The Australian Securities and Investments Commission (ASIC) is piloting a range of big-data policing tools including the use of “predictive analytics” to spot financial crime before it happens.
In a speech to the American Chamber of Commerce (AMCHAM) in Sydney last week, ASIC chair, Greg Medcraft, said predictive analytics would help the corporate cop in “understanding the future and can provide insight into the probabilities of future outcomes”.
“It can generate red flags that help identify risk areas to organise our resources – and help us to become more pre-emptive and proactive,” Medcraft told the AMCHAM audience
While the ASIC technology isn’t quite at the level imagined in the 2002 Steven Spielberg movie ‘Minority Report’ – where pool-based ‘precogs’ in togs dreamed up future murderers – he said the regulator had a “number of trials underway exploring the power of predictive analytics” such as “the use of cognitive sentiment analysis”.
“Firstly, using social media analytics to monitor sectoral hotspots in financial services,” Medcraft said. “Second, we are exploring how we can use natural language, voice and text algorithms to better detect and respond to misconduct.”
ASIC was also toying with automated ‘market announcement learning’ and a ‘hidden account trading’ detective tool that uses the sci-fi-sounding ‘Support Vector Machine’ “to read market data and detect previously unidentified accounts employed by fund managers to disguise their trading manner”.
Medcraft told the AMCHAM crowd the regulator could also go the whole cog as the “relatively new field” of ‘prescriptive analytics’ develops further.
“At its best, prescriptive analytics will predict not only what will happen, but also why it will happen, and will provide recommendations to take advantage of the predictions. This type of analytics can help to identify how to respond to the red flags we see,” he said.
“The evolution of this area of data analytics will no doubt provide us with new opportunities and capabilities in the future and will be an important development down the road.”
In the meantime, ASIC has already invested heavily in simpler big data tools covering ‘descriptive’ and ‘diagnostic’ analytics, Medcraft said, with the former including fund manager surveillance using data from Morningstar, Bloomberg and the Australian Tax Office.
He said the regulator currently uses diagnostic analytics to monitor the Australian equity market as well as for “investigations and enforcement”.
“For instance, we are implementing new third party software for use in investigation and enforcement matters, which allows pattern matching across our extensive evidence database, using algorithms,” Medcraft said.
“It allows us, for example, to map relationships of persons and entities and create chronologies using metadata, well beyond traditional word or enhanced word search capabilities.”
ASIC has also created a Chief Data Office to co-ordinate and drive the regulator’s new analytical offensive.