 We have one more panel before our closing conversation This will be our panel on AML and financial inclusion And I'm going to go ahead and introduce our moderator Joanne barefoot. Thank you Good afternoon I'm Joanne barefoot. I know it's getting to be late in the day But we are going to give you a lively panel and I'm excited about the fact that I Think our topic is going to weave together many of the things that we've been talking about over the last day and a half So maybe we'll be able to bring some general perspective to it I am CEO of AIR the Alliance for Innovative Regulation and a co-founder of hummingbird reg tech and I am thrilled to Introduce my panel and first to thank the University of Michigan Again, I'm Michael and Adrienne and their amazing team. I myself am an undergraduate Alarm of the University of Michigan. It's just been a joy to be back on campus And I also want to thank the Gates Foundation for the incredible Visionary work that you do in this area and everything else I'm going to start by introducing our panelists We are each going to make some opening comments and then we're going to have a conversation and my first guest far to the right is Jennifer Calvary my guests have long titles. I'm going to read them She is the global head of financial crime threat mitigation at HSBC and also group general manager based in London She runs a unified global capability that leverages analytics and technology To identify analyze and investigate financial crime risk to their HSBC group in 60 countries She and Aaron both have an amazing background in both the public and private sectors and Jennifer was previously the director of FinCEN the financial crimes enforcement network So she was the senior person in the United States running the anti-money laundering Organization in the Treasury Department prior to that she spent 15 years as a prosecutor at the Department of Justice Dealing with money laundering corruption fraud and organized crime my other guest is Aaron Klein Aaron is the Economic Studies Fellow and Policy Director of the Center on Regulations and Markets at the Brookings Institution Prior to that he ran the bipartisan policy centers work in this field and Before that he was Deputy Assistant Secretary of the Treasury for economic Policy in addition to having served on the Senate Banking Committee staff as did I but much earlier than you did So I'm going to ask them in a few moments to as I said give some opening comments but first I'm going to set the stage a little bit and Talk with a wide lens about anti-money laundering And as an issue in general and specifically for financial inclusion Bill Gates famously said We tend to overestimate the change that will happen in a year or two and underestimate the change that will happen in 10 and here we are looking at this 50 year Time frame and I think it's important to as we ponder this to think about the fact that as change occurs over that long Trajectory, it's not going to be a gradual linear slope It's much more likely that we're going to go through a series of sort of hockey stick Changes that are going to catch us by surprise There's a character in the novel by Ernest Hemingway who has asked how he went bankrupt And he says at first gradually and then suddenly So that's what we're dealing with a big shift to a new world that we know we're all working in because Finances digitizing and financial regulation is going to do the same The AML system that we have today. I'll be interested to see if my panelists agree Is broken and it's important that we fix it and we can fix it It there's a tendency sometimes especially in the financial world to think about Financial crime and money laundering is white-collar crime or victimless crime and it's really among the most terrible types of crime on earth it funds terrorism and it funds illegal trafficking and weapons and drugs and looted antiquities and endangered wildlife and in human beings and the human trafficking issue just to pick out one facet of this to focus on The UK financial conduct authority uses the numbers that there are 40 million people today enslaved as human captives More than all of the older history of the world combined 10 million our children and a million children are enslaved for sexual exploitation this is crime that's been growing and that we need to figure out how to stop and We work very hard to stop it the industry The UN estimates is that there's about 1.6 trillion dollars laundered every year and That we catch less than 1% of that with the tools that we use today and that's despite Spending tens of billions of dollars a year to try to catch it and despite all the efforts of the kinds of people in this room Who are working from a regulatory and enforcement standpoint to fix it? I You'll get my presentation. It's got some more statistics on it, but this effort is not working We have a you know, you can you can debate it at the margin But we've got basically a 99% failure rate and then beyond that AML is not just failing, but it's also doing harm it does cause financial exclusion and Makes it difficult for people to come into the financial system, which is the main focus of the gates work in this space as But has been discussed all day many people can't qualify They can't prove who they are easily enough or can't easily get into the financial system and have access to it And so there's been a lot of work looking at the impact of the de-risking process particularly what we've seen in part from the United States in fueling cutting off of whole sectors whole countries and even humanitarian crises again I've got a few quotes here From some of the people who have looked at these problems. So we ask ourself. Why aren't we doing better if we're spending so much? And I think there's three answers to that the first is that the forces of the financial world and the And of law enforcement and regulation are using old technology and the criminals are using great new technology more and more and more Secondly, we are using old analog era Identity systems that have just created a terrible terrible challenge with the know your customer rules It is important that people be identified to come into the system But whether you're in a developing country or a developed country Both models are broken in terms of people being able easily to prove who they are and begin to think about moving to a digital Identity system that could be both effective and efficient and the third thing that causes great difficulty in this system is That the forces for good need to protect the privacy of the individuals in the system and the criminals don't they share information freely they buy and sell data without any restriction on it and We on the other hand cannot easily share data with each other and Watchword that has developed to capture this point is the slogan that it takes a network to defeat the network And we are not well networked In fighting financial crime So the thing I'm going to going to focus on for a moment before I turn it over To Jen and I'm looking forward to a great discussion of this with Jen when we get into the panel is that there are a lot of people working on all three of those problems and The in particular people are working on the third one by trying to evaluate the potential of privacy-enhancing technologies to enable safe widespread sharing of information between Banks and each other banks and governments across country borders and so on and this issue was the focus of the Hackathon this year in July run by the Financial Conduct Authority in the UK To study the potential of privacy-enhancing technologies Jen was there. She and I were both at their similar event the year earlier to study new types of ways of Limiting the knowledge that's being shared if you think about a if you think about wanting to buy liquor And having to hand your driver's license to a clerk in a store There's no reason that that clerk needs even to know what your birthday is not to mention all the other information on your driver's license all he and or she needs to know is That you're old enough to buy it and so there's a lot of emergence of thinking about can we just give smaller amounts of Information that solve the problem at hand and not share everything the FCA Invented this technique called tech sprints. They are hackathons The FCA will tell you we are regulators, so we don't like the word hack and so we call them tech sprints And when they ran the one this year they asked air our new nonprofit to run a Washington satellite site for it And we had a very productive week Running the first-ever u.s. Tech Sprint Focusing on AML issues. We had a great turnout of large and small banks and fintechs and a great turnout of government agencies We had 65 regulators Participate in the u.s. Over the week in addition to a much bigger event that happened in In London and we had it key noted by the FDIC chairman Yelena McWilliams who has been a real driver for regulatory modernization in the United States and she also was a judge in the The program and we are excited to say that on Monday We are going to FinCEN to meet with the director and his direct reports with the teams that were in the hackathon and present their solutions to FinCEN and Model a new way of thinking about how to accelerate change by not just having working groups and conversations but actually getting people together across discipline and writing some computer code and Back to the point that these issues are coming fast This is a quote from our friends in the UK who said they realized that if they don't move forward given the pace of change in today's technology world in effect, they'll be accelerating Backwards and that they need we need to move forward even when we're not totally sure exactly what it is that we think we should do so many regulators throughout the world central banks and other regulators are working on these issues so our ones in the United States and I'll just offer this as something to think about and then turn to Jen Who can tell me whose picture that is? Barney Frank our former congressman Barney Frank of Dodd Frank co-sponsorship fame Brilliant man. I will say and this of course is a picture of Steve Jobs And we've been trying to ask the question. What would happen if you gave the same problem? If you had been able to give the same problem to Barney Frank and to Steve Jobs They would approach it completely differently Can we begin to create regulatory innovation models that are getting the best of both of these kinds of thinking combined? At the same table working on the same problems at the same time. I'll just mention very quickly I am a senior fellow at Harvard I have a series of papers coming out on these topics and I have a podcast show called barefoot innovation on these topics Which I commend to you I had wanted to show you a video from the FCA and we weren't able to Do it technically, but this is one slide from it I urge you to go to the FCA's website and watch the video of their seventh text sprint the one that they just held To talk about what they think they can do with technology to do better for anti-money laundering and The fairness and efficiency both of the system so with that I want to turn it over to Jen And she is going to talk to us for a little bit and then Aaron. I think you have some slides as well Jen. Yeah, thank you. Joanne. Good afternoon everyone I'd like to thank Adrienne and Michael as well and your team for the kind invitation to be here today I did not go to the University of Michigan, but I did grow up about a half hour away from here So it's it's nice to be home And I'll see if I can't even get a little bit of my prior Michigan accent to come out during the course of this Discussion it's mostly lost, but we'll see So as Joanne mentioned, I started my career my professional life as a prosecutor at the US Department of Justice And towards the end after 15 years there became an executive leading prosecutors My last job at the department was heading up The organization that was in charge of money laundering prosecutions recovering funds from things like kleptocracy abroad but also prosecuting financial Institutions that failed to have effective anti-money laundering programs in place or engaged in sanction stripping so from that position I moved over to FinCEN to be a regulator and to be the Part of an organization that collects all the data that banks file all those reports that banks file those go to FinCEN And they're the financial intelligence unit So did that and then moved over to HSBC where I work on the same issues I moved to HSBC and tell people that after 20 years in the US government and someone who cared a lot about Keeping communities safe that I thought I could make a bigger impact on the world for good at HSBC if I could help them To be effective at that at that aspect of their Responsibility then I could even in the US government and I have to say that hasn't changed now after three years at HSBC Although it does occur to me that there is now one position left in the anti-money laundering Financial crime space that I have not done and that would be to actually launder the money or help criminals to launder money So I'm going to keep that out there as a possible future job opportunity But for now focus I'm still trying not to say on the right side of the law so You know listening to the the discussions throughout the last couple days It occurs to me that as we talked about the financial crime agenda in the context of Financial inclusion it was so much of the conversation has been about how it is a hurdle to or obstruction to that end that end game that goal of financial inclusion And I can't help but wonder and have given quite a bit of thought to doesn't need to be that way Is that is that is that a fair? Assessment first of all and if so does it really need to be that way? And so I thought I'd just spend a few minutes talking about that that question Before we go into the over to Aaron and into the broader panel So is it an impediment? I Think it's probably fair to say that it is And maybe I can spend just a moment explaining why at least from the perspective of a large global Financial institution or in more than 60 countries. We have more than 30 million customers We are expected to and and want to Make sure that we don't have Criminals exploiting our financial services to harm our customers in our community So that can be everything from fraud to money laundering to sanctions to tax evasion Corruption human rights abuse human trafficking all the kind of ills. We don't want to be exploited for those ends But we need to try to do that by monitoring transactions and and understanding the individuals with whom we're Who we're banking and with whom they're transacting our customers are transacting So let me try to put a little bit of a color around that on a monthly basis we Screened 658 million transactions through 200 million accounts When we get a list of 10 names and asked, you know, do you bang these are 10 names of individuals? Do you bank any of these folks will tend to come back with thousands of Hits on those 10 names most of which of course are are not the actual 10 people And that's if we get 10 names and not hundreds or thousands And so the scale of trying to do under the best of intentions A good job at this is just daunting And the way that institutions have done it today is not particularly effective And this is across the boards the the system across the board the way we try to perform These responsibilities today is we look at transactions Individually and we try to understand if they are suspicious of financial crime So when you have that many transactions to start a transaction and screen that using rules-based systems means that the alerts to actual real suspicious activity Ratio is very low. So in the single digits Most of what we do is clear noise out of those High false positive rate systems and that's just in the AML space We have separate systems where we look at sanctions and clear out all the the false positives around name screening That we're doing we have separate systems that focus on fraud And try to do the real-time transactions and on screening of transactions to understand if there's fraud And so we go through all these different systems generate large amounts of false positives ultimately get to Reports that we do provide the government to help them keep communities safe In Europe. I've seen some studies saying that only about 10% of those are actually Used by law enforcement in Europe and we've all seen the Statistics that only about 1% of criminal assets are actually confiscated each year That's not to minimize some very good work that does come out of that and there is some really important and great work that that comes In that in the financial crime space But it's I don't think it's a system that any of us could look at We feel it's terribly effective at achieving the goals we want it to achieve So there is a large conversation a wide recognition I think globally amongst regulators amongst industry that we need to do something to achieve the ends that that that the financial crime Mandate is seeking to achieve and then when you put that next to what about the the Negative impacts it has on financial inclusion. How do you think about that? So the reason that we see the the negative impacts on financial inclusion is because like with a blunt system like that one I just described rules-based very blunt view to try to find Suspicion we're not very good at finding it which means that if we're operating in a high-risk jurisdiction In high-risk products high-risk clients we had run a very high Potential of missing the real financial crime and we face the very real possibility of having enforcement actions That levy finds over a billion dollars on institutions and so Trying to understand the the risk there you end up with Financial institutions all making and assessing the risk in the same way So you see financial institutions all saying the same jurisdictions the same products the same Classes of customers are risk and we're not sure how to manage it And so when everyone leaves that jurisdiction at the same time We end up with the de-risking debates with the financial inclusion issues And so that's that's what we see the impacts So then we turn to the question, but does it have to be that way? Is and I don't think it does so for many other reasons and the things we've been talking about today the Opportunities that technology provide for us. I think we can go a long way and being more effective at identifying financial crime and thus being far more Targeted in our actions Which number one means we don't have to de-risk entire categories of clients and Secondly can give us the confidence to go into jurisdictions or take on clients that we haven't had the confidence in the past because we Weren't sure we could actually identify risk when it occurred there We're doing a lot of thinking certainly at my institution, but others as well as how exactly do we get from here to there? and we're focused on looking Taking all the data essentially at our disposal that we already have Looking at Financial crime holistically so instead of doing it in silos fraud versus sanctions versus money laundering Look at a customer or their counterparty and try to understand the probability that this customer poses a significant financial crime risk to us today You updating that view dynamically as we get in new data each day And being able to understand down to a very high definition view What is the probability that someone poses a financial crime risk? And you could imagine if we were able to do that really zoom in almost like a camera and a high definition camera in And understand where there is risk and pinpoint that risk and take what the appropriate actions It means that we could also zoom out and understand what a risk is of a product of a jurisdiction of a sub-sub-sub jurisdiction and and Be much more able to go into places that we haven't had the confidence to go into in the past for a big bank like ours where our our business model is not mass retail at the at the at the lowest Social and social like economic and It doesn't mean that we're all of a sudden gonna gonna move into a different business model in that sense but what it does mean is that several of the the Whether it's fintechs or traditional financial Institutions who do operate in that space and eventually need access to the international finance system or products that we're willing to go in and engage those institutions or at least have The confidence that we would be able to do that So I do think then that there is an ability both to improve our financial crime outcomes make our community safer keep them safe While at the same time Making it so that financial inclusion and financial crime are mutually supportive and not at odds with one another In fact, it's fascinating to hear some of the ideas the policy ideas around financial inclusion Are some of the same ones that are needed? To implement the the vision that I just outlined so things like a digital identity are Absolutely at the core of being able to do some of the things that I'm talking about As is information sharing and the issue of data and sharing of data Cross-border within an institution between institutions with government But the issue of information sharing and data is absolutely at the core of this The only one we usually typically at HSBC we say there's three things we need to really go in this direction digital identity Information sharing and then the only one I didn't hear come up as part of a financial inclusion Discussion is central registries for beneficial ownership That one is more aligned to the pure financial crime discussion than I think the financial inclusion But the fact that we cross over on two out of the three major Policy solutions I think is is just indicative of the fact that we're there's more in common between these two policy goals than there is Intention between them. I guess the last thing I'll say before before turning over you Aaron is Also trying to think I guess as I was listening to the discussion over the last couple days What the significance to central banks? Might be to the extent that the central bank of the future offers Financial services such as a stable coin or or in that we're really working more on the payment rails And I wonder how much of the risks that I have to manage every day all of a sudden become your reality as well So I'm trying to deal with all of the cyber enabled Fraud looking at things like the the swift attacks on the bank of Bangladesh does that become your daily reality if you're now running the payment systems What about you keeping the the financial cry criminals from exploiting the products and services that that you're offering? So I think I would leave you with be careful what you wish for on that front And maybe it's a new career opportunity as I won't have to become a criminal I can come help central banks and protect on that that end But I'll leave it there and turn it over to you So Let me just start by thanking Adrian and Michael For having me and for putting this great event together and thank the Gates Foundation and echo comment Michael made last night about the fantastic diversity of Thought background experience geography and gender that's been represented at this conference It's really something that's quite impressive and I'm pleased to be a part of it I also Let me let me start by saying one thing that I'm excited about about a central bank of the future Is that central banks tend to be If not dominated by economists have a strong voice from economists economists hate unindexed numbers The current $10,000 threshold level for Currency transaction reports is hardwired into law and is unindexed and I think it's very useful I Made this graphic when I was at the bipartisan policy center and started delving into this to really think back about what this system Was designed to do when it was enacted in the late 60s And why the system is broken per Joanne's question. I will answer it. Yes, the system is broken One of the reasons the system is broken is because the system designed to do everything Kind of does nothing when this system was designed in the 70s. It was designed to catch tax cheats and organized crime Moving large sums of money through the financial services system $10,000 I usually like to focus that you could buy a fully loaded Cadillac and cash and not trigger a CTR today There's not a single car, but since we're here you could have walked into the University of Michigan in 1972 and paid your entire annual tuition at the average Public university today if you're out of state you can't at Harvard You could have walked in and paid your entire annual tuition prior any the average private university In cash no CTR in part of this is Because over time this system has been bootstrapped to catch other types of criminals in the 80s It was moved to catch drugs and drug money It was then again changed after 9 11 to catch and focus on terrorists Who use the financial system for radically different purposes at radically different dollar amounts? And so, you know Part of this question then becomes who commits I think Jen put it right the real financial crime So let's look at financial criminals who went to jail so The first one is Denny Hastert who's in jail for anti-money laundering particularly for structuring The actual crime I think he's meant to go to jail for was molestation of a child while he was a wrestling coach But the statute of limitations for that was a long time ago and the crime that he was easily proven of Was AML and structuring? Since I'm in a college audience I'm trying to to relate to the current generation a little bit more and for those of you that knew that the Jersey Shore Mike the situation was the smartest member of the Jersey Shore So he knew that there was a $10,000 reporting limit He was also part of the Jersey Shore So he thought if he continually put $9,999 into his account from DJ gigs that he wouldn't have to report taxes on that Now they are financial criminals Al Capone was a huge criminal the gravity of what Al Capone did is stunning He came along before AML, but obviously was brought in for tax evasion, which kind of AML was essentially a tool to catch and so But are these the people? Are these the high priority out of the hundreds of millions of transactions that Jen's organization is looking and the Scarce amount of resources we can dedicate are these the people that we want to use the AML regime and task the central bank to commit to prioritize Because they're criminals So here's that somebody else who's been snared in AML One I thought in the afternoon we wouldn't mind looking at at Leo But beyond that he'd gotten some money from a Malaysian film financier whose assets were eventually seized and you can kind of go through The story and realize he had nothing to do with this, but ironically enough the Wolf of Wall Street may have been financed By criminally laundered funds Which then when you go into asset and part of the the reason I pull this up here is because I think there's a thesis that often goes Unstated in AML reform that conflicts with the modern reality of What it have how it actually works on the ground and I did give some difference to Jen on this to Agree or disagree with me and you have more experience in it But the thesis kind of goes as follows Criminals are operating in the bottom of the ocean, which is super dark and hard to find Criminals generate financial profits particularly a lot of the heinous crimes that Joanne was talking about human trafficking, etc that profit like money kind of bubbles up to the surface and When if you could find the bubbles because it has to go through the financial system to be transmitted particularly for international crime Which began with the mafia returning money or drug cartels expropriating money out of the United States, but either way money needs to move around borders That then you could find the crime bubbles the money would be easy to find and then you could trace it down the water and catch the bad guys What I think instead we've actually ended up with is more of a system where law enforcement kind of finds bad guys on their own and Then often kind of queries the SAR database to look at the bubbles up because at the end of the day That's an easier case to prove Joanne and one of her papers had a nice quote from a former law enforcement official that said you know One of the things is you tend to prosecute the easier cases We catch the stupid criminals And it helps kind of pause this other question. Why are we only catching 1% of the money, right? The money isn't Isn't necessarily where we're looking, but I mean these are quote-unquote real financial criminals Here is a graphical Thing of just the amount of suspicious activity reports filed by depository institutions after Sarbanes after the Patriot Act of 911 We expanded the categories of filers, but so this tries to hold for that constant You can see this radical increase particularly recently from 2015 to 2018 even if you say that's a one-year transitory It's gone up about 30 40 percent There are a couple of theses one. There's that much more crime going on two, there's a ton of over reporting three, there's a lot We were never reporting the right amount to begin with and There's some tension between these theses But let me offer an alternative and this gets back to my first point if you do everything you do nothing Well, if you're looking for a needle in a haystack, we've really succeeded in throwing a lot of hay on the stack and The question again becomes this prioritization of quote-unquote the real financial crime I like the way that you said it and ask yourselves What is the central bank or somebody else who's in charge of monitoring AML? What crime should they prioritize? using their limited resources in terms of going after folks and I want to go in and zoom in on one type of financial crime that's exploded Since 2015 that is so prevalent. I think there are four operations Within a walking distance of this building the state of Michigan like Many other states to the point we're one out of five Americans live in a state that have created state licensed cannabis Notice I choose my words carefully here. It is not legal cannabis. The sale cannabis is illegal under federal law and Distinguished law professors far beyond me that pointed out that the theory of state nullification is not as true in cannabis today as it was when James C. Calhoun put forward in the United States. It is a federal crime and the financial institutions are processing The sale of cannabis That is federally illegal It is a crime. Are these the real financial criminals? Are SARS the most effective way to catch them? I knew the number of stores around here because there's a really easy way to find them. It's called Google Maps You if you really want to go to the state capital you can go and get a registry of all of the ownership You mentioned beneficial ownership, which is a critical issue. We had an event in that and Brookings Carolyn Maloney has a bill that just has a lot of bipartisan support And I think it's important But this is actually an industry where the beneficial owners have licensed themselves at the state level and you can walk into any State capital and get the beneficial ownership registry for these activities easier than for anybody else Our filing SARS really the most effective and efficient way And by the way, here's an interesting fact The state of Michigan is deriving significant revenue from this which their deposit it often in cash Is the state of Michigan committing a financial crime? Colorado's deposited a billion dollars in revenue since 2012 from their taxation and licensing fees, etc from cannabis Ironically state and local governments are exempt from currency transaction reports or you would find a lot of these but on a broader level if you replace the state of Michigan with a large private person deriving revenue from Contributions of the sale of cannabis wouldn't they be a financial crime? What does it mean for a state government to get a SAR file by their bank because they are banked by someone? and so I Kind of flag this because it also to me illustrates this question of We're using our AML to go after the wrong folk and when you preclude cannabis from banking which happens a lot What do you end up with you end up with cash on businesses which attract large amounts of crime by excluding them from the financial system? You have created the a lot of crime that the voters are attempting to mitigate through decriminalization This is just as true for cannabis as it is for other people who seek to access the financial system And it gets to the final tension in my in my talk which it revolves between this goal of Using the financial system to catch bad guys which is an inclusive Concept in which you actually want the criminals in the financial system So that the financial system can report Right that is a more effective way than having the cash carried physically on airplanes and having TSA try to Find the physical money verse the idea of Excluding the criminal systems because you don't want the criminals to ride the rails of our financial system and be able to Benefit and conduct crime more easily because of banking and there's an inclusion exclusion tension in a paper I wrote with with Michael bar and Karen Gifford. We kind of take the I think belief that there's a win-win outcome For enhancing financial inclusion through AML reform that has two kind of core elements behind it one is More fully leaning into the idea of an inclusive financial services sector Which then provides information to allow more efficient and effective law enforcement To catch the real financial Criminals at the same time by enhancing financial inclusion Reduces the real crime that occurs due to financial Exclusion right because what in a mausoleum's hierarchy of crime It's not clear to me that financial crime is better or worse than robbery assault other types of theft in this situation You can have tremendous savings of the amount of money spent Tracking other types of financial crime or trying to exclude people from the system and plow some of that cost savings into more effectively and efficiently Serving lower cost consumers one of the earlier panelists talked about the radical decrease of cost of onboarding customers due to technology from AML and The final point and I'll close my presentation with this, which is that? multinational financial Regulate not regularly but organizations FADF BIS global coordination have a huge role to play in this because they can set standards that allow individual countries to harmonize and prioritize crimes in criminal activity to Move central banks so as I think of what the central banks of the future will be plural I hope they will be working together with a common set of objective standards and goals in both financial inclusion and AML and have those divisions and Core objectives more intertwined because a financial system that includes more people is a more effective way to Reduce total sum crime, which is I think the policy purpose of all of this. Thank you Thank You, Aaron So I'm gonna put you both on the spot. I'm gonna throw the script that we had out the window And let's just talk because we already answered the first couple of questions. We had had thought about outlining So the bad news is the system we have now is not scalable There's no possibility that you could spend enough money to scale it up and make a dent in the problem The good news though as you and you said it so well is that more data Can be a solution to this if we can figure out how to manage it safely and and accurately So what I want to ask you both is for starters if you Could do one thing on the policy side Thinking of yourself as a central bank of the future that had the ability to reform this system And that would impact financial inclusion as well as the crime. What would be The most high-impact thing we could do Jen, yeah Well, I think I mentioned two already, but if I had to pick one of them its information sharing and having Abeling Those in the system who need it regulated institutions government To have the data at their disposal to understand the risk Ensuring that information is shared cross-border As we look at the world becoming more and more fractious more disputes between on trade and other issues We see it playing out in the data realm as well and we see a rise in data nationalism I think that undercuts both the financial inclusion and the Financial crime agendas so to do that. What is how would we go about that? How can we share data more widely and safely so Aaron referenced? The FATF already and as the international standard-setting body and the financial action Taskforce I'm declaring an acronym for you down I will define all acronyms for now So it's the the international standard-setting body for anti-money laundering and countering the financing of terrorism and Pretty much every nation In acts into law the standards that they set They have put some things out in guidance around information sharing What probably need to strengthen that and I think central banks play a role in enacting the laws that are encoded in that guidance and hopefully recommendations to enable information sharing This is tricky because there's some kind of low-hanging fruit solutions like beneficial ownership in the US which is a Gen mentioned, but I'm gonna try and think think big. I was motivated by the morning keynote to think think big I would try to solve The identity issue. I would try to have a globally accepted easily accessible form of identification for this which I think would simultaneously Do wonders for financial inclusion in terms of lowering costs to operate and reducing that long-list barriers and Simultaneously reduce the cost of AML compliance Systems as well And I think that's a really devilly dish difficult problem. I mean I'm reminded, you know taking a Provincial view of the United States The United States has no federal identification system beyond a social security number Which I think is somewhat easily accessible for all of us in terms of hacked and leaked information and For the 38% of Americans who've chosen to get a passport Your passport ID For the other 62% of Americans who don't have a passport That's it. We have an identification system at the state level Which is just a world of problems. If you talk to the sophisticated FinTech They and to the sophisticated banks to they will tell you that the the rules that we have for identifying people so name address Social security or other government-issued numbered and so on That doesn't tell you anything you you have to comply with that But then you have to go and actually Gather the data to figure out who whether people really are who they say they are and where they I mean it's just it's a relic Yeah, it's it's Putting together someone's identity So again for a bank in more than 60 countries and we often will bank someone in more than one country We often don't realize that we're banking the same person in more than one country once you put together the difficulties of Identification and the elements that that we piece together to to understand identification and data privacy laws We can't even understand our own customers and more than then one jurisdiction You can imagine the frustration when you bank with us in the US and then you come to London and want to talk to us And we don't know who you are because we don't even understand that you bank with us. It's getting better. I think it's you know Where we and other institutions are doing more and more to piece that together But it's it's nowhere near where it needs to be and then when you try to put on The ability to detect financial crime in that kind of context or or the ability to have Enable inclusion and bring costs down Not what you would invent. I think that has a digital idea on their agenda this year And China is chairing it this year. So I know there's a lot of focus on whether there can be some progress on that front So where does Use of machine learning and artificial intelligence come in to solving this problem We have all this data. We can't we don't know what it's telling us. There's too much of it. It's inaccurate and on and on How do you how are we going to use machine learning? So we are thinking about it and we do use different elements of machine learning now Amongst the different branches of artificial intelligence machine off focus on machine learning Although natural language processing that is probably pretty important as well But in that space when we're trying to understand if a customer poses a financial crime risk, we essentially want to know Do you have any indicators red flag indicators that we know and governments have told us our indicators of you being a criminal? So we want to look at that. We want to know how is your activity? Changed over time and are there any anomalies that would suggest your your engaged in financial crime We want to know how you compare to your peers And if you're an outlier if you're a flower shop Do you have a lot more money moving through your flower shop than any other flower shop that kind of question? And then we want to understand everyone you transact with And whether that tells us anything about your risk to put all of that information together and under with an overlay of What's the probability that you pose financial crime risk to us today? We start needing to bring in machine learning to be able to do that at scale to do it dynamically to update Number one to process the pure volumes, but you don't have to have machine learning to solve the volume problem But you do to start doing those volumes around unsupervised techniques where you're looking for anomalies and outliers And so we use and are increasingly looking to use machine learning in those contexts Which means that a big issue for us these days is data ethics And watching what regulators and central banks and other governments are putting out there on the topic of data ethics I think Singapore's put out some interesting Guidance in that respect and developing our own internal principles and practices to make sure that we're comfortable doing the things We do and want to do with data so I'll have a little fun and kind of raise my concerns and flags on false pause on AI and ML Issues in part because the I think they're two core issues. One is what is the mission the machine? What is the machine trying to accomplish and goal right AI and ML are fantastic at finding new Ways to accomplish a mission that wouldn't have occurred to you or I or a person right go visit six. What's the move and go? The famous AI the game. Yeah Yeah, I think it's moved 26 right on the other hand the machine beat the human go On the other hand We don't know what we want to accomplish out of a out of ML right when you said the flower shop The first thought that came to my mind was the slide behind me Which is I know exactly who's selling flowers and generating an inordinate amount of money. It's the one selling the cannabis flower and like You know that we don't have a system They could detect and find a lot of things that maybe you know, maybe they could find a lot of Denny Hasterts But it could potentially Target those scarce human resources to check that question. Yeah Well, perhaps or a different way it could generate a tremendous number of false positives, right? And so so improve on the false positive front and but I take your point on the prioritization so if you if you start to have An ability to find more financial crime faster and to understand in more detail what types of financial crime you're finding now You'd have an ability to take Direction and pointing from governments that say you know what we only care this year We're gonna focus on these types of financial crime give us anything that you bet So but but let me let me flip it a little bit in terms of proxying and Discrimination and inclusion issues, right? You know if you take as a given people from a certain background or national origin or last name, etc. Are higher risk right, but you at a financial institution would never Purposefully discriminate or have a process by which you say well We don't really want that kind of person because that's clearly Unethical getting back to your data ethics point now the AI ML says no no It's not that kind of person. It's a person who Subscribes to this, you know totally weird thing or eats at this restaurant, right? so like suppose you're really concerned about Somali people in Minneapolis and You know you you're trying to figure it out and it comes to shop at this kind of grocery store And the AI flags this is being correlated with being Somali and maybe the person sitting behind the desk On the other side of the world recognize that or maybe they don't and the institution starts to say well look our customer base It's causing us a lot of cost Has this come thing from a ML. So let's shy away Let's not target our ads for services to people who've shopped at these stores so that's I Think you're right does come really back to Ensuring that we're thinking about the ethics in this both in terms of bias on the front end So I think you're speaking of bias. So what elements and data are being used to make a decision Can you explain the decision was that was made so you have explainability around your models? And then the outcomes regardless of whether you can do that either are the outcomes still in some way biased or discriminatory Which can you could have back testing or independent back testing to take a look at? In our discussion questions in the pre-conference yesterday one of them was would we Support having central banks acting as the KYC know your customer utility And the central repository of data that could be a trusted Source what where do you to ball in that whether we could have central banks is the KYC utility? Absolutely. There's no question why that we Whether we could do it Should we should we I? Think there's a lot of positives actually so The danger for central banks is the one of the many is the Question of having to manage some of those risks of taking on that service Responsibility working through the conflicts of offering a service while also be for central banks that also are supervisors and regulators but those could be dealt with and Thinking about I go back to one of our speakers earlier today saying how many tasks do you want central banks to take on? Should they be more narrow and focused should they take on several tasks? I think those are some of the the policy Risks and challenges to think through but could a central bank do it? Would it be helpful to have a centralized utility in this sense a government back central bank take on that responsibility? I think it could show some promise so I Would I would tend to agree with all of those Points I think there's a lot to be gained from it particularly as it as it would lower cost for for industry as a trusted central Repository I do get a little concerned about the tension that Jennifer mentioned about Size of mandate, but I also get concerned with another tension, which is independence the central bank of the future to me needs to be an independent entity in this conduct of monetary policy and If you're independent in one sense then you tend to be independent in others And if I'm a registry of all of my customers and I won't use the US although I think one could easily see the applicability, but suppose I'm in England and suppose this is a dystopian future like that show years by years, right? and You know they're trying to find all the people who illegally stayed after Brexit Hmm, and they want to go to the know your customer Repository and most of the people illegally stayed or probably engaged in international money transmission because they stayed because that's where their job was in their family And so do you know what happens when the immigration department calls up the central bank and Says give me your list How independent is the central bank? How independent do we want them to be from an AML national security government perspective and you know Communities are smart if they know that getting a bank account leads to potential Threatens they'll leave the financial system and that's attention. Yeah, I would offer and I'm gonna open this up to questions in a sec here that I think there's a lot of interest in instead of using a central data utility for KYC to adopt what people are calling the traveling algorithm leave the data decentralized and and create an algorithm that can go analyze it With in an encryption format Keep it anonymized so that the machine doesn't need to see that the words That we need to see to look at information So you could potentially encrypt it in a way that it doesn't create a central honey pot But there is an ability to go look for patterns financial crimes have typologies That machines often if they have enough data can recognize the human trafficking looks different from drugs and so on So I want to ask you about Blockchains, but I want to open it to the To the group first and if we don't get questions, I will ask you about blockchains Chris Hi Chris club from the Gates Foundation. So before I entered philanthropy I was on the front lines writing enforcement actions against institutions that had failed to comply with US laws and regulations I don't know my no laundry and often it wasn't necessarily that they had actually laundered money It was that they had fallen the foul of the standards that were expected and those standards are quite high and demanding If we know that things are not working that well What in your minds are the biggest blocks to reforming them? Is it that there's a small group of people who think that actually the system works? Well, and we just need to push it harder or is it their fear of doing something different? Or are we afraid to think about new ways using using technology and so on But what do you think are the biggest blocks and that we should be focusing on? I think at the moment a lot of the energy is focused around Optimizing the current way of staying Regulatorily compliant Which is not focused on an outcome of identifying financial crime But is focused on an outcome of doing what my regulator expects cheaply So I think we need to really change Collectively our mindsets around focusing on outcomes and using the technology and Opportunities it provides to to improve our outcomes all together on the on the kind of Across I guess across the whole system. There's the fear of the change fear of the unknown What do you mean data has to be put in a cloud to give you the the? Processing power to enable this kind of compute We're not sure how we feel about clouds yet. What do you mean? You need cross-border information sharing We're not sure how we feel about that. What is this AI of which you speak? I'm not sure I understand how it works and I've heard really scary terms like black box and unsupervised learning And that doesn't sound like anything that a regulator should be behind So there's a bit of a fear of the unknown. I think there's a need to continue to To educate and to focus on what is the outcome we're truly trying to achieve So I had this weird Epiphany experience. I was discussing this issue with somebody from law enforcement. I was going on and on about the rise in SARS and He said well, you know, I don't see what your problem is I said well each of these is costly and there comes a point where the marginal cost exceeds the marginal benefit, right? I mean, you know That's all an economist says repeatedly when posed with any question And the law enforcement said what do you mean marginal cost? There's no cost to this and I said, you know, you know and all the I said, yeah Where do you think it came from? He said why it doesn't cost, you know, it's in the database It's sitting there. I can use it whenever I want. It's of no cost to me the more things in the database the better and That to me was the fundamental Mindset problem, which was from his perspective. There was no marginal cost to additional SARS and From my perspective There's somebody out there who wasn't included in the financial system because the cost of onboarding him was greater than the Expected economic return that that person's account would generate And I would add to that. I think you're exactly right and in speaking with Regulators AML regulators former colleagues now as a banker, right? I'm on the other side There's just a lack of trust when banks say The amount of money we're we're spending to produce this are isn't getting the outcomes Yes, we'd like to spend less money, but like let's not even start there Let's just talk about putting money towards something that would have produced the better outcome the more effective outcome There's so much suspicion as to the motivation of the financial industry that there's a fundamental Failure and trust between two major players in a system that are meant to be working towards the same goal So there's certainly a challenge there to overcome one of the things I covered in my Harvard papers is this is the single most Expensive compliance area and the most dangerous for the industry So it's a huge deterrent the fee the industry is not willing to take risk in this space And they're not willing to cut money even if it might be more effective for fear that they'll be criticized Having said that I'm I think it is going to change I think there's a sea change sweeping over us with people realizing that there's better tech We had a question we had questions over here Did Okay So Aaron your point about what kind of crime are we trying to? Solve for I Actually want to hear you to answer that question I Recognize this isn't necessarily a central bank question But I want to posit the idea that if what we want out of this is not just to Reduce the amount or solve the amount of financial crimes, but we want to Figure out how to do it in a cross-border fashion because it's going to be more efficient Etc. It seems to me that stopping terrorism It's kind of a good One to think about and it's part of the it's I would say the big reason why at least in the US We've ended up where we've ended up right But terrorism is not it's clearly not a US only problem and so I'd like to hear your perspective on that but also if Jennifer just from your perspective You've also been saying that different kinds of crimes Financial crimes have different kinds of patterns And I'm wondering whether you think that the way that SARS worked today and the $10,000 limit and all the rest of it If that's actually effective if we said we want to we want to focus on Reducing terrorism like what would we need to do if that were indeed the Target, what does that look like in terms of a trumpet transactions perspective and behind the scenes perspective? for terror terrorism fundamentally looks different It's changing too so We had a time period where you had financing Big financiers of terrorism That came external to the group then you had groups that took over Territory and earned their own money from the populace and took money from by Breaking banks is that are breaking in and actually taking cash out of banks Which is a fundamentally different funding model and then you have the problem of terrorist fighters people who travel all around the world to engage in Terrorism and so each of those has a different kind of financial footprint to them Generally speaking if I was going to talk about what's the best? How should we change the CTR to help with with terrorist finance? I would say go to a zero threshold and Give international transfers at a zero threshold Provide that data. It's something that already happens in Australia Canada has a thousand does the same thing but has been a thousand dollar threshold But I I want to take it back to your initial question is is that in fact the top crime? We should be focused on it's it's scary It takes lives. It has political impacts, but it's pretty low incidents. What about fraud? It's almost every what about elder abuse and people who lose their money through fraud. That's high Volumes of folks who who experience that should we focus there instead? What about corruption and grand corruption in it where you have? Nation states that fail because high-level political officials loot the country dry and move that money abroad Should we focus our efforts there? I think there's a number of good places But I agree. You can't be everything to everyone So if we really want to make an impact for good, I do think it would be helpful to have Governments prioritize I think central banks can play a role in Stimulating that conversation because you need to bring together policy makers enforcement and security services regulators into a conversation that to be willing to set forth a Combined view of what the priorities are which is you know, never an easy thing so Corruptions valid by debate. I think the next person to enter my slide Show is Paul Manafort Who had I think nine? Shell companies in Delaware without beneficial ownership and also among the other crimes that again One of the reasons is such a powerful tool for law enforcement is whatever you think about what Manafort did the Easiest crimes to commit convict him off were the ones where it was just very very clear illegal use of money so what we have is a Money laundering has become a tool for almost all generic forms of law enforcement For a wide set of potential crimes because it is far more cut and blank than things about intent and in other areas Elder abuse is a big deal I think I had a SAR filed on me for elder abuse this year Which was a an illustrative point and a totally separate story, but I'm glad they filed frankly I was kind of waiting for somebody to call me and investigate the situation I can't be positive, but I bet if you query the database, I'll show up in it and and and that's actually The one of the problems is that fit my category of crimes. It had been occurring, but nobody been reporting I don't think I think the rate of elder abuse was constant through that time horizon. Yes, demographically We've had an increase the number of elders, but there's been a hockey stick of filings on it that being said Jen It's unfair to ask me to answer my own question, but I will which is that number one is terrorism But I think Jennifer's point is right. Terrorism isn't necessarily 9-eleven. I think the San Bernardino shooter Could have been found in a traveling if the traveling algorithm were running around here You know each of it Here's a person maxing out all these weird sources of credit and then wiring all this money one way To Saudi Arabia and then you know a woe going to a gun store and all these different The financial thing that came out of that from the limited amount I saw in real time because the other question is do you want to use AML to prevent The crime which is a hidden in terrorism. I think we're trying to prevent In kind of minority report style Very different than the other ones. We're trying to somewhat catch Right system is currently set up in the AML and terrorism space It is an after the fact reporting system It is not live fraud fraud payment screening like it when you get the text is that really you buying the sweatshirt at the store That's that's live Screening Sanctions screening is live, but AML is not so I would also posit that the incidents of San Bernardino type terrorists may be higher in the future Then the other I could be wrong number two would be sex trafficking. I just that's a mausoleus hierarchy personal choice You know in terms of you know crime Number three would be high net worth tax evasion. I say that as a public finance economist who you know You want to raise revenue as efficiently as possible and then spend revenue for for valuable purposes There's a lot of high net worth tax evasion The AML system I think the audit rate is is now equal for the bottom 10% of income as it is for the top one right with a marginal benefit to society to catch a billionaire cheating taxes versus You know somebody earning minimum wage. I think you generate a tremendous amount of revenue that could be used for other social good and purposes But that that would be my three, but that's a policy choice, and I'm not sure the central bank is The entity that we should have in our political system to rake those three over a plethora of other positions Organized crime for example I'm gonna say I think we were gonna take one more. Is that okay? Go ahead. I Don't know the history of this field. So I'm just curious how we came here I was grappling with the first principles question in the field of privacy. We have a very important concept that Liberal democracies do not run mass surveillance systems the United States government would not be able to Screen every email or every text message or run searches inside it or store copies of it You would need a law enforcement Agents would need a warrant to go into Paul Manafort's email and That would be based on some other information that you'd have to show a judge and you've got that whole machinery of Privacy, why do we not see any of that around finance? Why are we running a mass surveillance system on finance? Yeah, I think it's a great question and it's definitely Policy makers choice on where they wanted to set the balance on the spectrum between Data privacy at one end of a spectrum and security at the other and we've seen that where Different countries and communities come out on that issue change over time and it come across different ends of that spectrum and certainly in response to to Different prompts when something really bad happens on the security front We move towards greater security and more surveillance Requirements on financial institutions When we have significant data privacy concerns of post-snowden something that like to that effect and then it moves in the other direction But those are policy choices made by governments that financial institutions then and are required to implement So one reason I support raising the currency transaction threshold to its indexed amount Which is somewhere between 50 to $60,000 depending on what you want to index it for is Exactly that which is this question about what level of information should be allowed to be found What level right because ultimately this is an incredibly powerful tool for conviction purposes I mean people can do structuring of financial situations and somewhat unknowingly one of the one of the examples of somebody didn't make the hall of fame is Bob Dole Bob Dole got caught for structuring at Riggs Bank. Those of you who know Washington This was like the preeminent Abraham Lincoln's Bank and Bob Dole for whatever reason Sarah Dole from like cash and Riggs Bank ultimately went down over an AML violation. Essentially they got bought by by a competitor But I think it was Saudi Arabia or something some Middle East country They were doing some really bad shady stuff like there's some really bad stuff But in this deep dive that came out it turned out that Bob Dole liked to have cash and every Friday He withdrew like 8,000 bucks Or every other Friday and if you go to the bank and withdraw $8,000 consistently over 10 years You were supposed to be filed on But the bank looked at the you know Senate majority or minority leaders. We are not gonna file a SAR on on This powerful senator who you know, I think we can all realize just happen to pay for things in cash right and so This kind of history of it and the judgment of it is gets to be very tricky And I don't think there's enough of your voice and that position being heard in the general debate And instead there's a tremendous amount of the the SAR database is free for me Put as much as you can there because I you know could use it to stop the next 9-11 or Stop catch some really bad guy or girl So law enforcement tells us that the average price to buy a human being in the United States is $7,000 that's under the CTR So this has been a fantastic discussion. I think we're gonna have to figure out how to use this data It's gonna get used and our mission is how are we going to regulate that make it safe? How do you process around the use of it? But I think these are just the urgent questions of our time and I cannot thank you both enough for all your comments So please join me in thanking