No one could accuse the Japanese of being backwards at coming forward when it comes to the robotic revolution. This is the country that brought us the world’s first robot hotel, the world’s first “humanless warehouse” and, more recently, a robotic pet designed to offer its human overlords love and affection.
Now Japanese bank Nomura is getting in on the act, investing in a start-up that deploys artificial intelligence (AI) in financial markets.
Named Aim2, the start-up was set up by Brevan Howard back in 2015 to help the hedge fund with its alpha investment strategies, ie to find ways of making money based on the particular profile of an investment rather than through market exposures.
Nomura — which is understood to have taken a 22 per cent stake in Aim2 — will use the same technology to price trades and markets in its wholesale business.
“Typically an electronic trading engine or trader makes prices based on certain tiering and logistics,” said Jez Mohideen, chief digital officer at Nomura’s wholesale business, describing a pricing process that involves “using large sets of data, understanding behaviours of the market, the market impact of clients and so forth”.
Machines, of course, excel at making sense of big data sets, and intelligent ones have the added edge of learning how market behaviours and trends evolve. Hence Nomura’s desire to deploy AI so it can find a “probabilistic price” that will generate maximum revenue from a client while also helping Nomura to maintain its target market share.
Nomura’s roll-out of its robot traders will be initially focused on its fixed income, currencies and commodities division, or FICC as its known in banking circles, since it’s far more reliant on human traders than equities, which has been largely electronified. Within FICC, fixed income is the main target since it has been less automated than foreign exchange.
“We’d gradually look at foreign exchange and equities,” Mr Mohideen added, as he outlined a plan to deploy the robots across Nomura’s trading business within two years.
While the prospect is new within Nomura, and may be greeted with concern by traders who have already seen some of their human colleagues laid off this year, bigger investment banks have been using AI in their fixed income businesses for some time.
Most banks didn’t want to go public on exactly how they use AI, but informed sources said robots have been used in some form or another in the FICC businesses at JPMorgan Chase, Morgan Stanley, Goldman Sachs, Citigroup and Bank of America for a few years now.
Europeans have been at it too. Barclays uses AI for everything from forecasting business volumes to recommendations engines that help the bank target its client pitches. Chris Purves, head of UBS’s strategic development lab, says the Swiss bank is using AI across “most” of its fixed income business lines, both for executing client orders and for market making and pricing. “The key to us is that it’s explainable,” he adds. “Can you say why something made the price that it did?”
Certainly, the commercial imperative is there.
Banks have had a bruising few years in FICC, with annual revenues across the big five US banks falling 15 per cent since 2013, and smaller players faring even worse. The industry’s most senior executives have described the pivotal role technology will play in determining the winners and losers in FICC’s next era, arguing that the biggest banks will fare best because they can invest the necessary sums for true innovation.
Nomura, a sizeable bank but not big enough to make the global top 12 investment banks as ranked by industry monitor Coalition, is getting around the problem by investing in a cutting-edge AI firm.
Mr Mohideen expects other smaller investment banks to join them in deploying third-party technology to bridge the gap between what they spend on R&D and the $11bn technology budget of JPMorgan. Indeed he’d even welcome them as clients of Aim2. “The way we look at it, the toolkit is beneficial for the industry as a whole,” he said. “Going forward the world is getting to a place where that (exclusivity) is less of a concern.”
The prospect of wider deployment of robot traders raises the thorny question of what happens to the human traders. The obvious conclusion — which bank executives go to great lengths to avoid acknowledging — is another wave of lay-offs.
It might not be as bad as some fear though. Earlier this year, Japan’s pioneering robot hotel admitted it had sacked more than half of its 243 robots because they were more hindrance than help. Humans: 1, Robots: 0.
One to watch: Hundsun Technology
Remember Hundsun Technologies, the Ant Financial-backed fintech that ran into trouble for providing software that enabled speculative margin trading by retail investors during China’s 2015 boom and bust?
From that most unlikely of starting points, Hundsun has created a booming business selling software to help financial institutions strengthen their risk control, writes Gabriel Wildau in Shanghai.
Net income rose 37 per cent to Rmb233m (approx $35m) for the first three quarters of 2018, and in a few weeks’ time we’ll find out if the full-year results are as encouraging.
The seeds for its growth spurt were sown in 2017, when Hundsun shifted towards technologies that match regulators’ emphasis on controlling financial risk.
The crown jewel is the O32 software package that offers a comprehensive risk management system spanning everything from monitoring irregular transactions to real-time tracking of profits and setting trading limits.
Hundsun says O32 can avoid trading errors resulting from computer errors, like the 2013 incident that triggered a Rmb523m fine for brokerage Everbright after mistaken orders were placed by a trading algorithm in its proprietary trading unit. Fund managers are a key target market for O32 as well.
It’s all a far cry from 2015, when Hundsun was fined Rmb400m because providers of grey market finance used Hundsun’s HOMS software to allow hundreds of retail investors to trade separately — and access margin loans — using a single brokerage account.
Number of the week: $8bn
The valuation of Ping An’s OneConnect unit, which the Chinese insurer is preparing to list, according to Bloomberg.
Further fintech fascination
Trend watch: It’s been a long time coming; now Hong Kong is finally opening up its banking industry to online competition. In the next few weeks, six of the new licences are expected to be issued to regional powerhouses in digital banking including Tencent, Ant Financial and insurer ZhongAn. Western digital banks may also ultimately join the rush to win share in the potentially lucrative market, where 66 per cent of all lending is done by just four banks (HSBC, Bank of China, Hang Seng Bank and Standard Chartered).
Regulators’ advance: Most countries would be glad to secure a fintech with the profile of Revolut, the foreign exchange app that’s fast becoming an online financial services supermarket and commands a valuation north of $1.7bn. Not Lithuania though, where Revolut secured a banking licence last year. Senior politicians in the Baltic country are now voicing fears that its central bank has bitten off more than it can chew.
Follow the money: Funding for insurtech companies reached a record $4.1bn last year, easily eclipsing the $2.7bn raised in 2015, according to new data from Willis Towers Watson, the insurance broker. The fourth quarter of the year was particularly busy, with $1.6bn raised in 63 deals. They included $68m raised by Germany’s simplesurance, which specialises in cover for products sold online, $116m for Italian motor insurer Prima Assicurazioni and $500m in a SoftBank-backed funding round from Cambridge Mobile Telematics.
New frontiers: Want to protect yourself from losing a fortune if the head of your chosen crypto exchange dies suddenly and takes your access codes to the grave with him? Then look no further than the Lloyd’s insurance marketplace, which has sold crypto custodian BitGo Inc a policy of as much as $100m for losses, including the loss of access keys, an unforeseen risk that has left clients of Canadian exchange QuadrigaCX out some $150m. More here from Bloomberg.
AOB: Only four fintechs in the world have raised more than $1bn in equity capital, find out how they fared since then in this post from Axios***The FT’s Matthew Vincent opines on cryptocurrency’s enduring fascination for the rich***It’s official, fintechs have a 10 percentage point lead over traditional banks in the US’s unsecured personal loan market, according to the latest data from TransUnion***