The market regulator has asked for the details of the trades, and has initiated investigations into factors that led to unusually high volumes. “Trade annulment is a serious issue,” a senior Sebi official noted. “This is the first time an entire session has been cancelled in so many years. We are looking into it.”
The regulator is also likely to look into deficiencies, if any, in the process of approvals and risk management systems, followed by the exchanges for such algorithmic trades, officials said.
|Flawed algorithm by Delhi-based Share India generated heavy volumes|
|Algorithm was executing both legs of the trades|
|Share India’s User-id disabled and enabled several times|
|Losses could have been manifold in regular session|
|Sebi probe into risk management and approval processes|
On October 26, BSE said it had cancelled all derivatives trades during the Muhurat session after volumes shot up to around Rs 27,0000 crore. This is almost ten times the average daily volumes clocked on the BSE’s cash platform. Blanket annulment of trades can put even genuine traders and investors who have hedged positions at risk. Also, there was a high movement in the Sensex from 17,000 levels to 14,000 levels. If the event had happened on a normal trading day, losses could have been manifold, say experts.
It’s now a closed matter, adds Ashish Kumar Chauhan, deputy CEO, BSE. “We cannot comment further on surveillance issues.”
Initial inquiries have revealed that the massive volumes were a result of a flawed algorithm used by Share India Securities, a Delhi-based entity for its proprietary trades. Share India had entered an immediate or cancel (IOC) order for a certain number of Sensex Futures. Without waiting for a confirmation a market order for the same trade was placed, resulting in a double entry. Due to the imbalance created, the algorithm then repeated the trades several times, resulting in the unusually high volumes.
BSE had suspended the broker from taking any further proprietary positions with immediate effect. Yashpal Gupta, a member of promoter family of Share India, said the suspension has not been lifted yet and directed further queries to his brother Praveen Gupta, the chief executive. Praveen could not be reached for comments.
Share India was earlier known as FMS Securities, and had settled charges of irregularities on several counts on a consent basis with the market regulator last year.
“Exchanges have limits for each member crossing which his user ID will be disabled,” said the official. “In this case, apparently the member could trade again once the positions were brought down to permissible levels. This was repeated several times. That led to the trouble.”
Experts said the same algorithm seemed to have executed both legs of the trade resulting in these extraordinary volumes despite heavy movement in the index.
“In a genuine live trade, where two parties are involved, the member would exhaust his limits,” notes Sandeep Tyagi chairman and managing director, Estee advisors, a firm that specialises in algorithmic solutions for wealthy investors. “This would act as a natural risk control. But when the same algorithm was buying and selling, it would not see much losses. Limits would come down only to the extent of the transaction costs.”
He describes these as “teething troubles” in building a robust algorithm. “Extremely strong risk management needs to be built and necessary checks need to be put in,” adds Tyagi. “The risk controls required for algorithmic trading are same for a normal trade. Only the speed is different and the data needs to be accurate.”
Algorithmic trading has been a contentious issue ever since it was introduced in 2009. While BSE openly accused rival National Stock exchange (NSE) of deliberately not clearing inter-exchange algorithms, NSE had expressed concerns over security measures, risk management aspects of algo trades. While NSE had demanded demonstrations and testing of these algorithms, BSE had offered to give applicants approvals on the same day.