By Lionel Laurent
LONDON (Reuters) - The financial industry's multi-billion-dollar technological race for stock-market supremacy is petering out.
After years of boosting profits with super-fast and super-efficient trading algorithms, deemed better and cheaper than human alternatives, banks and brokers are now facing persistent cost pressures, sceptical clients and more regulation that have effectively put a cap on the rewards from more spending on technology.
The industry is gradually evolving from a race against the clock - with banks trying to keep up with the demands of high-speed automated, or algorithmic, stock trading - into a less frenzied market where trading systems are increasingly generic and where competitive edge is sought in non-technological areas.
"The algorithms today are all very similar and increasingly homogenised," said David Miller, Head of EMEA Trading at Invesco Perpetual. "There are some differences in the specific strategies ...(But) the technologies are very similar."
It's a trend that some investors say is encouraging.
"The latency (transaction-speed) race is done," said Edmund Shing, a portfolio manager at BCS Asset Management in London.
"What matters more is liquidity and access to risk capital. And for the big asset managers that may mean one day cutting out the middle man entirely."
Trade-execution technology, on which the financial sector spends $2.8 billion per year, has become highly efficient: the time needed to process a trade has fallen by 90 percent over the past decade, according to consultancy GreySpark Partners.
But shaving yet more fractions of a second off trades costs a lot of money and means diminishing returns in a trading environment where - despite soaring stock prices - commission fees have been flattened, volumes have yet to significantly pick up and volatility is at its lowest in seven years.
Even high-frequency traders who fuelled the race by demanding ever-faster routes for their own algorithms are finding it harder to squeeze out gains: high-speed-trading revenue in U.S. equities is down from $7.2 billion in 2009 to an estimated $1.3 billion in 2014, according to research firm TABB Group.
"The differentiation between the algorithms is starting to fade," said Frederic Ponzo, managing partner at GreySpark, who said European equities markets would soon be as commoditised as the U.S. market.
"(Cutting microseconds) costs an arm and a leg compared with what you are able to charge your clients. The arms race has run its course."
This does not mean that banks are packing up their computers and hiring back seasoned traders.
On the contrary, some bankers say that the shift in client concerns away from speed and towards better access to deeper pools of investors - which can be a challenge now that buyers and sellers are spread across various opaque "dark pools" as well as regular exchanges - still requires bespoke technology.
"The variance among algos has definitely gone down," said a bank executive who asked not to be named. "But we still invest and find ways to improve performance across the board. And sometimes that improvement is significant."
But even those who praise the quality of banks' trading algorithms as cheaper and better than the manual trading that was once the norm say pressure is now rising on the industry to outsource more technology and find other ways to compete.
For Devesh Vishwakarma, head of hedge fund Premaeus Investments, choosing a brokerage is as much about geographic breadth and diversity of products as it is about trading speed.
This may not be too hard for the dominant actors, who can soak up market share as more marginal players drop out of the equities business, as demonstrated by Royal Bank of Scotland (RBS.L) and Credit Agricole's (CAGR.PA) respective cutbacks.
But finding a niche is not always easy. Investors are awash with research, analysis and trading ideas.
"A few years ago it was a great thing for us to say to our clients that we were using algos, they were a great selling point," said Mark Ward, head of execution trading at brokerage Sanlam Securities, who works with bigger banks to route trades.
"Now, everyone is selling algos ... It's affecting brokers as well as the banks. It just makes life a little more difficult."
In the end, the question remains as to how long banks can defend a sector that is increasingly encroached upon by others.
Rival trading venues like Liquidnet now offer asset managers the chance to trade with each other without going through an intermediary. This is not without risks: a U.S. regulator on Friday fined Liquidnet $2 million to settle charges that it improperly used subscribers' information to market its services, though Liquidnet says it has taken steps to tighten procedures.
Bankers dismiss the idea that fund managers will control the future of trading. But with banks facing more regulation and under pressure as a result of their own high technology spending, there is more scope for competition from outside.
"(Tech) has come back to bite the banks," said a trader.
(Reporting by Lionel Laurent; Additional reporting by Clare Hutchison; Editing by Giles Elgood)
- Banking & Budgeting
- trading algorithms