Sebi building team to police algo trades

Regulator to hire cyber experts to stay ahead in the game

Markets regulator Securities and Exchange Board of India (Sebi) is in the process of building a team for efficient monitoring and surveillance of algorithm trading, which makes up for a significant portion of cash and derivatives market volumes.

“The market is fast evolving. Traders are creating new algorithms to maximise gains in the market. Sebi needs to stay on top of the game to prevent abuse of the marketplace. With this objective, we are trying to build a large team of software experts,” said a senior Sebi official in the surveillance department.

Sources said the Sebi is screening internal candidates who would suit the profile. The candidates need to have awareness and knowledge of new-age trading techniques such as robo advisors, artificial intelligence and high frequency trading (HFT). The regulator might also be open to hiring external candidates with proficiency in Python, a commonly-used coding language for stock market algos, said another source.

Under the current regulatory framework, the onus is on the stock exchanges to vet the algo program used for proprietary trading. Going ahead, the Sebi intends to more proactively monitor this fastevolving domain.

Although the Sebi receives several complaints on the algorithms used, in a majority of the cases, the regulator has been relying on the information provided by the stock exchange as it doesn’t have a decided team of experts to analyse the data.

“In a majority of the algo cases, the traders claim that the algorithms used were stock exchange-approved. Due to lack of expertise, there is rarely any penal action. Building an inhouse expert team is essential for the Sebi to keep up with the changing contours of trading and prevent any mishap due to these technologies,” a Sebi official said.

Algos, along with other hi-tech tools such as robo advisors and HFT have become very popular in the past few years.

According to industry estimates, nearly 80 per cent of the trade orders placed in domestic markets come from these automated tools. But, algo trades account for only 40 per cent of the trades executed. Developing expertise will also help the Sebi to decide on maintaining an ideal order-to-trade balance.

Algo tools have become so common that brokers are increasingly offering them, even to retail investors.

“Technologies like algo trading help in lowering the trading costs and deepening the liquidity in markets. Despite these benefits, automated trading systems have capacity to disrupt the market if someone misuses it. It is difficult to completely prevent such disruptions; however, you can put systems in place to help mitigate these risks. Having an in-house expert team would help Sebi analyse the trade surveillance data better and in turn improve the surveillance standards,” said Richie Sancheti, head of funds practice, Nishith Desai Associates.

In the past few years, the Sebi has been making attempts to bring appropriate checks and balances to regulate the new-age tools. Last year, Sebi had brought robot advisors—unregulated until then— under the purview of the investment advisors rules. Further, the regulator had also floated a discussion paper proposing curbs on algo trading. In the discussion paper, the Sebi has put forth several proposals, including a minimum resting time for orders, speed bumps to delay order matching, randomisation of orders and review of tick-by-tick data.

Experts say monitoring algos is also needed to keep a check on artificial inflation of stock prices. For instance, consider an algo that is programmed to buy a stock “X” if it falls below say ~10. As soon as X falls below ~10, the algo triggers an order. If there are many algos with a similar trigger, it could lead to a surge in orders. This could drive up the stock price. Also, as the price moves up, it could also trigger cancellation of orders. The Sebi had, in the past, come across several instances where the order-to-trade ratio, which is the proportion of orders placed versus trades executed, was as high as 10,000:1.