Remember me

Register  |   Lost password?

The Trading Mesh

Predictive Analytics: Fighting Financial Services Fraud

Mon, 10 Aug 2015 04:11:01 GMT           

By Nigel Farmer

 

By combining trade and data monitoring with human surveillance and predictive models, financial services firms can anticipate and prevent fraud, errors and other market-moving events. 

When major financial markets fraud happens it sends a shiver through the industry. The world’s financial services industry is huge, estimated to comprise anywhere from 12% to 19.5% of the total global economy. So when something big does occur, be it a stock market crash or a country’s bond default, the reverberations are felt around the world.  Not only can careers be destroyed and firms sunk but markets can be demolished, taking economies down with them.

Despite efforts to regulate and punish, fraud continues to occur right under our noses. In 2012, a trader known as the London Whale hid his positions and lost JPMorgan Chase at least $6.2 billion before he was found out. 

That same year we discovered that a cadre of interest rate traders conspired to manipulate the London Interbank Offered Lending Rate (LIBOR). They were using instant messaging platforms and chat rooms, where they barely tried to hide their efforts. This activity went on for about seven years before someone blew the whistle. 

In another private messaging chat room, foreign exchange (FX) traders conspired to move the FX benchmark price for about four years before regulators caught on in 2013. More recently in Brazil a similar investigation is taking place as authorities look into alleged FX manipulation, naming 30 bankers as part of the probe who may have “shared competitively sensitive information in chatrooms with the aim of fixing prices to make bigger profits,” according to the Telegraph.

Major events other than fraud can also shake the foundations of financial markets. When Switzerland de-coupled the Swiss franc from the euro, banks lost tens of millions of dollars and some smaller companies went under.

Many regulators and financial services firms are taking steps to monitor trades and traders to spot aberrant behavior and anomalous trades, using streaming analytics. They look for patterns of misbehavior so that crimes can be spotted, stopped and tracked back to the perpetrators. This is all good, but it often means that these crimes have already been committed. 

In future, pre-trade risk and trading algorithms could employ predictive analytics to monitor volatility, P&L, etc. to detect/predict abnormal market activity, and adjust risk controls and trading algorithms in real-time to minimize losses. 

Using predictive analytics models on top of continuous streaming analytics means that regulators and financial services firms can predict with a good deal of certainty when something bad might happen. Insider trading, market manipulation, even money laundering could be predicted and halted before markets are affected. 

For example, using the fast Big Data streaming into and around an institution and pairing it with historical data from a firm’s internal operations can help a bank to predict what activity is likely to be going through its trading systems at any given time. If the actual activity on its systems is anomalous, and the predictive model signals that the chances are high there is insider trading going on, the bank can swoop in and stop it. 

Visibility is the key. To avoid the next flash crash or to prevent market abuse and risk-management disasters, regulators and financial services firms should be hungry for visibility, constantly monitoring positions for risk on a real-time basis and observing traders and back-office staff for anomalous behavior. Only by being vigilant and applying predictive analytics can they stop financial crimes before they happen. 

, , , , , , , , , , , , , , , , , , , , , , , ,