Big Data in Financial Markets Regulation – Friend or Foe?
Morgan Deane, member of the Board and International Head of Legal & Compliance for the Helvea-Baader Bank Group.
January 18, 2015
The Swiss National Bank’s decision to remove the Swiss Franc/Euro peg this month provided an abrupt reminder of how fragile financial markets can be. A simple, albeit unexpected, decision by a major central bank led to mayhem in the markets.
Hundreds of millions of dollars were lost in one day and some brokers were forced into bankruptcy. One of the big questions in the aftermath was how the risk management divisions of the banks and brokers did not see it coming?
It is interesting to consider these events in the context of Big Data.
Why? Because, on this occasion, the mayhem was not caused by Big Data, but Big Data may have been able to help insulate firms from the fallout.
Big Data is viewed by sections of the financial community as a problem child.
Not by the business and revenue generators, but by the functions whose job it is to provide control and manage risk.
While the trading floors of institutions around the globe have explored the benefits of Big Data – generating better returns and performance for investors – Risk Control and Audit functions have viewed the developments with a greater degree of trepidation. Technology continues to push the frontiers of finance, whereas control functions often scramble to keep up.
If one looks at the number of books published to simplify the mechanics of Bitcoin, or at the reams of articles pointing out the perils of regulating advanced financial technology, one can appreciate how unsettling big data may look to some people.
Public perception hasn’t helped either. Scepticism has grown out of the negative press which big data has received.
High Frequency Trading is known for the effect it had on the financial markets in 2010, when the “flash crash” led to the biggest intraday decline witnessed in the history of the Dow Jones Industrial Average.
Scepticism has also been fostered, somewhat unfairly, by success. A number of hedge funds have developed algorithms which analyse market sentiment. They gather and process huge volumes of online information and social media, enabling the funds to enter or exit market positions sometimes minutes before the general market can react to news. Consequently, funds such as MarketPsy Capital’s returns were viewed with suspicion by those who didn’t understand the technology.
Books such as ‘The Fear Index’ further served to create a public perception which isn’t always deserved.
A more balanced view of Big Data is needed because it can serve as a friend of regulation. Rather than being seen as something which can’t be controlled, it should be seen as a method through which control and regulation can be achieved in a way never possible before.
Some regulators have signalled their intention to use analytical methodology to determine what risk profiles to assign to firms under their supervision. Such an approach is a sensible one as it steps away from trying to “pre-regulate” technological innovation. Instead it allows firms to continue advancing technology in their business, and rather than stifle it, regulation can then use similar methods to control it as necessary.
Within the industry itself, firms could employ Big Data for surveillance which is already required and performed. For years, financial institutions around the globe have relied on clumsy tools to identify market abuse activity and to prevent money laundering. These systems have laboured with keyword searches to identify if inside information is being shared between parties on opposite sides of the Chinese wall. Surveillance of trading activity has looked for obvious red flags but has often produced vast amounts of “false positives” which need to be further reviewed by teams of control staff.
The “three Vs” of Big Data – Volume, Velocity and Variety – hold huge opportunities in this regard. Instead of monitoring emails with the naïve expectation that someone will actually use phrases such as “insider trading” or “don’t tell anyone”, Big Data can scan through millions of emails for trends, unusual communication patterns, and compare these to any number of investigative benchmarks. All in a matter of seconds.
Similarly, surveillance of transaction activity can be deepened to include comparisons with historical activity, behaviour from other participants in a much broader field of reference. Again, all in real time.
These examples are new solutions to relatively old and simple challenges. But it shouldn’t stop there. It is not impossible to envisage new ways in which Big Data could identify trends, or indeed scenarios, which would have helped brokers and banks around the globe to be better prepared for the fallout of the Swiss National Bank’s decision. Firms employ some of the smartest people on the street to manage their risk, but even these people have their limits. Equipped with the power of Big Data, however, these people could achieve their goals in a highly effective manner.