High-frequency trading (HFT) firms suffer out-of-control algorithms more often than commonly thought and must impose stricter controls and better testing procedures to limit risks in the future, according to a study by the Federal Reserve Bank of Chicago.
The study also finds that HFT firms may at times be unaware of their overall risk exposure and often skip certain recommended checks that could prevent or limit erroneous trades so as to avoid slowing their trading speeds.
The Chicago Fed interviewed more than 30 technology vendors, broker-dealers, futures commission merchants (FCMs), proprietary trading firms and exchange and clearing house professionals as part of the study, which was launched after erroneous trading at Knight Capital Group on August 1 inflicted a $440 mn loss on the company within 45 minutes.
‘Two of the four clearing broker-dealers/FCMs, two thirds of proprietary trading firms and every exchange inter¬viewed had experienced one or more errant algorithms,’ Chicago Fed senior policy specialist Carol Clark writes.
Researchers find that out-of-control algorithms are more common than anticipated prior to the study and that there are no clear patterns as to their cause.
The study concludes that the Knight Capital erroneous trading could have been prevented through profit and loss limits, a limit on the number of orders that can be sent to an exchange within a given period of time, implementation of a kill switch to halt trading, or intraday limits to maximum positions a firm can take in a single day.
‘Many firms fail to imple¬ment all the recommendations or rely on other firms in the trade cycle to catch an out-of-control algorithm or erroneous trade,’ the study notes.
‘In part, this is because applying risk controls before the start of a trade can slow down an order, and high-speed trading firms are often under enormous pressure to route their orders to the exchange quickly so as to capture a trade at the desired price.’
The study also shows that trading firms that design their own trading systems often apply fewer checks to some trading strategies than to others ‘in order to reduce latency’. Some firms also lack sufficiently thorough processes to develop, test and deploy code used in their algorithms, the study adds.
‘A few trading firms interviewed said they deploy new trading strategies quickly by tweaking old code and placing it into production in a matter of minutes,’ the study says.
Chicago Fed researchers cite the example of one firm that suffered two rapid-fire incidents of out-of-control algorithms: ‘To address the first occurrence, the firm added additional pre-trade risk checks,’ they write.
‘The second out-of-control algorithm was caused by a software bug that was introduced as a result of some¬one fixing the error code that caused the first situation.’
