The importance of artificial intelligence, particularly machine learning, was underlined earlier this week by a Refinitiv survey among over 400 banks and asset managers. It shows financial institutions are further along in machine-learning deployment than is widely known.
It did not address the question of how A.I. is governed. It noted that the majority of C-suite executives now consider machine learning a core strategy. Refinitiv didn’t ask them if they, or their data scientists, know how to explain a machine’s recommendation or decision.
Given the still-nascent use of A.I., such questions may not come to the fore yet. But they will.
Deployment, more and more
Refinitiv’s survey suggests M.L. is being applied mainly by buy sides (particularly in the U.S.) to manage risks, analyze performance and generate trading ideas.
DigFin, anecdotally, finds banks especially keen on A.I. for operational efficiency, sometimes with regulatory encouragement.
Tarrill Baker, chief data officer at HSBC, says banks are racing to deploy A.I. to meet compliance requirements, such as liquidity reports, or to deter crime.
Many aspects of finance could be driven by these modelsHuayi Dong, Daiwa Capital Markets
But various A.I. techniques will be applied to an increasing range of services, perhaps sooner than we think.
“As A.I. becomes more integrated with our lives, many aspects of finance could be driven by these models,” said Huayi Dong, managing director and global head of electronic trading solutions at Daiwa Capital Markets, speaking at a recent conference organized by industry association ASIFMA.
Such decisions will touch both institutional and consumer relationships. User behavior, captured by a firm’s algorithms, will help machines determine everything from allocations to whether to short a stock; from much to charge for insurance or whether to grant a loan.
Opening the black box
Right now surveys such as Refinitiv’s focus on the efficacy of A.I. For example, data scientists are most concerned about the quality of data inputs, or how readily they can combine unstructured data (news articles, research reports, transcripts) with structured data (financial and company information).
But if the data inputs are bad, can the algo still be used? If it is, and a customer doesn’t like the result, or a client’s trade goes wrong, or a chatbot delivers faulty information, who gets blamed? The computer coders, the machine trainers, the data suppliers, the portfolio manager or loan officer or sales trader, the CEO?
A.I. governance can be sliced in many ways but the essential question is: can you explain why the machine responded a certain way? This will have an increasing bearing on client relationships, brand protection, and legal liability.
We can’t stifle innovationTarrill Baker, HSBC
For example, Dong says best practice requires data scientists to understand how the computer’s model works. “The results should be replicable, and audited,” he said, noting firms should expect to reproduce the results before a regulator. The same goes for when a financial institution uses a third-party fintech: the firm needs to understand how the fintech’s models work.
One way to go about this is to only deploy an A.I. after considering the increase of value for the customer against the potential risk, says HSBC’s Baker. “Customer protection is paramount but we can’t stifle innovation,” she said.
Underestimating the consequences
Andreas Burner, Vienna-based chief innovation officer at SmartStream Technologies, a longstanding vendor to financial institutions, worries that data-science teams don’t understand the consequences of their use of A.I., particularly among fintech startups.
“If the risk in a use case has a regulatory impact, don’t apply deep learning,” he said, which includes natural language processing (NLP). “It’s okay if you’re using it for things like monitoring liquidity detection or identifying other patterns – things that won’t impact the market.”
Most finance applications should rely instead on techniques in machine learning that allow a machine, after a lot of data feeding, to do a regression analysis, generalize data, and detect patterns. Some of these techniques rely on mathematical formulas to vote on outcomes; others rely on massive amounts of data and vast computing power to fuel a data scientist’s algorithm.
Deep learning is a no-go for KYCAndreas Burner, SmartStream
These methods are more reliable, provided the data inputs are valid; but Burner reckons too many financial applications are being tested using more exotic versions of A.I. that could end in tears.
Startups using, say, facial-recognition software based on machine learning are making themselves vulnerable to fraudsters, Burner says: “Deep learning is a no-go for KYC.”
More human than it looks
Nor should machine learning be taken for granted as an arbiter of truth.
“Statistics are probabilistic; A.I. is a black-box model based on prior experience or training data,” said Jason Tu, CEO of MioTech, an A.I. company that sells products to hedge funds and investment firms. Citing the example of AlphaGo’s surprise moves in its famous matchup against Go master Lee Sedol, he said, “In financial services, unstructured data is more art than science.”
Take sentiment analysis around a piece of news, a tool that more fund managers are eager to use. The same news might be considered positive or negative by different people. Training data creates choices that over time will deliver different results. “It’s not just the algos,” Tu said.
Unstructured data is more art than scienceJason Tu, MioTech
For now, banks and fintechs say they are only deploying A.I. in cases that won’t hurt customers. Tu, for example, says his team backtests its models against human experience, so they can discard any unusual results.
In other words, there shouldn’t be room in financial services for an AlphaGo-like surprise, particularly if the outcome could have a material impact if things go wrong, either on customers or on a firm’s regulatory obligations. Financial institutions should retain human supervision for such cases.
“If we use A.I. for a marketing campaign and we lose money, well, okay,” said Shameek Kundu, Singapore-based group chief data officer at Standard Chartered. “But if we’re an insurance company and we make biased or wrong decisions, that would impact the customer.”
Who, or what, is accountable?
But fears over machines run amok may miss the broader point: governance is not just about computers.
In Singapore, for example, banks have been shown to have a lending bias against Malays. The same thing happens to minorities in other countries. That’s true of human loan officers today. Is a machine going to replicate that bias? Probably, if it’s trained with the same inputs.
At some point in the future, when computers achieve general artificial intelligence, they may need to be held separately accountable. But the industry is not there yet.
We have to have the same accountabilityShameek Kundu, Standard Chartered
In the meantime, financial institutions need to extend their standards to A.I. “We have to make human efforts,” Kundu said. “We have to have the same accountability, whether we train people or machines.”
That sounds like a tall order, given the recent history of financial institutions. The need for better oversight may not be specific to A.I. (just as the need for supervising computers isn’t an issue just for finance), so perhaps banks and asset managers don’t need new or special governance for these tools.
But the speed, frequency and scale with which new sources of data are being applied to decisions are going to test today’s protocols for accountability. At some point, someone will want to take the gloves off and really see what their A.I. can really do.
Startups from Finovate: the good, bad & ugly
DigFin’s entirely subjective review of recent pitches by fintech startups.
Finovate is a global conference organizer that provides a platform for fintech startups to pitch to an audience of VCs, incubators and accelerators. DigFin attended the latest Asia leg, in Singapore.
This is our take on a selection of pitches. To get the full exposure to fintechs, you have to attend the event! (Note: DigFin was a media partner to Finovate Asia but has no commercial relationship with the organizer or any fintechs involved.)
Unless stated otherwise, DigFin’s view is based solely on what was presented on stage. Here’s a roundup of what we liked, what we didn’t like, and what might be good but needs more information to judge.
Best in show
PearlPay – core banking systems for rural banks
This Philippines-based company is using technology to address real pain points among rural banks, helping these banks better serve their customers. Digital payments in the country is nascent. Of the 577 banks licensed in the country, 450 are rural, while 94% lack digital infrastructure and 99% of transactions remain based on cash or check. Rural banks are well placed to advance financial inclusion but lack the means to do so.
PearlPay’s business is to provide them with end-to-end core banking systems to enable them to reach the unbanked. What the company offers is basic, using Facebook-like interfaces to enable banks to set up new services. By digitizing basic services, these banks might be able to scale into branchless banking. One way PearlPay can help them scale (and generate fees) is by creating an online marketplace among participating banks to trade foreclosed assets, in a transparent venue.
The company’s software also acclimatizes customers to managing financial affairs via their mobile phone. The fintech founders note that Filipinos are digitally literate, as 75% of the population uses Facebook. PearlPay’s app is a Facebook feed to introduce people to the idea of an e-wallet.
The business model concept is relevant beyond the Philippines, so this fintech has the potential to scale.
DigFin’s view: This was the best pitch because it uses tech to address real problems, promotes financial inclusion, facilitate new markets, and can be applied to other countries. PearlPay’s success going forward is probably down to whether the team can execute.
Emotics – compliance analytics
Every bank has to train its people ethics. Compliance training is now de riguer but few employees enjoy it or take it seriously – and many people cheat, by claiming to have sat in on boring videos when instead they’re doing something else.
Emotics’s software aims to change behavior to improve compliance training. This is often more carrot than stick, however, such as using webcams and facial recognition to make sure people are at their desks actually watching the required videos. It helps compliance managers make sure people are paying attention, not checking the phones. The data is aggregated to give compliance teams a view of both where entire departments are in breach, or other red flags.
Emotics’s biggest value-add is it uses all of this data around how people respond to training videos, including facial responses and other clues to attentiveness, to figure out what parts of the training material are effective, or not – and thereby create better content that, hopefully, will engage people. And maybe even instill awareness of proper conduct?
DigFin’s view: Banks waste tons of money and time on compliance training, trying to instill ethics, etc. Maybe Emotics’s solution will help, if it generates meaningful feedback on the content. Its Big Brother aspects could backfire, though. And it’s not clear whether this is the most effective way a bank might perceive how it could spend budget on risk, when it faces spending pressure on cyber security, for example. To succeed, Emotics needs a sufficiently large bank to test and deploy its solutions, pour encourager les autres.
Worth a look at?
Comarch – SME banking
Comarch provides digital banking services to SMEs and therefore makes these companies “bankable”, allowing banks to better track loans, invoices, payouts, and so on. SMEs can manage their affairs easily via a mobile app.
DigFin’s view: The idea is good. There’s a lot of competition in this space, however. These services have been unbundled already, with players from accounting software providers to invoice financing platforms repackaging them, sometimes with a bank partner. Not sure how Comarch stands out.
Exchange Connect – equity capital markets
This fintech aggregates capital-markets data from sources such as FactSet to make it machine readable and consolidated. Its target is listcos, buy sides, and stock exchanges, with the aim to make compatible and connected a stream of actions, from pre-transaction documentation, to roadshows, meetings, and analyst transcripts.
DigFin’s view: Digitizing workflows makes sense. To make it work, though, requires building a network. The value adds up if there’s either a lot of buy-side demand or if a stock exchange decides to base its activities on Exchange Connect’s platform. A big if!
Fcase and Compliy – cyber security
Two startups aimed at helping banks automate risk management and regulatory workflows. Fcase finds correlations in a bank’s various departments and siloes to flag potential fraud. Compliy offers a searchable library of regulations and uses NLP to identify the most relevant clauses.
DigFin’s view: Banks should want to take a look. But there’s a lot of RegTech and cyber-security solutions out there. Banks are usually bad at working with these companies and pinch the best ideas for themselves. DigFin lacks the knowledge to tell whether Fcase or Compliy is top of the class.
ChintaMoney, FinBit.io, NewWealth and PayOK – personal finance software
These companies are in similar fields but in different countries: ChintaMoney is in India and PayOK is in Indonesia, while NewWealth and FinBit.io are taking a B2B approach to building clients across Southeast Asia.
Chinta and PayOK seek to be the Mint or Plaid of their markets, aggregating customer information to create holistic views of savings and spending, with budget and investment goals and advice. In India, Chinta can connect to the government’s Unifed Payments Interface, and it has a partnership with Kotak Mahindra Bank. PayOK says it’s in PoCs with local banks.
NewWealth has a multi-market strategy aimed at banks, not users, which sets it apart. It services a B2C wealth-tech company in Indonesia called Savio, providing a chat-app like environment to build personal finance goals.
FinBit.io appears to have the most sophisticated technology, and says it has over 50 fintechs, P2P lenders, and banks in Southeast Asia and India using its software to generate credit scores.
DigFin’s view: There are a lot of personal-budget apps and probably one room for one, maybe two, to dominate a given market. They can’t scale beyond the home market unless they develop a tech good enough to actually sell to banks on a white-label basis. So there’s a cap on the valuation.
B2B2C players have more scope to operate, but they run into robo-advisors and other types of competitors. This might keep a company such as NewWealth stuck servicing lower-tier banks unless it piggybacks off a Go-Jek or the like. FinBit.io appears to have the broadest reach, using machine learning to add credit scores to its end users’ activity, which will help its bank, P2P lending and fintech clients develop better, more personalized products. Worth a look but more detail needed on its actual deployment and business segments.
Hodlnaut – crypto wallet
Hodlnaut encourages bitcoin investors to hold their positions by offering interest on bitcoin placed on deposit. It calculates interest daily and pays monthly in bitcoin. Assets are kept in cold storage or are put out for lending, so liquidity with withdrawals is limited.
DigFin’s view: This is a VC-backed company but any investor would need a crystal clear understanding of where its deposits are kept, how they’re managed, what’s on reserve, what happens when bitcoin suffers a steady loss of value, how that interest rate is guaranteed, how liquid is the bitcoin lending/repo market, who are the counterparties holding bitcoins on loan, what kind of spread Hodlnaut is making… If you believe that the innate scarcity of bitcoin means you’re gonna lambo, baby, then making an easy x% per month on your deposits is very attractive. But read the fine print.
Xen – private asset trading
DigFin is cheating on this one. Xen made a presentation about “democratizing access and liquidity in alternative investments” like private equity. Its founders didn’t mention blockchain or tokens once, but DigFin interviewed them last year and this is the underlying tech. There’s no problem with not wanting to talk crypto: in fact it’s probably a smart move to keep the pitch on the problem and the outcome rather than the underlying gears and levers.
This is a bet on whether tokenization is going to be “a thing” (unless Xen has completely changed its business model). Xen’s particular insight is to avoid lazy assumptions about Libra-like “financial inclusion” and develop something for rich people. If you like these two ideas, you’ll like Xen.
Back to the drawing board
Libertypool – crypto
It’s an app to enable people to invest in crypto and tokens issued by blockchain companies, using a QR scan to connect to a US dollar e-wallet. A crypto custodian of unknown provenance stores your assets.
DigFin’s view: These apps have existed in the market for years and offer far more sophisticated offerings, from VISA-branded credit cards to VIP offers to trading, execution, advice, the works. What’s different about Libertypool? Nothing, except it’s been barely developed, at least as far as the pitch suggested.
Upgrade Pack – wealth management
This app allows wealth managers to acquire and retain high-value customers with flight and hotel upgrades. It buys unbooked airline and hotel inventory at the last minute, and uses the customer’s bank-issued credit card to purchase them at a discount.
DigFin’s view: Rich people don’t need to wait until the last minute to see if they can upgrade their economy booking to business class. Private banks catering to HNW give away flights and hotels, they don’t sell them at a discount.
This might work at a more retail institution, but then how many people would then need such a service? And there are other venues to get the same kinds of deal. This merely integrates it with a credit card. We don’t see this as a meaningful edge when banks are looking to go full-in ecosystemy with airline carriers and travel agencies, so the fintech is back to second- or third-tier banks…which again begs the question of what kind of customer base they’d have.
Finally if this is such a great idea, then Revolut or Grab or Zhong An can replicate it even more easily. [See below for Upgrade Pack’s response.]
Pocket Money – e-wallet/lender
This Singapore-based company wants to extend credit to the unbanked in places like Uganda, where its owners are located. It seems to be a gig-economy platform that adds credit, payments and interbank forex rates on top to its roster of service providers. The revenue model is meant to be based on advertizing. The founders are looking for advertizers, investors and technology partners to enter new markets.
DigFin’s view: Get the advertizers and tech partners on board and prove the concept before you ask VCs or incubators for money. The focus on advertizing revenue begs the question of where the founders’ loyalties lie. This feels very undercooked.
Letter to the editor from Upgrade Pack
We believe your description provides an inaccurate portrayal of our app, the value we’re bringing to the market, and in particular who will be using, and funding, Upgrade Pack.
Our app is provided free of charge to users via our clients (banks and credit card issuers rather than wealth managers) who fund access to the app as a loyalty benefit aligned with their premium account and/or possession of a given credit card.
We do not use user’s credit cards to purchase anything. In fact we don’t purchase anything – we merely facilitate a transaction between our airline and hotel partners with a closed user group who want to purchase upgrades at discounts not available elesewhere.
We’re currently in advanced conversations with over 30 banks and card issuers around the world about adopting Upgrade Pack.
––Louise Naqvi, head of public relations, London
Busting six myths about China’s e-RMB, part 2
There’s a lot we assume about the PBoC’s digital yuan – and we’re often wrong.
While politicians and central bankers in the U.S. and Europe wrangle over Facebook’s proposed Libra coin, one government is moving to seize the initiative: China.
The People’s Bank of China has been studying central bank digital currencies (CBDCs) for several years and probably has the greatest technical understanding of any public institution. Introduction of a digital yuan could come any day now.
There are a lot of unknowns and misconceptions about this, however. Here is the second set of three out of six myths about the digital yuan that tend to crop up in media, conferences, and cocktail conversation. See here for the first three.
A digital RMB would crush the payment rails of AliPay and WeChat.
The thinking here goes that by making cash digital, the PBoC would be reinstating the primacy of China’s commercial banks, or bringing the tech platforms to heel.
But these are different things. AliPay and WeChat Pay are examples of electronic money. They are in the same category as, say, Kenya’s M-Pesa. It is a means of payment wholly backstopped by a private company, limited to users and services on that company’s platform. A digital yuan is cash that is accepted anywhere the renminbi is legal tender.
If the PBoC wanted to use bank money to compete against e-money, it could, but the evidence so far suggests the opposite, particularly as the PBoC has said it will restrict a digital yuan to wholesale institutions, rather than let it run in the wild for all consumer payments.
The potential to compete against e-money is there. Uruguay’s experiment (see previous story) set its digital peso in direct competition with the likes of local WeChat users. But China’s approach has been convivial, with the government cooperating with its biggest tech players.
Notably, although use of cash by many Chinese people has declined, cash levels on balance sheets have not. Internet companies are basing their e-money on users’ bank account balances. So in the Chinese example, cash has already become a largely interbank concern – so digitizing it wouldn’t have a big impact on the relationship between cash and e-money.
What might change, however, is the way that e-money platforms compete, both which one another and with banks. One potential use of a central bank’s digital currency is to require compatibility among e-money players if they are to offer customers access to digital yuan. Right now, Alipay, WeChat etc are closed loops. If a digital yuan is restricted to interbank markets, this may not change, at least not for a while. It’s a bigger question mark if the government decided later it wanted consumers to pay with digital yuan.
A digital yuan restricted to wholesale (interbank) channels is no big deal to consumers or merchants.
So far it seems the biggest impacts of a digital currency are negligible because the PBoC has said it wants to keep its CBDC limited to the interbank market. So is this a giant yawn?
First of all, the wholesale market may be just a stepping stone. The PBoC hasn’t ruled out allowing (or encouraging) banks to disseminate digital currency to individuals.
How the central bank structures its digital cash is all-important. How much of the central bank’s liability can be lent out using digital currency? How much value in the form of digital currency is stored on the central bank’s balance sheet (i.e., reserved)? These things help determine the extent to which digital cash can be lent out, or if it is just used as a means of payment.
(Another question is whether the central bank issues these in the form of tokens valued at par, or if it creates a synthetic stablecoin to invest in a portfolio of cash-like instruments, a sort of proprietary/national Libra.)
Other technicalities that will impact use of a CBDC is whether it is just for interbank settlement, or if it is similar to cash (usable by anyone), or as a policy tool (to increase or decrease the money supply by making digital cash more or less appealing to hold), or as a deposit currency for accounts with the central bank.
Although the PBoC has indicated it will take the safest route and limit its digital yuan to interbank settlements, it hasn’t stated clearly how universal it wants its e-currency to be, or what goals it has in mind. Who will be allowed to access e-money in the interbank market: state-owned enterprises? Foreign bond fund managers? Domestic mutual-fund company portfolios?
CBDCs won’t matter to bitcoin or other private crypto projects.
Bitcoin enthusiasts might think a digital yuan that is limited to the wholesale market won’t impact public crypto-currencies. That is probably mistaken, although the impact will depend on how the central bank designs its system.
Even if the central bank’s goal is simply to do away with physical cash and encourage payments via e-currency, that is already a use case for bitcoin in many poorer countries. CBDCs are traceable, which might make bitcoin the preferred choice for users – but not necessarily for the payment rails they still need to connect to the traditional world.
The fact is that CBDCs are going to have a massive impact on private crypto. It could be positive, by legitimizing digital cash and making room for assets denominated in crypto tokens to flourish. It could help spur constructive regulation around the entire digital-asset environment.
But digital assets in what denomination? Because central banks, once they enter the digital-cash game, can also make life hard for private crypto. They could for example provide a yield on holding the government’s digital currency. They could encourage banks or companies to lend their digital currency. Or they could require banks, companies or private investors to place some of their digital assets on reserve with the central bank. Finally, the great mass adoption of crypto that many in the bitcoin space are working towards might be fully realized – when the coin in question is the government’s.
Of course, if states also choose to use digital currency to exert social controls over their citizens, and obliterate privacy, private currencies may become fashionable. But central banks have the option of wielding many carrots and sticks.
This whole crypto thing may have gotten started with Satoshi Nakamoto but it is going to be governments that determine how it ends. Expect CBDCs to evolve into competitive tools by governments to woo capital, punish capital, or channel capital.
Korea’s fintech scene growing fast: C.S.
A Credit Suisse report highlights activity in fintech by internet companies, startups and banks.
Customer demand, regulatory support and advancing technology are fuelling a boom in fintech in Korea, according to an equity research report by Credit Suisse.
Fintech is part of a broader trend in tech startups and unicorns (startups valued at $1 billion or more), of which there are now eight in Korea, according to the August 29 report, written by lead analyst Park Jeehoon.
Biotech and retail services are the hottest areas winning the most investment. But fintech investment is also thriving, especially in payments, crowdfunding, and virtual banking.
Today there are more than 9 million daily users of mobile banking services in Korea, about double the number from 2013. Even more striking: 90% of banking transactions in Korea are today made online.
Show me the money
This helps explain the ongoing flood of money going into the country’s startup scene.
The Korea Venture Capital Association says startups received W3.4 trillion of investments in 2018 and are on track to receive W4 trillion this year. In the first half of 2019, 826 startups received investment, up 16.3% year on year (there are now over 300 fintech startups in Korea). Funding for startups is also shifting from mainly early stage to also include growth-phase companies.
Within fintech, payments and money transfer account for 32% of investment (see chart, main image); ‘techfin’, that is investments from internet companies like Kakao and Naver, is 25%; lending and credit startups are getting 24% of the money; digital wealth 16% and insurtech 3%.
Korea’s fintech leaders
Credit Suisse cites several fintechs as new standardbearers for Korea. The biggest startup in terms of valuation is the unicorn Viva Republica, whose backers include PayPal and Altos Capital. Valued today at W2.7 trillion, the 2013-vintage company operates a popular financial platform app called TOSS.
Others include Wadiz, a crowdfunding platform that provides seeding solutions for startups and new business ventures; and insurtech Bomapp, supported by conglomerate Lotte Group and KB Financial.
The leading banks have also teamed up to jointly fund Honest Fund, a digital wealth and lending platform; its shareholders include KB Investment, Hanwha Investment, and Shinhan Capital, a Who’s Who of Korean private-sector banking.
Just as notable, Korea’s traditional banks are also busy with their own fintech initiatives.
Hana Financial is setting up a globally integrated platform based on blockchain; Shinhan is operating a mobile financial service that combines banking, insurance and wealth management; and Nonghyup Financial is the market leader in open banking, using APIs to connect with affiliates and third-party companies to share customer data (on customer request).
Oddly, the Credit Suisse report omits analysis of the hugely successful Kakao Bank or its fellow virtual bank, K-Bank (our story on VBs touted Kakao as the most important model for incoming VBs in Hong Kong and Singapore). It does list out the 33 tech investments made by internet parent Kakao into startups, from payments to crypto to gaming and healthcare.
But it does mention that the regulators in Korea are preparing to license a third virtual bank in 2020.
This is just one aspect of how regulators in Seoul have supported fintech, and why it continues to attract new players and financial backing.
The Financial Services Commission has taken steps over the past several years to support innovation. It has eased regulation on how firms authenticate customers, legalized crowdfunding and virtual banking (aka neo-banking), and blessed a blockchain-based authentication platform called BankSign.
This year, in April, the FSC launched a fintech sandbox giving companies a four-year forgiveness period to test new products. This coming month, it will set tougher mandates for open banking, to be followed next year with the debut of “MyData”, requiring open APIs between banks and fintechs to provide new financial services based on customer personal data.