Bank 4.0 Page 7

  The reality is that Bitcoin has a design problem that prevents it from being the first truly digital, global currency—and that is the current trend of hoarding Bitcoin3 and speculating around its possible future value. A ton of Bitcoin holders persist in the belief that Bitcoin’s scarcity is one day going to drive the price of a single Bitcoin up to $100,000 or even $1 million, and so they hold on to it like Gold or Apple Computer Inc. stock, waiting for the right day to cash out. This behaviour significantly undermines the likely viability of Bitcoin as a pure currency for now, because no one wants to spend it. As a result Bitcoin has very low utility compared with fiat currencies like US dollars. This could be perceived as a design flaw, but primarily this has emerged as an issue through user behaviour.

  Unless Bitcoin crashes in value, finds some stable level of value that means it behaves more like paper money, and then people start spending it again, Bitcoin will likely act as a learning foundation for a future digital currency somewhere down the line that will be even more disruptive and ubiquitous.

  At the time of writing there are roughly 4,500 cryptocurrencies and altcoins to choose from. If regulators attempt to outright stop BTC, ETH, XRP and others, then their markets become unattractive for investors and entrepreneurs alike.

  The development of autonomous networks, smart contracts and smart assets and infrastructure is likely to involve the creation of new methods of value exchange optimised around platforms. For example, the Sun Exchange, a solar start-up on the African continent, not only allows you to buy solar panels for local villages using Bitcoin, but the return that is generated off each kilowatt of energy produced is captured in a ‘solar coin’. This is a template for the sort of optimised value exchange systems we’re likely to see in the smart world. Artificial Intelligence, the blockchain and the need for smart contracts will lead to IP-optimised value exchange systems that circumvent currency controls by necessity.

  If a regulator inhibits cryptocurrency models or blockchain deployment, by necessity their economies will start to slow.

  The Decentralized Autonomous Organisation (DAO) and ICOs

  The DAO is a computer code-based venture capital investment fund or organisation sitting on the Ethereum blockchain. With no typical corporate governance structure, it used a cryptocurrency called Ether for its underlying asset architecture. Technically the DAO was an AI-based, or automated smart contract between the community members participating in the experiment, but from a regulatory perspective it was, well, an absolute nightmare. It represents a template for future business operations that will be commonplace in 10–15 years, and yet current regulation would prohibit these types of entities in most global jurisdictions.

  The DAO doesn’t have a management board, a charter or business license. As a proxy for a modern corporation, AI-based or not, it is technically illegal. There are no officers of the “corporation” to hold ultimately responsible for decisions executed within the DAO’s operating structure, which is entirely built in code. There are no by-laws, no management structure, no employees and no governance, there’s just code that executes instructions. There was no CFO or CEO who took fiduciary duty on the financial governance of the company. No one who could be sued for breach of tax laws. In fact, the DAO operated without any conventional revenue sources or income, so technically would not have paid tax either.

  From an investment perspective, while the DAO creators or programmers did explain that investments of Ether coins (the underlying altcoin behind the DAO) carried risk, this investment approach breached securities rules in most developed markets. Investors that invested in the DAO didn’t do so through a registered investment fund, they didn’t receive advice from an investment advisor, and there was no securities commission oversight in respect to the investment process. No risk profile questionnaires were done, no signatures accepting risk were received. That didn’t stop $150 million worth of Ether being deployed within the DAO’s engine in the first weeks of operation, and given the increase in price of Ether, we’re now talking well over $1 billion of investment in today’s terms. A very significant investment pool indeed.

  This investment in a blockchain startup using cryptocurrencies, was not the first such unregulated, crowdfunded coin offering—that honour goes to Mastercoin in 2013. Unless you are living under a Wi-Fi barred rock today, you’ve heard the buzz on ICOs or initial coin offerings. Essentially Mastercoin, the DAO, Ethereum, Blockchain Capital4 and a plethora of others have all raised billions through cryptocurrency based, over-the-counter trades or issues of ICOs.

  Technically considered by law to be a securities issuance (at least in SEC terms5) as of the first half of 2017 we’ve been seeing about 30–40 new ICO offerings per month. Will the SEC, FSA, MAS, HKMA, ASIC and others permanently outlaw ICOs? Will ICO creators be fined for being in breach of securities law?

  If regulators want to encourage capital flows and investment in entrepreneurial endeavours, shutting down ICOs wouldn’t encourage the free movement of capital, it would restrict it for some of the most innovative startups in their space—risks to consumers aside. The bigger problem is that ICOs could essentially be issued without jurisdiction, making enforcement a sticky problem. A startup that is incorporated in one country, operated in another, with a coin offering taking investment from investors in cryptocurrencies all over the world, could simply move their cryptocoins to another jurisdiction without any operational impact. The only issue would be cashing out their coins. However, if staff, contractors and suppliers are willing to accept altcoins instead of cash, this would be virtually unstoppable.

  This doesn’t mean regulators won’t try to stamp out ICOs (the SEC are making it very tough to legally issue tokens). It does mean that doing so might present significant challenges at law, and may make the markets where cryptocurrency-based offerings are illegal, far less competitive and attractive for startups. A regulator that want’s to encourage rapid innovation and investment in the emerging FinTech ecosystem would be much more likely to take a light touch on ICOs from a regulatory perspective so that both funding and innovation keeps flowing. A regulator that outlaws ICOs would be heavily limiting its options at the market level for participation in the future of financial services, as the best innovations that get the most ICO-based funding would simply flee their jurisdiction. More on this in Chapter 5.

  The underlying legal problem with the DAO as an AI-based company is that it wasn’t a business subject to human laws in the traditional sense; it was a construct operated via instructions in its code base. While law is how we get our codes or ethics as humans and as corporations, the DAO essentially used its code as law for the internal operation of the business. For machines code is law, for us humans, law is our code. When we mix up these concepts, we wind up with situations like the DAO, which doesn’t fit any of our current definitions of investment, companies and risk.

  Ultimately a flaw in the DAO smart contract allowed certain programmers to siphon off one third of the DAO’s value held (roughly $50 million). Thus, it has been framed as a failure by many, an experiment that came to nothing. However, we’ve learned from the DAO experience—and it’s very likely not the last AI-based company.

  Regulators might want to stop future instances of AI-based corporations or smart contracts like the DAO, or force the programmers who write the code to be held to securities law and provisions for regulation of investment businesses, but that would be a mistake. It’s clear that in the future we’re going to have more and more AI-based execution of smart contracts, particularly as trading floors disappear and are replaced with code. In that instance, any regulator who bans AI-based platforms like this might find their market woefully uncompetitive in a world where AI execution is becoming increasingly normal.

  Will AIs trading in the future have to pass a FINRA Series 7, OFQUAL or SFC licensing exam? When code is executing investment decisions en mass, will we still be insisting financial advisors are licensed and leaving code to run amok? Giving preference
to humans because they’ve sat some licensing exam isn’t the answer either. As we’ll learn later, robo-advisors are already performing at a similar level to human advisors and will likely exceed them for general portfolio management in the next few years.

  A flawed approach to financial crime and KYC

  Almost 30 years ago6 the Financial Action Task Force, a body attached to the OECD and sponsored by the G7 member governments and central banks from 37 countries, put in place 40 recommendations on combating money laundering (and nine on terrorist financing). These recommendations are now enshrined in law in major financial centres around the world. Amongst these changes were the requirements for banks to report suspicious transactions that could indicate money laundering. The definition of these suspicious transactions were in themselves a little problematic.

  If a financial institution suspects or has reasonable grounds to suspect that funds are the proceeds of a criminal activity, or are related to terrorist financing, it should be required, by law, to report promptly its suspicions to the financial intelligence unit (FIU).

  —Recommendation 20, The FATF Recommendations (2012)

  Banks find themselves today being the unwitting police force for a global anti-money laundering machine that is amazingly ineffectual.

  The AML (anti-money laundering) laws contain three elements. First, banks must authenticate the identities of people opening accounts under complex “Know Your Customer” or KYC rules. Then they must monitor customers’ transactions to understand normal patterns, including cash handling, and detect anomalies. Third, they must investigate anomalies and if necessary, report them by filing Suspicious Activity Reports, or SARs.

  Despite partial automation, these efforts are still configured largely as they were from the beginning, in an era when the best way to find and report suspected money laundering was to have a bank teller or analyst fill out a form. Not surprisingly, they are failing.

  The United Nations reports that financial crime today amounts to two to five percent of global GDP—as much as US $2 trillion annually—and that current AML efforts catch less than one percent of current illicit financial flows7. Unfortunately, that miserably ineffective result comes with an enormous price tag for banks. In the United States alone, banks spend an estimated $50 billion collectively each year on AML compliance alone8. This means that to catch all the financial crime that happens today would require spending the annual equivalent of the GDP of the United Kingdom just on AML policing. The current model is not scalable, nor is it effective.

  AML compliance costs are compounded by massive risks for banks—fines have exceeded a billion dollars for a single institution.9 Fear of aggressive enforcement encourages over-reporting of suspicious activity, which in turn mires law enforcement in low-value data and impedes detection of serious crime. In the United States, major cities have inter-agency task forces that print out reams of SAR reports each month and gather around tables laden with stacks of paper and yellow highlighters, searching for meaningful information. This is probably why the reporting is so ineffective. The hard end of enforcement comes down to yellow highlighters!

  Another unintended consequence of this low-tech system is it can block whole sectors of the economy from financial access due to regulatory mandates for “de-risking”. Customers whose industries, locations, and circumstances are potentially high-risk are screened out simply because, under the current KYC procedures, banks find it simply too difficult, too costly, and too risky to accurately sort the law-abiding people from the bad. This is a major concern of policymakers in the developing world.10 In the US it has been called the “new redlining”.

  One side effect of this simple process today has meant that some laws have rendered entire populations of customers unsustainable. The US FACTA (Fair and Accurate Credit Transactions Act of 2003) policing provisions require banks, wherever they are in the world, to report to the IRS when they onboard a US citizen as a customer. This has simply led numerous banks around the world to reject any US citizen even applying for a bank account11.

  AML challenges will continue to mount.

  Evolution of the payments space means regulators will have to deal with increasingly diverse types of value stores and payments vehicles, many of them that fall outside of regulation. For example, consider Bitcoin, Ether or XRP coins, Starbucks or Xbox credits—if someone transfers more than US$10,000 using these, they currently wouldn’t be reported in a STR12. In China today, 90 percent of mobile payments run across Alipay and Tencent’s WeChat network, and the trillions of annual payments flying across those networks today are virtually impossible to monitor from the reference point of a traditional banking system.

  The process of reporting a US$10,000 transaction that falls outside of predicted patterns is woefully ineffective at finding money launderers today. What we really need is a system that monitors flows of funds, looking for patterns where those funds converge. This requires an AI-based monitoring capability at a minimum on a country-wide basis, but probably on a global basis with coordination between different authorities. Such a transaction flow monitoring system would, on an aggregated basis, be much more effective at finding where money laundering is taking place and the identity of the players involved than the current reporting system.

  New technology throws into doubt even the core logic of AML regulation. The system is designed to keep criminals and terrorists out of the financial system, but at a time when arguably technology should be deployed inside regulators for sophisticated data analysis that could be applied to detecting, monitoring and catching them, we are policing using paper-based reporting and human eyeballs. According to a recent University of Chicago report, in their “generous” assessment they estimated that only 0.2% of money that is laundered is successfully seized. That means for every dollar caught by AML regulation today, $499 is still successfully laundered. We’re spending globally $50–100 billion dollars each year for a 0.2% success rate in AML. That’s appallingly ineffectual. Massive regulation, billions of man hours and efforts expended, customers disrupted and accused, regulatory enforcement action take, and it simply doesn’t work.

  The technology already exists to make AML efforts effective and efficient. The system needs updated reporting designs and norms, greater sharing of secure data, greater information security, faster and more accurate pattern analysis, and tools that remove manual work and free bank specialists and law enforcement. Some countries, such as Singapore, are exploring creating a shared data utility between government and industry for KYC. More regulators will need to engage.

  Current KYC laws are a path to exclusion

  But let’s think about KYC moving forward for a moment. While Uber has been making some big losses the last few years, things seem to be on the up13. In Q2 of 2017, Uber’s bookings were up 17 percent, and they were up 10 percent in the quarter before that, with almost $9 billion in revenue. For the first 20 years of Amazon’s life it made a loss, so it appears Uber’s investors are willing to trade off losses for growth for now. But as Uber grows it is definitely changing driving habits for millennials in particular. My daughter, Hannah, is 17 now and when living in New York it became clear she really had no intention of getting her driver’s license; in fact, when I talked about getting her a car she said, “Don’t bother Dad, just give me an Uber allowance.”

  As autonomous vehicles kick in and as services like Uber abound, the expectation is clearly that our sons and daughters will be driving less than we did. A report from the Frontier Group in the United States in 2016 showed the six-decade-long driving boom seen in the United States is already over14; based on other factors, Uber will just accelerate this decline.

  The Driving Boom of the second half of the 20th century coincided with rapid economic, cultural and demographic changes in the United States. Those changes largely pointed in the same direction: toward a more automobile-oriented society. Many of those trends, however, have either reached their natural limits or have reversed direction… those
trends point to the conclusion that the trajectory toward increased per capita driving that prevailed during the Driving Boom has likely reached its end.

  —Frontier Group assessment on the Future of Driving in the United States

  Figure 2: Less miles, less drivers, less driver’s licenses for KYC.

  Thus, when we combine lower incentives or tendency to drive, the increasingly ubiquitous nature of shared transportation services like Uber, and the medium-term impact of autonomous or self-driving cars, one thing is abundantly clear: less drivers means less driving licenses, means less identity qualifications, and this means greater financial exclusion based on current KYC rules in markets like the United States15.

  In developing economies like sub-Saharan Africa, branching as a mechanism for financial inclusion has stalled. Research from Accenture and Standard Bank showed that 70 percent of the currently unbanked on the African continent would be required to spend more than an entire month’s salary just to physically get to an available bank branch16. Something similar was found in India. Initially the Reserve Bank required banks based in India to deploy a minimum of 25 percent of new branches in rural areas to focus on unbanked customers; however, this measure didn’t affect inclusion significantly enough because of the inability for the unbanked there to meet account opening identity requirements. This is why India’s initiative to deploy the Aadhaar card has been such a boon for financial inclusion—it changed the game. As of 15 August 2017, more than 1.171 billion have been enrolled in the Aadhaar card program. That’s 88 percent of the Indian population.