It’s time for a new era of anti-money laundering intelligence
The ‘Forty Recommendations’ published by the Financial Action Task Force (FATF) are essentially the foundations upon which all countries have based their anti-money laundering (AML) laws and regulations. Indeed, through a system of mutual evaluations the FATF assesses each country’s level of compliance with the Forty Recommendations. However, David Lewis the Executive Secretary of the FATF has concluded the present evaluation process is failing and will be replaced in 2020.
In a recent media interview, he went further and asserted the global AML community were all. “doing badly, some not as badly as others”. In a critical way, this is a call for improvement and we need to meet this with a positive response. Collectively we need to generate better returns and outcomes from the resources we allocate to AML and our collective endeavours.
The blunt truth is banks and firms continue to pay huge fines for AML failings and all the while, on an annual basis, we seize less than 1% of the funds laundered in the world. The effect of this is to encourage money launderers, who do not fear losing their money or being caught. What all of this means is people suffer, people are killed and communities are devastated by the criminals who generate the funds which need to be laundered. The FATF have made a lot of references to the harm caused by money laundering and have now adopted the mantra. “Stop money laundering, save lives”.
This is a call to action, which aligns with the slogans adopted and promoted by governments confronting Covid-19, to demand/encourage compliance with restrictions, designed to save lives. With the threat of the virus, governments could not sit back and do nothing; they took actions, which have indeed saved lives. Thus, now is the time for the AML community to take action, because the greatest money laundering threat we face is to sit back, do nothing and tolerate the continued failing status quo.
The new ERA
To win the battles and the war against the money launderers we need to collaborate and at the same time combine three forms of intelligence:-
• E- Emotional intelligence, incorporating passion, enthusiasm, drive, desire, commitment and the moral crusade of AML professionals seeking to defeat the money launderers
• R- Real intelligence, which is derived from experience, expertise, knowledge and training, applied to the role and simultaneously developed on an ongoing basis, every day
• A – Artificial intelligence, providing the ability to replicate the thinking and analysis of the experienced AML professional, effectively creating cloned AML experts which increases the capacity to analyse huge quantities of data
These three forms of intelligence are in no way unique to the world of AML, on the contrary they feature in all walks of life and areas of business. Businesses succeed when ethical, passionate, smart people collaborate with enthusiastic experienced people who apply artificial intelligence to enhance processes, profits and outcomes. The success of this intelligence collaboration is all the more likely because passionate people with the relevant experience and expertise are better equipped to engage with and effectively train the artificial intelligence.
Notwithstanding the above, there is a need to fully understand what the AML professional does, why and how. Historically far too many AML professionals have operated with inadequate real intelligence and expertise, as well as a fear of ‘getting it wrong’. This runs contrary to the application of a risk-based approach, which implicitly accepts there will be failures, mistakes will be made and ultimately money will be laundered. Some AML regulators and AML compliance professionals seek zero failures. As such they seek to be all things to all risks, all customers, all transactions at the same time and they generate a failure rate of 99%.
As a result of the above, limited positive results are achieved. Thus, there is a need to change this strategy, a rethink of tactics and better focusing of finite resources, based upon intelligence extracted from the data. This approach will generate a reduced number of false alerts, because the scatter gun approach of being all things monitoring to all transactions will disappear.
So, what is the data; where is it the data; how should the data be applied? To some extent, data can be limitless, but of itself, this is a problem and therefore firms do need to limit the data used and the data/transactions monitored. As a result, it is necessary to understand both the data to be used and the data/transactions to be monitored. Historically banks and other regulated financial service businesses have only considered internal data, for both application and monitoring. Put simply, they have applied data/transaction monitoring to their customers and the customers’ accounts/transactions. In far too many instances, this has been applied to all transactions, all of the time.
This crude data/transaction monitoring approach generates significant numbers of alerts, which identify unusual, but not necessarily suspicious transactions. Consequently, on average 90%+ of the alerts are ultimately classified as false. The figures reference the difficulties presented when assessing and investigating an unusual transaction, which may be suspicious and indicative of money laundering, but is actually nothing more than a transaction or even a series of transactions which are different to those normally paid into or out of a customer’s account or accounts.
The present data/transaction monitoring processes are highly speculative and commonly assess the value, volume and frequency of transactions, against sets of rules, including peer groups and historic/perceived normal account activity. In some instances, the jurisdictions related to a transaction or transactions are considered, as are some high risk business types and high risk banking services/products, such as cash.
Some transaction monitoring models also identify high value, round figure transactions as both unusual and potentially suspicious. The logic being criminals do not pay tax percentages upon transactions and commonly make payments in round figures. Nonetheless, not all high value, round figure payments are suspicious, indeed, some legitimate trading businesses, only execute high value, round figure transactions.
These methods of analysis do not apply and adequately exploit the intelligence from external data which reduces the speculation and increases the likelihood a transaction or series of transactions being classified suspicious. By excluding the application of external data, banks and firms are not identifying transactions which reference data and specific characteristics, found to have featured in prior money laundering allegations, investigations and prosecutions.
Moreover, in some instances the application of external data presents the potential to identify transactions which go beyond unusual and are automatically determined to be suspicious. This new approach combines the energy and passion of the AML compliance professional who is seeking to make a difference here and stop money laundering, together with their expertise and artificial intelligence which can identify data/transactions which match with the referenced data from prior money laundering cases.
Imagine for a moment the current AML transaction monitoring process was transposed into the monitoring and searching of luggage within airports. It is likely some practitioners would stop and search all suitcases and bags, lest one or more of them contain illegal drugs. Well, it is likely a considerable number of passengers will be smuggling drugs or other contraband in their luggage, but the professionals at the airport acknowledge it is not possible to stop and search all passengers and all luggage, because were they to do so, the airport would grind to a halt.
Even if the authorities received intelligence suggesting there were several kilos of heroin within a black Samsonite suitcase, they would not stop and search all such cases. For sure, using intelligence, they would search some, in an effort to seize the drugs, but simultaneously they would accept, it may not be possible to identify the offending case. In the event AML professionals from a bank were faced with such a situation, it is highly likely some would seek to search all black Samsonites, with the obvious collateral damage to the business as usual operations of the airport.
The airport authorities, in particular customs officers, do make multiple drug seizures and it is seldom by luck. On the contrary, they are intelligence led, which means their finite resources are focussed. Thus, the AML professional must learn from this; regulators should publicly state a risk-based approach does not mean a zero tolerance of failure; AML compliance officers, should operate without fear and we should move away from the current transaction monitoring models which are failing all of us.
Resource spread to thinly are far easier to defeat and are far more likely to randomly examine innocent, albeit unusual activity than suspicious activity. All the while an AML analyst is assessing an unusual transaction which proves not to be suspicious, he/she is not looking at a suspicious transactions undertaken by a launderer. Conversely, when a customs officer is searching luggage which does not contain drugs, the drug trafficker is a beneficiary.
Thus, what is the intelligence which helps the customs officer to identify the luggage containing the drugs and in the same context can help the AML professional to identify the suspicious transaction? The answer is a simple one, it is the characteristics and sometimes the characters themselves, which have featured within prior drug trafficking/money laundering allegations, investigations and convictions. In the event a lady is found to be carrying a substantial quantity of heroin, when flying from Istanbul to London, the investigating officers will undertake an investigation in order to secure both evidence and intelligence, including:-
• If the lady was not travelling alone, who was she travelling with?
• Did she travel direct from Istanbul?
• How was her ticket paid for?
• Was London her final destination?
• Were there people at the airport to meet with her upon arrival?
• Was the luggage adapted with hidden compartments?
• What was the method of smuggling?
• Where does the lady reside?
• Does she have a mobile telephone?
• Are there UK numbers in the telephone data?
• Are any of these numbers know to Customs?
• Has the lady made this journey before; when; how many times?
• Anything and everything else which may help to convict her and establish who she was working with/for.
Now imagine, the following month the lady’s sister makes the same journey and her ticket was bought by the same method/source, would you stop and search the lady and her luggage? Would you be failing in your duty if you dd not do so? In the world of drug trafficking investigations, the next arrest/seizure very often originates from the last arrest/seizure. The success of the customs officer is enhanced by intelligence and this presents risks to the drug trafficker who constantly seeks to change, including the method of smuggling, in order to avoid detection.
In contrast, money launderers make very few changes, because the laundries have not been broken, the methods neither compromised or identified. Worst still, those that are identified are used again, because the AML professionals do not share and use the intelligence. These issues are often compounded by poor intelligence sharing between the public and private sector as well as peers within regulated financial service businesses. In addition, deferred prosecution agreements hinder the sharing of intelligence, because so much is withheld from the formal written agreements.
Nonetheless, there is a lot of intelligence out there and when applied in a focused manner results will improve.
The ‘black box’ AML data
When an aeroplane crashes, there is always an inquest which seeks to establish, what happened and why? Importantly the airline industry uses the data recovered from each plane’s black box data recorder to inform future decisions, safety, security and airline development. The industry does not solely rely upon this black box data, but it does fully exploits it in order to reduce risk and save lives. In other words, “Stop aeroplanes crashing, save lives”.
Over the prior twenty five years there have been numerous, high profile money laundering allegations, investigations as well as prosecutions and in some instances, the same characteristics and characters have appeared time and time again. The question is, how and why has this been allowed to happen? Perhaps the answer is, “It’s (the laundering process) not broken and therefore it does not need to be fixed (by the launderers)”. Consequently, the criminals have carried on laundering and whilst drug traffickers constantly change in order to manage risks and avoid detection, the risks for money launderers have changed very little.
Therefore, we need to make the launderers change; we need to change the risks and stop the same characteristics and characters featuring time after time. Some AML professionals ponder where and how to apply this new thinking, whereas others are looking to vendors to incorporate this ‘black box’ data into new, innovative transaction monitoring systems. Vendors have the research capacity to collate the ‘black box’ data, which will include:-
• The names of persons, companies, service providers and banks alleged to be connected to or engaged in the laundering process
• The names of individuals who acted as directors, shareholders or nominee parties for connected companies and partnerships
• The names of other companies performing the role of shareholders, partners or controllers
• The addresses of the connected parties
• The names of other parties providing services to the alleged launderers, wittingly or otherwise
The Laundromat investigations undertaken by the OCCRP over the prior 12 years referenced the constant presence of Limited Liability Partnerships (LLPs) and Limited Partnerships (LPs) incorporated in the UK; as well as server and IP addresses in Ukraine; corporate service providers from the UK, Belize, Panama, the British Virgin Islands, the Marshall Islands and other offshore jurisdictions. In some instances the same nominee parties and addresses have featured in several laundromats.
Thus, the intelligence is available and can be used to change the numbers, the effectiveness and importantly the risks for the launderers.
Fear and ‘What if?’
Fear of getting it wrong; failing to identify a beneficial owner hiding behind a nominee; fear of missing a suspicious transaction and being possessed by the notion of what if some money laundering takes place inhibit a lot of AML professionals. ‘What if’ presents infinite possibilities and undoubtedly contributes to a desire to monitor all customers and all transactions at all times, but it fails to exploit and apply intelligence.
It is the combination of courage and intelligence which will deter, detect and frustrate money launderers. It incorporates the courage to break away from accepted, but nonetheless, failing transaction monitoring practices and the wisdom to apply focused AML intelligence.
The rewards which flow from this new approach include a deep sense of accomplishment, a pride in making a difference, an acknowledgment of the value and the impact of powerful AML intelligence. It is a move away from process for the sake of process to thinking about saving lives. Pursuant to which we will take a look at speed, because speed kills.
Governments around the world impose legally binding speed limits on roads in order to reduce speed and save lives. Notwithstanding this, many people continue to be killed because too many people drive too fast. Behold the speed camera; the replication of the police officer with the speed gun and the artificial intelligence of road traffic management, speed monitoring.
The speed camera enables authorities to police roads without deploying an actual police officer, but they do not operate on the basis of ‘what if’, or through fear of missing a speeding driver who subsequently kills someone. No, they operate using real intelligence to train and focus the cameras/the artificial intelligence. The point being, speed cameras are not deployed to generate revenue from penalties imposed upon drivers who breach the speed limit, no they are strategically placed at ‘accident blackspots’, commonly where speeding motorists have previously caused serious accidents and/or killed people.
Speed cameras save lives, because they are used intelligently, as proven within a study by the London School of Economics who found, that from 1992 to 2016, speed cameras reduced road accidents by 17-39 per cent and fatalities by 58-68 per cent within 500 metres of a deployed speed camera. Now think of this in the context of money laundering transaction monitoring. Does your firm apply speed cameras/artificial intelligence/equivalent to your transaction monitoring? In the event the answer is yes, where are they located? Have you tested their effectiveness? Has the London School of Economics or equivalent undertaken a study to establish the impact of the deployment? Do you use intelligence to place the cameras/artificial intelligence? Now, bluntly, how many lives has your money laundering transaction monitoring saved? All being well this paper has provoked some more thinking and the next time you drive past a speed camera you may ponder, how many lives has it saved; how satisfying is that to the people who concluded that was the right place to deploy the camera and how can you transpose this thinking into your money laundering transaction monitoring?
It’s time to change, it’s time for a new ERA of intelligence within money laundering transaction monitoring.