HSBC claims quantum edge in bond market

• Key insight: HSBC and IBM showed a quantum computer could better predict winning bond trades in opaque markets.
• What's at stake: Success could mark the beginning of a new frontier of computing in financial services, moving quantum from a future concept to a near-term application.
• Forward look: While the results are promising, HSBC does not appear to have taken the step of trading real bonds with the help of these quantum machines.

Overview bullets generated by AI with editorial review

HSBC on Thursday said it had evidence that algorithmic trading enabled by quantum computers could outperform its existing algorithms that solely use classical computers.

The bank and IBM, the computing partner on the project, described the evidence as coming from a "promising trial" and "experiment," according to a press release from the two companies. Specifics on the methods and results remain proprietary, and the companies did not execute live trades as part of the trial.

In the experiment, a computing approach that used both quantum and classical computers delivered "up to 34% improvement" over classical-only methods in predicting the likelihood that a proposed price for a trade in the European corporate bond market would be accepted, according to IBM and HSBC.

In other words, the quantum computer helped improve HSBC's ability to determine how likely it was that a trade would be filled at a quoted price.

HSBC

"This achievement is important because it's the first example of using quantum on real industry data," said Jay Gambetta, vice president of IBM Quantum.

The companies also claimed in the press release that this was "the world's first-known empirical evidence of the potential value of current quantum computers for solving real-world problems in algorithmic bond trading."

How the experiment worked

The trial explored the ability of quantum computers to optimize bond trading by estimating the likelihood that a trade would be filled at a quoted price. Specifically, the trial involved predicting winning customer inquiries on the European corporate bond market.

This improvement optimized so-called requests for quote in over-the-counter markets, where financial assets such as bonds trade without a centralized exchange. Without this centralized source of information, over-the-counter markets are less transparent, making it harder to obtain pricing information.

Because pricing data is opaque and fragmented in the corporate bond market, predicting which quote will be successful is a complex problem with many hidden variables.

IBM and HSBC said their experiment proves what quantum computing experts have long claimed to be theoretically possible: A hybrid quantum computer can analyze complex relationships and probabilities more effectively than a classical computer, giving them a predictive edge in a market defined by incomplete information.

Where HSBC expects to go from here

HSBC

HSBC emphasized the immediate competitive advantage the technology offers. Philip Intallura, HSBC group head of quantum technologies, called the result a clear signal for the financial sector.

"This is a ground-breaking world first in bond trading," Intallura said. "It means we now have a tangible example of how today's quantum computers could solve a real-world business problem at scale and offer a competitive edge, which will only continue to grow as quantum computers advance."

Intallura said the focus remains on the near-term application of quantum technology.

"We have great confidence we are on the cusp of a new frontier of computing in financial services, rather than something that is far away in the future," he said.

HSBC

The complexity of the problem underscored the power of the quantum-enabled approach, according to Josh Freeland, global head of algo credit trading at HSBC.

"At one point there were 16 physicists and AI machine learning researchers working around the clock trying to achieve the same thing that the quantum computer did," Freeland said.

HSBC used IBM Heron, a quantum computing system, to augment classical computing in the trial. This hybrid approach — using both quantum and classical computing — is how many experts expect banks and other companies to integrate quantum computers into their systems.

This hybrid approach enabled the system to better unravel hidden pricing signals in noisy market data compared to standard, classical-only approaches HSBC used.

How HSBC's results compare to other banks'

HSBC focused its experiment on increasing prediction accuracy to achieve increased margins and greater liquidity; other banks have demonstrated gains in risk management and processing speeds during their own trials with quantum computers.

Ally Financial in 2022 partnered with Multiverse Computing and Protiviti to use a hybrid computing approach to optimize investment portfolios. Their research showed the method built portfolios that outperformed the risk profile of a target index by up to two times.

The quantum-enabled algorithm also dramatically reduced complexity, requiring four times fewer stocks for the Nasdaq 100 fund and ten times fewer for the S&P 500 fund compared to traditional portfolios.

Researchers noted that a portfolio optimization calculation that sometimes takes up to 30 hours on classical computers runs "almost instantaneously" on a quantum computer.

Similarly, JPMorgan Chase last year said it demonstrated a quantum speedup on a scientific problem known as random circuit sampling. The bank is heavily focused on optimization problems and preparing algorithms for derivatives pricing, portfolio optimization and fraud detection, but executives said at the time that they are still waiting for the hardware to become powerful enough for real financial problems.

Most financial institutions that have put effort into researching and developing quantum computing have so far come up with promising results but little in the way of practical applications.

HSBC's results announced on Thursday mark another promising result for quantum computers that exist today, but the bank has not yet executed live trades with the help of these machines.