A.  Goals

Taking stock of how money laundering controls should work must begin with system’s goals. They are multiple but broadly consistent. There is a marked difference between the goals as articulated by the major institutions, however, and those stated or implied in the small research literature. The institutional pronouncements emphasize threats money laundering poses to the integrity of the financial system. Consider the following statements:

Note that none of those statements gives much emphasis to reducing the extent of crime or even of terrorism. The only explicit goal is to prevent the use of the financial system for furtherance of crime or terrorism. A lower-level FATF objective (Intermediate Objective 3) does include “Depriving criminals of illicit moneys and otherwise sanctioning criminals.” The US National Strategy for Combating Terrorist and other Illicit Financing for 2024 mentions reductions in predicate crimes only late in the introduction and without giving it emphasis: “Increasing transparency in financial services, including for professionals who provide financial services, can help prevent illicit actors from distorting U.S. markets, driving up the costs of real estate and goods, stealing money from ordinary American citizens and companies, financing terrorist attacks, the proliferation of weapons of mass destruction, drug trafficking, and state-supported corruption” (US Department of the Treasury 2024, p. 8).

Other bodies have been more explicit in their view that AML has aims other than the reduction of money laundering. For example, the British Financial Conduct Authority, a major AML agency in the United Kingdom, links money laundering controls to the prevention and reduction of financial crime but not to predicate crime generally.¹³ The Canadian FIU also gives more emphasis to crime fighting.¹⁴

Most scholars assume that AML is a tool for reducing harms associated with crime, terrorism, and proliferation of weapons of mass destruction (Levi and Reuter 2006; Pol 2018; Levi and Soudijn 2020). This reflects in part the lack of compelling evidence that money laundering by itself causes much harm. Ferwerda (2013) list 25 possible harms from money laundering, none of which have ever been measured and many of which seem fairly implausible or strained. Protecting the integrity of the financial system is often presented as a primary goal of AML.

This claim invites critical examination. What does the “integrity of the financial system” mean when so many leading financial institutions have faced accusations of misleading or even defrauding their customers?¹⁵ If these institutions routinely engage in behavior that undermines trust—through deceptive practices, complicity in illicit financial flows, or outright cheating—how credible is the claim that AML protects system integrity? Does AML genuinely serve to uphold this integrity, or does it function primarily as a compliance mechanism that papers over more fundamental failings? Addressing these questions could provide crucial insights into the true purpose and effectiveness of AML. For now, however, we set these broader questions aside and concentrate on AML’s role in mitigating crime-related harms.

B.  Mechanisms

How is the system supposed to achieve its goals? The AML system imposes a set of obligations on financial institutions, broadly defined, to assess the origins and purposes of funds, in order to identify funds that are criminal in origin or being used to finance terrorist activities. Failure to implement those obligations faithfully can lead to fines or criminal penalties against the institutions and, in theory, individual executives in them.

The mechanisms by which this is supposed to reduce the mingling of criminal earnings and terrorist finance with legitimate funds include financial institution reputational risk and conventional cost-benefit calculation of penalties that are imposed only with some probability, conditional on violation. Banks are considered to be more concerned about reputation than other business enterprises, since they are dependent on customers trusting them to hold their assets. Banks can disappear quickly when their soundness is questioned, as occurred with Silicon Valley Bank in 2023.¹⁶ In theory, if a bank is found to have violated its money laundering obligations, it risks being judged unsound, which can significantly affect its value, as is illustrated by the Danske Bank scandal. In 2018, following revelations of massive money laundering control violations, Danske Bank lost 11,000 Danish retail customers and saw its annual profit fall from 20.11 billion Danish Kroner in 2017 to 13.91 billion Danish kroner.¹⁷ However, the lasting effects remain unclear, as Danske Bank is now set for record profits in 2024 after settling the charges related to the 2018 money laundering scandal in September 2024.¹⁸

A broad array of authorities, including bank regulators and law enforcement agencies, have the authority to assess the financial institution’s compliance with the increasingly complex laws and regulations that constitute the AML system. However, banks have incentives to ignore the AML rules, just as they might be tempted to ignore any other business restrictions. Fundamentally, they seek to minimize expenditures on compliance. They also benefit from accepting funds with criminal origins; these provide, as do any funds, the basis for money-making loans. Thus, in deciding how faithfully to meet their AML obligations, financial institutions must balance the probability of detection and any subsequent penalties against the foregone revenues from AML expenditures and lost deposits.

For nations, the same pair of mechanisms apply. A nation that flouts the international rules runs the risk of earning a bad reputation, which might lead to less cooperation from other countries on other issues. There are also penalties that can be imposed by the systemically important countries, in particular by the United States refusing to allow the offending nation’s banks to conduct dollar transactions. For example, Nauru in 2002 was designated a money laundering concern by the US Department of the Treasury. As a result, financial dealings between US financial institutions and any Nauru-licensed institution were prohibited.

Overall, there is a plausible theoretical explanation for why nations and financial institutions should make good faith efforts to impose a credible set of money laundering controls and why those controls should reduce the extent of money laundering and the flow of funds to terrorists. In Sections III and IV, we examine how well this theoretical explanation works.

A.  SARs and Convictions for Money Laundering in the United States

Federal, state, and local law enforcement agencies supplement on-going criminal investigations by querying the Bank Secrecy Act (BSA) database of the FinCEN for name matches to existing suspects and their known associates.³⁷ For example, if the US Drug Enforcement Administration is investigating a specific individual in a narcotics case, agents would likely query the FinCEN database by name to identify additional leads, such as bank accounts, individuals and business associates, geographic locations, or aliases. A local prosecutor making an organized crime case might do the same. According to the latest data from FinCEN, over 25,000 authorized personnel from 472 federal, state, and local law enforcement agencies have access to BSA data, and in fiscal year 2023 they conducted over 2.3 million searches using FinCEN Query (FinCEN 2024).

Figure 1 shows the number of SARs filed with the FinCEN (left hand axis) and the number of offenders sentenced in the federal system for money laundering (the right hand axis) from 1996 to 2023.³⁸ The number of SARs registered an astonishing 71-fold increase, rising from little more than 50,000 in 1996 to almost 4 million in 2023. However, interpreting changes in the number of SARs filed is challenging. A rise in SARs may indicate an increase in money laundering activities, heightened vigilance by AML obliged entities, or both. The COVID-19 pandemic likely accounts for most of the sharp increase after 2020, given the recorded rise of fraud and cybercrime worldwide during this period (Buil-Gil et al. 2021).

It is reasonable to infer that the level of scrutiny by financial institutions has intensified since the late 1990s, especially since no specific event would account for such a long-lasting increase. The list of obliged entities has consistently increased over time, with AML obligations being imposed on new industries (e.g., the cryptocurrency sector). Even so, depositary institutions (e.g., saving banks, commercial banks, credit unions) still file most SARs. For example, in the period 2014 – 2022, depositary institutions filed 10,154,148 SARs. That is 48 percent of the SARs received by FinCEN (21,144,577).

Moreover, there has been a rise in the sophistication of corporate AML programs, along with advancements in detection technology, increasingly stringent regulatory priorities, and increased scrutiny by enforcement bodies. For example, in addition to rule-based monitoring systems, obliged entities have started to use advanced monitoring systems that incorporate artificial intelligence and machine learning to identify more suspicious activity while supposedly reducing false positives. The global market for AML solutions provided by vendors is growing rapidly, with a staggering forecast of $11.9 billion by the year 2030, sustained by tightened regulations and expanding compliance obligations around the globe.³⁹

There is a face plausibility to the claim that the number of people prosecuted and sentenced for money laundering reflects the effectiveness of the SARs system. One test of that hypothesis is whether the number of money laundering cases and convictions showed a similar increase. They did not. Convictions for money laundering decreased slightly between 1996 (827) and 2020 (755), then increased about 50 percent between 2020 and 2023 (1,312). The ratio of money laundering convictions to SARs filed decreased from 0.016 in 1996 (one money laundering conviction for every 63 SARs) to 0.0003 in 2023 (one money laundering conviction for every 2,904 SARs). On this measure, the system seems to be getting less effective.

How to explain high and increasing numbers of SARs compared with low and stagnant numbers of convictions for money laundering? This discrepancy, which was identified by Unger and van Waarden (2009) in their comparative analysis of Dutch and US SARs figures spanning 1998 to 2007, seems to underscore a persistent systemic trend. There are several potential explanations. First, SARs record a diverse array of suspicious activities affecting obliged entities. The spectrum is broad and includes fraud (such as check fraud, mortgage and consumer loan fraud), embezzlement, identity theft, insider trading, bribery/gratuity, and computer hacking, among others. Moreover, in the United States, SARs are not limited only to suspicious activity by customers. For example, if an obliged entity believes an employee is engaging in insider trading, it must file a SAR to the FinCEN. Cyber-attacks or other unauthorized access by internal or external attackers trigger a reporting action. Although these instances constitute a tiny minority of the total,⁴⁰ the figures suggest that SARs have become multi-purpose instruments, extending far beyond money laundering.

Second, another possibility is that the small number of money laundering prosecutions is driven by the low quality of SARs. For example, many obliged entities consider incomplete customer information as sufficient grounds to report a suspicious transaction. Moreover, many SARs are filed when analysts cannot determine funding sources, find legitimate reasons for unusually large deposits or withdrawals, or identify relationships between parties that could explain out-of-pattern activity. They are more like “incomplete transaction reports” than “suspicious activity reports.” This information overload negatively affects resource-limited governmental authorities and law enforcement agencies’ capability to investigate money laundering activities effectively. For example, given the very large number of reports being received annually, FinCEN is not able to comprehensively analyze each SAR (FATF 2016). Instead, it prioritizes specific types of SARs for further analysis based on evolving parameters reflecting national strategic priorities and law enforcement agencies’ feedback. This is not surprising as FinCEN had only 269 staff members as of December 2021; in 2023 there were 14,162 SARs per employee. Not all staff members review SARs, so the workload is even heavier.⁴¹ This problem seems to have been recognized. The 2023 federal budget allocated $210 million for FinCEN, $83 million above the 2021 enacted level and $49 million above the 2022 enacted level, to increase oversight of the financial sector, strengthen corporate accountability, and provide adequate support to law enforcement and investigative entities, increasing its staff by more than 50 percent (from 269 to 420 staff members).

In the current AML framework, obliged entities may have a perverse incentive to report as many transactions as possible – the so-called “crying wolf syndrome” or “defensive filing” – rather than carefully select them for authorities’ scrutiny (Takats 2011; Gara and Pauselli 2020; Dalla Pellegrina et al. 2023). Failing to report transactions that are found ex-post to involve money laundering (i.e., false negatives) leads to significant financial fines for AML obliged entities. In addition, employees can be held criminally liable for not sharing their suspicion, or actual knowledge, of an illicit transaction with their institution’s money laundering reporting officer or with the national FIU directly by filing a SAR (e.g., Section 330 of the 2002 Proceeds of Crime Act in the United Kingdom).⁴² Conversely, no fines are imposed for reporting false positives, namely transactions which turn out not to involve money laundering. This issue is far from being merely theoretical. Dalla Pellegrina et al. (2023) empirically found that, in the years immediately following the introduction of the risk-based approach in Italy, financial intermediaries lowered the reporting test required to report a transaction as suspicious. They found that the observed increase in SAR activity in relation to total bank transactions could be explained by a decrease in the reporting test adopted by the intermediaries. Conversely, Gara et al. (2019), using data from Bank of Italy and the Italian Financial Intelligence Unit found that, ceteris paribus, inspection by authorities indeed increased the number of SARs filed by the banks by 18 percent, but the effect was not limited to low-quality reports. The authorities’ on-site controls increased a bank’s propensity to identify and report risky transactions.

Over-reporting is more likely to happen within the United States because of its adversarial legal system, characterized by stringent public rule enforcement and significant criminal and administrative penalties (Unger and Van Waarden 2009). This issue calls for examining, in addition to the number of filings, also the extent to which they have resulted in the detection and punishment of money laundering offenses. Despite being a cornerstone of the current AML system, information on the effects of SARs on law enforcement efforts remains scarce. Obliged entities receive minimal, if any, feedback from law enforcement or other government bodies on how these reports are used. FinCEN itself gathers limited data from law enforcement agencies regarding their use of BSA reports or outcomes influenced by them, as these agencies rarely collect such information routinely.⁴³ Moreover, FinCEN does not consistently share the few metrics available from agencies that do generate them and does not provide institution-specific feedback (US General Accountability Office 2022).

However, a little evidence on the use of BSA reporting by law enforcement agencies is now available. A US General Accountability Office (2020) survey of six federal law enforcement agencies found that from 2015 to 2018 more than 72 percent of their personnel used BSA reports to investigate money laundering or other crimes.⁴⁴ Investigators who used BSA reports most frequently found them useful for identifying new subjects for investigation or expanding ongoing investigations. In its year-in-review for fiscal year 2023, FinCEN reported that nearly 15.4 percent of active FBI investigations were directly linked to BSA data. In particular, BSA data was used in 33.8 percent of active complex financial crimes investigations, 22.1 percent of active transnational organized crime investigations, 27.6 percent of active public corruption investigations, and 14.8 percent of active international terrorism investigations (FinCEN 2024).

Additional evidence emerges from other agencies. The Internal Revenue Service’s Criminal Investigations unit disclosed that the primary subject of 88 percent of investigations opened during fiscal year 2023 had a Bank Secrecy Act (BSA) filing. From fiscal year 2021 to fiscal year 2023, BSA data was instrumental in securing average prison sentences of 39 months and seizing a total of $7.4 billion in assets tied to criminal investigations. BSA data also resulted in restitution orders totaling $434 million and forfeited assets totaling $629 million, nearly double and triple the amounts, respectively, from fiscal year 2020 to fiscal year 2022 (Internal Revenue Service 2024). Concurrently, the Organized Crime Drug Enforcement Task Force, an office within the US Department of Justice that focuses on transnational criminal organizations, used BSA information in numerous investigations closed between 2018 and 2022. Among these investigations, BSA information contributed to 89 percent of those resulting in financial convictions, 82 percent in other convictions, 86 percent in disruptions, and 87 percent in dismantlements (US General Accountability Office 2024).⁴⁵

These indicators show that AML filings do enhance law enforcement against a variety of crimes. While most individual SARs never lead to criminal convictions for money laundering, the cumulative flow of information as incremental intelligence is nonetheless of value to law enforcement agencies. SARs rarely serve as the initial trigger for investigations; instead, they often enhance existing cases by adding detail or corroborating evidence, particularly when law enforcement already has suspicions about a target. They help law enforcement and prosecutors build stronger cases against specific criminals, offering insights that may not have been available otherwise.

If a primary AML goal is the reduction in predicate crimes, the data on the use of BSA filings is more important than a measure of the number of money laundering prosecutions. While the FATF calls for a number of money laundering convictions consistent with the significance of the problem (see, e.g., the 2016 United States MER), it is worth noting that obtaining money laundering convictions is complex for several reasons. First, even if international standards do not require the establishment of the precise predicate offense for a conviction for money laundering, national supreme courts have increasingly emphasized that prosecutors must identify the predicate offense as well. This variability in legal interpretation significantly affects convictions, as prosecutors may hesitate to initiate money laundering investigations when this information is not readily accessible (Eurojust 2022).

Second, money laundering charges may be dropped if the defendant pleads guilty to an equally serious crime and cooperates with authorities in providing evidence against co-conspirators and higher ranking members in the criminal enterprise. For example, according to the Executive Office for US Attorneys CaseView data,⁴⁶ US Attorneys collectively charged 2,100 to 2,500 defendants annually for money laundering in fiscal years 2018–22. However, as Figure 1 showed, only 820 to 1,200 defendants per fiscal year were found guilty. This represents many decisions to drop money laundering charges, not mostly failures to convict (US General Accountability Office 2024).

Third, several significant issues potentially arise when prosecuting transnational money laundering cases. A conviction for a money laundering offense is usually impossible when the offender is a foreign national residing outside the country who cannot be located or extradited. Also, money laundering may not constitute an offense irrespective of the jurisdiction where the predicate offense was committed or the predicate offense in question may not constitute a crime in the other country but merely an administrative offense. While this was likely a more significant issue in the past, it has been mitigated over time, partly due to FATF pressure. Certain cases, such as tax avoidance or some environmental offenses, may still pose challenges. However, the remaining discrepancies in the treatment of money laundering reflect differing legal traditions and social values across jurisdictions, rather than a defect that requires remedy through harmonization.

B.  Forfeiture of Crime Proceeds

Asset forfeiture is a powerful tool in law enforcement, enabling authorities to identify, seize, and ultimately forfeit the assets of criminals and their organizations. By depriving criminals of their illicit gains, asset forfeiture serves as a critical mechanism not only for punishment but also for deterring future criminal activity. The rationale is straightforward: if criminals are unable to retain the financial benefits of their crimes, it reduces the incentive for others to engage in similar illegal activities. Criminals often care more about losing their assets than about facing temporary imprisonment. For many, keeping these assets—whether for personal use after release or for family support during incarceration—was a major motivation behind their criminal actions. To ensure effective punishment and deterrence, prosecutors frequently prioritize asset forfeiture as a central element in criminal case settlements.

Beyond its punitive aspect, asset forfeiture plays a crucial role in incapacitating criminals. By seizing assets, authorities do not just inflict injury on criminals personally; they aim to cripple the infrastructure of their illicit enterprises. Imagine confiscating vehicles used for drug smuggling and the large sums of cash meant for buying the next shipment. Without the means to finance or support their operations, individuals and groups are less capable of continuing their criminal endeavors. However, this incapacitating effect is limited if suppliers are willing to extend credit again, particularly if they do not view forfeiture as a consistent risk. While asset forfeiture may reduce the attractiveness of criminal operations, it does not necessarily incapacitate criminals in the long run if they can re-establish their financing.

In addition to punishment, deterrence, and incapacitation, asset forfeiture also serves other important functions in the justice system. It can facilitate restitution to victims by returning forfeited assets or their equivalent value to those harmed by criminal activities. Moreover, asset forfeiture contributes to community protection. Forfeited funds can be allocated to law enforcement agencies to support their operations, while some assets may be transferred to community organizations for social benefit.

However, a common, indeed almost universal, critique of the current AML system is the low level of forfeiture of the proceeds of crime (Pol 2020). A 2011 report by the United Nations Office on Drugs and Crime (2011, p. 7) estimated that “much less than 1% (probably around 0.2%) of the proceeds of crime laundered via the financial system are seized and frozen.” Over a decade later, the European Union Agency for Law Enforcement Cooperation (Europol), with the support of contributing member states, conducted an ad hoc data collection for the first European Financial and Economic Crime Threat Assessment. It estimated that the assets law enforcement successfully confiscates from criminal networks amount to less than 2 percent of the annual proceeds of organized crime (Europol 2023).

Although these figures are subject to significant uncertainty, largely due to difficulty in accurately estimating total illicit proceeds, they reflect an order of magnitude that is widely regarded within policy circles as plausible. They provide a rough but consistent indication of the effectiveness or lack of effectiveness of current AML efforts.

The limited effectiveness in asset forfeiture is acknowledged by the FATF: “Just 19% of the 120 assessed jurisdictions are demonstrating high or substantial levels of effectiveness in investigating, prosecuting, and convicting money laundering offenses and confiscating the proceeds of crimes” (FATF 2022, p. 38). Unsurprisingly, improving asset recovery has become a priority. In November 2023, the FATF introduced amendments to its standards focused on enhancing asset recovery efforts, specifically recommendations 4 and 38. Among other changes, the updated recommendations now require countries to establish asset recovery as a national priority, create a non-conviction-based confiscation regime within their legal systems, implement extended confiscation measures, and recognize other countries’ preliminary and final court orders related to assets subject to confiscation.

Confiscations of crime proceeds could, besides their use in obtaining money laundering convictions, be used as another proxy for the effectiveness of AML efforts. However, confiscation can take place many years after the original offense or even the initial criminal charge. This delay highlights the importance of examining the effects of asset seizure and freezing, rather than focusing on actual confiscation. Freezing involves temporarily prohibiting the transfer, conversion, disposition, or movement of assets or temporarily assuming custody or control of assets on the basis of an order issued by a court or other competent authority. Seizure means temporarily restraining an asset or putting it into the custody of the government and may apply to physical assets. Generally, these measures are used to prevent the movement of assets pending the outcome of a case.

Figure 2 shows the number and dollar value in billions of US dollars of seized assets (on the right-hand axis) associated with money laundering-related statutes from 2018 to 2022 in the United States. Both series exhibit a declining trend. Overall, the number of seized assets decreased from 1,621 in 2018 to 854 in 2022 (minus 47 percent). The total value decreased from $488 million in 2018 to $414 million in 2022 (minus 15 percent), after peaking at $1.6 billion in 2019; the post-2019 decline may reflect the Covid-19 lockdown. These figures are not adjusted for inflation, which means the decline in value is even greater. By any measure of the proceeds of predicate crime, the value of seized assets is negligible.

The results remain relatively consistent, even with a broader perspective. Figure 3 illustrates the number and dollar value (in billions of US dollars) of forfeited and seized property in the Department of Justice Assets Forfeiture Fund, managed by the Federal Asset Forfeiture Program, from 2018 to 2023. This covers all offenses, not only money laundering. The number of newly forfeited and seized assets increased modestly, by 8 percent, over this six-year period. However, the total value of assets increased a significant 324 percent, a rise largely driven by the 2023 data. Excluding 2023, the increase was a 43 percent, indicating a more stable trend up until that year. Overall, the main conclusion is unchanged: the value of seized assets is relatively negligible.

Several significant obstacles help explain why freezing, seizing, and ultimately confiscating criminal proceeds remains so challenging. First, the increasing number of tools used to disguise ownership of assets makes tracing financial flows difficult. Shell companies, numbered bank accounts, and strawmen are but a few such tools (see, e.g., Paquet-Clouston et al. 2024). Second, financial investigations often involve multiple countries, necessitating extensive international cooperation. Even when assets are located, linking them to suspects or defendants can be difficult, especially if they are in jurisdictions with strong banking secrecy laws (e.g., Liechtenstein) or if information about the beneficial ownership of trusts and companies is not readily available or accessible.

C.  Money Laundering: Threats

Outputs of the AML system do not provide a basis for assessing whether laundering is more expensive, difficult, or risky than before the creation of the current regime. There is no time series on average costs of laundering criminal proceeds. Moreover, those costs might vary by crime type to reflect the higher risks associated with, say, laundering drug money as opposed to money from antiquities theft. Overall, though the rhetoric of MERs says otherwise, FATF efforts have been almost entirely focused on measuring formal compliance with FATF standards, rather than looking at program effectiveness and outcome effectiveness (Halliday, Levi, and Reuter 2020). FATF has long incentivized countries to identify compliance metrics as a measure of success. However, this is inherently flawed as the opposite is also formally correct: if AML policies are effective in deterring offenders from laundering illicit proceeds, there should be small numbers of arrests, confiscations, and SARs filed by obliged entities.

A more useful approach for assessing whether the system is affecting its immediate target would be to look at how offenders launder their illicit proceeds. Offenders need to be only as sophisticated as AML controls in place require them to be to avoid detection (Levi 2015; Levi and Soudijn 2020). By looking at offenders’ observable behaviors, we may be able to infer the potential challenges they encounter in laundering their illicit proceeds.

Over the last two decades, criminologists have shown a modest but slowly growing interest in the process by which illicit proceeds are converted into forms that allow criminals unrestricted opportunities to spend and invest in the legal economy. Empirical knowledge on the behaviors and choices of money launderers is, however, still scarce (see, for an overview, Table A1 in the Appendix). To date, there is minimal public knowledge about challenges offenders face in laundering money, including their perceptions of risk associated with various laundering methods (Berry, Salinas, and Gundur 2023). Journalistic exposés are important but can be misleading since they accentuate the dramatic and complex; see, for example, Cox (2024) for one of the better such accounts.

According to the popular “arms race” metaphor in this domain, the growing reach of AML efforts should make money laundering more difficult and expensive for offenders by forcing them to use ever newer and more sophisticated methods and rely more extensively on professional money launderers (PMLs), namely individuals or groups of people who provide money laundering services in return for a market-based fee. We review the evidence provided by the existing literature on these dynamics below.

D.  Use of Ever Newer and More Sophisticated Methods?

Low-level offenders often do not launder their illicit proceeds as they use them directly for daily expenses or more hedonistic purchases to sustain a lavish lifestyle (Petrunov 2011; Caulkins and Reuter 2022; Berry, Salinas, and Gundur 2023). Likewise, offenders can spend their illicit proceeds in the criminal economy to pay wages to their affiliates, finance territorial expansion against rival groups, or keep their criminal business up and running (e.g., payments for new drugs shipments) (Soudijn 2016; Krakowski and Zubiría 2019; Levi and Soudijn 2020). Despite being classified as money laundering under the national laws of most countries worldwide, these actions rarely require any active laundering activities (such as efforts to conceal origins) to make the proceeds expendable in the legal economy (Van Koningsveld 2013). For purposes of this essay, we do not regard such expenditures as money laundering.

After direct spending and reinvestment in criminal activities, only the remaining funds need to be laundered. There has been much debate about what constitutes the process of money laundering starting with a well-known three-stage model that “has become synonymous with defining the modus operandi of all money laundering methods” (Gilmour 2016, p. 2). Although the model’s origin is uncertain,⁴⁷ it has been strongly endorsed by international and national agencies worldwide since the 1980s (Levi and Soudijn 2020). Money laundering is seen as the sequence of three different phases: placement, layering, and integration. Placement involves the deposit of illicit cash into the financial system to facilitate its movement (Van Koningsveld 2013). Layering involves multiple financial operations aimed at severing the connection between money and its criminal origin (Villányi 2021). Integration encompasses the incorporation of illegal proceeds into the legal economy, primarily through investments in real estate, legitimate companies, high-value goods, and financial assets once the funds cannot be traced back to their illicit origin (Soudijn 2016).

Scholars have long criticized the three-stage model for being incorrect and outdated (Van Koningsveld 2013; Cassella 2018; Levi and Soudijn 2020; Matanky-Becker and Cockbain 2021). First, not all three phases are consistently required in money laundering schemes. For example, financial fraud often involves illegal proceeds being drained from victims’ bank accounts and transferred into the offender’s accounts. In such cases, placement does not take place as the illegal proceeds are already within the financial system. Matanky-Backer and Cockbain (2021) reviewed 31 money laundering investigations carried out by the United Kingdom’s tax enforcement agency between 2016 and 2019. Only 29 percent of the cases featured at least one of these three allegedly fundamental stages of money laundering.

Second, the model was developed in the early 1980s for the fight against drug trafficking and inevitably overemphasizes the role of cash (Levi and Soudijn 2020). Many schemes (e.g., financial fraud, cybercrime, high level corruption) often do not involve cash at any stage of the laundering process (Nazzari 2023a; Costa and Jancsics 2024).

Unlike many other types of crime, money laundering “is notable for the diversity of its forms, participants, and settings” (Levi and Reuter 2006, p. 312), resulting in a variety of different conducts ranging from simple to highly complex (Arnone and Borlini 2010; Gilmour 2016). For both research and policy it is important to acknowledge and investigate the heterogeneity of money laundering from different criminal activities (Nazzari 2023b). The array of predicate offenses has broadened significantly beyond drug trafficking by the mid 1990s to include, among others, serious organized crime, fraud, tax evasion, terrorism financing, human trafficking, state-sponsored and corporate bribery, and cybercrime.

Laundering a single set of crime proceeds can involve multiple methods. The FATF investigates money laundering methods and trends to assist countries in identifying, assessing, and understanding their money laundering risks by publishing ad hoc reports in its “Methods and Trends” series. The methods include, among others, the electronic transfer of funds, use of cash-intensive businesses, real estate, trade-based money laundering, cash smuggling, and loan-back. The techniques employed to launder funds have also been extensively discussed in the literature. However, the research field has been mostly dominated by highly descriptive studies (or “typology” studies) on how illicit proceeds can be laundered from a theoretical point of view (see, e.g., Teichmann 2017, 2020; Wronka 2022; Gobena 2023).

The few empirical studies consistently find that offenders employ relatively simple strategies (Steinko 2012; Kruisbergen, Kleemans, and Kouwenberg 2015; Soudijn 2019; Matanky-Becker and Cockbain 2021). Tried and true methods keep returning. It is often assumed that money laundering methods are in constant evolution, but use of new and more sophisticated methods (e.g., cryptocurrencies) remains limited and is mostly used classes of offenders, such as cybercriminals, who have additional technical skills (Nazzari 2023b).

Complex money laundering schemes certainly exist but are probably not the rule and contribute to the persistence of the “ingenuity fallacy,” namely that “the situation is imagined to be more complex than it really is” (Levi and Soudijn 2020, p. 6). It is important to note that these findings of simple schemes are often criticized for being based on police files and court documents, which are inevitably biased by the priorities, constraints, and resources of law enforcement agencies (see Cockbain, Bowers, and Vernon 2020). Consequently, detected money laundering patterns may “unconsciously reflect what we have been able to detect in the past, and these can be ossified into routine detection and control practices” (Levi 2015, p. 9). Because of the “streetlight” effect (e.g., searching for keys where there is a streetlight rather than where they fell), simpler cases may be easier to catch. For example, given the well-known challenges of cross-border police cooperation,⁴⁸ law enforcement agencies may prioritize national cases over transnational ones which are potentially more complex. Likewise, law enforcement agencies may fail to effectively identify PMLs within criminal networks when they do not adopt a financial perspective from the beginning of the investigation (Soudijn 2014).

Cases that come to light are thus, at best, only a small and unrepresentative sample of the general population of money laundering schemes and caution is needed in drawing conclusions about overall patterns. These criticisms are certainly pertinent, but it is worth noting that basic money laundering strategies also emerge when using alternative data sources. Several scholars have analyzed money laundering behaviors of cybercriminals by exploiting blockchain data and have found that even offenders who seem to succeed in their illicit activities, having never been arrested, still resort to basic money laundering methods (Paquet-Clouston, Haslhofer, and Dupont 2019; Nazzari 2023a; Trozze, Davies, and Kleinberg 2023). Others, after interviewing active or convicted offenders, provide similar findings (Matrix Knowledge Group 2007; Berry, Salinas, and Gundur 2023; Cretu-Adatte et al. 2024).

E.  Growing Reliance on Professional Money Launderers?

Stricter controls (e.g., higher probability of detection, longer sentences in case of conviction) may encourage offenders to choose PMLs over self-laundering to lower their risks of detection and arrest. However, most empirical evidence shows that self-laundering is still the most prevalent type of laundering (Malm and Bichler 2013; Matanky-Becker and Cockbain 2021; Berry, Salinas, and Gundur 2023). The limited involvement of PMLs in money laundering schemes may have different explanations. First, most low-end offenders in several illicit markets have relatively low earnings that simply do not allow, or justify, the payment of PMLs’ fees (Berry, Salinas, and Gundur 2023). Second, entrusting illicit proceeds to a third-party inevitably involves a plethora of potential principal-agent issues that can significantly increase the risks faced by offenders (Levi and Soudijn 2020).

In addition to looking at offenders’ behaviors, one could use prices of money laundering services as “a performance indicator as to how well the control system works” (Levi and Reuter 2006, p. 320). Based on the risks and price framework elaborated by Reuter and Kleiman (1986), prices of goods and services in illegal markets should compensate offenders for the risks they are exposed to. From a theoretical point of view, stricter enforcement should increase the risk of detection and arrest by adding an extra cost to the suppliers of money laundering services who, assuming an efficient market, should raise their commission fees to compensate for the increased risks they face. In turn, predicate offenders should reduce their engagement in criminal activities because of the lower net returns after the costlier money laundering process.

However, scholars have generally overlooked the price dynamic of the money laundering market and information on the prices of money laundering services is minimal and ambiguous (Levi and Soudijn 2020). Despite a growing interest in PMLs within the criminological literature (see, e.g., Levi 2020; Kramer et al. 2023; Soudijn 2024), only two studies have investigated the price of money laundering services. Reuter and Truman (2004) analyzed eight cases involving PMLs worldwide finding an average commission fee of 9 percent for their services. Soudijn and Reuter (2016) analyzed six cases of bulk cash smuggling used by Colombian drug traffickers to move illicit proceeds from the Netherlands back to Colombia. Their results showed that such services are quite expensive, with cash smugglers charging between 10 and 17 percent of the total amounts.

Fragmentary information also emerges from other academic studies. Soudijn (2012) recounts a criminal case involving a Dutch cashier in Amsterdam who charged a British drug trafficker a 2 percent commission for exchanging pounds into euros. Kruisbergen et al. (2019) analyzed 30 organized crime cases to investigate the role of new technologies in money laundering. One case featured a PML who charged a commission fee of 7 percent to convert Bitcoins to cash–a much higher fee compared to those of legitimate providers. Farfán-Méndez (2019) analyzed the risk appetite of Mexican drug organizations laundering illicit proceeds. In one of three case studies examined, PMLs laundered drug proceeds by buying jewels and gold, reselling the high-value goods later to refineries mostly located in Florida and then smuggling the resulting cash to Mexico. This process cost the drug traffickers only 2 to 4 percent of the illicit proceeds handed over to the PMLs. Lastly, Benson (2018) analyzed 20 cases involving solicitors or chartered accountants convicted of money laundering in the United Kingdom. She found that 80 percent (16) were remunerated with fees that did not exceed the typical amount of money they would have received if the transaction had involved legitimate clients and funds. While this is surprising, it is not entirely unexpected. Charging significantly higher-than-usual fees could strengthen the evidence for prosecution by undermining the ignorance defense.

The market for money laundering services can be significantly influenced by policy measures that seem unrelated to laundering itself. This is illustrated by considering the consequences of the Chinese government’s 2015 imposition of new limits on the value of assets that its citizens could send overseas, equivalent to about $50,000 per annum.⁴⁹ China’s growing class of wealthy households, concerned about shifts in government attitude towards corruption and entrepreneurs’ initiatives, have sought ways to move more of their wealth beyond the reach of Chinese authorities. This has created a demand for money laundering services which, as it happens, match the needs of criminals in North America (and perhaps Western Europe). The scheme is only moderately complex. It involves PMLs transferring dollars from US drug dealers to Chinese individuals, who then transfer equivalent amounts to the drug traffickers’ accounts in Mexico via the purchase of Chinese goods exported to Mexico. This matching of the two groups allows for very low-cost laundering services for the drug dealers, reportedly as low as 2–5 percent (Langdale 2021). It is often asserted that Chinese money laundering organizations now dominate the laundering market in the United States (US Congress 2023).

Unfortunately, drawing meaningful conclusions from the available data on money laundering service prices is challenging. Research in this area has been limited, and no time series data are available to track changes over time. For instance, the only reference point for prices in the United States comes from the 2002 US National Money Laundering Strategy, which reported that prices for money laundering services ranged between 4 and 8 percent, with a high of 12 percent (US Department of the Treasury 2002). However, this figure is based on anecdotal evidence, as there is no public information on the number or types of cases underlying this estimate.

There is no compelling evidence to suggest that money laundering is on the decline or becoming more costly and challenging. New players entering the market may have driven prices down while enhancing service quality. For instance, the Black Market Peso Exchange has been a go-to method for Mexican and Colombian cartels for years, but it is a complex process that can take considerable time. In contrast, Chinese money laundering organizations now offer mirror transactions that provide pesos to Mexican cartels almost instantly, with minimal commission rates. This, along with the widespread use of low-sophistication schemes by criminals who self-launder, suggests that the current system is not highly effective. In the next section, we explore the potential reasons for this lack of efficacy.

A.  Defection by Countries from the FATF Rules

The FATF was formed in 1989 by the G7 when the United States was dealing with a severe drug problem, particularly associated with crack cocaine. FATF’s initial mandate reflected US concerns, though the United Kingdom and France were enthusiastic allies. These three countries have continued to be leaders in the effort to create a strong global AML system (Nance 2018). Despite their leadership, even these nations have not fully adhered to all of FATF’s 40 recommendations. As of March 2024, the United States is compliant on nine of the 40 recommendations and largely compliant on 23. It remains partially compliant on five and noncompliant on three.⁵⁰

For example, the United States has not required the disclosure of beneficial ownership information for establishing corporations (violating rec. 24) until recently. The 2016 MER for the United States, repeating what had appeared in the previous MER, stated: “The U.S. should introduce beneficial ownership preventive measures as soon as possible, continuing previous efforts to bring these regulations into force” (FATF 2016, p. 118). In its 2022 Financial Secrecy Index, the Tax Justice Network, a transparency-focused nongovernmental organization (NGO), rated the United States as the leading country globally in terms of financial secrecy.⁵¹ The US Department of the Treasury has been willing to admit these deficiencies in its 2022 National Risk Assessment:

Since 2018, the federal government has made significant improvements, prompting the FATF to upgrade its assessment of the US in 2020 from partially compliant to largely compliant on recommendation 10 (customer due diligence). More recently, in 2024, the US improved its rating on recommendation 24 (beneficial ownership) from “noncompliant” to “largely compliant,” thanks in large part to the enactment of the Corporate Transparency Act in 2021, which officially took effect on January 1, 2024.

However, significant implementation challenges remain. Since its enactment, the Corporate Transparency Act has faced litigation questioning its constitutionality. On December 3, 2024, the US District Court for the Eastern District of Texas issued a preliminary injunction, suspending enforcement of the Corporate Transparency Act and its implementing regulations nationwide. The court ruled that the act is unconstitutional, asserting that it exceeds Congress’s power.⁵² On December 7, 2024, FinCEN confirmed that reporting companies are not required to file beneficial ownership reports while the injunction remains in effect.⁵³

The FATF Recommendations also require that DNFBPs be subject to AML requirements (recs. 22 and 23); lawyers and accountants are among the designated professions. However, the 2016 MER noted that many professions in the United States, notably including lawyers, are not covered. Canada, another one of the FATF founding members, also failed to include lawyers or accountants in their AML system. In 2024 another FATF assessment of compliance with the DNFBP recommendations found that the United States was one of three countries (along with Australia and China) that scored zero (against an average of 74 percent across all FATF members) in that respect (FATF 2024).

Switzerland, though not a founding member, is one of the leading financial centers globally. Its 2016 MER contained criticisms for a variety of important technical deficiencies, such as a too high limit on cash transactions (>$25,000 compared to the recommended $10,000) and a failure to impose appropriate sanctions for violations (rec. 35). These simple examples show how the system may fail because the most systemically important nations do not impose the regulatory system that the FATF has designed.

The reason for failure to obtain full compliance in these leading countries is no mystery. It often lies in the domestic politics and the political structure of the country, rather than an explicit government decision to defect. In the United States, for example, corporations are chartered by individual states rather than the federal government, leading to a “race to the bottom” in which states compete to offer the most lenient regulatory standards to attract businesses for incorporation (Rakoff 2022). For a long time, the winner of that race was the state of Delaware, which imposed relatively few restrictions.⁵⁴ Delaware did not require disclosure of companies’ beneficial owners and the federal government could not directly intervene. Nevada is another state with lax beneficial ownership requirements and many shell companies are registered there (Jancsics 2018). With Joe Biden representing Delaware in the Senate from 1973 to 2009—serving much of that time as chair of the judiciary committee—before becoming the 47^th Vice President and later the 46^th President of the United States, and with Nevada’s Harry Reid as majority leader of the Senate from 2007 to 2015, there was little political momentum in Congress to address these issues until recently.

For developing countries with weak governments, the failure to adopt necessary laws and regulations, and to establish the required institutions, may have a more straightforward explanation: a significant shortage of skilled professionals. Even when there is political will to comply with international standards, the lack of qualified personnel to draft, enforce, and monitor these regulations makes it extremely challenging to meet the rigorous demands set by the FATF. The FATF recommendations—and the entire framework itself—were originally developed by the high-income countries that comprised its founding membership. As a result, these standards are better suited to the capacities and characteristics of wealthier nations. Alarmingly, AML standards and mutual evaluation processes seem to disregard the structural and resource limitations faced by lower-income countries (Maslen 2023; Littrell 2024). The FATF framework, driven by its early rhetoric that the system is only as strong as its weakest link (FATF 2014b), requires that all nations must meet the same obligations, regardless of their vastly different contexts and capacities.

Given these circumstances, one might expect more equitable representation in FATF’s policymaking processes. However, this is far from the reality. Thirty-five years after its founding, FATF membership remains heavily concentrated in higher-income countries (Jones and Knaack 2019). Low- and middle-income countries engage with FATF primarily through FSRBs, but their participation is often hampered by significant disparities in resources, capacity, and influence. While regional bodies provide some level of inclusion, the ability of lower-income nations to effectively navigate and shape these mechanisms is highly uneven, as is their capacity to wield meaningful influence (Moraes 2022).

As a result, the FATF policymaking remains largely dominated by the priorities of wealthier nations, leaving poorer countries with little say but full responsibility for compliance. Nowhere is this disparity more evident than in Africa, the world’s poorest continent. Of 40 FATF members, only South Africa holds a seat at the table, meaning the rest of the continent has no direct voice in shaping the policies they are bound to follow. Adding to this imbalance, South Africa itself was placed on the FATF grey list in February 2023. With the addition of Algeria, Angola, and Côte d’Ivoire in October 2024, African countries now make up 14 of 24 grey-listed jurisdictions (nearly 60 percent) as of December 2024.

This disparity is a significant issue. The benefits of implementing FATF standards in low-income contexts are unlikely to mirror those seen in high-income countries due to vastly different economic, institutional, and political realities. At the same time, the proportional costs of compliance are much higher for low-income countries, where governments are often forced to implement policies poorly suited to local circumstances. For instance, the financial burden of AML compliance disproportionately affects small businesses—far more prevalent in developing economies than in wealthier ones. Similarly, challenges such as de-risking and de-banking—where financial institutions sever ties with certain clients or jurisdictions to minimize risk—are more acute in poorer countries, creating significant barriers to financial inclusion.

The example of South Sudan starkly illustrates the disconnect between the FATF’s policies and on-the-ground realities in fragile states. Despite being embroiled in a civil war, South Sudan was placed on the grey list in June 2021 and remains there, unable to meet the FATF action plan due to its ongoing violence and humanitarian crisis. In such conditions, it is difficult to imagine who would choose to launder money through a country with such weak financial infrastructure, insecure property rights, and minimal global financial connectivity. Yet, the country continues to face the consequences of grey listing, including reduced access to international finance and reduction in development assistance, further crippling its already fragile economy (De Koker, Howell, and Morris 2023).

This raises a fundamental question: does an effective global AML system need all nations to adhere rigidly to the FATF model? In cases like South Sudan, the supposed risks of weak AML systems seem exaggerated, especially when the local financial ecosystem is so underdeveloped and disconnected from global laundering networks. By imposing uniform standards, the FATF not only disregards the vastly different capacities of low-income countries but also amplifies the costs and harms of compliance in contexts where benefits are highly unlikely to materialize.

To its credit, the FATF has started to address this issue, long emphasized by scholars (see, e.g., De Koker 2024). In April 2024, FATF Ministers stated: “recognising the significant number of low-capacity countries among those jurisdictions with strategic deficiencies, we commit to refining the risk-based criteria to identify the countries that pose a higher threat to the global financial system” (FATF 2024c, p. 4). This statement marked a shift, and in October 2024, the FATF announced revised criteria for grey listing, designed to alleviate pressure on the least developed countries while focusing efforts on jurisdictions that present greater risks to the international financial system.⁵⁵

Under the revised criteria, jurisdictions will be prioritized for active review if they meet the referral standards and are a FATF Member, a country on the World Bank High-Income Countries list (excluding those with financial sectors of two or fewer banks), or a country with financial sector assets exceeding USD 10 billion (measured by broad money). Least developed countries, as defined by the United Nations, will not be prioritized for review unless the FATF determines they pose a significant money laundering, terrorist financing, or proliferation financing risk. Additionally, least developed countries that enter the review process may be granted a longer observation period—up to two years—to address deficiencies and make progress against their “key recommended action roadmap.”

B.  Implementation Failure

Even if a nation complies with the 40 recommendations, enacts the necessary laws, and establishes the required institutions, the AML system can still fail. This is recognized by the FATF through the addition of the 11 Intermediate Outcome (IO) measures to the MERs. In terms of what draws the little public attention that money laundering controls receive, it is mostly failure of implementation. For example, in seizing laundered assets, though that is one of the most fundamental methods of deterring laundering, the results are on their face almost breathtakingly poor (Pol 2020).

For example, the United Kingdom received a highly effective rating for IO9 (confiscation of proceeds, instrumentalities, and property of equivalent value) in the 2018 MER—one of the few countries to achieve such a rating. However, performance in subsequent years suggests otherwise. Between 2018 and 2024, median annual seizures in the United Kingdom remained below 240 million pounds (239.4 million pounds), despite frequent claims that over 100 billion pounds is laundered through or within the United Kingdom, including UK-registered corporate structures, each year (National Crime Agency 2024). Comparing the financial year ending March 2019 to March 2024, there was a 3 percent increase in recovered receipts, from £235.6 million to £243.3 million. However, in comparison to the financial year ending March 2023, there was a 29 percent reduction in recoveries. This decline is largely due to a few high-value cases in the financial yearending March 2023: a civil recovery case (£53.9 million) and a confiscation case (£93.5 million). In contrast, there were far fewer high-value cases in the financial year ending March 2024.

Part of the problem is interpretation of the numbers. Enforcement statistics are inherently ambiguous, as noted above. A small number of convictions for laundering might mean the system is effective because launderers are deterred. Alternatively, it could mean that enforcement efforts are failing. Given the revelations about successful launderers, the conspicuously small number of money laundering convictions suggests that agencies are relatively unsuccessful. In the United States, with its huge illegal markets and its role as the destination of so much laundered money, the total number of money laundering convictions, with money laundering as the primary offense, has for many years been about 1,000 annually.⁵⁶ That confiscations are rarely more than 1 percent of not-very-sound estimates of annual laundering flows reinforces that sense of failure.

The difficult of proving specific intent in a money laundering case is also thought to be a barrier to effective enforcement. This became evident in the case of United States v. Millender in 2018. The United States District Court for the Eastern District of Virginia emphasized that, despite jury convictions, the government must establish elements such as intent, which can be challenging to prove. In this case, Terry Millender was convicted of money laundering with intent to promote the underlying fraudulent activity, while Brenda Millender’s convictions were overturned due to insufficient evidence of intent. The court found that the government failed to demonstrate that Mrs. Millender had specific knowledge or intent regarding the laundering of the proceeds, especially as they related to concealment under 18 U.S.C. § 1956(a)(1)(B)(i). Without clear evidence of intent, even convictions for money laundering may be reversed, reinforcing the challenges faced by prosecutors in establishing guilt beyond a reasonable doubt.⁵⁷

Moreover, in many cases, the sentence for the predicate crime—typically drug trafficking—is already sufficiently harsh, reducing the perceived need to pursue additional money laundering charges. This can contribute to limited resources being allocated toward pursuing money laundering convictions, weakening overall enforcement efforts.

Supervisory agencies are also important for effective implementation of the AML regime (FATF 2021). In particular, the powerful set of regulatory agencies around the banking system in most countries can increase pressure to take AML seriously (Gara et al. 2019), with guidelines being regularly published for this purpose (see, e.g., Chatain, Van Der Does De Willebois, and Bökkerink 2022). They can, for example, require a bank to hire more staff or cut its lending because of AML deficiencies. However, bank supervision is generally opaque; little is revealed about the results of bank inspections (Prescott 2008).⁵⁸ Exceptions show massive failures. For example, in January 2013, a consent order from JPMorgan’s main federal regulator, the Office of the Comptroller of the Currency, cited the bank for “critical deficiencies” in its AML controls, including inadequate procedures for monitoring transactions at foreign branches.⁵⁹

AML is, however, just one of many responsibilities of bank supervisors and one that fits awkwardly with their general responsibilities, since the regulator and the regulated have conflicting interests. While research on the topic is sparse (see, e.g., Demetriades and Vassileva 2020), there is general skepticism that bank regulators aggressively enforce bank AML responsibilities. This skepticism extends to the oversight of regulated professions, such as legal and accounting sectors. For instance, the United Kingdom’s Office for Professional Body Anti-Money Laundering Supervision, established in 2018 to oversee the country’s 22 professional supervisory bodies, reported a lack of full and consistent effectiveness across the nine supervisory bodies assessed (Office for Professional Body Anti-Money Laundering Supervision 2024).

C.  Financial Institution Noncompliance

Almost every major bank in the United States and Western Europe has paid a large fine, often in the billions, for violations of money laundering regulations. Table 4 shows the ten largest fines paid specifically for money laundering violations, as opposed to sanctions violations, based on multiple sources available online.⁶⁰ Recently, Binance, a virtual asset service provider, was subject to an extraordinarily large fine ($4.3 billion) for money laundering violations.⁶¹ It is important to note that the true cost of these fines is likely much higher than the fine amount itself, given the additional expenses associated with remediation efforts (e.g., investments for new AML personnel). For example, HSBC, following the scandal that led to a $1.9 billion fine in 2011, announced in 2017 that it employed 6,000 employees worldwide on AML compliance. Saperstein, Sant, and Ng (2015, p. 4) noted that “HSBC recently estimated it now devotes $750 million to $800 million per year on compliance—an amount equivalent to one quarter of the operating budget of its entire U.S. operations.”

These fines have not been for isolated violations or mistakes made by poorly monitored low-level officials. In many cases, the violations have involved the bank’s board of directors. The most egregious example was the 2018 revelation that Danske Bank’s Estonian branch had accepted over 200 billion euros in questionable Russian deposits over a ten-year period. Not only was the bank’s board aware of these violations by 2014, but in 2013 they promoted the head of its international division, which included the Estonian branch and was responsible for much of the bank’s profits, to CEO.⁶² Box 1 provides some illustrations of the kind of blatant behavior that major banks have managed to ignore.

One theory behind creating a control system reliant on private enforcement was the belief that banks would strive to uphold their reputation for integrity. According to this theory, banks would avoid money laundering to prevent damage to their reputation and subsequent loss in market value. Bankers often cite this as their motivation for compliance (De Souza and De Souza 2024). Yet studies of share prices after a bank is fined and sanctioned for money laundering show that the stock market is generally sublimely unconcerned that the bank has been caught; or, paradoxically, is positively influenced. For example, Köster and Pelster (2017) built a unique database of 671 financial penalties imposed on 68 international listed banks over the period 2007 to 2014. Overall, they found a positive relation between financial penalties and bank stock performance, suggesting that investors may be pleased that cases are closed, that the banks successfully manage the consequences of misconduct, and that the financial penalties imposed are smaller than the accrued economic gains from the banks’ misconduct.

Given that almost every major Western bank has been fined large sums for violations of money laundering controls or for transactions that aid nations trying to obtain weapons of mass destruction, it is hard to believe that reputation is much affected by news of such events. To resort to caricature for a moment, does anyone think that London clubs will expel CEOs of banks caught violating AML rules or that these members are even snubbed in the dining room? It is not entirely frivolous to suggest that not being caught for money laundering might actually hurt a bank’s reputation, but not in the usual sense. A bank that has never been caught laundering money might be perceived as excessively rule-compliant and insufficiently willing to take risks, potentially hurting its share prices.

The government’s efforts aim at making money laundering a losing proposition for banks. A common criticism, however, is that penalties—though significant in absolute terms—are relatively small compared to the overall size of the penalized institutions.⁶⁶ While this argument may not always hold up when the total costs of remediation are taken into account, another criticism rings true: senior managers at major banks rarely face criminal prosecution for their roles in these schemes, even though some have lost their jobs.

Consider the recent case of TD Bank, the tenth-largest bank in the United States. In October 2024, it was fined $3.09 billion for pervasive and systemic failures to maintain an adequate AML program, including failure to monitor approximately $18.3 trillion in suspicious transactions between 2014 and 2023. The severity of these violations even led to the imposition of a $434 billion asset cap to the bank as part of the penalty agreement. Yet, no TD Bank executives were criminally charged in connection with the AML compliance deficiencies.⁶⁷ This is particularly striking given that TD Bank intentionally incentivized a “flat cost” approach to compliance, leading to chronic underspending on AML efforts—a shortfall that was known at every level of the bank.⁶⁸

TD Bank is far from an anomaly. When HSBC was fined $1.9 billion in 2012, no executives faced criminal prosecution then either.⁶⁹ A 2016 report from the Office of Senator Elizabeth Warren reviewed 20 major criminal and civil cases of corporate misconduct settled in the United States in 2015, finding that the federal government “failed to require meaningful accountability from either large corporations or their executives involved in wrongdoing” (US Senate, Office of Senator Elizabeth Warren 2016, p. 1). This lack of accountability extends beyond the United States. For example, in December 2024, Dutch prosecutors dropped a case against the former chief executive of ING, citing insufficient evidence to establish criminal liability for the bank’s AML violations between 2014 and 2020.⁷⁰

Two key factors may partly explain this pattern. First, disentangling individual accountability within a vast corporate structure is difficult, particularly in determining whether any executives had the requisite knowledge and intent to be held liable.⁷¹ Second, prosecutors often bring cases only when they are confident of winning. The complexity of corporate prosecutions, combined with the Department of Justice’s limited resources, makes securing a conviction against a high-profile corporate executive a rare and uncertain endeavor.⁷²

Yet, the current over-reliance on fines raises questions about their effectiveness. There is little evidence that banks are making lasting improvements in response to these apparently large fines, as changes in the private sector often appear to be temporary and primarily driven by the immediate threat of regulatory action. As noted above, the largest known instance of money laundering was unearthed in 2018, well after the issuance of huge penalties on many other banks. Commonwealth Bank of Australia was caught even more recently in gross violations that led to a fine of A$ 700 million (ca. US$ 500 million) for gross negligence in allowing money to be laundered through its “intelligent” ATMs. The violations were obvious and continued over five years.⁷³ The enduring nature of the violations is a recurring characteristic in these cases, with failings persisting over long periods before being detected by authorities.

The fines are never for a single or even a small number of offenses; in every case it was shown that deliberate behaviors, sometimes by the Bank’s board of directors, resulted in frequent offenses. For example, to take a fairly mundane example, ABN Amro, one of the three largest banks in the Netherlands, in 2014 opened a bank account to a client who had been known within the bank to be involved in fraud since 1995. Over €2.2 million was transferred through the bank accounts of the client’s private limited liability companies and the transactions on the bank accounts of the companies did not fit the business activities as registered with the Chamber of Commerce. This was just one of various misbehaviors that led the Dutch bank regulator to impose a fine of 480 million Euros in 2021.⁷⁴

Not only have these banks paid massive fines, but they have also repeatedly asserted their commitment to AML processes and compliance—yet many of them are already recidivists. For example, following its US fine in 2012, HSBC faced another penalty of $88 million by the U.K.’s Financial Conduct Authority (FCA) in 2021 for “unacceptable failings” in its AML controls from 2010 to 2018.⁷⁵ In the same year, the Guardian newspaper published a joint investigation with the Bureau of Investigative Journalism, revealing that, in 2016, HSBC uncovered a suspected money laundering network which received $4.2bn (£3bn) in payments, but failed to disclose this information to US monitors who were overseeing its compliance reforms.⁷⁶

Most of the largest fines have been for violations of rules related to sanctions of specific countries rather than criminal money laundering. Major players in the international banking community were fined for processing illegal transactions with sanctioned countries, such as Cuba, Iran, Sudan, and Syria. For example, Credit Suisse settled for $536 million in 2009; Barclays agreed to pay $298 million in 2010.⁷⁷ These sanctions can be seen as an expression of US extra-territorial power. Not every European bank felt that it was morally obliged to comply with sanctions against, for example, Iran. That was, of course, no defense once the violations were detected. The fine was usually accompanied by a requirement to implement appropriate changes to ensure that such violations would not occur in the future. For example, following its $1.9 billion fine in 2012, HSBC committed to increase its compliance staff from a few hundred to several thousand by 2017 to increase the likelihood of detecting a money laundering violation. Regulators oversee obliged entities’ implementation of orders stemming from fines. In 2023, the Federal Reserve imposed a $186 million fine on Deutsche Bank for its tardiness in implementing the requested changes to its AML controls.

Fines represent just one tool among many available to regulators for addressing AML deficiencies in obliged entities. For example, in 2022, the Bank of Italy banned N26, a Deutsche bank from recruiting new customers after an on-site inspection revealed inadequate AML controls.⁷⁸ Moreover, regulators have the authority to restrict obliged entities from engaging in acquisitions and other business activities due to AML deficiencies, as evidenced by the case of Julius Baer in Switzerland in 2020. The Swiss Financial Market Supervisory Authority barred Julius Baer from pursuing large and complex acquisitions in response to significant AML failures between 2009 and early 2018, linked to alleged corruption involving PDVSA, the Venezuelan state oil company, and FIFA, the global soccer federation. These measures, akin to fines, aim to encourage compliance among obliged entities by targeting their primary business operations and profit-making endeavors.

It is hard to avoid a cynical interpretation of the record of bank violations. Banks have shown blatant disregard for the goals of AML, routinely violating rules when it serves their financial interests. Table 4 shows that six out of the ten largest AML fines were imposed in 2019 or later, providing little evidence that systematic violations are becoming less common. In 2023, ComplyAdvantage surveyed 800 C-suite and senior compliance decision makers across North America, Europe, and Asia Pacific, asking if they regularly consider the risk of or choose to incur AML fines and violations with respect to their business decisions and compliance investment. Overall, 79 percent of them “choose to incur AML fines and make violations ‘all the time’” (ComplyAdvantage 2023, p. 95).

However, fairness demands that we recognize the broader context in which these failures occur. Banks operate within a regulated environment, and their compliance shortcomings often reflect regulatory failures as well. When the private sector repeatedly flouts AML rules, it is reasonable to conclude that relevant government agencies have also fallen short in enforcing compliance or deterring misconduct. Weak oversight, insufficient resources, and limited accountability for executives all contribute to an environment in which AML violations persist. After all, a compliant private sector can only be expected when relevant government agencies adequately supervise, monitor, and regulate financial institutions, as emphasized by FATF’s IO3.

The lack of incentives for banks to treat AML compliance as a core responsibility is perhaps the most fundamental flaw in the current regime. Many financial institutions view AML obligations as a burdensome cost they would gladly eliminate if given the opportunity. Third-party policing is unlikely to succeed when the third party is either disengaged or, worse, benefits from enforcement failures. We turn in the next section to review costs and errors of the current AML system.

1.  Costs to Governments

Several costs can be included in this category, such as operating the Financial Intelligence Unit (FIU) (e.g., handling incoming SARs and related required reports), conducting money laundering investigations, and carrying out associated prosecutions and incarcerations. However, if money laundering charges are ancillary to charges for predicate crimes and the predicate crime prosecutions are possible only because of SARs, as discussed above, then AML may actually reduce the cost of enforcement against predicate crimes. To our knowledge, there are no serious estimates of what governments spend on AML. Reuter and Truman (2004) offer an estimate of $7 billion for the US government in 2003. So much has changed since then that the estimate is only of historic interest. Ferwerda (2018) attempted to obtain estimates for EU member countries in 2016, but almost no government had any estimates to report. The cost-benefit analyses of specific AML rules in the United States, required by law for all major federal regulations, are good faith efforts to assess costs but are very narrow.⁷⁹

However, the costs of an AML system do not stop at setup and maintenance. As we have discussed in previous sections, falling short of international standards is more than just a misstep—it is a serious risk with heavy consequences. A noncompliant country risks being included on the FATF blacklist or grey list, potentially facing reputational and financial costs, at least from a theoretical standpoint (Riccardi 2022). Over the past decade, these policy instruments have sparked increasing attention in academic circles (see, e.g., Eggenberger 2018; Morse 2019; De Koker, Howell, and Morris 2023). However, evidence on the effectiveness of AML blacklists remains mixed and inconclusive (Case-Ruchala and Nance 2024).

For example, blacklisting prior to 2010 had no apparent effect on the exchange rate between the US dollar and the currency of listed jurisdictions (Nance 2015). Balakina et al. (2017) tested whether being included and later excluded from the FATF blacklist is an effective measure that influences countries’ cross-border capital flows. Using annual panel data for the period 1996-2014 for 126 countries, the authors did not find any evidence that a stigma effect exists. In contrast, Morse (2019), using quarterly data from the Bank for International Settlements, found that FATF listing results in a 14 percent decrease in the country’s bank liabilities. However, Case-Ruchala and Nance (2024) replicated Morse’s study with a different model specification and found no association between blacklisting and financial harm.

Regarding grey listing, the IMF, using a sample of 89 emerging and developing countries, found that being grey listed results in an average 7.9 percent decline in foreign direct investment (Kida and Paetzold 2021). De Koker et al. (2023) employed a pooled cross-section and time series approach with fixed effects, analyzing data from 177 countries across 3540 country-years (2000–2020). Their findings indicate that grey listing significantly reduces development assistance during the grey listing period, with these reductions persisting even after a country is delisted. However, the authors caution that these correlations do not establish causation and may instead reflect co-determination rather than direct causal relationships.

2.  Costs to Obliged Institutions

Many of these institutions have complex internal surveillance systems for on-boarding new customers, monitoring transactions, and reporting suspicious activities. AML responsibilities may be little more than additional programming to those systems. Penalties for AML violations are also a cost to corporations, though in this case a benefit to the government. While it is easy to identify theoretical second order effects, such as reputational cost when a bank is caught for failing to fulfill its AML obligations, or board time spent on AML issues rather than improving the financial performance of the firm, these are likely to contribute little to the total cost.

Very high estimates of compliance costs occasionally capture headlines. For example, an organization called the Financial Crime Academy claimed—without providing the basis for the figure—that global spending on AML compliance by financial institutions alone was estimated to be more than $100 billion in 2019 (Financial Crime Academy 2019). LexisNexis Risk Solutions, a much more prominent organization, estimated that banks in the United States and Canada spent $61 billion on AML compliance in 2023 (LexisNexis Risks Solutions 2024).

These large numbers are consistent with some other fragments of data. In 2018, the Bank Policy Institute (2018) surveyed 19 US banks with asset sizes ranging from approximately $50 billion to over $500 billion to better understand the resources they devote to Bank Secrecy Act (BSA) compliance. Overall, respondents reported that they employ over 14,000 individuals, investing approximately $2.4 billion and using as many as 20 different IT systems per institution to assist them with compliance obligations.

Resources allocated to AML efforts are by no means secondary, especially to regulators. In December 2021, for instance, FinCEN fined the CommunityBank of Texas $8 million for BSA violations that occurred between 2015 and 2019. FinCEN deemed the bank’s allocation of resources to AML efforts insufficient, as the AML office was understaffed. The three BSA analysts responsible for reviewing case reports (of eight analysts) handled an average of 100 case alerts per day, frequently overlooking supporting documents (e.g., cash deposit slips, wire transfers, check images), even though this information was readily accessible.⁸⁰

Of all AML obligations, SARs are often regarded as having the most significant impact on the overall compliance costs for obliged entities. FinCEN recently estimated that it takes 1.98 hours to file each SAR as part of its effort to renew the form used by US financial institutions for reporting under the Paperwork Reduction Act. However, in response to FinCEN’s notice, several trade groups referred to a survey of 15 major US banks with assets exceeding $100 billion, which found that banks actually spend 21.41 hours per SAR filed.⁸¹ With an estimated hourly labor cost of $50, this translates to approximately $1,070 per SAR. Given that around 4 million SARs are filed annually, the total cost amounts to roughly $4.28 billion. While this figure is substantial, it must be weighed against the critical role SARs play in supporting law enforcement efforts, as discussed in Section IV.

However, all these figures come from interested parties; showing high AML costs helps make the case against more stringent regulation. The US General Accountability Office has conducted a study of AML costs in a small “non-generalizable sample” of 11 banks. It found that AML “costs comprised about 2 percent of the operating expenses for each of the three smallest banks in 2018 but less than 1 percent for each of the three largest banks in US General Accountability Office’s review” (US General Accountability Office 2020, p. 3). That suggests a smaller figure for total AML in US banks, perhaps around $10 billion in 2019.

Recently, New Zealand has provided a unique case study to collect some evidence on AML compliance costs to obliged entities. The Phase I of the New Zealand’s Anti-Money Laundering and Countering Financing of Terrorism Act 2009 came into force in 2013, and covered banks, insurance companies, financial services (e.g., investment advisers), money remitters, and casinos. Phase 2 aims at extending the scope of the AML legislation in the country to cover additional sectors, thus bringing New Zealand in line with its FATF obligations. To support the roll-out of this phase, a cost-benefit analysis was carried out in 2017. Overall, it is estimated that the cost to new regulated business sectors could be in the order of $0.8 to $1.1 billion over ten years (Ministry of Justice 2017).

AML costs are not fixed but tend to increase periodically due to evolving regulatory requirements and the ongoing need to update and enhance monitoring systems. According to a recent survey conducted by PricewaterhouseCoopers (2024), which involved 396 financial institutions (with 52 percent being banks) across 40 countries in the EMEA region (Europe, Middle East and Africa), more than half of the respondents (51 percent) reported that their AML compliance costs had risen by over 10 percent in the last two years. Among these, banks saw the largest increase, with 62 percent experiencing significant rises. On average, AML costs have gone up by 14 percent in the region. Key factors contributing to these increases include additional staffing and investments in new digital tools aimed at improving compliance measures.

Ironically, regulators seem to treat the hefty expenses borne by obliged entities as a sign of success in combating money laundering.⁸² The logic appears to be that the more obliged entities spend on AML efforts, such as hiring compliance officers or deploying new monitoring systems, the more effectively they are tackling financial crime. The above-mentioned fine imposed on CommunityBank in Texas for insufficient AML resources illustrates this dynamic, with similar examples found elsewhere. For instance, UBS Financial Inc. was fined in March 2018 for failing to provide “adequate resources to the AML Compliance Officer” (FinCEN 2018, p. 8), while the US National Bank Association was penalized in January 2020 for employing “a woefully inadequate number of AML investigators” (FinCEN 2020, p. 5). The assumption that more money spent equals better compliance ignores whether these efforts lead to meaningful outcomes in reducing money laundering and financial crime. This raises a critical question: are institutions merely throwing money at the problem to avoid fines, or are they genuinely working to curb money laundering? Without proper scrutiny of outcomes, the AML framework risks becoming a costly compliance exercise with little real impact on criminal activities.

3.  Costs to the General Public and Corporations

Financial institutions may shift their AML costs to customers through charges for banking services. For example, at the end of December 2022, the four biggest Dutch banks introduced a new banking fee for their corporate customers to cover AML compliance costs (DutchNews 2022). Though banks sometimes impose explicit charges for transactions (e.g., international transfers) that require more stringent AML review, the charges are more often buried. More important still are the costs of false positives i.e., the excluding of customers from access to their accounts (temporarily) and from the institution (longer term) as a result of risk aversion by the institution. Both the transfer of costs and the errors can affect marginalized communities most heavily.

Stories of unexplained and apparently unjustified closings of accounts, probably for money laundering concerns, is now a staple of financial reporters. For example, the New York Times in 2023 provided detailed reporting on three customers who had suddenly found their accounts frozen by major banks and could not get explanations.⁸³ After HSBC was sanctioned for the many violations of its recently acquired Mexican affiliate, many US banks expelled business customers on the south side of the border, even those with longstanding banking relationships and no indication of being used for laundering. There are not even estimates of the frequency with which customers are rejected, let alone of how costly that is to them in terms of reputation and obtaining an account with a different bank.

Nor is this just a US phenomenon. Debanking, as it is called, is a problem in the United Kingdom as well. In 2021–22, UK banks closed 343,000 accounts, up from 45,000 in 2017 when the Money Laundering, Terrorist Financing and Transfer of Funds (Information on the Payer) Regulations 2017 came into effect. In about half of these cases, the closures were due to the bank’s inability to verify, under the standards established by the new law, that the customer was not involved in money laundering or terrorist financing (Whyte 2024). Likewise, eight major UK banks closed more than 140,000 business accounts belonging to small and medium enterprises in 2023.⁸⁴

An associated but distinct problem is what is called “derisking,” when banks drop classes of customers, rather than making individual risk decisions. This has attracted a great deal of attention in the context of transfers of remittances to developing countries with strong terrorist groups. Somalia, for which remittances provide a major source of income (one third of GDP by one estimate) and for which the United Kingdom houses a large share of expatriate workers, was affected in the mid-2010s when all the major British banks refused to provide services for the Money Transfer Organizations (MTOs) that were used to send money to relatives still resident in Somalia. The British banks argued that the business risk (possible leakage to a terrorist organization) was too great, given the opacity of the Somali financial system. The problem also manifested itself in the United States and other countries (Ramachandran, Collin, and Juden 2018).

That soon afterwards cascaded into a broader international problem. Major Western banks, particularly in the United Kingdom and the United States, started cutting the number of foreign banks for which they provided “correspondent banking” relationships. One estimate is that the number of correspondent banking relationships fell by 22 percent between 2011 and 2019, a period in which the international finance system was expanding.⁸⁵ Banks that had been heavily fined for AML failures were particularly likely to cut their correspondent banking. For example, following its involvement in the Danske Bank scandal, Deutsche Bank reduced its correspondent banking portfolio by 40 percent and the portfolio of potentially high-risk correspondent banking customers by 60 percent. The bank has cut its active correspondent banking ties in Russia by about 75 percent since 2016, and it has stopped doing business in Moldova, Estonia, and Latvia.⁸⁶

Bank regulators claim to be concerned that banks focus on the specifics of the customer and its transactions rather than use a broad associated characteristic to exclude classes of customers. However, given the cost of scrutinizing each customer in the suspect class individually and the high costs of failing to detect laundered money or money that might be funding terrorist activities, banks have strong incentives to make broad categorical decisions. Recent advice from the US Department of the Treasury on this matter was very nuanced; it pointed to the problem created by categorical derisking but also stressed the banks’ AML obligations.

There is, however, a missing element to these stories that creates a doubt as to whether these decisions by banks are a major problem. In none of them does it turn out that the affected party is without resource, the ability to work around the problem. The customers whose problems were highlighted by the New York Times or other major news outlets always seem to have found another financial institution with which to bank; they may have been inconvenienced but no more than that. Perhaps that is simply a sample selection bias; anyone who is able to reach the New York Times has already shown a certain resourcefulness. The low-earning Pakistani immigrant in the United States who is told by her bank that she can no longer bank there, with no explanation, may be much more seriously harmed.

It is also true, though, that one response, illustrated by the refusal of the United Kingdom banks to handle Somali MTOs, is to work around the system. Somali MTOs made use of the hawala system, a highly efficient method of moving money internationally. It is unregulated and of course lacks the protections offered by the highly regulated formal banks. However, there is little evidence of hawala operators defrauding customers or facing insolvency. These systems often operate within strong cultural frameworks that impose informal but robust obligations to protect clients. As a result, the harm caused by debanking and derisking may be less severe than initially appears, as the informal financial sector often fills the void effectively, albeit outside formal regulatory structures.

However, this simple example highlights another unintended consequence. As noted by the US Department of the Treasury (2023), derisking can be detrimental to AML efforts. AML controls work best when illicit funds enter and remain within the regulated financial system. As the cost and complexity of using the regulated financial system increase, more individuals who cannot access it may resort to unregulated financial channels for receiving and moving money. As a result, derisking can inadvertently drive illicit activities towards unregulated financial services, making it more challenging to detect and prevent the flow of illegal proceeds.

Last but not least, the primary outcome of the wide array of AML controls is, fundamentally, information on clients and transactions that, although collected for a legitimate purpose, can be misused. For example, two well-publicized federal criminal cases in the United States demonstrate the potential misuse of SARs confidential information. In August 2019, an IRS analyst pleaded guilty to making unauthorized disclosure of five SARs to an attorney and was sentenced to five years of supervised probation.⁹³ Likewise, in June 2021, a former Senior FinCEN employee was sentenced to six months in federal prison for unlawfully disclosing SARs information to a reporter.⁹⁴

In addition to wrongdoing by public servants, AML information and regulations can be—even more worrisome—abused for surveillance and suppressing political dissent by governments (Reimer 2024). In January 2024, Ranking Member Tim Scott, the top Republican on the Senate Committee on Banking, Housing, and Urban Affairs, sent a letter to the US Treasury Secretary and FinCEN Director to request information on allegations that FinCEN instructed financial institutions to surveil their customers’ transactions using specific search terms such as “MAGA”, and “Trump” as well as transactions for the legal purchase of firearms.⁹⁵ In the official response, the US Treasury Department acknowledged that, in the wake of the violent assault on the US Capitol on January 6, 2021, the FinCEN convened a series of FinCEN Exchange events with government and private sector representatives. A document distributed on January 15, 2021 suggested that banks could review payment messages for indications that an individual participated in the assault on the Capitol and included terms such as “antifa,” “MAGA,” “Trump,” “Biden,” “Kamala,” “Schumer,” and “Pelosi,” along with terms indicating an intent to do violence, such as “shoot,” “kill,” “murder” and “storm the Capitol”.⁹⁶ However, FinCEN highlighted that words or transaction codes included in these typologies were not designed to be used in isolation to “flag” or “target” any individuals but rather to be used alongside other factors and data obliged entities regularly use in carrying out their AML duties.⁹⁷

Authoritarian governments have found the FATF rules useful for suppressing dissent. Reimer (2024) provided an overview of how FATF rules can be weaponized for malicious purposes. For example, an NGO that is critical of the government and is in receipt of foreign philanthropic support can be easily monitored or accused of money laundering. An instance was discovered in Uganda in 2019 when the government, through the country’s FIU, asked the Equity Bank Uganda to disclose financial details (i.e., account opening documents, bank statements for the last three years) of 13 NGOs that have been critical of the government’s political and human rights records.⁹⁸ According to the government, the NGOs were under-declaring their funds and were spending more money on what it called “subversive” activities, including money laundering.

This troubling trend is not limited to developing nations. For example, in July 2020, Serbia’s FIU, citing the new Law on the Prevention of Money Laundering and the Financing of Terrorism, demanded sensitive financial data from banks regarding dozens of organizations, including approximately 50 NGOs and media outlets critical of President Aleksandar Vučić.⁹⁹ The move provoked strong backlash from UN experts¹⁰⁰ and hundreds of NGOs,¹⁰¹ who raised concerns about the chilling effect on civil society and freedom of expression. By July 2021, even the Egmont Group raised concerns, reiterating that FIUs must avoid abusing their powers for political purposes.¹⁰²

A.  Reform

We have found no documents that seriously outline major reform proposals. One could argue that this is, at least in part, due to the widespread belief that the likelihood of significant reform is so low that meaningful change becomes difficult to envision. This pessimism may function as a self-fulfilling prophecy, with the perception of reform being unattainable discouraging efforts to propose or pursue it. As this mindset spreads, it further diminishes the possibility of reform, reinforcing the idea that change is not worth attempting, and thereby perpetuating the status quo.

However, attempts at reform have not been entirely absent. The Clearing House (2017), a banking association and payments company owned by the 22 largest commercial banks globally, came close to articulating an alternative vision in its 2017 assessment of the US AML framework. Its critique was scathing. For example, it argued: “Financial institutions devote vast resources to activities that could easily be performed centrally by government or some other party or not at all – for example, constant monitoring of media for adverse stories about customers, or multiple firms engaging in customer due diligence on the same customers” (p. 7). It also criticized the supervisory process as stultifying and rigid: “Although financial institutions have been developing innovative methods for identifying criminal behavior, they face regulatory criticism for taking unconventional or innovative actions that seemingly deviate from policy and may not be readily auditable” (p. 7). Furthermore, it highlighted a lack of clarity in goals and the absence of clear priorities as key issues.

To address the shortcomings of the US regime, the Clearing House identified seven areas for immediate reform, namely:

The Clearing House’s policy recommendations focus on refining the existing system. It is perhaps unsurprising that an industry group would not propose an entirely new approach.

The AML system aims to cast a wide net, striving to block every conceivable method of laundering money across the globe. The rhetoric behind this effort is clear: “The chain we build is only as strong as its weakest link” (FATF 2014b, p. 1). Yet, the drive to tighten regulations often collides with the ongoing push for financial innovation and efficiency. As the world becomes more interconnected and financial transactions more instantaneous, the tension between securing the financial system and facilitating legitimate global commerce becomes more pronounced. In many ways, AML is waging a battle against the relentless tide of globalization.

Given these challenges, it may be necessary to set more realistic goals for AML, scaling back controls and cutting costs where possible. Without questioning its overall mission, it is not hard to imagine a different approach. For instance, what if the responsibility for KYC checks was shifted from individual banks to a central government agency—much like issuing a passport? This approach could establish a consistent standard and create a centralized database of high-risk individuals, although it would also raise significant privacy concerns.

As criminologists, we do not have the technical expertise to design a new AML framework from scratch, but it is evident that reform is urgently needed. The disproportionate impact on disadvantaged communities, alongside the systemic inefficiencies of the current system, underscore the need for a comprehensive reevaluation of its structure and goals. This is not just a matter of regulatory fine-tuning; it is an opportunity for criminologists to engage in a wider conversation about justice, equity, and effective crime control.

B.  Research

A clear sign of the limited research focus on money laundering and its control is the existence of only a single regular conference on the topic. Since 2020, the Central Bank of the Bahamas has sponsored a two-day conference dedicated to empirical research on money laundering.¹⁰⁶ Despite its significant implications for policy and research, money laundering has so far remained off the radar for criminologists and other social scientists, with only a handful of exceptions discussed in Section V. Yet, this is a critical area of study that criminologists, economists and political scientists can no longer afford to ignore, given its significant implications for both research and policy. While skepticism about the current AML system’s effectiveness is warranted, exploring how money laundering controls can be leveraged to reduce these crimes is undeniably important. Section IV, for example, demonstrated how SARs—central to AML—have enabled law enforcement agencies to build databases that facilitate a high percentage of investigations, particularly in the United States.

A major obstacle to advancing research on money laundering is the lack of robust and accessible data. Unlike other criminological areas of interest such as terrorism and extremism (LaFree 2022), there are no easily accessible, and comprehensive, data sources available on money laundering. No credible estimates exist for its scale, either nationally or globally. The existing monitoring system generates very noisy data; vast numbers of false negatives and a fair number of false positives. The real problem? We seldom know whether a report represents a false alarm, a missed red flag, or an incident that will remain undiscovered.

This scarcity of reliable data makes conducting quantitative research especially challenging, deterring many criminologists who favor data-driven approaches from engaging with the topic; economists, who play an increasing role in the study of crime, are even more averse to working on topics without quantitative data sources. Those few scholars who pursue research in this area often rely on case studies or painstakingly gather their own limited datasets to navigate these data constraints (see, e.g., Kruisbergen, Kleemans, and Kouwenberg 2015; Farfán-Méndez 2019; Matanky-Becker and Cockbain 2021; Riccardi and Reuter 2024). Given the resource limitations faced by individual researchers and small teams, it is not surprising that much of the existing empirical evidence on money laundering is anecdotal and descriptive.

Despite these challenges, the lack of existing research presents social scientists with a fertile ground for exploration. Many critical questions remain unanswered. For instance, while the global AML system is widely seen as essential for combating financial and political instability, its purported benefits have yet to be empirically tested (Gerbrands et al. 2022).

Moreover, money laundering, with its multifaceted nature, is exceptionally well-suited for interdisciplinary research. Economists and political scientists have laid a foundational understanding of its economic and governance dimensions (Ferwerda 2018; Nance 2018; Nielson and Sharman 2022). However, a comprehensive understanding requires collaborate across disciplines, including criminology, breaking down silos to build on existing knowledge.

We single out criminology, while encouraging other disciplines as well. Reforming AML aligns closely with the core principles of criminology: advancing social justice, addressing inequality, and examining the broader societal implications of crime control policies. Criminologists are uniquely equipped to pose critical questions: Are current AML measures genuinely effective in disrupting criminal networks? How do socio-economic conditions exacerbate the vulnerabilities of marginalized populations to AML enforcement? What criminological insights can inform the design of more equitable and impactful regulatory frameworks? A largely dysfunctional regulatory system offers numerous opportunities for constructive research.