For some years, I’ve been heavily involved in high frequency trading and quantitative trading, areas that are highly automated and pretty complex. I don’t blog about it much here, as it’s not appealing to the everyday person. It involves too much dialogue explaining acronyms used in investment markets and acronyms used in technology. In other words, like bitcoin, it combines everything you don’t understand about money with everything you don’t understand about technology. Nevertheless, I do spend time understanding it as it is usually the place that I first see new technologies used in finance. It was in investment markets that I first learned about neural networking to do pairs trading; it was the first place using software to trade beginning with programmed trading in the 1990s to algorithmic trading in the 2000s to high frequency trading (HFT) in the 2010s. It also introduced me to the old joke about the future of the City being one man and his dog. The man is there to feed the dog and the dog is there to stop any human touching the machines. It used to get a big laugh a decade ago or two but, today, no one laughs because it has come true. This was brought home to me in reading a fascinating article in Bloomberg Markets, a magazine that purposefully produces only a printed version with no digital copy available. A bit strange for a global market of people using technology, but it is done on purpose to actually make us read it.
The magazine is released six times a year and the April/May copy caught my eye as it shares the history of the battle between Goldman Sachs, Morgan Stanley and JP Morgan for the prime brokerage market in equities trading. Equities trading has been more and more automated ever the 1980s, but really took off when BATS got a chance to set up equities trading on their servers in a garage in Kanas in 2005. Over 10% of US and 25% of European equities and options trading now runs through their platforms.
However, that is not the focus of this blog update. You can put the origins of this major move to automation down to the 1998 regulation from the SEC that allowed a new breed of electronic trading networks to trade, weakening the role of human traders. This was followed by another kick towards automation in 2001 when the SEC forced firms to quote all exchange trading prices in pennies, rather than as fractions. In 2007, they followed up by demanding that trades take place on whichever venue offers the best price. All of this pushed trading inevitably towards software and servers.
I often make a comment in my speeches today about how AI gets rid of humans, and how Goldman Sachs and UBS’s trading desks are now empty of most human involvement. Vikram Pandit, the former CEO of Citibank and head of equities trading at Morgan Stanley, puts it succinctly:
“We realised that the human touch was interesting but actually a hindrance to what it took to really trade these markets correctly. The only thing you could do is figure out how to automate all the human aspects of trading, understanding what drove stock prices, and then used those algorithms to make markets.”
This is the reason that Pandit believes Morgan Stanley successfully took over the #1 spot for equities trading in 2014, just a few years after the bank almost disappeared. As one executive from a rival bank put it: “In 2010 I remember telling people ‘these guys are roadkill’”. During the 2008 subprime crisis, Morgan Stanley was pushed to edge of failure. They only avoided it thanks to a $9 billion investment from Mitsubishi UFJ Financial Group Inc. And yet, by 2014, they had taken over as the preferred prime brokerage for equities trading all thanks to their long-term investment and understanding of using technology to trade equities.
“Imagine the accumulated learning of having done it for het last 20 years,” Pandit says. The algorithms, he adds, are rules that essentially gather “all the things you did wrong that informs what you should do today that’s right. They’re so far ahead of the game, it’s going to give them an edge for a while.”
I agree. In fact, it’s interesting that the Vampire Squid they call Goldman Sachs had lost sight of the game. According to the Bloomberg article, they believed in high-touch trading with a lot of human support, and it was embedded in their culture. As a result, Goldmans didn’t pay much heed to the rise of HFT and Quant trading, and put their automated trading systems in New Jersey away from the Wall Street trading rooms. It was a chasm that pitted the humans against the technologists and guess who won? The technologists of course.
In fact, it’s interesting to see that the 2000s saw Goldman Sachs making billions in revenues from fixed-income trading ($33 billion in 2009, compared to $13 billion in equities). It made them fat and happy. So much so that they didn’t take electronic trading seriously. As the financial crisis hit and regulation along with risk management meant a major switch from fixed income to equities, Goldmans lost out to Morgan Stanley. The latter had low-touch automated trading, that was far simpler to link and run with than Goldmans human traders.
When they did finally respond, it was a major renovation and rewrite of nearly all their systems in 2014 so that they can now process market data every 20 to 50 milliseconds – half the time it takes you to blink your eye – and make instantaneous decisions about where to route orders for best execution.
Meanwhile, because Goldman Sachs were snoozing, guess who stepped into the fray and took over as another key equities trading contender? Yep, good old JP Morgan. JP Morgan Chase Inc were in the top three of nearly all markets they traded, except equities. Jason Sippel, a 16-year veteran with the company, remembers it well: “Clients used to beat us up when we saw them. We didn’t have a strong risk offering for bigger trades, we were far behind in electronic trading, and we didn’t have a low-latency (HFT) offering.”
Jamie Dimon wasn’t going to allow that to continue and saw an opportunity post-crisis to hire Frank Troise, an electronic trading specialist formerly with Lehman Brothers and Barclays Bank. He was charged with building their new electronic trading division and, in order to protect this new business from the voice traders that would become redundant, Troise was made a direct report to the Global Equities Chief.
It worked. As the overall pool of trading fees went down, JP Morgan more than doubled their market share of global equities trading from 5% in 2006 to 10.3% in 2017.
These learnings amongst the top-tier investment banks is what makes them lethal when they take these technologies to retail and commercial markets. It is because investment markets learn how to deal with revolutionary new technologies first – such as AI and Quantum Computing – that they are the true banks that understand everything they need to know about technology and money. The fact that most others don’t, including many other big bank names I could mention, is what makes JP Morgan and Goldman Sachs truly digital banks.