Traders see big money in software-based trading

Algorithmic trade accounts for a fifth of derivatives volumes on NSE.
Algorithm trading, or programme trading, common in the US and Europe, has swiftly gained ground in India over the past year, as volumes picked up after the crunch.

Algorithmic trading uses strategies that exploit short-lived market opportunities and depend highly on execution speed. Essentially, set software programmes decide when, how and where to trade, without the need for human intervention.

While the monopoly stock exchange for derivatives in India, the National Stock Exchange (NSE), is yet to make public data on algorithmic trading, conservative estimates by some top brokerage houses reveal that programme trading may account for 15-20 per cent of the total equity derivatives volumes. Brokers said by next year, algorithm trading could account for over 30 per cent volumes.

According to official figures, programmed trading generated nearly 30 per cent volumes on the New York Stock Exchange this year. The average daily turnover in the equity derivatives segment on NSE over the past month has been nearly Rs 100,000 crore, while the combined cash market volume on NSE and the Bombay Stock Exchange (BSE) is over Rs 15,000 crore.

Rashesh Shah, chairman of one of the countrys top broking firms, Edelweiss Capital, which generates over Rs 4,500 crore volume in domestic equity markets daily, said: Algorithmic trading has picked up recently. Nearly 20 per cent volumes in the options segment and 10 per cent in futures come from algo trading. Shah said algorithm trading was behind the recent sharp rise in options volumes. Options generate nearly half the derivatives volumes in the country.

DK Agarwal, director of SMC Global, another top arbitrager, said average daily algo trading volumes at their company were Rs 200-400 crore. SMC has spent over Rs 10-15 crore for setting up algo trading facilities so far over the past one year and intends to scale up significantly, he said.

SMC had a manual arbitrage desk of over 350 people. However, after it entered algo trading, only 15-20 people were engaged in manual trading and over 70 per cent of the arbitrage business was through algo trading, said Agarwal.

Algo trading in India is not as advanced as in the US, where buying or selling through software is pegged to movements in currency and commodities markets and even other corporate news. The commonly-used strategies in India include selling of stock futures quoting at a premium to the spot, and buying the shares, thereby locking in the difference. Another strategy is to sell shares on the exchange where the price is higher and simultaneously buy back when prices fall, or vice-versa. For brokerages, this generates meaningful profit, as securities transaction tax in the derivatives segment has been abolished.

NSE has already signed with 60 members for a co-location facility, whereby they can place their trading servers close to the exchanges engine for Rs 22.5 lakh on a first-come-first-served basis. Co-location saves crucial milliseconds from the time it takes to place an order and its receipt at the other end. The broker with his server next to the exchange engine gets a price feed that is updated every three-four milliseconds, while a broker at a remote place will get this feed updated every 30-40 milliseconds. SMC, which had applied for four rack spaces with NSE, was allotted two. It would be allotted the other two soon. Each rack can easily handle two servers, each of which can handle orders worth Rs 200 crore.

Well, I am a little confused due to this line.. "For brokerages, this generates meaningful profit, as securities transaction tax in the derivatives segment has been abolished."

Any comments ?