Vendors of trade surveillance software are not generally regarded as quasi-godly entities. They scrutinize their customers’ transactions, and send alerts—often, too many alerts—when something anomalous is spotted. Perhaps a trader is using inside information, spoofing, or hiding losses. Most of the time, the trader will be doing nothing untoward at all.

But that describes a vendor’s work with an individual client. What about the vast trove of data that is created when all of those clients are aggregated—especially when the software is used by asset managers with almost a fifth of the sector’s total assets? Suddenly, the humble provider of surveillance tools has a shot at omniscience.

That’s the position TradingHub occupies, and the opportunity it is currently considering.

“The more people have commented on this concept [to us], the more we have thought that the market could be improved by us producing some kind of index of market abuse risk across asset classes,” says David Hesketh, co-founder of TradingHub. 

Analyzing aggregated, anonymized data from consenting clients would enable TradingHub to see if there is a spike in alerts in a particular asset or group of assets, he continues, or track deviations from industry medians.

“That is incredibly valuable information from the client’s perspective, as they can proactively increase surveillance to respond to increasing levels of risk rather than reacting after the event,” Hesketh says. 

Buy-side firms see the potential.

Victoria Paris, head of monitoring at Aviva Investors and a user of the firm’s surveillance service, says she would be interested in any insights TradingHub could derive from aggregated data that might help Aviva strengthen its controls. “If there is anything, I would absolutely love to take it,” she says. 

It needn’t end there, of course. Another way to use the aggregated data would be to compare the prices that individual firms are getting for specific transactions—a best execution tool that would be driven by market-wide analysis, rather than each firm’s experience.

TradingHub is already working on it.

“Most people will look at their transaction execution costs and then also compare themselves to themselves—perhaps look for outliers relative to their typical execution cost—to see whether they’ve achieved best execution. But we can work out the average execution cost for a transaction across the whole market, and the distribution of those costs. And then we can say, ‘Well, is your execution good compared to what other people are paying?’ And that’s a feature people don’t have access to right now,” says Edward Selby, the firm’s head of products.

This service—dubbed Transaction Efficiency and Accuracy Monitor, or TEAM—is currently being trialed with a small number of clients, ahead of a full launch in 2021.

Implausible success

TradingHub didn’t start life as a surveillance play. Hesketh’s original plan—after spending three years as a derivatives lawyer, and a further three as a structured credit trader at Merrill Lynch—was to help buy-side firms measure the performance of their traders. During one uncomfortable client meeting, an epiphany struck—traders that were implausibly successful might be tilting the odds in their favor.

“I remember one client in the early days where we were looking at trader performance; we asked the client why a desk was so profitable. After an awkward silence, somebody said this individual was trading a very illiquid market, had a very large market share and was essentially able to control the price. This gave us the confidence our algorithm could spot potential market abuse by looking at an individual’s trading patterns,” he says.

• Read More: Technologies meant to detect and stop financial crime may add more headaches than they relieve. Others say that the best tech can’t overcome how bankers use these tools. The AML dilemma likely falls somewhere in the middle. Click here to read more.

The firm went on to do some work for a US regulator in 2013, when its software was still in development. TradingHub reviewed data on about a billion trades, provided by the regulator. Not only did the software find cases of market abuse that were not public knowledge, but it also unearthed instances of wrongdoing the regulator itself did not know about—triggering fresh enforcement actions, Hesketh says. 

“I couldn’t really wish for a better endorsement,” he says, adding jokingly, “I would have preferred a contract.”

Known today as Market Abuse Surveillance Tool, one of the service’s distinguishing features is the risk score given to each questionable transaction, Hesketh says. Another is a reduced number of false positives. 

MAST uses a traffic-light system to indicate the risk of each suspicious trade, which it derives from Monte Carlo simulations. The higher the tool’s confidence that the transaction was part of a banned practice rather than just a fluke, the higher the risk score. This enables analysts at bank and buy-side clients to focus on the most alarming cases for further investigation.

For Alan Lovell, global head of surveillance at HSBC, MAST is “game-changing”. The bank uses the software for foreign exchange trades and is rolling it out for equities and fixed income. 

“It’s game-changing because the legacy systems are based on very basic triggers, which are just going off all the time. We get hundreds of thousands of those alerts and every single one of those in the old systems is of equal value until a human has looked at it,” Lovell explains.

The question now is whether TradingHub—and its aggregated data—can change the game a second time.