Banks look to 'big fast data' to meet regulator demands

Increasing pressure from regulators to improve data monitoring capabilities in the light of several recent market crises is leading Australian banks to invest in real-time market surveillance tools.

Software AG vice president Dr John Bates, who is in Australia talking to banks and others about ‘big fast data’, said the group had recently sealed a deal with a large Australian bank.

“Like many banks around the world they are getting a lot of regulatory pressure to make sure that they police their own systems well and run an orderly trading floor,” Bates told iTnews.

“We’re doing real-time market surveillance, allowing banks, exchanges and regulators to find patterns that indicate that things are going wrong, such as rogue traders, market abuse or things going outside their normal operating parameters, and actually stop it before its too late.”

Earlier this year, the Australian Securities and Investments Commission introduced new rules requiring banks and stockbrokers to boost their monitoring and data storage capability, as part of a review of dark liquidity and high-frequency trading.

At the time the regulator described algorithmic trading as “the new normal".

Now, the same technology that has been used by banks for algorithmic and high-frequency trading, as well as compliance reporting, is being used to help identify and prevent rogue traders.

In June, Software AG acquired Apama, a company originally co-founded by Bates and later sold to Progress Software. Apama specialises in complex event processing, analysing large volumes of data and looking for patterns and correlations.

Apama already counts the Commonwealth Bank as a customer of its platform for algorithmic trading.

“Businesses increasingly have a shorter and shorter time window to make high quality decisions before the data goes stale,” Bates said.

Apama has built up a library of around 50 well-known patterns that help organisations identify things like insider trading, but Bates said such systems were continuing to evolve as they relied more on self learning.

“The key thing is new (patterns) are emerging all the time, and some can be specific to geographies, so we’re learning that some things are specific to Australia, as they always are.”

As real-time processing of data improves, monitors can incorporate real-time market data, real-time news, information shared on messaging networks and even the keystrokes and logins of traders.

“What you can do with all that information, as well as large quantities of historical information, is say what is normal. We know what time traders come in, what time they login, what they trade.

“Once you know the benchmark of normal and are continuously updating that in real-time you can spot something abnormal when it happens.”

Real-time reporting

Wolters Kluwer Asia-Pacific finance market manager Matthias Coessens said his firm was also seeing increased demand from clients for real-time reporting for compliance purposes.

Wolters Kluwer recently signed ANZ Bank up to its data analytics and reporting platform.

“A core issue for reporting is the mapping and categorisation of the organisation’s data to the many different reports required by the various regulators and other stakeholders,” Jack Cornford, head of Markets Finance Systems at ANZ said in a statement.

“With this solution, core mapping is done once by Wolters Kluwer Financial Services, saving us and their other customers the time required to duplicate this high risk and high cost exercise.”

Coessens said ANZ would initially be using the platform to deliver a single view of transactional trading data across its risk and finance units, informing things like hedging strategies and budget forecasting.

He said in the future ANZ could use the platform for real-time tracking of anomalies, but for now it was focused on adjusting its source systems to deliver real-time data, and improving data quality.

“Most of the institutions face data quality issues and also most of the time resources are spent purely on process and not that much on making decisions,” Coessens said.

“We start off with one central data architecture and on top of that the platform, then run different sets of calculations that go across risk and finance, such as liquidity.”

Coessens said with increased reporting and monitoring was coming increased volumes of data.

“We can’t overwrite any data so the need for big data sets has increased. We’re working very closely with Microsoft to handle the big volumes nowadays.

“On top of that the trend in the industry for real-time information makes it more complex…the speed of being able to do things and make decisions is very important, things like intra-day liquidity and intra-day P&L use allow people to react as soon as possible.”