If you’ve worked in banking, you might be familiar with Monkey Business, a book about life in an investment banking division (IBD) as a Wall Street associate in the late 1990s. As a snapshot of 20-hour days spent on Excel models and interminable pitch books, plus inappropriate activities on the desk, it remains unsurpassed. Much has changed since it was written, except the use of Excel and PowerPoint.
Throughout the years, Microsoft maintained its grip on junior bankers… but not anymore. The coming generations are all about Python.
Already, around 50 percent of incoming analyst classes have some knowledge of Python coding. So suggested Adrian Crockett, a former managing director at Credit Suisse who now advises banks on digital innovation: “The industry always asked juniors to do less glamorous work, but now the juniors who think in computational terms are just saying, ‘That’s fine, I’ll delegate this to Python.’ They see an army of bots as their worker bees.”
The new “computationally literate” graduate recruits are driving a revolution in the way banks operate, Crockett added. Whereas in the past, bright young graduates were happy to work until 1 A.M. crunching models and making endless minor tweaks to pitchbooks, today’s juniors have a much lower threshold for tedious work. “Someone who thinks in computational terms will get bored much more quickly than someone who doesn’t,” Crockett continued. “Banks need to take advantage of the new generation of computational thinkers or they will find it much more difficult to hire the generation of bankers after that.”
The biggest banks are ahead of the curve. JPMorgan set up a team to ‘digitize dealmaking’ under ex-DCM banker Huw Richards in summer 2018. Richards has said he wants to ditch pitchbooks for data and insights on how past deals performed, in addition to the likely implications for deals in the future. His 40+ strong team includes includes analysts who began in traditional M&A roles and developed an interest in analytics, plus former employees of Palantir Technologies, the secretive data science company used by the likes of the CIA. In London, Richard’s team works closely with Dan Zinkin, a technologist who also manages IT for the global investment banking division. One of the first initiatives is a data analytics tool to predict the impact of activist shareholders.
Other banks are on the same journey. Goldman Sachs has spent the past few years building an ‘IBD strats’ team in India to automate investment banking functions. Lazard, that most old-school of banks, wants to build an automated activist defense tool powered by artificial intelligence.
Juniors who don’t know Python are being brought up to speed. Raoul-Gabriel Urma, an Imperial College graduate and Cambridge University PhD who specializes in training bankers in Python and data skills, says he’s trained 1,000 people in Python in 2019 and expects to train far more in 2020. Urma’s clients include major U.S. banks.
Python: Accelerating Banking Change
Python has already replaced Excel on trading floors, Urma added, echoing Matthew Hampson, deputy chief digital officer at Nomura: “If you want to work on a quick file, Excel is fine, but for anything more interesting Python is needed.” Senior traders are alert to their ignorance: “The MDs are saying they want us to train them too.”
As automation sweeps the front office, however, some are sounding a note of caution. Crockett believes most banks are still in no position to deliver on the expectations of their new Python experts: “When I arrive at my desk as an analyst who is using Excel, then I am good to go to do simple things like modeling an M&A target,” he said. “But if I want to model ALL likely M&A targets using predictive analytics, then I will probably find there is no infrastructure to help me. There’s not the data and there’s no access to the cloud. Banks need to do more than just enable today’s analysts to use Python if they are going to truly benefit from this next generation of bankers.”
For this reason, Crockett thinks banks need to make much broader changes than simply hiring juniors with Python expertise. New data infrastructure needs to be built, with the support of senior staff, which is easier said than done. “The politics are huge and the permafrost at the top is the biggest obstacle to digital transformation,” related one managing director at a large bank. Nonetheless, changes need to be made with some urgency: The new Python experts won’t stick around if they’re made to perform manual processes until 2 A.M. each morning. They’re simply not the same as their Excel predecessors.
A modified version of this article originally appeared in eFinancialCareers.