LONDON (Reuters) - Carl James became a trader when he was 19, in the boisterous telephone dealing rooms of the 1990s.
Now, he and his colleagues at Pictet Asset Management are about to start learning computer coding to help them understand the technological changes sweeping through one of the final frontiers of ‘old-style’ trading.
Stocks and currencies mostly moved to electronic and automated trading platforms over the past 15 years but the opaque $100 trillion-plus world of government and corporate debt has been a laggard.
It is catching up. Nearly two-thirds of government and corporate bond orders in the United States and Europe are now placed electronically on average, according to 12 large asset managers, banks and trading venues interviewed by Reuters. Volumes are also rising in more specialized markets such as junk debt, emerging markets, repurchase agreements and swaps.
The change has accelerated since new European Union rules in 2018 made it more cumbersome to carry out trades over-the-counter, they said. Electronic platforms make the required time stamping and data reporting easier.
This has contributed to a reduction in staff in trading rooms and a shift toward hiring quantitative analysts, often mathematicians. It has also dented fixed income revenue at some banks - the middle men who have oiled bond market trading.
James is Pictet’s head of fixed income trading and as recently as 2016, his team conducted four-fifths of his deals the old way - entrusting contacts at banks to quote him competitive prices by phone.
Four in every five trades at Pictet, which manages around $200 billion, are now placed electronically - tapped into a machine aggregating prices and analyzing data from dozens of banks and trading venues.
“You can make decisions based on the data rather than just the relationship,” said James.
Eventually, “the relationship won’t matter at all,” he said.
He has hired a data expert and the coding trading should help the traders to better understand the new technology they are using.
Some fund managers have gone one step further since the introduction of the EU’s new MifiD II rules to improve transparency. They have built systems where computers replace humans in deciding when to buy and sell.
Proponents say the shift to electronic and automated trading should make the bond market more transparent, lower costs for the ultimate buyers - asset managers like pension funds - and give fund firms power held by banks to get the best prices.
Along with lower market volatility and tighter regulations, the shift to electronic trading has contributed to a reduction in staff on trading desks at banks and fund firms, banks and asset managers said.
Analytics firm Coalition estimates the 12 biggest investment banks employed 1,400 fixed income sales and trading staff at end-2018, down from 2,300 in 2010 and with a year-on-year fall in each of the past eight years.
Dutch lender ING last year rolled out Katana, an artificial intelligence tool to identify and trade small emerging market bond deals.
Katana is designed to free up traders to focus on higher-value business rather than replacing them, said global head of credit trading Santiago Braje, but he said such tools allow desks to do more with fewer people. He said Katana had increased ING’s deal “win rate” in emerging markets by 20 percent.
The sources at the banks, asset managers and trading platforms told Reuters there was a shift toward hiring quants, who usually have a degree in maths, physics or computer science, or staff with a data background. Sometimes they are trained as traders.
One head of multi-asset trading at a German asset manager said that of the nine traders on his team, at least five have hybrid roles that have a focus on technology.
Fixed income revenues, headcount: tmsnrt.rs/2Elxkws
The shift has also impacted investment banks. Fixed income revenue is down at the top twelve, due in part to the shift to electronic trading. They earned $5.4 billion in 2018, down from $6.3 billion in 2016 and $9.1 billion in 2010, Coalition data shows.
“On the issue of lower profits in fixed income among the world’s biggest banks, the growth of electronic trading has played a major role,” said Coalition research director Amrit Shahani.
Ed Wicks oversees Legal and General Investment Management’s trading desk. Britain’s biggest fund manager, it invests more than 160 billion pounds ($202 billion) in fixed income assets.
Wicks says the buying and selling of corporate debt is now done in seconds whereas before it took hours.
Three-quarters of the fixed income trades by ticket on his desk are now made electronically. He also expects 15-20% of his bond book to be automated - that means trades selected and executed by computers - this year, from less than 10% today.
U.S. primary dealer holdings: tmsnrt.rs/2Wg9o7S
Daily trading volumes for government bond and credit products on electronic platform Tradeweb rocketed to $368 billion on average in 2018, up two-thirds from 2016.
“The first battle was between getting people off the phone and clicking a mouse,” said Billy Hult, president of electronic platform Tradeweb, which last month went public with a valuation of nearly $10 billion.
“Now the mouse is going away and it’s about engaging with clients through algos and more sophisticated tools.”
Additional reporting and graphics by Saikat Chatterjee; Editing by Anna Willard