The AI revolution comes to investment banking

This article was originally written by Paul Clarke and published on Financial News on January 7 2019.

Tobias Taupitz beat thousands of applicants to land a job in Barclays’ mergers and acquisitions team in London after graduating with a masters in management from University College Dublin, and expected a high-flying and lucrative job.

The pay lived up to expectations, but the work did not. “I spent so much time updating Powerpoint slides, or hours in Excel — and I’m not talking about the funky stuff,” he said.

He soon discovered that the majority of the job’s long hours would be spent crunching numbers for pitchbooks, sales documents that contain information on the key risks of a deal and valuations data in order to justify why clients should choose a bank’s services.

“I ended up sweating in the printing room at 2am, queuing behind a line of junior bankers,” Taupitz said. After just two years, he quit to launch his own fintech firm, Laka, which is looking to shake up the insurance industry.

In the past, bankers like Taupitz with little or no experience would have been advised to view the drudgery as a kind of apprenticeship, a necessary prerequisite to promotion and more interesting work.

Now, as investment banks struggle to retain juniors who are often quickly lured away into innovative tech jobs or private equity, the end of grunt work is looming. The waves of automation that crashed over supermarket cashiers and bank tellers reached City trading desks a few years ago; today, a handful of investment banks are investing heavily in technology that can do in seconds tasks that would have taken teams of analysts and associates days to perform previously.

The technology promises nothing less than a radical shift in what it means to be an investment banker, junior and senior.

“There’s going to be more thinking, less doing,” said Huw Richards, head of digital investment banking at JPMorgan.

Richards moved into his role in May, leading a team of around 40 senior bankers, technologists and data scientists looking to automate parts of JPMorgan’s dealmaking process.

Automation is not entirely new in the investment bank. Junior bankers working in Excel sometimes use complex macros — a kind of program-within-a-program — that help with financial modelling. Bringing in fresh technology is expected to open up new opportunities, however. Richards’ team’s initial projects include using predictive analytics to help identify the best time for a transaction, and pitching deals to clients digitally using software with an interactive component, rather than static Powerpoint presentations.

“It’s on their desktop, they can do the analysis themselves,” Richards said. He pointed out that paying analysts and associates to produce presentations is expensive.

“That’s not a good return on investment,” he said.

Goldman Sachs has also been turning to technology in its investment bank. It has 50 people working within an M&A solutions team, led by partner Steven Barg in New York. For two years, the group has been developing a system called Jupiter to automate parts of the initial public offering process, help assess threats from activists, and model the impact of a merger or acquisition on a company’s business and share price.

“We can have a company call up and say, 'I’m thinking about raising my dividend, how is that going to be received by our shareholders?' Our analytical tool can provide an answer in 30 seconds,” said Barg. “It would have taken a team of analysts days to do this before.”

Other banks are also taking tentative steps to automate elements of their investment banking advisory functions. UBS hired Ronald Jansen from Goldman Sachs in August to head up a new investment bank data and analytics lab; Lazard, the independent bank, has created a data analytics team; and Jefferies has moved tech investment banker John Riccardi into a role leading its digital investment banking strategy. Barclays is believed to be using technology to speed up client onboarding. It also uses artificial intelligence to help clients defend themselves against activist investors and to value companies to find M&A targets.

Spokespeople for Morgan Stanley, Bank of America Merrill Lynch, Credit Suisse, Deutsche Bank, HSBC and Barclays all either said no projects were underway in this sector or declined to comment.

Some in the industry fear that automating part of the work done by junior bankers will lead to fewer jobs. Bank executives admit privately that analyst classes are already smaller and further thinning out of the herd is likely. A 25% reduction is probable within five years, said one senior banker. Former Credit Suisse managing director Adrian Crockett, who now advises investment banks on digital strategy through his firm Fingital, predicts analyst classes will shrink by 30%.

Publicly bankers insist the ultimate goal is to free up staff to pursue more opportunities.

“Talking about having fewer junior investment bankers is not really the point,” said Jansen. “Right now, it’s the other way around — we have more potential opportunities than the resources to be able to action them. Technology allows us to free up capacity.”

Goldman Sachs is pushing for more deals in the middle market — defined as $500m-$3bn in size — and other big banks are following suit, according to sources. Larger banks previously shunned deals that did not come with blockbuster price tags because they were viewed as not profitable enough to justify the expense.

“Investment banking is a human-capital intensive business, and banks have to manage that resource where it’s going to generate the biggest fees — namely, big deals,” said Crockett. “Technology will allow banks to increase the number of deals they do exponentially, because the cost of executing a deal will go down.”

Technology could level the playing field for smaller banks that do not have the firepower to bring in hundreds of juniors every year. Boutique firms LionTree Advisors and William Blair are using technology in the dealmaking process, largely to uncover potential buyers and sellers for clients.

Vikram Pandit, the former chief executive of Citigroup who now runs Orogen Group, the fintech-orientated investment firm, told Financial News that boutique banks will use technology to “scale more quickly”, allowing “smaller players to compete with large legacy banks”.

Though banks say they will not need fewer juniors, the talents they are expected to bring to the job will undoubtedly change. The shift towards automation in trading has increased the demand for recruits with mathematical and programming skills, and a similar dynamic is likely to emerge in advisory work.

“It will drive a dramatic rethink of how we staff the deal teams,” said Richards. “In five years’ time, it’s impossible to think that most of the junior bankers we hire won’t have some form of programming experience.”

“We are a team of investment bankers who are also data scientists,” added Jansen.

JPMorgan’s Richards said: “Many of the people interested in getting into finance are going to be computer science minors, or have joint business degrees.”

The idea of hiring “bankers who code” is gaining traction. Jansen said that an analysis of the students applying for roles within its investment bank revealed that up to 40% had knowledge of the core coding languages used by banks: Python, Java or C.

“It’s really up to us to develop platforms that allow them to utilise those skills,” he said.

Crockett said: “Excel is a time-saving tool, but even assuming that juniors use Excel VBA to write macros — which the majority of them can’t — using Python is an entirely different world. The quantum of data it opens up and the longevity of that is completely different."

Goldman is so keen to ensure that bankers use its new Jupiter system that it has become a mandatory part of its graduate training programme. At the same time it is aiming to ensure that junior bankers get client exposure within two years, rather than around the current five, in an attempt to boost retention rates.

“For junior bankers it means they focus on higher-quality, more value-add functions earlier,” said Barg. “We want to replace the needless repetitive work analysts and associates do, automate that, so that they can start talking to clients earlier.”

Crockett said: “The real challenge [for technology teams] is delivering some small package of value that impresses senior bankers. The aim, ultimately, is to change the culture to the extent that a dealmaker comes to you and says ‘I need something’.”

The question now is how far technology will go in automating a process that has largely relied on human relationships and expertise. Goldman’s Barg believes technology will not make investment bankers extinct any time soon.

“One can’t automate good advice and good judgment,” he said.