Just because you have a first class economics degree from Harvard doesn’t mean that you’re ready to work for a hedge fund now. Graduates are usually missing one key ingredient – data science.
Point72 Asset Management – which hires less than 1% of the students who apply to its academy – now puts all of its analyst recruits through the sort of technical training it believes is key to gaining an edge in the future.
“Every analyst we hire is now required to undertake some basic data science and programming training,” said Matthew Granade, managing director and chief intelligence officer at Point72. “Most economics graduates coming out of Wharton and the like do not have much focus on statistics or programming. These are the foundations of the skills we require.”
Point72, like many discretionary hedge funds, is at an inflection point. Currently, around two-thirds of its investments are the traditional discretionary style, and the rest is quant investing. Much of this is down to Steve Cohen’s belief that highly-skilled people should be picking individual stocks using a bottom up approach. But this is shifting.
“When you think of the finance careers of the future, it’s going to be much more like the Social Network and less like Wolf of Wall Street,” said Granade speaking at the Newsweek Artificial Intelligence and Big Data in Capital Markets conference yesterday.
But this doesn’t mean that Point72 is going out to hire PhDs or computer science graduates. It still wants investment expertise, but expects its portfolio managers to have the skills to work with huge datasets.
The future will be less about “ sleepless analysts reading the tea leaves”, he said, making investment decisions based on sell-side research and accounting data. Point72 is instead now using the huge quantities of data coming out of third-parties, such as credit card receipts or consumer geo-location data. Every day, two or three new sources of data are added to its investment process, said Granade.
Therefore, Point72 is trying to figure out the “limits of the human brain” and how best to use people and computers for its investments.
“We’re relying heavily on alternative data sources and quantitative techniques, but it can’t be fully-automated,” he said. “Right now, these data sets are too fragile to just put through an algorithm. Human judgement is important – the logic and decision making for investments has to be laid out clearly. But we can leave portfolio construction and trade execution to the computers.”
Photo: Getty Images