Investment banks to sell technology as conventional margins are crushed

Voices at the top of the banks keep pushing the notion that they are really technology companies, perhaps in a desperate effort to get their stock prices motoring with a message that the chronic underperformers are honestly growth plays. It used to be almost funny. But it could be about to get serious.

Euromoney sits down with the chief executive of one large European bank who agrees to share some thoughts not for attribution.

“Intellectually, I see no problem with being a technology provider,” he says. “We have to think differently about new ways to generate revenues. Technology is the most exciting part of banking and we have invested a lot in development ourselves, bought from tech vendors and adapted, invested in some fintechs and acquired others outright.”

And where does he see the chance to generate revenue from such investments?

“We are right now looking to rent out spare capacity on our processing platforms for a fee and to sign service-level agreements with customers. I could certainly see us providing technology to smaller banks, for example around regulatory compliance and know-your-customer, maybe even whole operating systems.”

There are examples here: Live Oak Bank set up in Wilmington, North Carolina in 2008 to provide quick loan decisions, credit and other banking services to small business owners that the big US banks seemed to be abandoning after the financial crisis.

It started out focusing on veterinary practices and soon moved into agriculture, pharmacy, healthcare, restaurants, bowling alleys, fitness centres, funeral homes, plumbers and electricians and grew into one of the biggest originators of small business loans in the country.

The secret sauce to making commercial lending much more efficient was new technology and an integrated cloud-based bank operating system. When other banks came calling on Live Oak to ask how it managed this, the bank’s founders realized a whole new business existed in providing this operating system to other lenders.

nCino was hived off into a separate business in 2012, with backers including Wellington Management, Insight Venture Partners, Salesforce Ventures and Bessemer Partners. It now has more than 250 bank customers. Live Oak still operates as a bank, but perhaps its biggest contribution to transforming the banking industry will have been spawning nCino.

But selling banking technology to other banks is not quite the same as selling software services to traditional customers, corporations and investors. Can a big established bank really do that?

Before long the chief executive of our European national champion is thinking ahead to the potential limits bank regulators might put on such expansion into non-financial businesses.

“We are not going to become Accenture,” he admits, “and if such technology businesses grew as a proportion of what we do, there might be the question of whether a bank is the best shareholder.”

That sounds like a first-world problem, Euromoney suggests. He laughs and nods: “It would be.”

This chief executive has spent a large part of the meeting talking up his bank’s sophisticated procedures and tools for risk management as evidenced by a cost of risk well below peers. If he wants to generate revenue by selling technology services, surely he should be selling versions of the bank’s risk management software?

Suddenly, he sounds less sure: “Risk management is so core to what we do. Yes, we are experts. But should we sell that? And what would be the potential liabilities if we did?”

On his final earnings call as chief financial officer in the third quarter last year, before stepping back into the securities division of Goldman Sachs, Marty Chavez – who made his name as the firm’s chief information officer – sounded much more confident.

Chavez talked up the firm’s credit trading algorithm as a source of efficiency in handling electronically trades of up to $2 million in some 10,000 investment-grade bonds, before looking across fixed income, currencies and commodities (FICC), equities and risk management.

He said: “We’re leading in those extremely competitive businesses with content, scale and making it all client-centric and investing to modernize it with digital access, digital formats of many kinds, digital user experiences over the web, for the same tools that our people use, deploying them to clients, also giving clients the abilities to plug directly into our platform through APIs [application interfaces], which is very much a theme for us, as well as all companies that are building and deploying technology for their clients.”

Euromoney sits down with Jim Esposito, global co-head of the securities division at Goldman Sachs with Chavez, who explains a little more about where the firm sees opportunities to monetize some of its investments in technology, for example in all the market data it has collected and stored going back for decades.

If the firm is not quite moving into software as a service, it may be moving into something close to that.

“Data as a service and risk management analytics are becoming more important for us to deliver to our clients,” Esposito tells Euromoney. “Whether we charge separately for these tools or provide them for free is still evolving, but there is no debate these activities will be core to ensuring we can service our clients’ needs in the future.”

Quantitative hedge funds and more conventional investors too are increasingly asking to access stores of data across which to run analysis of the historic skew on the volatility surface for specific stocks and other securities looking back over long periods.

Goldman does have that, Esposito says: “We need to externally package and efficiently deliver what is a very valuable data library and set of risk analytic tools. We have over 25 years of experience building and using these same sets of tools for our own risk management needs. Our brand has been synonymous with thoughtful risk management. Now we’re pivoting to deploy these same tools to support our clients’ needs.”

Banks are only just beginning to mine such stores of internal data to inform the prices they themselves make for securities. As trade execution services reach scale efficiency and further compress commissions and spreads, the value of data once generated as a consequence of the core business is becoming a business in its own right.

Euromoney decides to dig a little deeper into an area close to our hearts, FICC and debt primary markets, to check how far new technology is transforming the markets and whether or not the banks that have developed a lot of this technology for their own use can now monetize that investment by selling versions of it.

There are almost as many opinions on whether or not banks really are going to source much revenue from selling pricing engines, smart order routers, risk-management systems and the like to clients as there are people to ask.

“Banks as tech providers is no longer science fiction,” insists Fabian Vandenreydt, executive chairman of B-Hive, a membership group that promotes collaboration between banks and fintechs.

Vandenreydt spent 10 years as head of capital markets and Innotribe at Swift; he suggests: “Many banks are now seeking to follow the example of BlackRock with Aladdin.”

BlackRock has been providing a hosted solution, first developed internally for its execution desks to source liquidity between different BlackRock managed funds, for example, and which it has since turned into an operating system for other investment managers, complete with risk analytics, portfolio management, trading and operations tools.

In 2018, technology services brought in $785 million of revenue, equivalent to roughly 7% of what BlackRock makes from managing money. That is not a game changer for the world’s biggest asset manager, but it is a decent boost in a business where margins are getting crushed.

“If you look at some of the leading firms in investment banking, they’re doing a lot in the technology area now,” says Vandenreydt. It’s not exactly a stretch to wonder why.

“What you are picking up on, below the grand statements of CFOs and chief executives, is one of the most significant changes in the whole structure of the capital markets business,” says a former banker now running a bond trading platform. “Banks that have built good systems for themselves are looking for new ways to monetize them as traditional revenues shrink.”

But not everyone is convinced.

“Banks as technology providers? It’s a bit like teenagers and sex,” says another senior banker. “There are almost certainly many more talking about it than actually doing it. The few I see taking the lead are the usual suspects among the top handful of US investment banks.”

But others are also joining in. ING, for example, is starting to act out the fantasy.

ING is one of the more innovative European banks in all the usual areas – open banking and personal finance tools for retail customers, trade finance on blockchain for corporates and investing in faster cross-border payments platforms for small and medium-sized enterprises. But it has also been investing in more unusual areas closer to the capital markets, such as artificial intelligence for predicting the appetite for syndicated loans.

In December 2017, ING announced – to almost no fanfare – Katana, a tool employing predictive analytics to help its bond traders make sharper prices to institutional investors.

“Katana was developed within ING wholesale banking advanced analytics,” says Santiago Braje, global head of credit trading at ING. “To start with, our focus was on improving our own performance in market making in bonds by identifying the key components of the decision-making process and finding ways to use artificial intelligence and analytics to support that.”

ING particularly wanted to help its credit traders respond better to requests for quotes (RFQs) as they came in from asset managers. Bond market making was hit badly by post-crisis regulation and the allocation of risk-weighted assets and expensive funding and capital costs.

With reduced risk appetite among banks, very soon after the collapse of Lehman Brothers a liquidity crisis blew up in the bond markets, with asset managers desperately spraying RFQs around to find the other side of their trades as they sought to radically rebalance portfolios. Only the downward pressure on yields from concerted central bank buying contained this.

ING was interested in what benefit it could bring from more comprehensive delivery, interrogation and analysis of both structured data – observed prices on actual trades – and unstructured data.

“We aimed to bring in real time all the information a trader might want before responding to an RFQ,” says Braje. “Obviously, it’s important to have the recent history of prices in liquid bonds, but also it would be useful to combine that with internal data, for example what prices our traders may have previously quoted on the same bonds when that quote didn’t lead to a trade and by how much we may have lost out to the order that eventually went through.

“It would be good then to overlay all that with information on present market conditions and with analytics that help predict what might be the winning quote. Most importantly, you then have to deliver that on an interface designed to be useful for traders who may have only seconds to respond.”

The outcome is a predicted price at which the client will trade that also takes account of where competitors will likely quote in that specific bond given prevailing market conditions.

Katana is targeting a very narrow and precise range. Dealers may want to win the trade but in the most efficient way possible: that is at just a fraction ahead of the next closest bid or offer. And there are also times, of course, where any bank might want to lose a trade but at least look respectable for an important client in not losing by too obviously wide a margin.

“It requires quite a diverse range of sources to arrive at a trade decision,” says Braje. “And there is a lot of information that cannot be captured as data and translated into code, such as from human conversations and the human sense of the client relationship. But Katana makes a difference by providing a baseline for the decision that is objective.”

And while Braje won’t go into precise details, he says that Katana has proven itself internally at ING and achieved the original aim of improving traders’ performance.

“We have analyzed data for traders using the tool and not using it and have found that it improves both our hit rates – how often traders win the deal – and their trading cost efficiency – that is the difference between our price and the second-best quote – by around 20% in each case.”

For most of the 12 months following the launch of Katana, ING used it in emerging market credit. When Euromoney speaks to Braje in February the bank is preparing to relaunch it with new features as “a more solid, production-grade piece of software”.

And here’s the intriguing thing: ING is making a different version, called Katana Lens, available to asset management clients.

“Let me be clear. This is not simply the other side of the same tool only this time for the buy side,” Braje says. “Rather, we took a step back and tried to focus once again on the decision-making process on the buy side and how to support that with augmented intelligence. We found a great partner in PGGM [the Dutch pension administration co-operative] and started exploring the most interesting ideas and important problems for the buy side. From two or three candidates, we chose to focus on a relative value discovery tool, working along similar principles to Katana, aiming to support the process of deciding what trade ideas to put on.”

Whenever a long-only active bond manager wants to buy one bond, it generally requires selling something else, unless the manager is investing newly arrived cash or cash held in expectation of market falls. Worried about the rise of cheap passive index-tracking funds, active managers don’t like to hold too much cash for fear of exacerbating their usual underperformance after fees.

Bond investors typically grow used to thinking about associated pairs of bonds and look for periods when they diverge from the traditional value relationship and one trades cheaper than usual to the other, suggesting a good moment to sell the expensive bond, buy the cheap one and wait for the mean reversion to realize a gain.

For years, emerging market investors monitored Argentina versus Brazil, for example. But if an asset manager’s investment team is surveying a universe of 2,000 bonds in emerging market credit globally, that is actually two million pairs potentially to trade. An asset manager covering 3,000 bonds in investment-grade credit globally has potentially 4.5 million pairs to consider.

The first challenge is to work out which of those possible trade ideas are really worth analyzing in depth, not just by looking at prices but also by combining changing market data with a lot of unstructured data, for example from conversations between portfolio managers and issuers, as well as the sell side. When an asset manager is seeing a lot of suggestions from dealers to trade in a certain bond, then it may well be one to look at but perhaps in combination with a non-traditional paired buy or sell.

“There may be a lot of news around Brazil one day or Russia the next,” says Braje. “But there is limited value in looking at one bond in isolation or just against the overall market index. One of the valuable things our Katana Lens algorithm does is throw up for asset managers trade opportunities in pairs of bonds they don’t normally consider, either because they have never thought of them as a pair, none of their providers of trade ideas have suggested them or they don’t have a tool that spots deviations from the trend quickly enough to take advantage of.”

This is what artificial intelligence can bring to financial markets. It doesn’t just analyze the price history of Argentina versus Brazil. From an addressable universe of thousands of bonds, it can analyze everything against everything else and throw out suggestions for particular trades within a framework of how they look against every other possible idea.

Looking for such unusual correlations is not new. It was the aim of cutting-edge hedge funds like DE Shaw 20 years ago. Cheaper computer power, more digitized data, advances in machine learning and open-source software mean that it is now within reach of more participants.

There are two obvious questions: how often does Katana Lens suggest trade opportunities that asset managers then put on and how do those trade ideas work out?

“Our clients give us feedback, which is very encouraging, that it throws up trade ideas to put on positions at interesting levels which are profitable and which have improved their performance,” Braje reports. “Just as importantly, it does so in different market conditions. Its success is not correlated to overall market direction. And this alpha capture accords with our own back testing, though all back testing is theoretical of course.”

Katana Lens emerged as a prototype in the third quarter of last year and ING began on-boarding more buy-side clients to trial it in the fourth quarter. Euromoney wonders if its usefulness will diminish as more institutional investors start using it?

“Each asset manager will have their own particular mandate, focus and biases with different universes of investable securities and portfolio constraints,” Braje counters. “So the same signals will not necessarily be relevant to all users.”

ING has different user groups now looking at how to use the tool in new markets beyond emerging market bonds, including investment-grade credit and developed-market government bonds.

These are very different businesses. European government bond markets for example have only a handful of issuers, but each with very large numbers of ISINs [International Securities Identification Numbers]. That is almost the opposite of emerging markets with far more issuers, each with fewer bonds outstanding. So it is intriguing to hear the same algorithm being re-purposed with different parameters.

ING is also seeking to integrate new features, such as optimization of trade ideas for payoffs over different time horizons.

“This tool is still at an early stage in its development,” says Braje. “We are also looking to embed some elements for collaboration, so that a buy-side analyst or portfolio manager might curate a list of intriguing trade possibilities and ask colleagues to provide their feedback.”

The big question, it seems to Euromoney, is how does ING get paid for providing this technology to investors? Will it just be through the traditional means of order flow? Analysts at banks are measured in large part by their capacity to produce trade ideas that prompt investors to buy and sell. The banks profit from turnover: that’s how analysts get paid.

Katana Lens embeds inside customers a tool for suggesting trade ideas to act on. That is good for them if they are good ideas; it’s good for the banks too, whether they are or not, just as long as they are credible enough to induce activity.

Is this, perhaps, too cynical a way to look at it?

“Although Katana Lens is fully functional, we still consider it a prototype for now,” Braje tells Euromoney. “How we may eventually get paid for providing this is not yet defined. We are exploring a range of methods from integrating it into our current client offering all the way to offering it as a standalone product with its own fee income, and various combinations of those options.”

Does he worry that other banks will seize this new business opportunity?

“A lot of banks are making big investments in technology, but from our understanding the list of those coming up with tools like this is not a long one. Of the 30 or so investors we have shown Katana Lens to, 80% have nothing like it and none has told us that any other banks are offering anything similar. The other 20% are using tools they have built internally themselves.”

He adds a big qualifier, however: “There is certainly a big change underway in capital markets technology, which is shifting from development in closed environments to much more open development thanks in large part to the low cost of cloud hosting and the emergence of libraries of open source software.

“We are seeing platforms develop tools for very specific jobs that are lighter to use and on-board than in the past and can be integrated more easily with other tools. And the pace of development is really accelerating.”

Vikram Pandit, the former chief executive of Citigroup and today chief executive of Orogen group, an investor in financial technology, thinks back to the 1990s and the transformation of the US cash equity markets, an area in which he ran Morgan Stanley’s leading franchise for a number of years.

“The theory, pushed by regulators, was to democratize the equity markets at a time when the marginal cost of processing was coming down close to zero – and the marginal cost of communications and of computing capacity,” Pandit says.

“It really was rather like the advent of cloud computing today. And as those costs became de minimus, the question around markets structure was how that should translate to commissions. One choice was to maintain large specialist market makers on the NYSE, but our bet was that equity trading would increasingly be electronic.

“We put our own money to work on this in what we called our equity trading lab, applying the technology first to futures and options before taking it to clients in the cash markets,” he recalls. “The promise was that instead of per share commissions, we would build a pipe that clients could push as much business through as they wish at a fixed charge.”

Other markets are still building their pipes. The algorithms are ready to flush a lot of business through them.

This article appeared in April 2019 Euromoney magazine with the title 'Generation tech rent'