Religion and the Money Laundering Risk

Abstract

In this paper, we investigate the impact of religion on the money laundering risk, focusing specifically on Protestantism compared to other religions. Protestantism is often associated with greater individual self-discipline and stronger institutional economic governance. Analysing data from 27 EU member states, we find that Protestant countries exhibit a lower risk of money laundering. Additionally, our findings indicate that Protestantism exerts a distinct influence that is separate from the overall religiosity levels of countries. Our results remain robust across various estimation models and when incorporating additional governance and cultural control variables. This study enhances our understanding of the significant role that religion plays in shaping individual behaviour toward financial fraud, particularly money laundering.

1. Introduction

Globally, money laundering, the process of disguising the illicit origins of funds, carries a significant cost. Estimates suggest that it costs the world economy between USD 1.4 trillion and USD 3.5 trillion annually (EY, 2023). Large-scale money laundering erodes governance, fuels corruption, weakens public services, widens income inequality, and destabilises political systems, ultimately hindering sustainable development. Despite ongoing efforts by international organisations like the Financial Action Task Force (FATF) and the International Monetary Fund (IMF), the average global money laundering (ML) risk rose from 2022 to 2023, highlighting potential shortcomings in the current regulatory frameworks and their implementation.

The gravity of this problem has spurred research into factors influencing the ML risk across nations. A key area of focus is the role of informal cultural institutions. Studies like that of Yamen et al. (2019) suggest a lower ML risk in countries characterised by higher uncertainty avoidance, individualism, and longer-term orientation. Similarly, Mejri et al. (2022) link tighter cultures with a lower ML risk compared to looser ones. Extending this discussion, we investigate the influence of another cultural facet—religion—on the ML risk. In particular, we examine whether countries with dominant Protestant religious values exhibit a lower money laundering risk.

Religious values exert a profound influence on both individual ethics and societal governance. Social norm theory posits that individual behaviour is shaped by a complex interplay of psychological factors and the broader sociocultural context (D. T. Campbell, 1975). Bisin and Verdier (2001) further demonstrate the intergenerational transmission of cultural values, including ethnicity and religion, through marriage and socialisation, which significantly impact individual preferences. Religious beliefs have been shown to strongly influence economic growth through character traits such as honesty, work ethic, and thrift (Barro & McCleary, 2003, 2019).

Drawing on institutional theory, North (1990) argues that dominant religions often play a pivotal role in shaping legal norms and enforcement mechanisms, fostering both formal and informal governance frameworks within communities. Williamson’s (2000) institutional framework echoes this sentiment, highlighting the influence of cultural values as a primary informal institution that shapes lower-level formal institutions, including governance structures and human behaviour. Religious commitment, characterised by moral conduct, teamwork, and adherence to the rule of law, has been shown to contribute positively to both governmental effectiveness and economic progress (Barro & McCleary, 2003).

Max Weber’s (1905) seminal work highlighted the significance of the ‘work ethic’ in Protestantism, particularly its role in shaping values conducive to economic development. The characteristics of the Protestant work ethic, such as diligence, moral rectitude, and self-discipline, align closely with the development of capitalism. Arruñada (2010) further supports the role of Protestantism, proposing an alternative ‘social ethics’ argument. He contends that Protestantism promotes a social environment where individuals monitor each other’s conduct, support political and legal institutions, and adhere to more homogeneous values. In his analysis, Protestantism appears to facilitate capitalist development not through the Weberian psychological work ethic but by fostering a social ethic that enhances impersonal trade.

Recent research suggests that Protestant-majority societies generally exhibit stronger legal frameworks, higher regulatory quality (La Porta et al., 1999), growth in formal financial institutions, and transparency (Stulz & Williamson, 2003), as well as enhanced legal accountability and integrity (Licht et al., 2007).

Based on the preceding literature, we hypothesise that Protestant-dominant societies may exhibit a lower money laundering (ML) risk. This anticipated effect is attributed to both individual-level factors, such as heightened ethical behaviour and the avoidance of financial crimes like ML, and institutional-level factors, including stronger governance structures and regulatory frameworks, often associated with a Protestant influence.

For our empirical analysis, we use data from 27 European Union (EU) member states from 2012 to 2022. EU member states provide an ideal context for our analysis because they share a common anti-money laundering legal framework at the union level, which relies heavily on country-level formal and informal institutions for implementation. Therefore, the varying levels of money laundering risk among EU members are likely due to differences in these institutions, rather than differences in the definition or regulation of money laundering. We measure the risk of money laundering in each of the EU member states with the annual BASEL AML index. The main independent variable is a binary indicator, coded as 1 if the dominant religion in a country is Protestant.

Our findings indicate that the risk of money laundering is significantly lower in cultures with a Protestant majority. We confirm that our results are not influenced by general levels of religiosity. Our analysis supports the theory that cultural and religious factors, particularly Protestant ideals, are important for more effective anti-money laundering systems.

Our research contributes to the existing literature in two significant ways. Firstly, we show the substantial influence of religion on the money laundering (ML) risk, expanding upon recent studies that explored cross-country cultural determinants of the ML risk (Yamen et al., 2019; Mejri et al., 2022). Specifically, we demonstrate that the ML risk is lower in protestant countries.

Secondly, our study adds to the body of research examining the role of religion in financial crime prevention. While most prior studies have focused on the relationship between religion or religiosity and tax evasion, our work extends this inquiry to the realm of ML risk. These previous studies have yielded mixed results. Some have found lower tax evasion rates in countries with specific religions or among religious individuals (Richardson, 2008; Schneider et al., 2015; Nurunnabi, 2018), while others have reported unclear findings (Boone et al., 2013; Khalil & Sidani, 2020) or even higher tax evasion rates (Cason et al., 2016). Notably, Ben Othman et al. (2024) observed a general decline in tax evasion associated with religiosity, although this effect was not apparent in developed countries.

This paper is organised as follows. In Section 2, we briefly review the related literature. Section 3 presents our data collection procedures. Section 4 outlines the reasons to use the EU context. Section 5 introduces the empirical methodology and variables. In Section 6, we report the empirical results. The final section concludes the study.

2. Theoretical Framework and Hypothesis

2.1. Determinants of ML Risk

Money laundering, initially classified as a white-collar crime due to its association with high-status individuals, has evolved into a significant financial crime, involving money originating from illicit activities, including tax evasion, corruption, and terrorism financing (Gottschalk, 2010; Rusanov & Pudovochkin, 2021). The risk of money laundering varies across countries. Existing research has examined various factors influencing the money laundering (ML) risk, including cultural, governance, macroeconomic, and socioeconomic variables. Explanations for these variables are grounded in attitudes and opportunities, which determine the likelihood of becoming involved in financial crime.

Cultural values and norms represent acceptable behaviours within a social group and may influence the extent of financial crimes. Existing studies have examined the influence of the cultural dimensions of Hofstede (Hofstede, 1980, 2001; Hofstede et al., 2010) and the cultural tightness variable (Gelfand et al., 2011) on the ML risk. For instance, Yamen et al. (2019) found that countries with higher levels of cultural values such as uncertainty avoidance, individualism, and long-term orientation tend to have a lower ML risk. Similarly, Mejri et al. (2022) linked tighter cultural norms to a reduced ML risk.

Regarding governance factors, research has highlighted the importance of democratic accountability and the rule of law in mitigating the ML risk. Kalokoh (2024) demonstrated that democratic accountability can decrease the ML risk. Reganati and Oliva (2018) further emphasised the role of governance by highlighting regional disparities within countries, with a higher ML risk observed in areas characterised by corruption and organised crime.

The interplay between culture and governance has also been explored. Al Qudah et al. (2019) found that the control of corruption and rule of law variables moderate the influence of culture on terrorism financing. In another recent study, AlQudah et al. (2022) found that the direct relationship between culture and the ML risk diminishes when public governance is considered as a mediating factor. This suggests that the effect of culture on the ML risk is channelled through the institutional environment of a country.

Economic factors might affect the likelihood of economic and financial crimes. Achim and Borlea (2020) argue that underdeveloped countries have higher levels of financial crime. Specifically focusing on the ML risk, Ghulam and Szalay (2023) examined the influence of various macroeconomic variables, including the net primary income from abroad, GDP growth rate, export volume, and exchange rates. While the net primary income and GDP growth rate were found to have no significant impact, they observed a positive correlation between the export volume and ML risk and a negative correlation with exchange rates.

Socioeconomic factors, such as education, have also been linked to the ML risk. Reganati and Oliva (2018) found that higher levels of educational attainment are associated with a lower ML risk, suggesting that education can play a role in reducing the vulnerability to ML activities.

2.2. The Influence of Religion on Financial Crime

Religious beliefs significantly influence human behaviour. As Adam Smith emphasised, religious morality can motivate individuals to act in socially acceptable ways, fostering feelings of pride for virtuous actions and guilt for transgressions (T. Campbell, 2014).

Recent interdisciplinary research on religion and economics suggests that religion plays a significant role in shaping an individual’s financial and economic behaviour (Guiso et al., 2003; Brammer et al., 2007). However, the direction of this effect is not straightforward. While almost all religions often emphasise fundamental moral values like honesty and fair dealing (Vitell, 2009; Walker et al., 2012), they vary in their religious organisations. Some are more hierarchical than others, with an emphasis on group belonging and material benefits. Higher religiosity also comes with some drawbacks, such as lower levels of general social trust among members of a society (Berggren & Bjørnskov, 2011) and religious in-group favouritism (Dunkel & Dutton, 2016).

The literature on the effects of religion on economic and financial crime issues, including corruption, the shadow economy, and tax evasion, presents mixed and sometimes contradictory findings.

Regarding corruption, Xu et al. (2017) identify a negative correlation between religiosity and bureaucratic corruption in Chinese provinces, observing that this relationship is stronger for Taoism and Buddhism than for Christianity and Islam. In contrast, Gokcekus and Ekici (2020) report that higher levels of national religiosity are associated with increased corruption, regardless of specific religious denominations, such as Catholicism, Islam, or Protestantism. Their findings rely on World Values Survey data to categorise countries by religious affiliation, a method that may introduce classification challenges. Additionally, Niu et al. (2022) observe that, in countries with high levels of religiosity, borrowers perceive greater corruption in bank lending.

In examining the shadow economy, often defined as unregistered or cash-based economic activities, researchers have also identified complex links with religiosity. Higher levels of shadow economy activity can facilitate tax evasion and raise the risk of money laundering. Schneider et al. (2015) find that general religiosity tends to increase the size of the shadow economy; however, this effect is somewhat attenuated in Protestant-majority countries, which they attribute to the replacement of informal religious norms with formal institutions that regulate transactions. Similarly, Achim et al. (2019) measure religiosity using World Values Survey data and conclude that higher religiosity correlates with a larger shadow economy across 31 European countries. In contrast, Dada et al. (2024) note that religious diversity appears to reduce the shadow economy’s size in 13 African nations.

Studies on tax evasion also present divergent perspectives among religions. While tax evasion is deemed unethical in Jewish, Mormon, and Baha’i teachings, it is not universally condemned in Catholic and Islamic traditions (McGee, 2006). Empirical analyses present additional complexities: Richardson (2008) finds that the tax evasion rates are lower in countries with higher levels of religiosity, while Nurunnabi (2018) shows that greater adherence to Shariah principles correlates with reduced tax evasion in 38 Muslim-majority countries. However, in Lebanon, Khalil and Sidani (2020) find no significant difference in tax evasion behaviours between Christians and Muslims.

These diverse findings reflect the complexity of religiosity’s influence on financial misconduct and suggest that further research is needed to understand how religious beliefs interact with socioeconomic and institutional factors in shaping financial behaviour.

2.3. Protestantism and ML Risk

Although organised religions share a foundation based on moral principles, they diverge in specific beliefs regarding God, the afterlife, heaven, and hell. These distinctions in doctrine can either reinforce or discourage particular behaviours. For example, Protestantism often promotes pride in virtuous conduct, while Catholicism tends to emphasise guilt associated with immoral actions (Weber, 1905). These contrasting ethical perspectives may lead to different motivations and experiences, even among individuals performing the same actions, regardless of their moral implications.

The literature has extensively explored Protestantism’s impact on behavioural and institutional outcomes, especially through the lenses of work and social ethics, which ultimately shape economic decision-making. Weber (1905) famously argued that the Protestant work ethic, marked by diligence, moral integrity, and self-discipline, aligns closely with the growth and success of capitalism. Building on this, Arruñada (2010) introduced the concept of a Protestant ‘social ethic’, suggesting that Protestantism not only fosters an environment in which individuals monitor each other’s behaviour to maintain shared values but also supports institutions that reinforce legal and political accountability. His analysis implies that Protestantism’s contribution to capitalist development may lie more in its social ethic, which supports impersonal trade, than in Weber’s psychological work ethic.

Social trust also appears to be higher in regions with Protestant majorities. For instance, in his study on development in Italy, Putnam (1994) attributes lower social trust in the South to the strong Catholic tradition, which emphasises a hierarchical relationship with the Church and diminishes peer-to-peer social bonds. Welch et al. (2004) found that members of mainline Protestant denominations tended to report higher levels of social trust than those in Pentecostal and other Christian groups.

Protestant regions, by contrast, tend to exhibit lower levels of corruption and smaller shadow economies. Treisman (2000) suggests that the Protestant tradition of challenging state-sponsored religious authority has fostered societies that are more likely to scrutinise and expose official misconduct. Schneider et al. (2015) further observe that Protestant-majority countries generally have smaller shadow economies than Orthodox Christian countries and that other religious denominations also correlate with lower shadow economies compared to Orthodox Christianity.

Institutional quality is generally higher in Protestant-majority countries as well. Research by La Porta et al. (1999) shows that nations with predominantly Protestant populations display stronger governance than those with Catholic or Muslim majorities, attributing this to Protestantism’s less hierarchical structure. Stulz and Williamson (2003) support these findings, noting that Protestant-majority nations tend to provide stronger legal protections for creditors than Catholic-majority ones. Likewise, Licht et al. (2007) document a positive correlation between Protestantism, the rule of law, and reduced corruption.

Protestant-majority regions also tend to have companies that are less likely to engage in unethical practices and generally perform better. Grullon et al. (2009) find that corporate misconduct, such as excessive earnings management, options backdating, and excessive executive compensation, is less common in US regions with a higher concentration of Protestants. Similarly, El Ghoul et al. (2012) observe that US firms in more religious counties benefit from lower equity financing costs, a trend that is particularly pronounced in areas with a larger proportion of mainline Protestants.

In summary, the Protestant emphasis on social ethics and communal monitoring (Arruñada, 2010), which culturally reinforces transparency and accountability, along with higher levels of social trust (Putnam, 1994; Welch et al., 2004) and generally lower levels of corruption and smaller shadow economies (Schneider et al., 2015; Treisman, 2000), may discourage illicit financial practices and limit the opportunities for involvement in money laundering activities.

Based on this discussion, we hypothesise that Protestant-majority countries will exhibit a lower level of money laundering risk. Our hypothesis is therefore formulated as follows.

Hypothesis 1.

Countries with dominant Protestant values are likely to exhibit a lower money laundering risk.

3. Data Collection and Sample

To empirically examine our hypothesis, we began constructing our sample by downloading data on the BASEL AML Index for EU member countries from the website https://index.baselgovernance.org, accessed on 10 October 2024. The AML Index is available annually from 2012 to 2022. Data for the Protestant variable were sourced from Nardulli et al. (2012). The religiosity data were collected from Meisenberg (2011). The annual GDP per capita in current US dollars was obtained from the World Development Indicators of the World Bank. Data on the rule of law and other governance indicators were downloaded from the World Governance Indicators database of the World Bank.

After merging all these datasets, we had an unbalanced panel of 27 EU member states, with 294 annual observations spanning the period from 2012 to 2022.

4. Why the EU Context?

We chose EU member states as our sample because they provide an ideal context for our analysis. The EU has used a common anti-money laundering legal framework since 1991. Member states are responsible for implementing this framework within their respective jurisdictions. The varying levels of money laundering risk among EU countries are likely due to the differences in member states’ formal and informal institutions, rather than the differences in money laundering regulations.

The evolution of Anti-Money Laundering (AML) regulations in the European Union (EU) demonstrates the progressive tightening of its legal framework to combat money laundering and terrorist financing. Beginning with the First AML Directive introduced by the then European Community in 1991, which mainly focused on financial institutions and client identification, the EU has successively broadened its scope and refined its approach. The Second AML Directive in 2001 extended the requirements to non-financial professions and introduced the identification of beneficial ownership, while the Third AML Directive in 2005 explicitly incorporated terrorist financing into its framework, introducing enhanced due diligence measures for high-risk clients and politically exposed persons.

Subsequent iterations of the AML framework addressed emerging risks and enhanced transparency. The Fourth AML Directive in 2015 introduced national risk assessments and beneficial ownership registers, while the Fifth AML Directive in 2018 expanded coverage to virtual currencies and high-value goods, increasing the scrutiny of high-risk third countries. The Sixth AML Directive in 2021 harmonised the definitions and penalties across member states and included modern predicate offenses such as cybercrime. Recent developments focus on the creation of a centralised Anti-Money Laundering Authority (AMLA) and new regulations to streamline compliance and supervision across the EU. These efforts, driven by financial scandals, technological advances, and global cooperation, reflect the EU’s commitment to combating financial crime in an increasingly complex financial landscape.

Member states have been largely responsible for implementing the provisions of EU AML directives in their respective jurisdictions. However, with the establishment of the Anti-Money Laundering Authority (AMLA), the EU will take on a more direct supervisory role, particularly over high-risk entities and cross-border financial institutions (Sunder, 2021).

5. Methodology

We specify the following pooled panel ordinary least squares model to examine the effect of religion on the money laundering risk:

$$Y_{j,t}={\alpha }_i+{\beta }_1{Protestant}_j+\sum _{l=1}^l{\beta }_l{X}_{j,t}^l+\sum _{t=1}^{T-1}{ϵ}_t{D}_t+\sum _{c=1}^{C-1}{ϵ}_j{C}_j+{\epsilon }_{j,t}$$

(1)

where the j and t subscripts represent the country and year, respectively.

The dependent variable, Y, stands for the risk of ML, measured by the Basel AML Index. α_i is a constant term. Protestant is the main explanatory variable. $X_{j,t}^l$ denotes country-level control variables, including the GDP per capita and rule of law. D_t is a set of year dummy variables, controlling for changes in the overall ML risk in EU countries over time. $C_j$ is a set of country dummy variables to control for other country-specific factors. ε_j,t is an idiosyncratic error term.

The Basel AML Index measures the money laundering risk based on countries’ vulnerability to money laundering and related financial crimes (https://baselgovernance.org/basel-aml-index (accessed on 10 October 2024)). The index is constructed using a composite methodology that incorporates 17 indicators from publicly available sources such as the Financial Action Task Force (FATF), Transparency International, and the Global Initiative Against Transnational Organized Crime. It evaluates countries across five critical domains that contribute to the money laundering risk: the quality of the AML/CFT/CPF framework, reflecting the effectiveness of measures to combat money laundering, terrorism financing, and proliferation financing; corruption and fraud risks, indicating the levels of corruption and susceptibility to fraudulent activities; financial transparency and standards, assessing the adherence to global financial standards and practices; public transparency and accountability, which measures openness and accountability in public governance; and legal and political risks, capturing the stability and reliability of the legal and political environment. Higher index values indicate a higher risk of money laundering in a given country, while lower values reflect a reduced risk.

The variable ‘Protestant’ is binary, coded as 1 if the dominant religion in a country is Protestant and 0 otherwise. This classification is based on data from the Composition of Religious and Ethnic Groups (CREG) Project by the Cline Centre for Democracy at the University of Illinois (Nardulli et al., 2012). The CREG Project compiles data on the percentage of the population identifying with major religious groups, including Muslim, Orthodox, Roman Catholic, Protestant, Jewish, Buddhist, Hindu, Bahai, Sikh, and nonreligious. For our sample of 27 European Union member states, we code the Protestant variable as 1 if the Protestant population exceeds 20% in a country. This classification relies on data from 2013, the latest year available in the CREG Project. Five countries are classified as Protestant: Denmark (87.3% Protestant), Sweden (84.2%), Finland (83.5%), Germany (39.7%), and Hungary (22.5%).

The GDP per capita (current USD) measures the average economic output per person in a country. We take the natural logarithm of the GDP per capita, Log(GDP per capita), to account for differences in economic development across the sample countries.

The rule of law variable measures perceptions of the extent to which individuals and institutions have confidence in and adhere to the rules of society. This includes the quality of contract enforcement, property rights, policing, judicial systems, and the prevalence of crime and violence.

6. Empirical Analysis

6.1. Summary Statistics

Table 1. Summary statistics.

Variable	Mean	S.D.	Min	Max
Basel AML Index	4.31	0.77	1.78	6.78
Protestant	0.19	0.39	0.00	1.00
Religiosity	7.03	1.17	5.07	9.18
Log(GDP per capita)	10.26	0.62	8.86	11.80
Rule of law	1.07	0.60	−0.16	2.12

Note: This table presents descriptive statistics for the key variables used in the analysis. The Basel AML Index measures the money laundering risk. Protestant is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Religiosity measures religiosity in the population. Log(GDP per capita) captures the level of economic development. Rule of law is a measure of governance quality. The statistics include the mean, standard deviation (S.D.), minimum, and maximum values for each variable.

presents the summary statistics of the main variables. The Basel AML Index has an average value of 4.31, with a standard deviation of 0.77. It ranges from 1.78 to 6.78, suggesting significant differences in the money laundering risk among the sample countries. The mean value of Protestant shows that 19% of the observations are from countries with more than 20% Protestant populations. The religiosity variable has a mean of 7.03 and moderate variation (standard deviation of 1.17), ranging from 5.07 to 9.18. Countries vary in their economic development, as shown by the 0.62 standard deviation of the Log(GDP per capita) variable. Finally, the rule of law variable varies from −0.16 to 2.12, with a mean value of 1.07 suggesting significant differences in legal institutions and governance. These statistics highlight the diverse cross-country dynamics in governance, religiosity, and economic development, which are essential in understanding variations in the money laundering risk.

6.2. Correlations

Table 2. Pairwise correlations.

		(1)	(2)	(3)	(4)	(5)
(1)	Basel AML Index	1.00
(2)	Protestant	−0.20 *	1.00
(3)	Religiosity	0.24 *	−0.46 *	1.00
(4)	Log(GDP per capita)	0.10	0.28 *	−0.22 *	1.00
(5)	Rule of law	−0.13 *	0.41 *	−0.37 *	0.82 *	1.00

Note: This table presents pairwise correlations between the key variables used in the analysis. The Basel AML Index measures the money laundering risk. Protestant is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Religiosity measures religiosity in the population. Log(GDP per capita) captures the level of economic development. Rule of law is a measure of governance quality. * indicates significance at the 5% level.

provides pairwise correlations between the main variables. The AML Index has a negative correlation with the Protestant dummy variable, indicating that the risk of money laundering is lower in Protestant countries. The coefficient is significant at the 5% level. The correlation between Log(GDP per capita) and rule of law is 0.82, showing that countries with better rule of law are more developed.

6.3. Multivariate Regression Results

Table 3. Effect of Protestantism on ML risk—baseline regression results.

Variable	Basel AML Index	Basel AML Index
	Model (1)	Model (2)
Protestant	−1.261 ***	−1.524 ***
	(0.000)	(0.000)
Log(GDP per capita)		2.410 ***
		(0.000)
Rule of law		−0.710 ***
		(0.007)
Year Dummies	Yes	Yes
Country Dummies	Yes	Yes
Constant	5.327 ***	−19.202 ***
	(0.000)	(0.000)
R-squared	0.739	0.788
Observations	294	294
Countries	27	27

Note: This table presents the regression results for the effect of Protestantism on the ML risk. The dependent variable is the Basel AML Index, which measures the money laundering risk. Protestant is the main explanatory variable of interest. It is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Log(GDP per capita) and rule of law are control variables. Log(GDP per capita) captures the level of economic development. Rule of law measures governance quality. Year and country dummy variables are included to control for time trends and other country-level non-variant factors, respectively. The results are estimated using a pooled panel ordinary least squares regression model, with heteroscedastic robust standard errors. *** indicates that the coefficient is significant at p < 0.01.

reports the main regression results. The coefficient for the Protestant dummy is negative and statistically significant, indicating that countries with a significant Protestant population experience lower levels of ML risk. This result is consistent with our hypothesis that Protestant cultural and ethical norms discourage financial crimes.

The results for the control variables are consistent with our expectations. The rule of law variable has a negative coefficient, underscoring the fact that stronger governance frameworks and legal institutions are effective in mitigating the ML risk. In contrast, the GDP per capita exhibits a positive and significant relationship with the ML risk, suggesting that economically developed countries face higher ML risks, potentially due to their larger and more sophisticated financial systems that offer greater opportunities for illicit financial flows.

We perform several robustness tests to further confirm our results. First, one potential concern with the results is that the Protestant dummy variable might be capturing the effects of general religiosity rather than the specific influence of Protestantism. To address this, we include religiosity as a control variable in the baseline model and re-estimate the results. The religiosity variable is collected from Meisenberg (2011), who created country-level trends in religious beliefs based on the data from the World Values Survey covering the period from 1981 to 2008. Higher values of the religiosity variable represent strong religious beliefs and vice versa.

As shown in

Table 4. Effect of Protestantism on ML risk—controlling for religiosity levels.

Variable	Basel AML Index
	Model (1)
Protestant	−3.040 **
	(0.040)
Religiosity	−0.616
	(0.304)
Log(GDP per capita)	2.410 ***
	(0.000)
Rule of law	−0.710 ***
	(0.007)
Year Dummies	Yes
Country Dummies	Yes
Constant	−14.488 ***
	(0.000)
R-squared	0.788
Observations	294
Countries	27

Note: This table presents the regression results for the effect of Protestantism on the ML risk, adding the religiosity control variable. The dependent variable is the Basel AML Index, which measures the money laundering risk. Protestant is the main explanatory variable of interest. It is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Religiosity measures religiosity in the population. Log(GDP per capita) and rule of law are control variables. Log(GDP per capita) captures the level of economic development. Rule of law measures governance quality. Year and country dummy variables are included to control for time trends and other country-level non-variant factors, respectively. The results are estimated using a pooled panel ordinary least squares regression model, with heteroscedastic robust standard errors. *** indicates that the coefficient is significant at p < 0.01; ** p < 0.05.

, the Protestant dummy retains its negative and significant effect on the ML risk after including the religiosity variable. The religiosity variable remains insignificant. These findings indicate that Protestantism has a distinct and significant effect.

Secondly, as mentioned above, existing studies, such as those of Reganati and Oliva (2018), Al Qudah et al. (2019), AlQudah et al. (2022), and Kalokoh (2024), demonstrate that formal country-level governance is important in lowering the ML risk. While the rule of law is included as a control variable in Equation (1), it is possible that the Protestant variable may capture the influence of other omitted institutional factors, such as democracy, political stability, government effectiveness, or regulatory quality. To address this omitted variable bias, we incorporate additional governance indicators from the World Governance Indicators dataset into the baseline model, including voice and accountability, government effectiveness, political stability, and regulatory quality indices. Due to the strong correlations among these variables, we include them in the baseline model one at a time. As presented in

Table 5. Effect of Protestantism on ML risk—adding governance control variables.

Variable	Basel AML Index
	Model (1)	Model (2)	Model (3)	Model (4)	Model (5)	Model (6)
Protestant	−1.524 ***	−1.459 ***	−0.842 ***	−1.545 ***	−1.352 ***	−1.352 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Log(GDP per capita)	2.410 ***	2.401 ***	2.436 ***	2.507 ***	2.334 ***	2.339 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Rule of law	−0.710 ***	−0.538 *	0.012	−0.610 **	−0.513 *	−0.386
	(0.007)	(0.063)	(0.961)	(0.017)	(0.060)	(0.195)
Government effectiveness		−0.394 *
		(0.085)
Control of corruption			−1.049 ***
			(0.000)
Political stability				−0.478 ***
				(0.010)
Regulatory quality					−0.463 *
					(0.069)
Voice and accountability						−0.975 **
						(0.011)
Year Dummies	Yes	Yes	Yes	Yes	Yes	Yes
Country Dummies	Yes	Yes	Yes	Yes	Yes	Yes
Constant	−19.202 ***	−18.804 ***	−19.262 ***	−19.908 ***	−18.076 ***	−17.661 ***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Observations	294	294	294	294	294	294
R-squared	0.788	0.791	0.811	0.794	0.791	0.794

Note: This table presents the regression results for the effect of Protestantism on the ML risk, including additional governance control variables from the World Governance Indicators dataset of the World Bank. The dependent variable in all models is the Basel AML Index, which measures the money laundering risk. Protestant is the main explanatory variable of interest. It is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Log(GDP per capita) and rule of law are control variables. Log(GDP per capita) captures the level of economic development. Rule of law measures governance quality. Government effectiveness, control of corruption, political stability, regulatory quality, and voice and accountability are added as additional governance control variables. Year and country dummy variables are included to control for time trends and other country-level non-variant factors, respectively. The results are estimated using a pooled panel ordinary least squares regression model, with heteroscedastic robust standard errors. *** indicates that the coefficient is significant at p < 0.01; ** p < 0.05; * p < 0.1.

, the Protestant dummy remains negative and statistically significant across all specifications, reinforcing its robust association with a lower ML risk. The governance indicators generally exhibit negative coefficients, indicating that stronger governance—reflected in higher political stability, effective government functioning, the tight control of corrupt practices, and better regulatory quality—contributes to reducing the ML risk. These findings highlight the independent role of Protestantism while emphasising the importance of governance in mitigating the ML risk.

Thirdly, as existing studies, including those of Yamen et al. (2019) and Mejri et al. (2022), highlight the significant role of the national culture in influencing the ML risk, to ensure that the Protestant dummy does not capture the effects of cultural values, we include Hofstede’s four most significant cultural dimensions—individualism, power distance, uncertainty avoidance, and masculinity—into the baseline model, one at a time, as additional controls. The results, as shown in

Table 6. Effect of Protestantism on ML risk—adding cultural control variables.

Variable	Basel AML Index
	Model (1)	Model (2)	Model (3)	Model (4)
Protestant	−0.042	−0.257 **	−0.242 *	−0.281 ***
	(0.761)	(0.033)	(0.071)	(0.003)
Log(GDP per capita)	1.213 ***	1.058 ***	1.072 ***	1.033 ***
	(0.000)	(0.000)	(0.000)	(0.000)
Rule of law	−1.103 ***	−1.118 ***	−1.115 ***	−0.793 ***
	(0.000)	(0.000)	(0.000)	(0.000)
Uncertainty avoidance	0.007 ***
	(0.001)
Individualism		0.006 **
		(0.024)
Power distance			−0.004
			(0.197)
Masculinity				0.015 ***
				(0.000)
Year Dummies	Yes	Yes	Yes	Yes
Constant	−6.919 ***	−5.137 ***	−4.728 ***	−5.540 ***
	(0.000)	(0.000)	(0.000)	(0.000)
Observations	272	272	272	272
Countries	25	25	25	25

Note: This table presents the regression results for the effect of Protestantism on the ML risk, including additional cultural control variables from the national cultural framework of Hofstede. The dependent variable is the Basel AML Index, which measures the money laundering risk. Protestant is the main explanatory variable of interest. It is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Log(GDP per capita) and rule of law are control variables. Log(GDP per capita) captures the level of economic development. Rule of law measures governance quality. Year dummy variables are included to control for time trends. The results are estimated using a pooled panel ordinary least squares regression model, with heteroscedastic robust standard errors. *** indicates that the coefficient is significant at p < 0.01; ** p < 0.05; * p < 0.1.

, remain consistent, with the Protestant variable largely retaining its negative and significant effect on the ML risk. This robustness check confirms that the observed relationship between Protestantism and reduced ML risk is not merely a reflection of broader cultural characteristics but rather reflects a distinct influence associated with Protestant values.

We also estimate Equation (1) using one-step system GMM regressions to further confirm that our main results are not biased due to the dynamic panel properties of our dataset. The dependent variable, the Basel AML Index, may exhibit persistence over time as the ML frameworks and practices do not change rapidly from one year to the next. For a dynamic panel dataset, system generalised method of moments (GMM) estimators (Arellano & Bover, 1995; Blundell & Bond, 1998) are considered appropriate.

As shown in Model 1 of

Table 7. Effect of Protestantism on ML risk—system GMM estimations.

Variable	Basel AML Index
	Model (1)	Model (2)
L. Basel AML Index	0.707 ***	0.625 ***
	(0.000)	(0.000)
Protestant		−0.506 **
		(0.020)
Log(GDP per capita)		1.387 ***
		(0.000)
Rule of law		−0.107
		(0.676)
Year Dummies	Yes	Yes
Country Dummies	Yes	Yes
Constant	1.241 ***	−13.375 ***
	(0.010)	(0.000)
Diagnostic tests
AR(1)	−4.33 ***	−4.27 ***
	(0.000)	(0.000)
AR(2)	0.50	0.36
	(0.618)	(0.718)
Sargan test	12.26	11.46
	(0.199)	(0.245)
F-test	25.92 ***	28.28 ***
	(0.000)	(0.000)
No. of instruments	46	48
Observations	267	267
Countries	27	27

Note: This table presents the regression results for the effect of Protestantism on the ML risk, es-timated with one-step system GMM estimators. The dependent variable is the Basel AML Index, which measures the money laundering risk. Protestant is the main explanatory variable of interest. It is a binary variable that takes the value of 1 for countries where the Protestant population exceeds 20% and 0 otherwise. Log(GDP per capita) and rule of law are control variables. Log(GDP per capita) captures the level of economic development. Rule of law measures governance quality. Year dummy variables are included to control for time trends. The results are estimated using a one-step system GMM model, with heteroscedastic robust standard errors. *** indicates that the coefficient is significant at p < 0.01; ** p < 0.05.

, the lagged Basel AML Index is positive and significant, indicating that the model has dynamic panel properties. We also observe that the coefficient of 0.707 lies between the values estimated with OLS (0.82) and fixed-effects (0.52) estimators. The Protestant variable is negative and significant in Model 2.

Diagnostic tests of the system GMM model validate that the model is correctly specified. We use the instrument collapse option, and the number of instruments—48 in Model 2—is lower than the number of observations, consistent with the advice of Roodman (2009). The AR(1) statistic is significant, confirming the first-order serial correlation in the residuals, while the AR(2) statistic is insignificant, confirming that there is no second-order serial correlation in the residuals. We use the lagged Basel AML Index as an instrument, and the Sargan test statistic is insignificant, indicating that the instrument is valid.

These results further confirm that our findings are not biased due to the dynamic panel properties of the data.

7. Conclusions

In this study, we examine the impact of Protestant religious values on the money laundering (ML) risk among European Union member states, where much of the formal institutional framework to combat money laundering is established at the union level. Our findings indicate that countries with significant Protestant populations exhibit a lower ML risk. This supports the hypothesis that Protestant values, which emphasise transparency, honesty, and accountability, contribute to reducing the ML risk. These values, particularly those related to self-discipline and economic governance, foster a culture of financial integrity. Our results underscore the influence of religious traditions on ethical behaviour in financial matters, reinforcing the existing literature on the role of cultural and institutional factors in economic governance.

Our results are robust. The inclusion of general religiosity does not diminish the significance of the Protestant dummy, suggesting that the effect is specific to Protestantism, rather than religiosity in general. Further, the influence of Protestantism stands when we include additional governance indicators from the World Governance Indicators or cultural dimensions from Hofstede’s framework of national culture.

Our results have important academic and policy implications. They strengthen our understanding that informal institutions such as religion play an important role in determining individual behaviour towards financial fraud, especially money laundering. For international organisations combating money laundering, addressing the ML risk requires not only robust institutional frameworks but also attention to the cultural and ethical norms underpinning financial behaviour in target countries.

Future research may extend this analysis to the sub-national level—for example, considering whether the ML risk differs in US states with dominant Protestant populations as compared to Catholic states. Likewise, this analysis can be extended to larger samples with more religious denominations.