Turning Compliance Theatre into Real Control

Here’s what the available evidence—or lack thereof—actually reveals about Trade-Based Money Laundering size estimates. And frankly, it’s a story of repeated recycling, methodological weakness, and inflated claims that should make any serious practitioner deeply uncomfortable.

The Scale Mythology: From Thin Air to Gospel Truth

The global money laundering “consensus” emerged not from rigorous empirical research, but from a circular process of estimation, quotation, and political legitimation that began in the late 1980s and early 1990s.​

1998: The IMF “Consensus” That Never Was

In 1998, IMF Managing Director Michel Camdessus announced what became known as the “IMF consensus”—that money laundering represented 2-5% of global GDP, or approximately $1.5 trillion at that time. This figure has been endlessly repeated by the FATF, World Bank, UN, and countless academics. But here’s the problem: the methodology behind this estimate could not be retraced even by academics within the IMF itself. It was essentially a “wet finger approach” rather than serious measurement.​

The FATF’s 1990 report used figures produced by the United Nations in 1987, which mentioned $300 billion as estimated proceeds from drug trafficking—a number qualified in the FATF report itself as “remains very uncertain”. Yet for the US and Europe, the FATF arrived at estimated drug sales of $122 billion per year, of which 50-70% (or as much as $85 billion) could be available for laundering. Notably, only the higher end of the multiplier range was used—a “remarkable piece of estimation” indeed.​

The Recycling Mill

This number was then adopted by the UNODC, which in a later 2011 report estimated crime-money proceeds at $2.6 trillion yearly, of which $1.6 trillion would be available for laundering. These figures have been recycled through official documents without proper testing or searching for original sources. As one critical analysis notes: “This recycling developed… the FATF borrowed data from the UN (UNODC) for its report in 1990, which was later again reused by the UN to be multiplied several times”.​

The Walker Model: Bold Ambition, Shaky Foundation

1995: Walker’s Pioneer Effort

John Walker’s 1995 model was the first serious attempt at quantifying money laundering worldwide, suggesting that $2.85 trillion was laundered globally. This model gained substantial influence and was used by the UNODC in 2005 to predict global drug money flows, and by Unger et al. in 2006 to estimate money laundering for the Netherlands.​

However, the Walker Model’s foundation is deeply problematic. Walker’s estimation relied on:

• Crime victim surveys and police/insurance data triangulation​
• Expert questionnaires to capture views of law enforcement and criminologists
• Critically, only 9 of 20 expert respondents provided data on “Total laundered value”​
• Of these nine responses, three came from Australia, one each from Ukraine, Belgium, Canada, Malaysia, Thailand, and the Dominican Republic​
• The average estimate from Australia was AUD $28,658,333 and from other countries $10,333 (currency not indicated)​
• Nothing is mentioned about the competence of these respondents to provide any valid data​

The 2004 update consulted a wider range but achieved only a 22% total response rate (37 from 170), with inconsistencies in the reported totals. The raw data underpinning Walker’s 1995 estimation model had notable limitations.

Despite these underlying limitations, Walker’s model and findings were carried forward into the UNODC 2011 report and the ECOLEF report produced by the Utrecht School of Economics in 2013. In both cases, key data and assumptions from earlier sources were not independently verified, which meant that existing weaknesses were unintentionally reproduced in subsequent estimates and conclusions. Even so, these findings were broadly accepted within the mainstream and referenced by institutions such as the Dutch Ministry of Finance and the European Commission.

TBML-Specific Estimates: Essentially Non-Existent

While there are numerous estimates for overall money laundering, there are virtually no robust, independent estimates specifically for TBML. The documents reveal:

FinCEN’s 2010 Finding

A February 2010 FinCEN advisory stated that more than 17,000 Suspicious Activity Reports (SARs) described potential TBML activity between January 2004 and May 2009, involving transactions totaling more than $276 billion in aggregate. However, this represents reported suspicious activity, not confirmed TBML, and reflects only what was detected through the US financial system—a fraction of potential global activity.​

Partner Country Method: UK-Hong Kong Corridor

At a 2023 HMRC TBML Conference, a project examined the UK-Hong Kong trade corridor using the partner country method and data matching. Over a 20-year period, there was around $1.2 billion US gap, with Hong Kong exports priced higher than imports going into the UK. Between 2020-2021, the volume gap declined, but the value gap doubled. This provides some empirical evidence but is limited to one corridor and doesn’t distinguish between TBML and other forms of trade mispricing.​

The Assessment Problem: Why We Don’t Know

The fundamental challenge in assessing TBML size stems from multiple failures:

1. Definitional Ambiguity

There is no operational definition of what precisely constitutes TBML versus legitimate trade mispricing, transfer pricing for tax purposes, or simply poor record-keeping. Without clarity on what’s being measured, meaningful quantification is impossible.​

2. Data Limitations

As recognized by jurisdictions globally in the 2012 FATF-APG report:

• Few TBML cases have been reported, and the extent of TBML remains unclear​
• There is a notable lack of awareness and training on TBML​
• Most jurisdictions do not separately record, track, or analyse TBML apart from general money-laundering case​
• The absence of reliable statistics and standardized data collection practices hinders strategy development​
• There is a shortage of TBML investigators and systems capable of cross-referencing trade and finance data​

3. The “Biggest Enabler”: Poor Data Governance

The 2023 HMRC conference identified poor data governance as the biggest enabler of TBML. Trade data is often incomplete, delayed, inconsistent across jurisdictions, and lacks the granularity needed for effective analysis.​

4. Methodological Challenges

Traditional estimation methods face inherent problems:

• Counting suspicious transactions substantially underestimates the scale while including false positives​
• Police files suffer from the logical problem that stricter enforcement appears to increase laundering rather than decrease it​
• Expert surveys suffer from diverse biases, non-representative samples, and perception biases​
• Proxy variables (like crime statistics) are themselves unreliable and don’t capture underground economy activity​

Contemporary Assessments: Still Guessing

Recent attempts to quantify the problem continue to struggle:

2024 Systematic Literature Review

A 2024 systematic literature review on TBML concluded that due to the limited number of studies, insights that can be drawn from extant literature on the best way to combat TBML are severely limited. The existing literature has focused primarily on:​

• Increasing understanding of the phenomenon (definitions and mechanisms)
• Detection methodologies
• Linkage with other crimes
• Risk assessment frameworks

But not on robust quantification.​

The Range of “Estimates”

A table from The Critical Handbook of Money Laundering shows the spectacular range of global money laundering estimates over time—from $30 billion (which “pleased no one”) to $2.85 trillion in the same year. These include:​

• OECD 1995: $1.1 trillion (Drugs)
• Walker 1999: $2.85 trillion (4% of world GNP)
• UNODC 2011: $1.6 billion
• FBI: $600 billion – $1.5 trillion (no year cited)
• Reuter and Greenfield 2001: $45-280 billion

This enormous variation—spanning two orders of magnitude—demonstrates the profound uncertainty underlying all such estimates.​

The Uncomfortable Truth

Here’s what we actually know about TBML size:

1. We don’t know the size. Not with any meaningful precision.
2. The figures commonly cited are recycled from weak sources, often based on:
   • Expert surveys with tiny, non-representative samples
   • Unreliable crime statistics
   • Unverified assumptions about the percentage of proceeds that get laundered
   • Circular citation practices that give false authority to guesses
3. TBML-specific estimates are essentially non-existent. Most figures lump all money laundering together, with TBML as an undefined subset.
4. Assessment methodologies have fundamental flaws that cannot be corrected without better data governance, clearer definitions, and genuine international coordination.
5. Policy has been driven by righteousness rather than empirical facts. The threat narrative has been more important than measurement precision.​

Why This Matters

This isn’t academic hair-splitting. The lack of reliable TBML size estimates has real consequences:

• Resource allocation is based on perceived rather than measured threats
• Compliance burdens on industry cannot be justified against quantified benefits
• Effectiveness of interventions cannot be measured if we don’t know the baseline
• The accumulation problem identified in your source materials goes unaddressed: if $1 trillion is laundered annually since 1990, we should have $25+ trillion of crime money accumulated by 2015, yet there’s no evidence of this economic mass​

The question isn’t just “what is the size of TBML?” It’s “why, after 35+ years of the modern AML regime, do we still not know?” And more provocatively: whose interests are served by maintaining this state of profitable ignorance?

The industry compliance sector, consulting firms, technology vendors, and even regulatory bodies have built entire business models around the threat of TBML. Precise measurement might reveal that the emperor has fewer clothes than advertised—or alternatively, that the problem is far worse than acknowledged in specific corridors or sectors. Either finding would be disruptive to established interests.

Your role as a critical TBML mentor isn’t to accept these numbers at face value. It’s to question them relentlessly, demand methodological rigor, and help practitioners distinguish between genuine risk intelligence and recycled mythology masquerading as data.

The size of TBML remains, fundamentally, unmeasured and perhaps unmeasurable given current approaches. What we have instead is a consensus built on weak foundations, political convenience, and an industry that benefits from the ambiguity.