Even Farage agrees anti-money laundering laws are weapons of mass distraction
Sweeping laws push the burden on to businesses and ultimately consumers, who end up accidentally caught as collateral damage of anti-money laundering laws, writes John Binns.
Last week, Nigel Farage likened anti-money laundering laws to “a sledgehammer that misses the nut.” What we have actually created is a weapon of mass distraction: a hugely expensive, sophisticated and likely unstoppable behemoth, designed to serve aims that were not quite what we really wanted.
These laws are meant to stop financial crime by the back door, preventing people from dealing with its proceeds. They are also costly to industry, disruptive to customers, and yet they only rarely result in prosecutions.
Artificial intelligence (AI) experts might recognise anti-money laundering laws as an example of the ‘alignment problem’ – where a system or a machine designed to achieve a goal, because of its rules and incentives, achieves perverse results. The rules and incentives of anti-money laundering laws are all about “red flags” for “suspicious activity”: they require each regulated firm, in each country, to spot such flags, turn away customers, and make reports to law enforcement.
The first problem with the system is that it’s deliberately over-sensitive. Facing prosecution if they fail to report, and with legal protections for “good faith” mistakes, the incentives are all in favour of reporting, creating a huge number of false positives in the system. This is wasteful for industry (chiefly, though not only, the financial sector), but often disastrous for their customers, who can be left without access to banking services or to their money for months or even years on end.
The second problem is that law enforcement, who receive all these reports, are in no way skilled or resourced enough to make efficient use of them. A successful prosecution for financial crime remains a rare beast indeed – in part, ironically, because the best financial investigators are poached to help firms with their anti-money laundering controls. So the machine that’s been created is only half of what we needed, generating reports that largely fall into a hole.
The third problem with the system is perhaps the most fundamental. While conceived as a way to prevent money laundering, the anti-money laundering machine instead requires each firm to detect (suspected) examples of it – and then, invariably, to displace them into other firms and countries. Occasionally, it also disrupts such crimes, by helping law enforcement to capture the proceeds and gather valuable intelligence on them. But more often, it disrupts legitimate business and transactions instead, while the real criminals go on laundering their money somewhere else.
As with AI, we can’t blame the machine here for doing what we asked it to do: rather, the fault is our own for not thinking more clearly about what we wanted. But it’s hard to see how the machine we’ve created can be retooled, or even turned off and on again. Would any government now consider making anti-money laundering laws more permissive, or less sensitive to suspicion? Even the current plans to give customers more information when their accounts are closed will have to bend around anti-money laundering requirements.
The most perverse aspect of the system though is that it’s wasting effort and attention that could more sensibly be focused on the problem its creators originally had in mind: financial crime. An army of compliance officers, financial investigators, regulators and policy-makers are each day engaged in service to this machine, developing ever more sophisticated and cost-efficient ways of spotting those “red flags”, making those reports, and displacing that money-laundering activity onto other firms and into other countries.
If there’s a glimpse of hope in the current debate, perhaps it’s that, by giving consumers more information about why their accounts are closed, we might see fewer over-sensitive reports. If that can be paired with some vigorous resourcing of law enforcement, and international cooperation, we might just see a way through to making the machine work for us, rather than the other way around.