Improving Anti-Money Laundering Models with Synthetic Data

As readers of this blog are well aware, an effective anti-money laundering (AML) regime is crucial for fighting grand corruption, as well as other organized criminal activity. A key part of the AML system is the requirement that banks and other financial institutions identify suspicious transactions and file so-called suspicious activity reports (SARs) with the appropriate government agencies. This is an enormous task, given the volume of financial transactions that banks need to monitor and the challenge of identifying which of those transactions ought to be considered suspicious. Banks spend billions on AML compliance every year, and have developed complex automated systems to assist them in flagging suspect transactions, but existing systems’ ability to efficiently sort suspicious from innocent transactions is limited by the sheer complexity of the task. (False positive rates with current systems, for example, frequently top 90%.)

Many believe that artificial intelligence (AI) systems, such as those employing machine learning (ML), hold enormous promise for improving AML compliance and reducing cost. ML algorithms scrutinize vast datasets to identify patterns that can be used to fashion predictive models. In the AML context, ML algorithms identify those transaction characteristics (or complex combinations of transaction characteristics) that are associated with money laundering, and use these patterns to more efficiently and effectively identify suspicious transactions.  

But some commentators have suggested reasons for skepticism, or at least caution. For example, Mayze Teitler recently wrote on this blog about a number of challenges to operationalizing AI-derived algorithms in the AML context, primarily those arising from limitations in the data on which those algorithms are based. As Mayze correctly pointed out, ML algorithms require vast datasets from which to learn, and the data demands are compounded by the relatively rarity of known money laundering cases in the existing datasets.

Despite these concerns, I am more bullish than Mayze regarding the promise of AI-based AML systems. Many of the challenges and concerns regarding the development of effective AI systems in the AML context can be overcome through the use of synthetic data.

Continue reading

ML for AML: Is Artificial Intelligence Up to the Task of Anti-Money Laundering Compliance?

Fighting corruption—especially grand corruption—requires effective anti-money laundering (AML) systems capable of efficiently and correctly flagging suspicious transactions. The financial institutions responsible for identifying and reporting suspicious transactions employ automated systems that identify transactions that involve certain red flags—characteristics like transaction amount, location, or deviation from a customer’s typical activity; when the automated system flags a transaction, this triggers further review. But—given the ever-increasing volume and complexity of financial transactions that occur each day, as well as the increasing sophistication of kleptocrats, criminal groups, and others in disguising their illicit activities to avoid the usual red flags—picking out the genuinely suspicious transactions can be extraordinarily difficult. Even the cleverest compliance system designer couldn’t hope to incorporate every potential red flag into the automated system.

The need to stay one step ahead of the bad actors has fueled greater interest in how new advances in data processing technology may help make automated suspicious transaction detection systems more effective. Techno-enthusiasts are particularly interested in deploying deep learning artificial intelligence (AI), as well as classic algorithms that fall under the machine learning (ML) umbrella, in the AML context. ML and AI systems extract patterns from training datasets, and “learn” (by induction) what data patterns are associated with particular identifiable categorizations. Email spam filters provide a simple example. A spam filter, which can be created to conduct a process known as classification, sorts input variables into two categories: “spam” and “not spam.” It makes its categorization based on individual characteristics of the emails (such as the sender, body text, etc.). In the AML context, the idea would be to train an algorithm with data on financial transactions, so that the system “learns” to identify suspicious transactions even in cases that might lack the usual red flags that a human designer would program into an automated system. Advocates hope that ML/AI systems could be used both to filter out the false positives (transactions which are flagged as suspicious but turn out, on review, not to raise any concerns—an estimated 99% of all flagged transactions), while also identifying unusual, potentially fraudulent behavior that may be overlooked by human regulators (false negatives). Indeed, industry experts are understandably enthusiastic about AI systems that will cut costs while improving accuracy, and proponents claim that “AI holds the keys to a more efficient and transparent AML stance[,]” urging that “[b]anks must take hold of this new [AML] weapon[.]”

To the extent that AI tools can improve upon the admittedly-clunky automated systems currently in use, it could be a step forward. But ML/AI systems have a less than stellar track record in other contexts, and a model targeted at AML compliance presents some unique challenges.

Continue reading

Guest Post: Every Bank Robber Needs A Getaway Car; Banker Held Accountable For Money Laundering

GAB is pleased to publish this analysis by Emile J. M. Van Der Does De Willebois, Coordinator of the World Bank/UNODC Stolen Asset Recovery Initiative, of the significance of a decision of the Gerechtshof Den Haag, the Dutch appeals court in The Hague. As he explains, for too long authorities in the developed world have ignored the role lawyers, bankers, and other “enablers” play in facilitating corruption in the developing world.  Let us hope that the court’s decision marks a turning point in holding them accountable for their role in corruption crimes.  

Last month, a Dutch appeals court ordered the public prosecutor to initiate the criminal prosecution of the former CEO of the nation’s largest bank. The court directed that Ralph Hamers be put on trial for money laundering and other crimes the Amsterdam-based banking giant ING committed during his sevenyear tenure as its chief executive. Financial and legal professionals are rarely prosecuted for crimes they facilitate, and it is even rarer that senior executives, as opposed to the institution they run, are targeted. Until this decision, the indictment of Goldman Sachs bankers for their role in the 1MDB scandal was a notable exception.

The culpability of those who, like the driver in a bank robbery, facilitate a crime is not particularly controversial. We all know that the corruption that happens “over there” needs the services of bankers, lawyers, accountants and other facilitators “over here.” We like to pay lip service to the idea that “it takes two to tango” and acknowledge, at least verbally, that the financial and corporate services in the financial centers of the developed world facilitate the corruption found in large parts of the developing world.

But whether those working on anti-corruption always act upon that notion is another matter. A quick look at the Transparency International corruption perceptions index helps maintain the illusion that the rich developed world is doing well on corruption, and that, looking at the bottom of the table, corruption is really a developing-country problem. We have not really internalized the lessons of the Panama Papers, 1MDB, Danske Bank and, most recently, the FinCEN files, which shone a spotlight on the services provided by banks, lawyers and other professionals in making corruption possible.

Continue reading

It’s Not Just the Corporate Transparency Act: Other Reasons To Welcome the Passage of the U.S. NDAA

Last week I posted about the Corporate Transparency Act (CTA), the new law requiring companies to provide the government with information about their ultimate beneficial owners. The CTA, which was passed (over President Trump’s veto) as part of the National Defense Authorization Act (NDAA), has been getting a lot of attention in the anticorruption and anti-money laundering (AML) community, and rightly so. The product of decades of tireless and shrewd advocacy, the CTA—despite its limitations and imperfections—will make it substantially harder for kleptocrats, terrorists, organized crime groups, and others to abuse corporate structures to facilitate their crimes and hide their loot. But the CTA is not the only part of the NDAA that may have a substantial positive impact on the fight against corruption and money laundering. And while it’s entirely understandable that most of the attention (and celebration) in the anticorruption community has focused on the CTA, I wanted to use today’s post to highlight several other provisions in the NDAA that may also prove important in combating corruption and money laundering. Continue reading

A Few Thoughts on the Passage of the U.S. Corporate Transparency Act

[Note: I drafted the post below earlier this week, before yesterday’s shocking events in the U.S. Capitol. I mention this only because it might otherwise seem odd, and perhaps a bit tone-deaf, to publish a commentary on new corporate transparency rules when we just saw an attempted insurrection incited by the siting U.S. President. I don’t really have anything to say about the latter events (at least nothing that others haven’t already said), so I decided to go ahead and publish the post I planned to publish today anyway.]

Last week, as I suspect many readers of this blog are well aware, the United States Congress enacted one of the most significant anticorruption/anti-money laundering (AML) reforms in a generation. The Corporate Transparency Act (CTA), which was incorporated as part of the annual National Defense Authorization Act (NDAA), will require—for the first time in the United States—that corporations, limited liability companies, and similar entities will have to provide the U.S. government (specifically, the Treasury Department’s Financial Crimes Enforcement Network (FinCEN)) with the identities of the ultimate beneficial owners of those entities. That beneficial ownership information, though not made publicly available, will be provided to law enforcement agencies, as well as to financial institutions conducting due diligence (with customer consent). This reform will make it substantially harder for kleptocrats and their cronies—as well as other criminals, including human traffickers and terrorists—to conceal and launder their assets in the United States through anonymous shell companies, and will make it substantially easier for law enforcement to “follow the money” when investigating possible criminal activity.

This important reform has already gotten a ton of coverage in the anticorruption/AML community (see here, here, here, and here), as well as the mainstream media (see here, here, here, and here), though mainstream coverage has understandably been overshadowed by both the coronavirus pandemic and President Trump’s attempts to subvert the recent election. And we’ve had quite a bit of discussion of the issue on GAB prior to the passage of the NDAA (see, for example, here, here, here, here, and here). So, I’m not sure I really have that much to add to what others have already said. Nevertheless, it felt strange to allow this landmark event to go entirely undiscussed on GAB, so at the risk of self-indulgence, I’d like to throw out a few additional thoughts and observations related to the CTA. Continue reading

Combating Money Laundering in Africa: John Hatchard’s Latest Guide for African Corruption Fighters

The war on corruption is being fought on many fronts. One where victory is especially critical is the battle to prevent leaders of poor countries from robbing their citizens blind, and nowhere will a victory be more welcome or more hard-fought than in Africa.   Seventy percent of the world’s poor live on the continent while, thanks first to colonialism and then to Cold War machinations, Africans are saddled with governments ill-equipped to keep greedy leaders in check.  Courts, legislatures, and other accountability institutions are weak; the media and civil society hobbled by repressive, non-democratic measures.

Not that in recent years there have not been promising developments. South Africa’s once powerful leader Jacob Zuma was forced to resign the presidency over corruption allegations for which he is now on trial.  Former Guinea Minister of Mines Mahmoud Thiam forfeited $8.5 million and was sentenced to seven years in prison for corruptly granting virtually the whole of his nation’s mineral sector to a Chinese conglomerate.  The son of former Mozambique President Armando Guebuza is one of over a dozen members of the country’s ruling circle facing trial for his role in the “hidden debt” scandal.

What will be required to continue this progress is the theme of John Hatchard’s latest book,  Combating Money Laundering in Africa: Dealing with the Problem of PEPs. Like his earlier ones on African anticorruption laws and institutions (here, here, and here), it’s a must have for African corruption fighters.

Continue reading

FACTI Background Paper: Analysis of the Different Peer Review Mechanisms for Ensuring Compliance with Anticorruption and Financial Integrity Norms

For two decades governments have been signing agreements where they promise to curb corruption and halt the international flow of illicit funds. A promise, however, is only as good as the method for enforcing it, and in the case of international conventions and treaties the only method available is the peer review.  Experts from neighboring or similarly situated nations review how well the government is keeping its promises, recommending ways it can do better and sometimes chastising it for breaking its promises. The theory is that threat of a bad review will put pressure on a government to live up to its commitments.

Peer reviews come in various shapes and sizes, and experience with ones has shown that some are more effective than others.  At the request of High-Level Panel on International Financial Accountability, Transparency and Integrity for Achieving the 2030 Agenda Financing for Sustainable Development (FACTI), Valentina Carraro, Lecturer in International Relations at the University of Groningen, and Hortense Jongen, Assistant Professor in International Relations at the Vrije Universiteit Amsterdam, reviewed the effectiveness of the peer review mechanisms of six of the most important anticorruption and financial integrity agreements:

  • the Implementation Review Mechanism of the United Nations Convention against Corruption,
  • the Follow-Up Mechanism for the Implementation of the Inter-American Convention against Corruption (MESICIC),
  • the Organization for Economic Co-Operation and Development Working Group on Bribery (OECD Antibribery Convention),
  • the Global Forum on Transparency and Exchange of Information for Tax Purposes,
  • the Inclusive Framework on Base Erosion and Profit Shifting,
  • the Financial Action Task Force and the Financial Action Task Force-Style Regional Bodies.

Their summary of their findings and recommendations is below. and their paper here.  (Background on the FACTI and a link to its interim report recommending changes in international and domestic laws to combat corruption and stem  illicit financial flows is here.)

Continue reading

Reforming the US AML System: Some Proposals Inspired by the FinCEN Files

Last week, I did a post with some preliminary (and under-baked) reflections on the so-called “FinCEN Files” reports by BuzzFeed News and the Independent Consortium of Investigative Journalists (ICIJ). These stories relied in substantial part on a couple thousand Suspicious Activity Reports (SARs) that had been filed with the U.S. Treasury Department’s Financial Crimes Enforcement Network (FinCEN), and leaked to a BuzzFeed journalist in 2018. The documents, and the reporting based on them, highlight the extent to which major Western banks assist suspected kleptocrats, terrorists, and other criminal actors move (and launder) staggering amounts of money all over the world, and highlight the deficiencies of the existing anti-money laundering (AML) system.

What can we do to rectify this depressing state of affairs? Much of the commentary I’ve seen so far (both in the FinCEN Files stories themselves, and commentary on the reporting from other sources) emphasizes the need for more individual criminal liability—putting bankers in jail, not just fining banks. Even when banks are threatened or hit with penalties, the argument goes, this doesn’t really have much of a deterrent effect, partly because even what seem like very large monetary sanctions are dwarfed by the profits banks stand to make from assisting shady clients with shady transactions, and partly because the costs of monetary sanctions are mostly passed on to the bank’s shareholders, and don’t really hurt the individuals responsible (or the managers who tolerate, or turn a blind eye to, misconduct).

I’m quite sympathetic to both of these arguments, though with a couple of important caveats. Caveat number one: The absence of individual prosecutions of bankers is sometimes attributed to the fecklessness—or, worse, the “soft” corruption—of federal prosecutors, but as I noted in my last post, I tend to think that the more significant obstacle is the fact that it is very difficult in most cases to prove beyond a reasonable doubt that the that bankers or other intermediaries had the requisite level of knowledge to support a criminal money laundering conviction. Caveat number two: I don’t think we should be too quick to dismiss the idea that levying significant monetary penalties on banks can affect their behavior. After all, these institutions are motivated overwhelmingly by money, so hitting them in the pocketbook is hitting them where it hurts. The problem may be less that monetary sanctions are inherently ineffectual in this context, but rather that they are too low and too uncertain to have a sufficient impact on incentives and behavior.

In that vein, I want to suggest a few legal reforms that might make the U.S. AML system function more effectively. I acknowledge that these are “inside the box” ideas, insofar as they seek to make the existing framework more effective rather than to drastically transform that system. That may make these proposals feel unsatisfying to some, though I suspect the proposals will seem radical, even outlandish, to others. I should also acknowledge that I am not at all an AML expert, so it’s quite possible that the discussion below will contain errors or misunderstandings of the law or the system. But, in the spirit of trying to stimulate further discussion by those who really understand this field, let me throw out a few ideas. Continue reading

New Podcast Episode, Featuring Frederik Obermaier

A new episode of KickBack: The Global Anticorruption Podcast is now available. In this week’s episode, my collaborators Nils Köbis and Christopher Starke welcome back to the podcast Pulitzer Prize-winning investigative journalist Frederik Obermaier of the German publication Süddeutsche Zeitung, who is also affiliated with the International Consortium of Investigative Journalists. A year and a half ago, I had the opportunity to interview Mr. Obermaier on the podcast about his work breaking the Panama Papers story, which shed unusual light on how corrupt officials and other criminals use anonymous companies to launder the proceeds of their illegal activity. In the new episode, Mr. Obermaier discusses the so-called FinCEN Files (which I blogged about last week): the leak of over two thousand suspicious activity reports (SARs) filed with the U.S. Treasury’s Financial Crimes Enforcement Network (FinCEN). Mr. Obermaier explains why and how the FinCEN Files reveal how badly broken the international anti-money laundering (AML) system is, the likely reasons for the ineffectiveness of the system, how the ICIJ and its journalistic collaborators handled such a sensitive story, and the possible political implications of the stories based on the FinCEN Files reporting.

You can find this episode here. You can also find both this episode and an archive of prior episodes at the following locations:

KickBack is a collaborative effort between GAB and the ICRN. If you like it, please subscribe/follow, and tell all your friends! And if you have suggestions for voices you’d like to hear on the podcast, just send me a message and let me know.

FinCEN Is Seeking Public Input on Proposed Amendments to Its AML Regulations. AML Advocates Should Comment!

In my last post, I discussed the so-called “FinCEN Files” (leaked Suspicious Activity Reports (SARs) filed by banks with the U.S. Treasury Department’s Financial Crimes Enforcement Network (FinCEN)), and the reports from BuzzFeed News and the International Consortium of Investigative Journalists (ICIJ) based on those leaked documents. This reporting highlighted serious weaknesses in the current anti-money laundering (AML) system, both in the United States and globally. Perhaps coincidentally (but perhaps not), just a couple of days before the FinCEN Files stories went public, FinCEN issued an Advanced Notice of Proposed Rulemaking (ANPRM), seeking public comment on various proposed changes to its current regulations implementing the AML provisions of the Bank Secrecy Act (BSA). The comment period will remain open until November 16th, 2020. Of course, it’s never clear how seriously federal agencies will take public comments, but in at least some circumstances sophisticated comments, supported by evidence and analysis, can move the needle, at least somewhat, on agency policy. So, I very much encourage those of you out there in ReaderLand, especially those of you who work at organizations that have expertise in this area and might be well-positioned to submit the sort of detailed, substantive comments that stand a chance of making some practical difference, to submit your comments before that deadline. (Comments can be submitted through the federal government’s e-rulemaking portal, referencing the identification number RIN 1506-AB44, and the docket number FINCEN-2020-0011, in the submission. The link above goes directly to the comment section for this rule, though, so you don’t need to enter that info again if you follow the link.)

The full ANPRM is not that long, but let me provide a very quick summary, highlighting the main proposal under consideration and the specific questions on which FinCEN is seeking public input. Continue reading