Top JPMorgan Traders Talk Robots, Brexit Impact and Remote Work

(Bloomberg) -- The next generation of traders will need quant, data and coding expertise to keep up with the needs of clients, two of JPMorgan Chase & Co.’s top electronic traders say.

“The trader of the future isn’t a person who’s pressing buttons to trade, it’s somebody who understands more of a macro view through the use of data and technology,” Sarah Heffron, global co-head of electronic client solutions and program trading for equities, said in a video interview.

Over the past 12 months the 22-year JPMorgan veteran, and Joanna Martin, head of global liquidity solutions and co-head of EMEA execution services, have been balancing the challenges of remote work and childcare, while also managing the post-Brexit transition and examining how automation can make trading more efficient.

Their comments to Bloomberg have been edited and condensed.

What do your clients worry about?

Joanna Martin: Every time we quiz clients -- Europe, U.S., Asia -- about what is your number one concern on a daily basis? It’s always access to liquidity, it’s always liquidity.

You ask if they feel like liquidity has decreased over the last two years. And they say, ‘Well, no, but the complexity with the regulation we’ve been through -- MiFID, Brexit -- add layers of complexity.’

Sarah Heffron: One of the things that we’ve been really focused on is the close, because the close as a liquidity event in Europe is something that is much bigger now than it was several years ago.

It can be the most expensive time of day to trade. So we want to help the clients manage cost and impact. But also, just in the most basic way, it’s 20% of European volume going off in five minutes of the day. For the dealer community and the investors, it’s an operational daily challenge. We think about how can we help manage that risk. You see what happens when one provider falls down on that, one exchange falls down on the close.

How has Brexit impacted your trading floor?

SH: We are changing some of the ways and the locations from which we’re servicing clients. And some of that has happened already.

Largely that’s meant moving some of our existing employees to Paris, complemented by some local hiring, resulting in a build out of our presence there and the trading floor. That’s a meaningful change for us. Ironically, Covid probably prepared us a little bit better for it. So the concept of a team being on a perpetual Zoom is not foreign to everybody. It happens already.

The thing that we’re going to be keeping an eye on is data center locations, because some exchanges for example, have their data centers in the U.K., which could change for them long term. We’re going to see what that means for the technology and infrastructure work. It’s the less exciting stuff. But unless we keep our plumbing and our infrastructure and all of our network connectivity at its best, you can very easily lose competitive edge.

JM: The change that we had to make for Brexit took a lot of resources. It diverted resources from potential innovation but it had to be done. It is a hard but crucial balance to strike. It does make it difficult for smaller players to keep up in that sense.

What skill sets are in demand ?

JM: When we poll clients, the number one skill set they’re adding to their desk is quant data scientists. And they can’t afford to have somebody just sitting there looking at data. A lot of the time these are quantitative people who also have to trade as well. So they’ve got coders who trade, and algo is code essentially. This is how everyone trades.

And machine learning?

SH: One of the things we’re looking at in our signals and modeling space is our traditional calibrated models, and how machine learning can help there. In many cases, that is the most successful way to calibrate to develop a model, we run a machine-learning version of that alongside of it.

We continually look to backtest those together to say which one is outperforming the other, because there are times when your machine learning will find something that you won’t find in your calibrated model.

JM: A human needs to be supervising 100%. And that’s why we run the two models or tests side by side, because things change and you can inadvertently import something into the middle that can skew you off and for accuracy you need that kind of human overlay to be constantly questioning and thinking.

We need the machine learning to scale our business, to be able to do more and more quicker and be dynamic. The industry thinks of it as a feature or a new algo, or a new funky toy. But it’s just there. It’s everywhere.

Does that mean fewer jobs?

SH: We become more and more automated every year. There’s not been one year that I have been here in equities where we have gone down in headcount.

How has remote working gone?

JM: We’ve been having a blend of work from home and office the whole time. To be in the office for us has been relevant and necessary. We both have children. I have preschool-aged children, so I feel somewhat lucky that I haven’t been dealing with home schooling. But obviously juggling childcare is a challenge.

SH: It’s a challenge in different ways, but I’ve learned that I’m definitely not a teacher. This is not my strong suit.

There was some chaos in the beginning personally, as for many I’m sure. We all got to know each other personally a little bit better, because we’re all looking into each other’s homes and seeing their kids.

What will the return to the office look like?

SH: We are coming back to the office because actually, ultimately we work better in an office, but we’ve adapted so much and done a lot of good work around understanding what makes people tick.

It’s not just the parental bit. It’s the fact that some of our people are younger and just out of university have had to share a space with their other flat mates at home. And that’s really stressful for some people.

Will it be the same for everyone?

JM: A lot of the sales traders have more of a need to be more office-based but they’re absolutely part of that increased flexibility in an organized and structured way. The other key consideration is whether we have a quorum of people. So we get that team vibe, we get that communal learning, especially for juniors or people who are new into the firm.

I’ve had a flexible arrangement for five years. I feel there’s a wider understanding that this can work really well.

Something we’ve been incredibly cognizant of is the mental health element, especially right at the beginning where people were adjusting to the new environment and we were insisting that people get out and go for a walk, take a few days off in the few weeks when it’s quiet. Sarah’s got a yoga session tomorrow night for the crew. I’m doing virtual cocktail making with a client.

©2021 Bloomberg L.P.