By Liz Hoffman and Telis Demos
Meet the straders.
Part risk-taking trader and part computer-whiz "strategist,"
they are prowling the halls at Goldman Sachs Group Inc., erasing a
once-religious line between the jocks and the nerds.
"You say 'trader' and I don't even know what we're talking
about," said Adam Korn, a 16-year Goldman veteran. "Everyone who
comes to sales and trading needs to know how to code."
Mr. Korn is the unofficial king of the straders, and an
evangelist for the financial world they represent. Across Wall
Street, traders who spent their formative years barking into phones
are signing up for coding classes. Engineers once relegated to the
back office are being empowered to try their hands in the
market.
It is upending the pecking order of the trading floor and is, in
large part, a concession to the reality that has set in a decade
after the financial crisis.
Wall Street traders buy and sell everything from stocks and
bonds to bundled credit-card debt and oil. Before the 2008
meltdown, they thrived on instinct and an informational edge. They
worked the phones, sussing out which customers were hungry and
which ones were desperate, and pounced on weakness. "There's blood
in the water," Morgan Stanley chief John Mack would tell his
traders. "Let's go kill." They did, and were richly rewarded for
it.
Those days are largely gone. Today, snippets of code, sometimes
called algorithms, do much of the job of a trader. They keep tabs
on the banks' positions, generate price quotes for clients, match
buyers and sellers, and flag unseen risks.
They are increasingly doing the job of a salesman, too: The
latest software can suggest which clients might be interested in a
particular stock or bond by analyzing their recent investments, the
same way Amazon.com can suggest items inspired by a customer's
purchases
The rise of automation is partly a response to the financial
crisis and the rogue trading scandals that followed, which
encouraged banks to take discretion away from error-prone and
ego-driven humans. It owes partly to a talent war between Wall
Street and Silicon Valley, with banks eager to stress their tech
bona fides.
It is also a response to the rise of computer-driven "quant"
funds. These investors ignore traditional stock-picking methods and
instead hunt for patterns in the market that signal a buying or
selling opportunity. When a bank like Goldman comes calling, these
clients more often want to talk about data processing than, say,
dairy production. "Being able to speak that language" is important,
Mr. Korn said.
For traders, succeeding in this world depends less on trusting
one's gut than being able to interpret what the computer is
spitting out -- and what ought to be fed in. "Think of it like
cruise control," said Matt Cherwin, a trading executive at JPMorgan
Chase & Co. "It can make the car go 55 miles an hour, but
someone needs to decide, 'well, is that the right speed?'"
Tech whizzes aren't new to Wall Street. They arrived in the
1980s to computerize the trading floor and program mathematical
models that could value new, complex instruments known as
derivatives. The code they wrote predicted how a drop in the U.S.
dollar might affect corn prices, for example, or how to value a
loan to Ferrari if the Italian government raised interest
rates.
But these programmers were distinct from -- and distinctly
subordinate to -- traders, who used their models to decide how much
corn or Ferrari debt to buy or sell. "Strats," as they came to be
known at Goldman, crunched numbers. Traders made money.
On Goldman's stock-trading floor in lower Manhattan, strats were
relegated to a corner. Today they don't just sit with traders.
Increasingly, they are the traders, licensed and empowered to put
the bank's capital on the line.
Their rise mirrors what is happening in the broader economy.
Technology can displace workers, particularly in rote tasks like
factory production. But more often, it changes what's expected of
them, thrusting employees once confined to the back office into
front-of-the-house roles and forcing those already there to
adjust.
Tax software has automated much of the grunt work for
accountants, who are now rebranding themselves as trusted advisers,
not bean counters. Architects have ditched their slide rules for
computer software, but big buildings don't get constructed without
them.
On Wall Street trading floors, quick thinking under pressure and
a deep rolodex are still prized. So traders haven't been replaced
by technologists so much as merged with them -- the "strader"
hybrid on the rise at Goldman.
"Ten or 15 years ago, the engineers were the ones who didn't
speak to anyone and maybe seemed like they hadn't showered that
day. The traders were the ones that looked like they stepped out of
a Brooks Brothers catalog," said Oliver Cooke, a financial-industry
recruiter at Selby Jennings. "That line has basically
disappeared."
It's not just the fashion. Trading floors were once a cacophony
of orders being shouted and phones being worked, and occasionally
thrown. Today they're surprisingly quiet, said Steve Grob, director
of group strategy at Fidessa, which sells trading systems to banks.
"Much less shouting, and much more thinking," he said. "If you put
someone in a time machine from 2006, they would be amazed."
At Citigroup Inc., traders work alongside coders from the bank's
quantitative analysis group. But in June, traders themselves were
offered a three-day introductory course in Python, a coding
language. The offering proved so popular -- even veteran traders
were willing to be away from the floor for three days -- that the
bank plans another session in September, and is also considering a
hybrid technology-trading training program.
"The tools continue to evolve, and continue to become more
sophisticated, " said Lee Waite, Citigroup's former head of North
American markets, who is now the bank's country head for Japan.
"That has us thinking about the future of trading, and what sort of
person we need in those roles."
Goldman started offering free computer-programming classes to
its trading staff last year through edX, an online classroom.
"Programming is going from a 'nice-to-have' to a 'must-have,'"
Mr. Korn said.
An applied math and economics double-major at Brown University,
he comes from a family of computer geeks: His father worked at Bell
Labs and in the 1980s helped develop the backbone of the Unix
operating system. He came to Goldman in 2002 and spent his early
years as a strat in stock-trading, among those stuck in the corner
on the 50th floor of New York Plaza.
That began to change in the early 2000s. The stock market became
increasingly electronic, setting off an arms race for code that
would give banks an edge. Strats weren't only building trading
models, but entire trading systems.
One early piece of software at Goldman split a big order to
smaller pieces and routed them off to different exchanges. They
called it "V.I. Joe," short for Virtual Joe, a nod to the human
whose job it did. Today, it's a standard order-management system,
and every bank has one.
As the technology became more sophisticated, Goldman's traders
understood less about what was happening under the hood. A trader
trying to fix a misfiring algorithm could do little better than a
befuddled homeowner when the WiFi goes out -- press reboot and hope
for the best. "That created a risk for us," Mr. Korn said.
The "flash crash" of 2010, when the stock market lost and then
recovered $1 trillion of value in minutes, added to the concern.
Across Wall Street, traditional traders were mostly helpless, while
the engineers best equipped to explain the whipsaw weren't in a
position to do anything about it.
Some executives are wary of blurring the roles too much, worried
that a few lines of glitchy code written by a relative novice could
wreak havoc. Firms as big as Knight Capital Group have been brought
down by computer snafus.
And remember the rogue traders that cost JPMorgan and UBS Group
AG billions of dollars? A rogue coder, deliberately planting
snippets meant to siphon off funds to a personal account or blow
past risk limits, could be just as dangerous.
"There must be a clear delineation of responsibility," said Mike
Dargan, group chief information officer at UBS. On the Swiss bank's
markets desks, "scrum teams" of traders and coders collaborate to
quickly to roll out new functions, but aren't encouraged to start
to do each other's jobs.
"We want traders to be educated" about how their software works,
Mr. Dargan said. "But we want them trading."
Regulators have raised similar concerns. In 2016, the Securities
and Exchange Commission changed its rules to require anyone
responsible for designing or overseeing a trading algorithm to
become licensed as a "securities trader." That industry designation
separates people who are empowered to put a firm's capital at risk
from those who aren't, and holds them accountable for positions
they take on.
The SEC was trying to limit screw-ups or malfeasance stemming
from the knowledge gap between people who write the code and those
responsible for managing the market risk that code generates.
Mr. Korn took the required exams, sending his wife and young
children on vacation for a week so he could study. He proposed a
new role to his bosses at Goldman -- strader -- and they went for
it. That working title has been replaced by the more approachable
"traders who code," and today the bank employs about 200 of
them.
Among them: Joe Montesano, whose order-routing job was
programmed into "V.I. Joe" back in the 2000s.
(END) Dow Jones Newswires
August 18, 2018 00:14 ET (04:14 GMT)
Copyright (c) 2018 Dow Jones & Company, Inc.
Goldman Sachs (NYSE:GS)
Historical Stock Chart
From Mar 2024 to Apr 2024
Goldman Sachs (NYSE:GS)
Historical Stock Chart
From Apr 2023 to Apr 2024