlled by the Chinese government. Therefore in only very few circumstances the courts will reject public prosecutions against corrupt cases. 10

SC R

IP

T

ACCEPTED MANUSCRIPT

NU

Fig.1. Crimes of corruption in China in the period 1998~2009. Note: Data sources: the variables in Fig 1 are collected or calculated by authors from China Procuratorial Yearbook, China Audit Yearbook and China Statistical Yearbook. The definitions of variables can be found in Table 2.

Region Guang

Corrup

Corrup

tion1

tion2

0.488

0.021

Region Henan

Corrup

Corrup

tion1

tion2

0.328

0.022

Region Corrup Jiangs

Corrup

tion1

tion2

0.264

0.013

u

0.466

0.050

Jiangxi

0.315

0.025

Hunan

0.257

0.018

Jilin

0.464

0.042

Hubei

0.310

0.031

Hainan 0.255

0.039

Liaoni

0.450

0.013

Shaanxi

0.310

0.032

Shang

0.254

0.021

0.251

0.025

0.246

0.022

0.241

0.015

jiang Shanxi

0.448 0.388

AC

Helong

CE P

ng

TE

Tianjin

D

xi

MA

Table 1 Statistics of average province-level corruption from 1998 to 2009.

Shando 0.375

0.028 0.042

hai Qinghai

0.306

0.022

Beijin g

Ningxia

0.301

0.043

Chong qing

0.023

Guizhou

0.295

0.016

ng

Sichua n

Xinjian 0.345

0.033

Zhejiang

0.280

0.015

Anhui

0.237

0.017

g Fujian

0.336

0.014

Yunnan

0.266

0.017

Gansu

0.210

0.022

Hebei

0.329

0.016

InnerMo

0.265

0.023

Guang

0.185

0.016

ngolia

dong

Note: Data sources: the data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook. The definition of variables can be found in Table 2. 3.3. Measures of Religion Religion is usually excluded from developing economics, especially in China. 11

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

This is because firstly, religion is a sensitive topic in China. Second, religions are hard to quantify. Previous studies usually adopt the number of the religious sites, the religious population proportion of the total population, or/and the extent of religious participation in a country or region as proxies for religion. It is also difficult to evaluate religiosity in China. Taking Buddhism as an example, most adherents’ activities are different from those of Catholic or Christian. A Catholic or a Christian goes to churches in a regular pattern (e.g., every Sunday), but a Buddhist does not go to monasteries regularly because religious ceremonies and rituals are irregular. It is also difficult to obtain accurate statistics on the numbers of the religious believers because quite a few religious believers are very conservative and discreet. Especially, people may hide their real beliefs for the reasons of being afraid of prejudice or not promoted. In short, to utilize the extent of religious participation as a proxy for religiosity in China is invalid. According to Stark and Finke (2000), religious culture is mainly determined by the supply side of religion, i.e. the amount of religious sites. Thus, as an alternative approach, we use the number of religious institutions as the measure of religion. A religious institution is an institution that is not-for-profit and established for religious purposes. It consists of churches, mosques and temples. In China, most religious activities are organized by the religious institutions, so more religious institutions means wider and deeper effect of religion on social life. Since the existence of religious institutions is objective, this statistical indicator reflects the religious belief more accurately than the number of followers (Wang & Lin, 2014). The data of religious sites comes from the system of Spatial Explorer of Religion.8 This system provides a detailed atlas which contains rich and reliable information of demographic and religious sites in all China’s geographies, including 31 Provinces, 345 Prefecture Cities, 2,873 Counties, and over 50,000 Townships. We calculate Religion1 as the number of religious sites in a given province, including Buddhist monasteries, Taoist temples, Islamic mosques and Christian churches. It is constructed using the starting year of the religious site, such that a site is counted if the current year is later than the starting year. Specifically we count the number of religious sites in province i and year t to measure the variable religionit. In addition, the last year for obtainable religious data in this database is 2004, while the period of other variables is from 1998 to 2009. Following previous studies, we linearly interpolate the data to obtain the values in the missing years (2005~2009). This method allows us the opportunity to study the time-series properties in our setting whereas the results still hold when we substitute the latest data on 2004 for the values in the next missing five years. Fig. 2 depicts the evolvement of the density of the religious institutions from 1998 to 2004. As seen, in a given province, there is an obvious time variance in the density, implying that there is sufficient within-province variation to drive fixed-effects estimation. In addition, in Qinghai, Xinjiang, Ningxia, Gansu, Fujian, 8

This system is developed by China Data Center of the University of Michigan (UMCDC), the Center for Religion and Chinese Society of Purdue University (CRCS), and the State Key Laboratory for Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS) of Wuhan University. Detailed information refers to http://chinadataonline.org/religionexplorer/. 12

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Zhejiang provinces, the number of the religious institutions is distinct from the other provinces, which illustrates significant between-province variation in religious density. Qinghai, Xinjiang, Ningxia, Gansu are major ethnic enclaves and large religious settlements in western China. In Zhejiang and Fujian, religions development has a longer history.9

Fig. 2. The evolvement of the density of religious institutions in provinces from 1998 to 2004. The vertical axis represents the number of religious institutions and the horizontal axis represents Year. Note: Data source is China Data Center at the University of Michigan. In China, Buddhism and Taoism are two most influential religions. Buddhism is China’s oldest import religion10 and Taoism is a genuinely Chinese religion with a 9

According to information released by the State Administration for Religious Affairs in P.R.C., Buddhism and Taoism were transmitted to Zhejiang province more than 1800 years ago. Among the “top ten fairy grottoes”, three are located in Zhejiang. While Fujian province does not harbor any significant religions, local folk beliefs predominant landscape is highly developed, including animals, Rain God, various folk beliefs such as ancestor worship. 10 Despite its origination in India, Buddhism is now generally recognized as native religion by 13

ACCEPTED MANUSCRIPT

SC R

IP

T

very long history. To assess the effects of these two religions on corruption, and also to improve the robustness of empirical results, we obtain additional data of religious sites from other sources. Since it is difficult to count all the monasteries and temples, regardless of whether they are big or small, notable or not, , we narrow sites for religious activities down to national famous religious sites, including 141 Buddhist monasteries and 21 Taoist temples. The data was obtained from the religious directory issued by the State Council of the People’s Republic of China in 1983. These monasteries and temples are considered to have more far-reaching influence because their historical development, religious heritage, and intergenerational inheritance in followers. We use these data to construct another alternative index to measure province-level religious density (Religion2).

NU

3.4. Measures of Other Variables

CE P

TE

D

MA

We obtain data of the variable of Law from the National Economic Research Institute’s Index of Marketization. Fan, Wang, and Zhu (2011) provides a comprehensive database on the marketization index and sub-indexes that proxy for the institutional development in a province or provincial municipal city in China from 1998 to 2010. These measures cover the following aspects of marketization: the relationship between the government and market, the development of non-state sectors, product market and factor market in the economy and the development of market intermediary and the legal environment. Similar with prior studies, we employ the sub-index of the development of market intermediary and the legal environment as the main measure of regional legal institutions. In addition, Wage is defined as the ratio of average official income to average salary of all employed staff in a given province; HC is defined as the proportion of the regional population that has completed a college degree. All of the variables are defined in Table 2. Table 2

AC

Variable definitions. Variable

Definition

Corruption1 The number of registered cases of corruption per 10,000 people in a given province Corruption2 The amount of money involved in all registered corrupt cases in a given province, and then divided by regional GDP Religion1

The number of religious sites in a given province based on the system of Spatial Explorer of Religion

Religion2 Law GDP

The number of national famous Buddhist monasteries and Taoist temples based on the list issued by the State Council A proxy variable for the degree of province-level legal environment, which is one sub- index of the Marketization index The annual GDP Per capita in a given province (Ten thousands RMB).

most Chinese people because of its long existing history and enormous adherents in China. Thus in this study, Buddhism is also classified into native religions. 14

ACCEPTED MANUSCRIPT

SC R

IP

T

All the monetary values were deflated with the base of year 1978 The ratio of salary of civil servants to average wages of staff and Wage workers in a given province The ratio of population with a college degree to total population in a HC given province Note: Data sources: the data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; Religion1 is extracted from the detailed atlas in the system of Spatial Explorer of Religion; Religion2 is obtained from the religious directory released by the State Council of China in 1983; Law data is from the Index of Marketization of China’s Provinces; GDP, Wage and HC data are from the China City Statistical Yearbook.

NU

4. Empirical Results 4.1. Descriptive Statistic

CE P

Descriptive statistics.

TE

D

MA

Table 3 repots summary statistics for main test variables. As shown in Table 3, Corruption1 and Corruption2 are reasonably distributed with mean of 0.313 and 0.025 respectively. The mean values of Religion1 and Religion2 are 1190 and 5.433, respectively. The variable Law has a mean value of 4.873 with a standard deviation of 3.097, which suggests that the province-level law environment during our sample period varies greatly. With reference to control variables, the mean values of GDP, Wage and HC are 0.175, 1.155 and 7.446, respectively. Table 3

AC

Variable N Mean Std Min Median Max Corruption1 330 0.313 0.175 0.154 0.293 3.017 Corruption2 358 0.025 0.021 0.0004 0.020 0.189 Religion1 360 1190 1289 5 661 5847 Religion2 30 5.433 5.090 0 4 14 Law 360 4.874 3.097 1.15 4.01 19.89 GDP 360 0.175 0.122 0.046 0.130 0.689 Wage 360 1.155 0.144 0.899 1.132 1.767 HC 360 7.446 5.592 0.5 6.01 35.91 Note: Data sources: the data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; Religion1 is extracted from the detailed atlas in the system of Spatial Explorer of Religion; Religion2 is obtained from the religious directory released by the State Council of China in 1983; Law data is from the Index of Marketization of China’s Provinces; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. 4.2. Multivariate Test of Hypotheses Hypothesis 1 predicts that religion is negatively associated with corruption. Table 4 reports the fixed effect (FE) regression results of corruption on religion and 15

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

other determinants. As shown in Columns (1) of Table 4, the coefficient on Religion1 is negative and significant at the 10 % level, suggesting that religion does matter and can reduce government corruption. This result is consistent with Hypothesis 1 and provides support for the following view: as an important implicit constraint, religious heritage can contribute to curbing official’s unethical behavior and alleviate regional corruption situation. Besides corruption frequency (Corruption1), we also test the impact of religion on corruption seriousness (Corruption2) which is shown in Columns (3) of Table 4. Similarly, the coefficient on Religion1 is significantly negative (-4.783 with t=-1.67). This result lends further support to Hypothesis 1. Hypothesis 2 predicts that regional legal institutions attenuate the negative association between religion and corruption. Table 4 reports the FE regression results. As reported in Columns (2) and (4) in which we adopt Law as the proxy for regional legal institutions, the coefficients on Regilion1 are all negative and significant which is again consistent with Hypothesis 1. Moreover, the coefficient on Law are all negative and significant at the 1% level (-0.123 with t =-7.81, -0.287 with t =-6.39, respectively), which indicates that external legal supervision is an important mechanism to restrict public officials’ rent-seeking activities and depress bureaucratic corruption. In addition, the coefficients on the interaction terms, that is, Religion1×Law are all positive and significant at the 1% or 5% level (0.216 with t=6.45; 0.203 with t =2.32, respectively), suggesting that the negative association between religion and corruption is weaker for provinces with strong legal institutions. These results provide strong and consistent support to Hypothesis 2. In terms of the control variables in Columns (1)–(4) of Table 4, GDP is significantly positively related to both Corruption1 and Corruption2; at the same time, the associations between GDP2 and Corruption1 and Corruption2 are both significantly negative. The relationship between the level of GDP and corruption is controversial in the existing empirical literature. Our estimates are consistent with Del Monte and Papagni (2007) who also identify an inverted U curve relationship between corruption and economic development. Similar with Herzfeld and Weiss (2003), our model shows that the variable of Wage has a significantly negative sign in both cases. These results provide some support to the prevalent notion that corruption declines with the level of civil servant income. Table 4 Fixed effects panel analysis of religious impact on bureaucratic corruption. Dependent variable: Corruption1 Dependent variable: Corruption2 Variable (1) (2) (3) (4) *** -1.612 (-6.37 Religion1 -0.206*(-1.88) -4.783*(-1.67) -1.520**(-2.34) ) Religion1×La 0.216***(6.45) 0.203**(2.32) w -0.123***(-7.81 Law -0.287***(-6.39) ) 10.830***(5.38 GDP 2.817***(4.78) 4.400***(7.19) 8.257***(5.03) ) 16

ACCEPTED MANUSCRIPT

0.004(0.73)

-0.005(-0.85)

-0.102***(-6.93 )

-0.943***(-9.20 )

-0.489***(-3.96 )

-0.911**(-2.05)

Yes

Yes

Yes

Yes

Wage HC Constant Province Fixed Effect Year Fixed Effect N

0.027(0.08)

-0.686**(-2.03) -0.032**(2.20)

-2.254***(-6.20)

Yes

Yes

Yes

Yes

330 358 358 10.525(<0.001 24.432(<0.001 F(p value) 6.929(<0.001) 24.506(<0.001) ) ) Adj_R2 0.224 0.343 0.691 0.542 Notes: (1) This table presents fixed effects regression results for the effect of religion on corruption using the following Models:

MA

NU

330

-1.156(-0.23)

IP

-4.746(-5.72)

-10.583***(-4.67 )

T

-0.480***(-4.48 )

-5.412***(-6.81 ) *** -0.403 (-3.70 )

SC R

GDP

2

Corruptionit  1 Re ligionit   2GDPit  3GDPit 2   4Wageit  5 HCit  i  t   it

D

(1)

Corruptionit   1 Re ligionit   2 Lawit   3 Re ligionit  Lawit   4GDPit   5GDPit   6Wageit

TE

2

 7 HCit   i   t   it

AC

CE P

(2) (2) All regressions include province fixed effects and year fixed effects, t-statistics based on standard errors which are clustered at the province level are reported in parentheses. The data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; Religion1 is extracted from the detailed atlas in the system of Spatial Explorer of Religion; Law data is from the Index of Marketization of China’s Provinces; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. *Represent the parameter estimator is significant at 0.1 level. **Represent the parameter estimator is significant at 0.05 level. ***Represent the parameter estimator is significant at 0.01 level. 4.3. Further Tests of the difference in anti-corruption effect among different kinds of religions The system of Spatial Explorer of Religion reports the concrete information on religious sites of different kinds of religions, including Buddhist monasteries, Taoist temples, Islamic mosques and Christian churches. To address the concern about whether different kinds of religions have asymmetric influences on corruption, we distinguish different religious types and re-estimate Eq. (1). Regression results are presented in Table 5. We include four religions in the same regression model simultaneously. As 17

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

shown in Columns (1) and (4) of Table 5, the coefficients on Buddhism are both negative and significant at the 5% level (-0.072 with t =-2.35, -0.180 with t = -2.49, respectively). However, the coefficients on other three religions (i.e., Taoism, Christianity and Islam) are mostly insignificant. It implies that the anti-corruption effects of Buddhism are stronger than those of other religions. In other words, only Buddhism matters and can reduce bureaucratic corruption. Taoism, Christianity and Islam have no significant influence on anti-corruption. Our findings are consistent with Du et al. (2013) who also suggested that different religions have asymmetric influence on curbing manager’s unethical behavior and among them Buddhism has the most significant positive value. Buddhism is China’s oldest import religion and Taoism is a genuinely Chinese religion with a very long history. Islam and Protestantism came to China later. We categorize Buddhism and Taoism as native religions, and define Islam and Christianity as foreign religions. Based on this categorization, we re-estimate Eq. (1). As shown in Columns (2) and (5) of Table 5, when Native and Foreign are included in the same regression, the positive role of religion in reducing corruption only exists for native religions. These results indicate that, native religions (i.e., Taoism and Buddhism) play much more positive role in anti-corruption than foreign religions (i.e., Christianity and Islam). Table 5 Further tests of difference in anti-corruption effect among different kinds of religions. Dependent variable: Corruption1 Dependent variable: Corruption2 Variabl e (1) (2) (3) (4) (5) (6) ** * ** Buddhi -0.072 (-0.054 (-1. -0.180 (-0.226***(sm1 2.35) 80) 2.49) 3.35) Taoism 0.027(0.77 0.011(0.37 0.068(0.82 0.094(1.31 ) 1) 0) ) 1 0.009(0.60 -0.046(-1. Christ1 ) 31) -0.025(-1. 0.070*(1.8 Islam1 57) 3) *** -0.038 (-0.166***(Native 3.02) 5.32) Foreig -0.011(-0. 0.071**(2. n 82) 08) *** *** *** 1.794 (2. 2.057 (3. 2.074 (3. 2.334(1.57 2.268(1.61 1.545(1.05 GDP 94) 67) 55) ) ) ) *** *** *** ** ** -3.679 (- -3.894 (- -3.871 (- -6.696 (- -6.732 (- -6.151**(2 GDP 4.42) 4.96) 4.77) 2.57) 2.58) 2.29) *** *** -0.120(-1. -0.132(-1. -0.114(-1. -0.861 (- -0.890 (- -0.874***(Wage 24) 40) 18) 3.64) 3.71) 3.68) 0.007(1.28 0.005(0.98 0.005(0.88 -0.015(-0. -0.010(-0. -0.007(-0. HC ) ) ) 84) 61) 411) *** *** *** *** *** Consta -1.283 -1.289 -1.310 -3.059 -3.048 -2.991*** 18

ACCEPTED MANUSCRIPT

NU

SC R

IP

T

nt (-10.35) (-10.44) (-10.65) (-9.73) (-9.64) (-9.51) Provin ce Yes Yes Yes Yes Yes Yes Fixed Effect Year Fixed Yes Yes Yes Yes Yes Yes Effect N 330 330 330 358 358 358 F(p 5.272(<0. 5.733(<0. 5.732(<0. 19.607(<0 21.567(<0 21.504(<0 value) 001) 001) 001) .001) .001) .001) 2 Adj_R 0.198 0.196 0.196 0.498 0.495 0.494 Notes: This table presents fixed effects regression results for the anti-corruption effect of different religions using the following Model: 2 Corruptionit  1 Re ligionit   2GDPit   GDP 3 it   Wage 4 it   HC 5 it + i   t   it

CE P

TE

D

MA

All regressions include province fixed effects and year fixed effects, t-statistics based on standard errors which are clustered at the province level are reported in parentheses. The data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; The data related to religion is extracted from the detailed atlas in the system of Spatial Explorer of Religion; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. *Represent the parameter estimator is significant at 0.1 level. **Represent the parameter estimator is significant at 0.05 level. ***Represent the parameter estimator is significant at 0.01 level.

AC

We go further to focus exclusively on the two native religions and examine their impact on corruption. As presented in Columns (3) and (6) of Table 5, the coefficients on Buddhism are still negative and significant at the 10 % level or 1% level (-0.054 with t =-1.80, -0.226with t =-3.35, respectively), but the coefficients on Taoism are not significant (0.011 with t =0.371, 0.094with t =1.31, respectively). These results confirm that only Buddhism matters and can reduce corruption. With respect to control variables in Table 5, the signs and significances are qualitatively similar to those in Table 4. These findings provide additional evidence for religious literature about whether different religions have asymmetric consequence on bureaucrat behavior. It is possible that differentiations among the doctrines of the religions and their prevalence in China can explain the differences. For example, the goal of Buddhism is transcendentalism, not secular interests. It focuses on freedom from suffering (Wang & Lin, 2014). It emphasizes “Karma” or the belief that kindness always begets kindness. These doctrines may restrain officials to behave, for example, they may avoid the expansion of material desire or the abuse of power, which will then reduce corruption. Taoism advocates the pursuit of healthiness of life and physical immortality which have less help for ethical human behavior. As to Christianity and Islam, they are not the mainstream of reliefs in modern China, which may mitigate their impact on the behaviors of Chinese people. Therefore, we do not 19

ACCEPTED MANUSCRIPT find the significant association between foreign religions and the reduction of corruption. 5. Robustness Checks

T

5.1. Discussion on Potential Endogeneity between Religion and Corruption

AC

CE P

TE

D

MA

NU

SC R

IP

One concern in this study is the potential endogeneity. For example, religion may be correlated with regional culture which affects corruption simultaneously. Unobservable time-variant or time-invariant factors may also be correlated with corruption. Furthermore, there could be reverse causality, in the sense that religion thrives where corruption is low. We address this issue as follows. First, the direction of causality between religion and corruption is not a major threat. This is for the following reasons: (1) because of the restriction of the census register system, population mobility or migration is not universal in China. Especially beside high-ranking officials who are appointed by the central government, the vast majorities of civil servants work and promote in their homeplace. Therefore, our sample naturally excludes the possibility that the behavior of local governments attract officials of certain faiths (or no faith) to work together. In other words, it is more likely that the religious make-up of the population cause officials to behave in a certain way (e.g., lower corruption in our case). (2) As noted by Miller (2000), being irreligious represents risk-taking behavior in Western societies, but in Eastern societies, it will not always hold. As the relations-based economy and the second largest economy in the world, religions in Chinese society tend to be less exclusive and the emphasis is on personal behavior. Therefore, it is impossible that a region’s behaviors attract officials with common religious belief to work together. Second, besides the simultaneous religious sites gathered from the system of Spatial Explorer of Religion, we also built an alternative measure of religious density (Religion2), using the religious information released by the State Council in 1983. The period of variable Corruption in our study is distributed between 1998~2009, which is far lagged behind the observation time of independent variable Religion2. To some extent, this approach overcomes the reverse-causality between religion and corruption. Since the variable Religion2 is time invariable, it eliminates the within transformation in a fixed-effect model. Referring to Wang and Lin (2014), we then use a fixed effect vector decomposition approach (FEVD) (Plumper & Troeger, 2007) to estimate Eq. (1). This method suggests a three-stage procedure for the estimation of time-invariant in panel data models with unit effects. The first stage runs a fixed-effects model without time-invariant variable to obtain the unit effects, the second stage breaks down the unit effects into a part explained by the time-invariant and an error term to obtain the error term, and the third stage re-estimates the first stage by pooled OLS including the time-invariant variables plus the error term of the stage 2. As reported in Column (1) and (5) of Table 6, the coefficient on Religion2 are negative and significant. Column (2) and (6) also show that the coefficient on Buddhism2 are negative and significant at the 1% level while the coefficients on Taoism2 are insignificant, which further indicates that Buddhism exerts more positive impact on curbing corruption than Taoism. 20

ACCEPTED MANUSCRIPT

(-6.96)

(-5.17) *

-0.077

-0.077

-0.209

***

***

**

4.887*

(-3.18) 0.006 (0.20) 4.887*

4.595*

(-2.67) 0.012 (0.02) 5.204*

14.740*

(-2.88) -0.033 (-0.03) 14.740

14.893*

-0.007 (-0.01) 13.870

**

**

**

**

**

***

**

**

(5.26) -6.232

(5.26) -6.232

(3.27) -6.040

(4.02) -6.382

(8.23) -24.180

***

***

***

***

***

(-4.10) 0.086 (0.65)

(-4.10) 0.086 (0.65)

(-3.36) 0.161 (1.10

(-3.87) 0.104 (0.83)

0.003 (0.46)

0.003 (0.46)

0.002 (0.32)

0.0004 (0.06)

(-7.80) 0.069 (0.25) -0.071**

-1.831

-1.944

-4.287**

***

***

*

Taoism2

GDP2 Wage HC

CE P

Buddhis m2

GDP

(-2.10)

AC

(-2.34)

TE

2

D

MA

NU

SC R

IP

T

Third, as to the way through which religion affects corruption, we do not think political effects exist in China. Under the leadership of the CCP, the political effect of religion is limited, or at least, it does not function as is in western countries. The central government in China has special religious policies in minority nationality areas to support the freedom of religion for the inhabitants. As known, the economic growth of minority nationality areas is lagged far behind the other regions of China. As pointed by many extant literatures, economic growth level is an important determinant of corruption. In order to exclude the effect of omitted minority nationality regions which affect both religion and corruption, we remove the four ethnic autonomous regions11 from our sample and re-estimate Eq. (1). The results are presented in Column (3), (4), (7) and (8) of Table 6, It is shown that the results are qualitatively similar to those reported previously and thus provide additional support to H1. Table 6 The test of religious impact on bureaucratic corruption using an alternative religion variable. Dependent variable: Corruption1 Dependent variable: Corruption2 Variable (1) (2) (3) (4) (5) (6) (7) (8) ** ** -0.015 -0.014 -0.044 -0.038 Religion ** ** * *

-1.804 Constan

***

t

(-10.2 6)

-1.895 (-0.85)

(-9.64)

-0.173 (-0.94)

(2.88) (7.79) (2.02) -24.18 -22.731 -22.16 *** 0** 8** (-2.44) (-7.15) (-2.00) 0.069 0.512* 0.068 (0.13) (1.67) (0.12) * ** -0.071 -0.090 -0.087*

*

*

*

**

(-4.84)

(-2.57)

(-5.62)

(-2.85)

(-3.52) (-15.08)

-4.603* (-1.75)

-4.807** *

(-15.05)

-4.465 (-0.90)

11

They are Xinjiang Uygur Autonomous Region, Ningxia Hui Autonomous Region, Inner Mongolia Autonomous Region and Guangxi Zhuang Autonomous Region. 21

ACCEPTED MANUSCRIPT Province Fixed

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

330 41.38 (<0.00 1) 0.680

330 41.38 (<0.00 1) 0.680

283 37.03 (<0.00 1) 0.672

283 50.17 (<0.00 1) 0.829

358 49.95 (<0.001 ) 0.695

358 29.29 (<0.00 1) 0.695

310 48.98 (<0.001 ) 0.682

310 28.72 (<0.00 1) 0.685

N F(p value) Adj_R2

IP

Effect

SC R

Year Fixed

T

Effect

NU

Note: This table presents regression results using an alternative measurement of religious density based on the following Model: Corruptionit  1 Re ligionit   2GDPit  3GDPit 2   4Wageit  5 HCit +i  t   it

CE P

TE

D

MA

All regressions include province fixed effects and year fixed effects, t-statistics based on standard errors which are clustered at the province level are reported in parentheses. The data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; Religion2 is obtained from the religious directory released by the State Council of China in 1983; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. *Represent the parameter estimator is significant at 0.1 level. **Represent the parameter estimator is significant at 0.05 level. ***Represent the parameter estimator is significant at 0.01 level. 5.2. Spatial Spillover effects from “Nearby” provinces

AC

Another concern of this study is the spatial spillover effects. As denoted by Gruber (2005), even though we control for provincial religious density, there are other spillover effects from “nearby” provinces. This can be tested either by including neighboring provinces' religious density in the regression, or by constructing some distance-weighted index of religious density. We adopt the former approach. As shown in Table 7, the coefficients on variable NeighborReligion1 and NeighborReligion2 are both negative and significant at the level of 1% and the level of 10%, which indicates that religious culture from nearby provinces also play a positive role in curbing corruption and thus provide some supportive evidence to spatial spillover effects. We also find that, most empirical results remain qualitatively unchanged even when we introduce the religious density of neighboring provinces into the models. These results suggest that our conclusions still hold after considering the spatial spillover effects from “Nearby” provinces’ religious sites. Table 7 Tests of spatial spillovers effects of religious density from “Nearby” provinces. Dependent variable:Corruption1 Variable (1) (2) 22

ACCEPTED MANUSCRIPT -1.035***(-4.3

Religion1

1)

TE

D

MA

NU

SC R

IP

T

Religion1×Law 0.142***(4.68) NeighborReligion1 -0.473***(-3.36) NeighborReligion1×Law 0.058**(2.39) Religion2 -0.022***(-3.24) Religion2×Law 0.002**(2.15) Neighbor Religion2 -0.003*(-1.71) NeighborReligion2×Law 0.001*(1.73) Law -0.162***(-6.98) -0.042***(-3.88) GDP 3.934***(6.507) 5.181***(11.58) GDP2 -4.753***(-5.845) -6.158***(-9.70) Wage -0.183*(-1.86) 0.055(0.74) HC -0.005(-0.96) 0.005*(1.94) Constant -0.410***(-4.17) -1.687***(-16.54) Province Fixed Effect Yes Yes N 330 330 Year Fixed Effect Yes Yes F(p value) 9.968(<0.001) 66.000(<0.001) Adj_R2 0.353 0.685 Note: This table presents fixed effects regression results after considering spatial spillover effects from nearby provinces using the following Model: Corruptionit  1 Re ligionit   2 Re ligionit  Lawit + 3 Neighbor Re ligionit   4 Neighbor Re ligionit  Lawit   5 Lawit   6GDPit

CE P

 7 GDPit 2   8Wageit   9 HCit   i   t   it

AC

All regressions include province fixed effects and year fixed effects, t-statistics based on standard errors which are clustered at the province level are reported in parentheses. The data related to Corruption are obtained from China Procuratorial Yearbooks; Religion1 is extracted from the detailed atlas in the system of Spatial Explorer of Religion; Religion2 is obtained from the religious directory released by the State Council of China in 1983; Law data is from the Index of Marketization of China’s Provinces; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. *Represent the parameter estimator is significant at 0.1 level. **Represent the parameter estimator is significant at 0.05 level. ***Represent the parameter estimator is significant at 0.01 level. 5.3. Re-estimate Eq. (1) and Eq. (2) using data between 1998~2004 In our study, the last time point of which we obtained the religious data in the system of Spatial Explorer of Religion is 2004, while the period of corruption and other variables is from 1998 to 2009. In order to match the data from different channels and time points, we delete the samples between 2005~2009 and re-estimate Eq. As shown in Table 8, the empirical results in most cases are consistent with Table 4, suggesting that our basic conclusions remain qualitatively unchanged when we adopt clean data. 23

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Table 8 Re-estimate Eq. (1) and Eq. (2) only using data between 1998~2004. Dependent variable: Corruption1 Dependent variable: Corruption2 Variable (1) (2) (3) (4) *** *** -0.471 (-3.30 -1.744 (-4.43 -10.545**(-2.30 Religion1 1.940(0.42) ) ) ) Religion1×La 0.223***(3.76) 0.833***(4.19) w -0.157***(-5.96 -0.416***(-7.47 Law ) ) *** -22.822 (-3.50 GDP 1.795***(2.70) 4.831***(5.61) -0.543(-0.12) ) -3.375***(-3.58 -5.648***(-5.17 GDP2 9.725(1.04) 2.015(0.34) ) ) Wage -0.094(-0.51) -0.181(-0.97) -1.911**(-2.12) 0.365(0.81) -0.043***(-2.79 * ** HC 0.006(0.78) -0.015 (-1.79) -0.048 (-2.11) ) *** -1.177 (-5.95 Constant -0.530**(-2.15) 1.847*(1.73) -0.959(-1.48) ) Province Yes Yes Yes Yes Fixed Effect Year Fixed Yes Yes Yes Yes Effect N 207 207 209 209 F(p value) 3.556(<0.001) 6.197(<0.001) 5.936(<0.001) 15.972(<0.001) 2 Adj_R 0.120 0.247 0.447 0.770 Note: This table presents re-estimation of Eq. (1) and Eq. (2) using data between 1998~2004. All regressions include province fixed effects and year fixed effects, t-statistics based on standard errors which are clustered at the province level are reported in parentheses. The data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; Religion1 is extracted from the detailed atlas in the system of Spatial Explorer of Religion; Law data is from the Index of Marketization of China’s Provinces; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. *Represent the parameter estimator is significant at 0.1 level. **Represent the parameter estimator is significant at 0.05 level. ***Represent the parameter estimator is significant at 0.01 level. 5.4. Re-estimate Eq. (1) and Eq. (2) using an alternative measurement of religious density It makes better sense to have the religious density measured by the number of religious sites per some population, instead of the total number of religious sites, because it aligns with the measure of the dependent variables and can remove the correlation that larger provinces have more temples/churches (province size may or 24

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

may not affect corruption). Therefore, we replace province-level religious density with the number of religious sites per 10,000 populations in a given province (marked as Per-Religion1), and then re-estimate Eq. (1) and Eq. (2). Table 9 shows that our main results are robust even when we change the measuring method of religious density. Table 9 Re-estimate Eq. (1) and Eq. (2) using alternative measurement of religious density. Dependent variable: Dependent variable: Corruption2 Corruption1 Variable (1) (2) (3) (4) *** *** -0.670 (-2.9 -0.159 (-5.3 Per-Religion1 -0.566(-0.89) -0.231**(-2.33) 5) 6) Per-Religion1×La 0.030***(4.56) 0.057***(2.72) w -0.127***(-8.3 -0.311***(-7.56 Law 7) ) *** *** *** *** GDP 4.689 (5.40) 4.318 (7.33) 14.515 (3.98) 7.884 (5.08) -5.648***(-4.9 -5.159***(-6.7 -23.974***(-3.7 -10.025***(-4.7 2 GDP 1) 6) 8) 0) *** -0.885 (-3.53 Wage 0.098(0.78) -0.054(-0.54) -0.393(-1.014) ) *** -0.073 (-2.90 HC 0.001(0.03) -0.005(-1.01) -0.035**(-2.50) ) -1.521***(-6.6 -0.890***(-6.6 -4.214***(-6.01 -1.880***(-5.65 Constant 4) 1) ) ) Province Fixed Yes Yes Yes Yes Effect Year Fixed Yes Yes Yes Yes Effect N 330 330 358 358 15.892(<0.001 26.313(<0.001 F(p value) 13.77(<0.001) 9.344(<0.001) ) ) 2 Adj_R 0.636 0.313 0.652 0.561 Note: This table presents re-estimation of Eq. (1) and Eq. (2) using religious density measured by number of religious sites per 10,000 populations. All regressions include province fixed effects and year fixed effects, t-statistics based on standard errors which are clustered at the province level are reported in parentheses. The data related to Corruption are obtained from China Procuratorial Yearbooks and China Audit Yearbook; Religion1 is extracted from the detailed atlas in the system of Spatial Explorer of Religion; Law data is from the Index of Marketization of China’s Provinces; GDP, Wage and HC data are from the China City Statistical Yearbook. The definition of variables can be found in Table 2. *Represent the parameter estimator is significant at 0.1 level. **Represent the parameter estimator is significant at 0.05 level. 25

ACCEPTED MANUSCRIPT ***Represent the parameter estimator is significant at 0.01 level. 6．Conclusions

AC

CE P

TE

D

MA

NU

SC R

IP

T

In the past thirty years, China’s economy has grown enormously. This growth is coupled with corruption. Despite the introduction of several anti-corruption campaigns, corruption in China is still common. Literature has addressed concerns for the influence of public governance, legal monitoring, anti-corruption investigation and other determinations which belong to formal system arrangement on corruption. However, little evidence has provided on whether and how religion, as one of important informal power, influence public officials’ inclination to corruption. Our study fills this gap and examines the impact of religions on corruption crimes in the context of China using provincial panel data. Our empirical results show that at the provincial level, religious density is significantly negatively associated with regional corruption, suggesting that religion, as an informal institutional factor and a part of social culture, has a significant impact on officials’ unethical behavior and can reduce the likelihood of them engaging in corruptive practices. Moreover, the negative association between religion and corruption is weaker in provinces with stronger legal institutions and the rule of law. It implies that implicit constraints from religions can serve as an alternative mechanism to prevent corrupt crimes in emerging markets like China where formal institution system and external monitoring mechanisms are incomplete. Furthermore, after separating the native religions from the foreign religions, our findings indicate that the anti-corruption effects of native religions (i.e., Taoism and Buddhism) are more prominent than foreign religions (i.e., Christianity and Islam). Our research has some practical implications. First, corruption has become a major public policy issue across the globe. Our findings suggest that religion, as an informal institution, can reduce corruptive crimes through restraining public officials from conducting unethical even illicit rent-seeking activities. Second, scholars have long focused on political system and considered that the improvement of legal supervision is the major channel to fight against corruption. Our study suggests that in emerging markets like China, where modern political systems, legal institutions, and codes of ethics for civil servants are being formed, religions can serve as an alternative mechanism to curb corruption. Our study has its limitations. First, it only examines the influence of four major religions (i.e., Buddhism, Taoism, Islam, and Christianity). We did not analyze the impact of other religions mainly because of the data constraints. Second, we use local religious density as the alternative variable for religiosity because the real data on the individual-level is not available. We concede that our measure may not fully capture the underlying spiritual status of Chinese civil servants. Nonetheless, this approach is relatively objective. It is difficult to assert people’s spiritual status, so information from personal interviews often suffers from subjective bias. In 2010, Chinese General Social Survey (CGSS) made a sampling survey of individual religiosity of Chinese citizens. Using this dataset which covers 11783 respondents cross provinces, we find

26

ACCEPTED MANUSCRIPT that individual religiosity is significantly related to local religious density,12 therefore, local religious density can be considered as a powerful alternative for individual religiosity.

SC R

IP

T

Acknowledgments We acknowledge the National Natural Science Foundation of China (approval number 71102063; 71232004; 71572019) and the Fundamental Research Funds for the Central Universities (approval number 106112015CDJSK02XK12; 106112016CDJXY020003) for their financial support. We are also especially grateful to editors and anonymous referees for their helpful comments and suggestions. Any remaining errors are our responsibility.

NU

References

AC

CE P

TE

D

MA

Aidt, T. (2009). Corruption, institutions and economic development. Oxford Review of Economic Policy, 25(2), 271-291. Aidt, T., Javasri, D., & Vania, S. (2008). Governance regimes, corruption and growth: theory and evidence. Journal of Comparative Economics, 36(2), 195-220. Anderson, G. M. (1988). Mr. Smith and the preachers: The economics of religion in the wealth of nations. Journal of Political Economy, 96(5), 1066-1088. Ashiwa, Y., & Wank, D. L. (2006). The politics of a reviving Buddhist Temple: state, association, and religion in Southeast China. The Journal of Asian Studies, 65(2), 337-359. Barro, B., & McCleary, R. (2003). Religion and economic growth across countries. American Sociological Review, 68(5), 760-781. Barro, B., & McCleary, R. (2006). Religion and economy. Journal of Economic Perspectives, 20(2), 49-72. Cai, H., Fang, H., & Xu, C. L. (2011). Eat, drink, firms and government: an investigation of corruption from entertainment and travel costs of Chinese firms. Journal of Law and Economics, 54(1), 55-78. Conroy, S. J., & Emerson, T. L. N. (2004). Business ethics and religion: religiosity as a predictor of ethical awareness among students. Journal of Business Ethics, 50(4), 383-396. Del Monte, A., & Papagni, E. (2007). The determinants of corruption in Italy: regional panel data analysis. European Journal of Political Economy, 23(2), 379-396. Dong, B., & Torgle, B. (2013). Causes of corruption: Evidence from China. China Economic Review, 26, 479-503. Douglas, S.B. (2007). Global corruption and religion: An empirical examination. Journal of Global Ethics, 3(1), 69–85. Dreher, A., Kotsogiannis, C., & McCorriston, S. (2007). Corruption around the world: Evidence from a structural model. Journal of Comparative Economics, 35(3), 12

Based on the survey data released by CGSS, we find that the correlation coefficients between individual religiosity and Religion1 and Religion2 used in this study are 0.642 and 0.337, and significant at the level 1% and 10%, respectively. 27

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

443–466. Du, X. Q. (2013). Does religion matter to owner-manager agency costs? Evidence from China. Journal of Business Ethics, 55(3), 373-386. Fan, C. S., Lin, C., & Treisman, D. (2009). Political decentralization and corruption: Evidence from around the world. Journal of Public Economics, 93(1-2), 14-34. Fan, G., Wang, X., & Zhu, H. (2011). NERI Index of Marketization of China’s Provinces 2011 Report. Economic Science Press. Flavin, P., & Ledet, R. (2013). Religion and government corruption in the American States. Public Integrity, 15(4), 329-343. Glaeser,E., & Saks,R. (2006). Corruption in America. Journal of Public Economics, 90(6/7), 1053-1072. Goel, R. K.,, & Nelson, M. A. (2010). Causes of corruption: history, geography and government. Journal of Policy Modeling, 32(4), 433-447. Gong, T. (2006). Corruption and local governance: The double identity of Chinese local governments in market reform. The Pacific Review, 19, 85-102. Gruber, J. (2005). Religious market structure, religious participation, and outcomes: Is religion good for you? Working paper, ssrn.com. Guiso, L., Sapienza P., & Zingales, L. (2003). People’s opium? Religion and economic attitude. Journal of Monetary Economics, 50(1), 225-282. Herzfeld,T., & Christoph, W. (2003). Corruption and legal (in)effectiveness: An empirical investigation. European Journal of Political Economy, 19(3), 621-632. Hull, B., & Frederick B. (1994). Hell, religion, and cultural change. Journal of Institutional and Theoretical Economics, 150(3), 447-464. Iannaccone, L. (1998). Introduction to the economics of religion. Journal of Economic Literature, 36(3), 1465-1495. Javorcik, B. S., & Wei, S. J. (2009). Corruption and cross-border investment in emerging markets: firm-level evidence. Journal of International Money and Finance, 28(4), 605-624. La Porta, R., Lopez-de-Silanes, F., & Vishny, R. (1997). Trust in large organizations. American Economic Review Papers and Proceedings, 87(2), 334–338. Liang, G. (1994). The practical encyclopedia of anti-corruption in China and foreign countries. Beijing: Xinhua Press. Lipset, S. M., & Lenz, G. S. (2000). Corruption, culture, and market. New York: Basic Books. Longenecker, J. G., McKinney, J. A., & Moore, C. W. (2004). Religious intensity, evangelical Christianity, and business ethics: An empirical study. Journal of Business Ethics, 55(3), 373-386. Mauro, P.,(1995). Corruption and growth. Quarterly Journal of Economics, 110(3), 681-712. Mensah, Y. M. (2014). An analysis of the effect of culture and religion on perceived corruption in a global context. Journal of Business Ethics, 121(1), 255-282. Miller, A. S. (2000). Going to Hell in Asia: The relationship between risk and religion in a cross cultural setting. Review of Religious Research, 42(1), 5-18. 28

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Miller, A. S., & Hoffmann, J. P. (1995). Risk and religion: An explanation of gender differences in religiosity. Journal for the Scientific Study of Religion, 34(1), 63-75. North, D. C. (1990). Institutions, institutional change and economic performance. Cambridge: Cambridge University Press. Osoba, B. J. (2003). Risk preferences and the practice of religion: Evidence from panel data. Working paper. Paldam, M. (2001). Corruption and religion: adding to the economic model. Kyklos, 54 (2–3), 383–413. Plumper, T., & V. E.Troeger, (2007). Efficient estimation of time-invariant and rarely changing variables in finite sample panel analyses with unit fixed effects. Political Analysis, 15, 124-139. Ramirez, C.D. (2014). Is corruption in China “out of control”? A comparison with the US in historical perspective. Journal of Comparative Economics, 42(1), 76-91. Sandholtz, W., & M. Gray. (2003). International Integration and National Corruption. International Organization, 57(4), 761–800. Shadabi, L. (2013). The impact of religion on corruption. The Journal of Business Inquiry, 12, 102−117. Stark,bR., & Finke, R. (2000). Acts of Faith: explaining the human side of religion. University of California Press. The World Values Survey. (2007). China 2007. Retrieved from http://www.wvsevsdb.com/wvs/ WVSDocumentation.jsp. Treisman, D. (2000). The cause of corruption: A cross-national study. Journal of Public Economics, 76(3), 399-457. Treisman, D. (2007). What have we learned about the causes of corruption from ten years of cross-national empirical research? Annual Review of Political Science, 10(1), 211-244. Wang, Q. Y., & Lin, X.Y. (2014). Does religious beliefs affect economic growth? Evidence from provincial-level panel data in China. China Economic Review, 31, 277-287. Weaver, G. R., & Agle, B. R. (2002). Religiosity and ethical behavior in organizations: A symbolic inter-actionist perspective. The Academy of Management Review, 27(1), 77-97. Weber, M. (1905). The Protestant Ethic and the Spirit of Capitalism. Routledge Classics. Wedeman, A. (2012). Double Paradox: Rapid Growth and Rising Corruption in China. Ithaca NY: Cornell University Press. Williamson, O.E. (2000). The new institutional economics: taking stock, looking ahead. Journal of Economic Literature, 38(3), 595-613. Xiong, P. (2014). Freedom of religion in China under the current legal framework and foreign religious bodies, Brigham Young University Law Review, 3, 605-618. Yang, F. G. (2010). Religion in China under Communism: A shortage economy explanation. Journal of Church and State, 52(1), 3-33. 29

ACCEPTED MANUSCRIPT Yao, S.T. (2002). Privilege and corruption: The problems of China's Socialist Market Economy, American Journal o f Economics and

AC

CE P

TE

D

MA

NU

SC R

IP

T

Sociology. 61(1), 279-299.

30

ACCEPTED MANUSCRIPT Highlights

SC R

IP

T

Bullet point 1: Religious culture plays an important role in curbing corruption. Using Chinese provincial-level panel data from 1998 to 2009, the empirical results show that, religious density is significantly negatively associated with regional corruption. It suggests that religion, as an informal institutional factor and a part of social culture, has a significant impact on officials’ unethical behavior and can reduce the likelihood of them engaging in corruptive practices.

NU

Bullet point 2: The negative association between religion and corruption is weaker in provinces with stronger legal institutions and the rule of law. It implies that implicit constraints from religions can serve as an alternative mechanism to prevent corrupt crimes in emerging markets like China where formal institution system and external monitoring mechanisms are incomplete.

AC

CE P

TE

D

MA

Bullet point 3: After separating the native religions from the foreign religions, our findings indicate that the anti-corruption effects of native religions (i.e., Taoism and Buddhism) are more prominent than foreign religions (i.e., Christianity and Islam). As to Taoism and Buddhism, Buddhism plays much more positive role in curbing corruption crimes in the context of China.

31