I. Introductions

Country compliance with the AML/CFT international standard plays an important role in enhancing the world’s financial system integrity. As money laundering and terrorist financing can occur through many different avenues in different sectors of the economy, the standard itself has become equally broad in scope. The scope and depth of AML/CFT assessments also reflect the natural evolution of the FATF’s peer review process, which over twenty years, has prompted the gradual expansion and refinement of the standard.

The paper argues that the compliance with the AML/CFT standard is low. It therefore becomes more necessary than ever to better understand the factors that challenge compliance with the standard. Once these factors are better understood, addressing them should be part of a comprehensive policy reform agenda. In particular, domestic governance has received considerable attention in the recent past as one of the factors hampering compliance. However, the empirical evidence on the effect of domestic factors on countries’ compliance with regulatory standards is absent in the literature.

This paper attempts to provide an answer to these questions by using an entirely new analytical approach to understanding the factors affecting countries’ compliance with AML/CFT international standard. The purpose of this approach is threefold: (i) review how well countries² have complied with the AML/CFT international standard during 2004–2011; (ii) analyze what factors explain countries’ compliance with the AML/CFT standard; and (iii) discuss the implications of these findings for the policy agenda.

The paper finds that few countries—in particular, advanced countries³— comply to a larger degree with the AML/CFT standard. The results analyzed in the paper also reveal a widespread weak compliance by financial institutions and the adverse impact on financial transparency created by the cumulative effects of poor implementation of standards on customer identification in the financial sector and FATF-designated non-financial businesses and professions (e.g., trust and company service providers), as well as standards on entity transparency (i.e., ownership and control of companies and trusts).

Moreover, evidence was found that the quality of the domestic regulatory framework helps boost compliance. On the contrary, high net interest margin and prevalence of corruption have the opposite effect. Other sets of factors, such as financial depth, country openness, and illegal activities, do not seem to have an impact on compliance.

Compliance is correlated with the countries’ economic development reflecting a designed weakness in the AML/CFT standard and the assessment of countries’ compliance with it. This is notwithstanding the fact that the approach in measuring compliance has evolved from the traditional “one-size-fits-all” basis into a more risk-based approach. We observe that the relevance of the implementation of AML/CFT measures in countries varies according to their ML/FT risk levels. These levels are not always a factor of the relative low level of development of jurisdictions and their financial systems. While some low-income countries are also at low risk for ML, this is not always the case.

Our findings carry potentially significant policy implications. At the national level, achieving high compliance requires implementing a combination of reforms, including strengthening domestic governance, and improving the coordination of efforts against ML/FT between countries, regulators and private actors. At the international level, the standard and the assessment of countries’ compliance with it need to be better designed to capture countries’ risk for ML/FT (by capturing threats as well as vulnerabilities).⁴

In the next section, the paper describes the theoretical framework on which this research is based. Section III analyzes the degree of compliance of 161 countries with the AML/CFT standard. Based on cross-country analysis, Section IV explores the determinants of countries’ compliance for 116 countries during 2004–2008.⁵ Section V describes the caveats of the methodology that we employ in this paper. Section VI presents our conclusions and policy recommendations.

II. Literature Review

The international standard is particularly rigorous and detailed. Established by the Financial Action Task Force (FATF) in 1990, the standard has been updated several times and now consists of 40 Recommendations dealing with anti money laundering and 9 Special Recommendations dealing with combating terrorist financing (the FATF 40+9). To assess a country’s compliance with the FATF 40+9, a team of 4–5 assessors will typically take two weeks on-site to evaluate over 285 essential criteria and judge not only if the legal and regulatory system conforms to the standard, but also if it is being effectively implemented. Each recommendation is rated on a four point scale - non compliant, partially compliant, largely compliant, and compliant. The methodology provides that a rating of compliant shall only be given if all essential criteria are fully met.

How well do countries meet the AML/CFT standard? Interest in this question has grown as the international community (the FATF, the FATF- Style Regional Bodies, and the International Financial Institutions) reached the end of the third round⁶ of peer assessments and the completion of 161 mutual evaluation reports. This paper examines countries’ compliance with the AML/CFT standard, and the effect of domestic factors on countries’ compliance with the AML/CFT standard. While the literature on how well do countries meet the AML/CFT standard is large, the literature on the effect of domestic factors on countries’ compliance with the AML/CFT is sparse. In the following, we discuss a subset of this literature.

A report on the overall compliance with FATF recommendations was prepared in 2005.⁷ For that paper, the overall compliance with the recommendations of 18 countries (seven high income, six middle income, and five low income) was assessed.⁸

Putnam (1988) argues that the effectiveness of an international AML/CFT regime largely depends on the effectiveness of its constituents and vice versa.⁹ Due to the global nature of ML/FT phenomena, the compliance of domestic regimes with the AML/CFT international standard is seen as the first mechanism for achieving the global effectiveness of AML/CFT regime against a global phenomenon.¹⁰

Several authors argue that domestic factors determine the likelihood for implementing international standards. A typical explanation often cited for the absence of countries’ convergence in international rules is that domestic differences persist.¹¹ The literature emphasizes the role of framework conditions such as cultural, institutional and socioeconomic settings as determining the likelihood of change to implement and harmonize international standards.¹²

Most of these studies consider that cultural factors determine the domestic opportunity for change. Political discourses—the ideas and narratives behind policies and policy change— are set within the broader culture of a country. ¹³ At the root of the problem of ML are governments’ attitudes towards ML, which dictate its level of acceptance and the extent of the involvement of the banking sector in this activity. Governments must recognize that ML poses a serious threat to democracy and to the soundness of the financial system (Johnson et al., 2002).

Another strand of the literature has examined how institutional factors create both opportunities for and impediments to change. Some authors stress that institutions provide the context in which policy changes are defined. There are limits to institutional explanations, of course. Institutions may accommodate, constrain, or refract policies but can never function as the sole “cause” of policy change.¹⁴ An effective domestic AML/CFT regime requires that certain structural elements be in place, such as a good regulatory framework, appropriate measures to prevent corruption, rule of law, and government effectiveness, culture of compliance, an effective judicial system. ¹⁵ The lack of such elements, or significant weaknesses or shortcomings in the general framework, may significantly impair the implementation of an effective AML/CFT framework.

The relationship between some economic and financial domestic conditions is somewhat less clearly established. Some socioeconomic and financial domestic conditions are rarely mentioned in the literature but are especially relevant for compliance with the AML/CFT regime. This paper examines the effects of domestic factors, specifically domestic economic and financial factors, on countries’ compliance with the AML/CFT standard.

III. How Well do Countries meet the AML/CFT Standard from 2004 TO 2011 ?

The countries’ compliance has been analyzed for the 161 countries assessed since the revision of the AML/CFT assessment methodology in 2004. Table 2 (Annex 1) shows the countries selected for this section. ¹⁶ The scores reflect the levels of compliance at the time these countries were assessed and do not take into account any progress made in addressing deficiencies since then.

Country compliance with each of the FATF 40 + 9 Recommendations is rated according to four categories: compliant (C), largely compliant (LC), partially compliant (PC), and non- compliant (NC). Another category, non-applicable (N/A), is not a rating. It reflects instead situations wherein “…a requirement or part of a requirement does not apply, due to the structural, legal or institutional features of a country.”¹⁷ These ratings represent per-recommendation summary assessments of the findings of detailed assessment missions of assessor bodies.

In general, progress has been made in each of the seven components of the AML/CFT compliance: (A) legal; (B) institutional; (C) financial institutions’ prevention; (D)DNBFPs’ prevention; (E) informal sector prevention; (F) entity transparency; and (G) international cooperation. Although there has been substantial improvement in recent years in many countries’ compliance with the AML/CFT components, the analysis indicates that the countries’ degree of compliance remains low (see Figure 1 in Annex 1).¹⁸ We can summarize the degree of compliance with the AML/CFT as follows:

Assessed countries’ compliance with the FATF Forty Recommendations—the FATF 40+9— is low. Applying a scale under which a score of 49 represents full compliance with each of the FATF 40+9,¹⁹ the average compliance score for the 161 countries assessed using the current methodology from 2004 to 2011 was 20.8 or 42.5 percent. Full compliance on any FATF Recommendation was rare, occurring in only 12.3 percent of the almost 7,889 observations in this dataset.²⁰ Countries achieved the second highest score, largely compliant, 25.5 percent of the time, partially compliant, 35.6 percent of the time, and noncompliant 24.9 percent of the time.²¹

Different elements of the standard pertain to different components of a country’s AML/CFT regimes. Figure 1 summarizes the AML/CFT compliance with its seven components during 2004-2011.²² See also Table 2 in Annex 1 on jurisdictions’ compliance with the various groups of recommendations.

One clear conclusion is that elements of the standard that have been in place longer have higher compliance ratings, for example:

• The AML Recommendations, which have been in place since 1990, were more highly rated than the CFT Special Recommendations which were introduced in October 2001.²³ The degree of compliance²⁴ for the 40 AML Recommendations was 45 percent and 31.5 percent for the 9 Special Recommendations on CFT.

• Recommendations concerning designated nonfinancial businesses and professions, which were made subject to the standard only in 2003, have had some of the lowest compliance scores, averaging only 12.1 percent of the theoretical maximum.

It also appears to be easier to enact legislation and set up government institutions than to ensure that the system functions well on an ongoing basis.

• The degree of compliance for those Recommendations evaluating the criminalization of ML and TF were relatively high at 45.1 percent, and a somewhat broader measure of the strength of the legal system was 41.4 percent. The legal measures are not sufficiently harmonized with the international standard. The wording of the Conventions is rarely transposed fully into domestic legislation.²⁵

• The Recommendations that assess the strength of AML/CFT institutions (financial intelligence units (FIU), the specialized supervisory bodies, the police, and the judiciary) show a degree of compliance at 50.6 percent. However, some countries fail to provide adequate resources for financial intelligence units (FIU) and regulatory authorities to supervise financial and other institutions for AML/CFT purposes.

• Compliance with the Recommendations that assess the strength of preventive measures in the financial institutions was lower at 40.1 percent, and, as noted above, compliance was much lower for designated nonfinancial businesses and professions. In particular, some countries maintain weak customer identification policies for preventing access by money launderers and terrorism financiers to financial systems through financial institutions, DNFBPs, and the informal sector.

• One particular weakness in system functionality was indicated by the significantly low compliance (22.1 percent of the theoretical maximum) that all countries received on Recommendation 5 which concerns customer due diligence measures in financial institutions.

• Countries were also relatively strong in the area of international cooperation, showing a degree of compliance at 56.3 percent on applicable Recommendations. However, some countries have made inadequate provisions for mutual legal assistance, information sharing, and cooperation both domestically and cross-border.

• Compliance with the Recommendations that call on countries to have adequate transparency of legal persons and arrangements is relatively high at 40.1 percent. However, in many countries the prevention of ML/FT is still hampered by difficulties in identifying or lack of commitment to identify the ultimate beneficial owners of funds/assets.

• A higher degree of compliance appears to be associated with a higher level of economic development. In the 46 advanced economies in our sample, the degree of compliance with the AML/CFT is 56.8 percent, whereas, in the 115 emerging economies in our sample, the degree of compliance was lower (37 percent), indicating that wealthy countries have more resources to devote to constructing and implementing complex AML/CFT systems.

• In many cases it does reflect the measures they have taken to strengthen their supervisory and regulatory systems in advance of, or as a result of, the mutual assessment reports. The financial sector makes an important contribution to the economies of many of the higher-income jurisdictions, and these jurisdictions have taken steps to protect the integrity of their financial industry and their reputations. In general countries seek to avoid the stigma of being perceived to be AML/CFT havens and possibly identified by the ICRG as having strategic AML/CFT deficiencies.²⁶

• The AML/CFT domestic policies have generally grown more alike over time, in the same group, but at the same time they have moved in an upward direction. For example, in the case of European countries, accession to and membership in the EU seems to be more effective for harmonizing the AML/CFT because the countries have similar regulatory frameworks,²⁷ and the majority of them have participated in the design of the standards, so that they already have similar regulatory frameworks in place. East and South Asian countries and sub-Saharan countries have the lowest degree of overall compliance.

• While the standard has evolved from a one-size-fits-all basis into a more risk-based approach, it still appears to have been designed mainly for formal financial systems, but the relevance of the AML/CFT measures implementation varies by risk of countries to ML/FT. The risk does not always depend only on the level of development of jurisdictions and financial systems, but rather also on the level of proceeds of crime in the country. While some low-income countries are also at low risk for ML, this is not always the case.

• Assessing the effectiveness of countries’ implementation of the AML/CFT standard is quite challenging for a number of reasons. In practice, the assessment of effectiveness depends on what the supervised entities and countries’ authorities tell the assessors and the measurement of effectiveness by the existing assessment tools continues to be quite rudimentary. The assessors’ judgment is heavily dependent upon the quality and quantity of the information that the entities and authorities are willing to share. Furthermore, there are other factors undermining the evaluation of effectiveness, including the absence of agreed definitions of effectiveness and rudimentary measurements including, for example, the correlation between suspicious transactions reports, investigations, and convictions. Overall, the entire assessment apparatus suffers from a lack of data and the difficulty of interpreting data meaningfully.

• The challenges in strengthening the effectiveness of the global AML/CFT regime result from four tensions. Three of these tensions operate at a domestic level: the first lies between financial regulation and political will; the second arises from the difficult interaction between the international standard and its domestic implementation, and the third occurs between government and financial and non-financial institutions. At the international level, the tension lies in the difficulties cash-based economies continue to face in implementing the standard. However, the low compliance with the standard achieved by cash-based economies does not mean that they face a higher risk for money laundering or terrorist financing than do formal economies.

However, all of these results do at least raise the question of whether focusing on AML/CFT compliance ensures progress toward mitigating real ML/TF risk. Fund staff has been working with the FATF to better understand ML and TF risks on the basis of analyzing threat, vulnerability, and consequence. In this connection, future research should include the development of different, broader metrics for analyzing the level of ML/TF risks and the level of criminal activity in a given jurisdiction as well as better methods of capturing and presenting cross-border flows.

In this regard, we have analyzed the relationship between our level of jurisdictions’ compliance with the AML/CFT standard (2004–2011)—as a measure of financial sector vulnerability—and staff’s ranking on systemically-important financial sectors (2009). Figures 2 and 3 in Annex 1 present that relationship. The negative relationship indicates that, in general, smaller, less-connected financial systems are also less compliant with the standard, which is consistent with the rest of the analysis. Looking more closely at the scatter plot, however, helps identifying those relatively large and interconnected financial systems that are also relatively vulnerable to ML based on their low compliance with the standard. This, in turn, can provide a basis for prioritizing the TA and surveillance agenda of the Fund in the area of AML/CFT.

These pictures contribute to the Fund’s analysis of risk, although, of course, they are only a point of departure. A more complete analysis would require both a better proxy for systemic vulnerability than can be provided by compliance scores and other variables which add to the threat dimension. In general, work remains to be done in testing the relationship of compliance to ML/TF risks and on measuring the effectiveness of AML/CFT regimes. Continued work in these directions could open the door to a more nuanced understanding of where the ML/TF-related threats to the international financial system lie than can be derived from simple comparisons of countries’ ratings on their AML/CFT assessments.

IV. Empirical Evidence: Determinants of the AML/CFT Compliance

V. Caveats

The caveats of this research may be grouped as follows: (i) caveats derived from the 2004 methodology⁴⁰; (ii) caveats that affected the sample, such as bias and self-endogeneity; and (iii) the difficulties encountered in when comparing different domestic data.

• The 2004 methodology is the main limitation to overall compliance analysis. First, we are assuming that all the reports have the same quality and consistency. However, in accordance with the requirements of the 2004 methodology assessors often base their judgment of effectiveness, on statistical information, which is open to divergent interpretations.

  • Some Recommendations, such as those relating to customer due diligence or supervision, require a single aggregated rating for compliance, but the aggregated score covers considerable information.

  • The great variety in the operational rules and guidelines adopted by countries in the implementation of the AML/CFT measures to DNFBPs should require assessors to review the issues in a more disaggregated manner; strengths or weaknesses observed under one Recommendation can have a cascading effect on a country’s compliance with other Recommendations.

  • Some features of the AML/CFT regime are rated under more than one Recommendation; some criteria have proven difficult to assess in largely cash-based economies.

These factors call for the exercise of caution in interpreting the data produced by assessments.

• Bias and self-selection limit the validity of the results. ⁴¹

  • The sample is not randomly selected. The 116 countries in the sample are countries that have voluntarily submitted themselves to mutual evaluations, through their membership in FATF or a FATF-style regional body, or to an assessment by the IMF or the World Bank.

  • The scientific challenge created by this situation presents certain challenges for which there is not an obvious solution at this stage. Therefore, caution is required in interpreting and extrapolating the results, as there may be unobservable factors involved and the available data is representative of countries that are motivated to report differently than some other countries.

  • Accordingly, the estimator is biased and does not, on average, predict the population parameter correctly. However, based on the expected systematic error that will be present due to the configuration of the sample (high compliance countries are probably reporting more often and more accurately than are noncompliant countries), the curves can still be considered as accurate depictions of the simultaneous relationships of the variables for the countries in the sample.

  • Notwithstanding their weaknesses, the analysis of the findings produced by the tests is necessary and relevant in the absence of an alternative source of data.

• The aggregation of cross-cultural and cross-country data limits a meaningful result.Each country displays its own peculiarities in defining criminal behavior, elaborating policies, and implementing the international standard, including the recording and reporting of suspicious activity. This creates significant challenges in accurately interpreting the resulting aggregation.

VI. Conclusions and Policy Recommendations⁴²

The paper contains two main conclusions: (i) a comprehensive assessment of countries’ ML/FT risk is overdue. Although some countries have initiated a dynamic process through which all the key components of compliance are improving, it is necessary to evaluate to what extent countries’ compliance with the AML/CFT international standard is effective to lessen countries’ vulnerability to ML and FT and (ii) the standard itself should incorporate more comprehensively the underlying risk of ML/FT, including all sectors that are vulnerable to ML and FT —the financial sector and the real sector (e.g., trade) particularly in relation to international trade. This in turn would allow countries to respond more effectively to existing risks and to adapt its AML/CFT policies to address those risks changes in ML/FT typologies.

The paper proposes that when a country’s compliance with the standard is insufficient and/or the country’s risk to ML/FT is high, the Fund should pay attention to the economic effects of ML/FT on both the domestic economy and across borders. Currently, the assessment of country compliance with the AML/CFT standard is not sufficient to establish the economic risks and consequences of ML/FT. The Fund should devote the resources necessary to better understand the phenomena of ML and FT, including through the development of a holistic ML/FT vulnerability assessment that would evaluate in addition to the existing areas, the more opaque areas of the financial and real sectors, the transmission channels, and the propagation mechanisms that could affect macroeconomic performance and financial stability.

Existing departmental fora—the preparatory processes for the World Economic Outlook (WEO), the Global Financial Stability Report (GFSR) work, and the Vulnerability Exercise (VE) for advanced economies and emerging market economies—could be adapted to this end. Moreover, these fora have the particular advantage of bringing global and regional developments to bear on the analysis of individual countries.⁴³ The VE⁴⁴ could serve as a central forum for interdepartmental discussion of emerging market economies that would ultimately feed into the selection and prioritization of AML/CFT issues into Article IV discussions. These exercises could contribute efficiently to the Fund’s macroeconomic analysis as it takes into account criminal and underground market developments. For countries vulnerable to ML/FT, the Fund could elaborate concise analytical briefs on ML/FT that would feed into the broader surveillance agenda. The briefs could serve to identify and prioritize key ML/FT issues to be covered in successive Article IV consultation cycles.