Securities and Exchange Board of India (Sebi) appears to be a little ahead of the curve when it comes to regulating algorithmic trading. The Indian regulator has been reviewing this segment since late last year and issued additional guidelines for trading members and exchanges in March this year. Some of its peers in the region are currently in the process of framing guidelines for algorithmic trading. Australian Securities and Investments Commission (Asic) and Securities and Futures Commission (SFC), Hong Kong recently issued draft proposals and have invited public comments by mid- and end-September respectively, after which it will frame final guidelines for the segment.

Shyamal Banerjee/Mint

Sebi must improve in the area of transparency and effective communication. As it is, there is a general mistrust towards algorithmic trading, and if the regulator is found to be averse to openly discussing the subject, it will only do more harm.

Having said that, Sebi has done reasonably well in framing guidelines for algorithmic trading. Hirander Misra, chairman of Forum Trading Solutions, which offers stock exchange systems and surveillance services, says, “Sebi has thus far applied a robust yet practical approach to algorithmic trading regulation. Its guidelines fall in line with international best practices and ensure the right level of monitoring and control to ensure market integrity at both broker and exchange level, without resorting to draconian measures such as slowing a system down."

Sebi has also done well to put much of the onus on exchanges to protect the market from rogue algorithms. Australia and Hong Kong are following a different approach, with the onus being on market intermediaries to ensure market integrity. To ensure market intermediaries have safety measures such as pre-trade filters etc. in place, these regulators are proposing to mandate annual testing of their systems and regular audits. In addition, Asic has proposed a fine of 1 million Australian dollars for non-compliance on any of its guidelines.

But monitoring a myriad number of trading firms can be a nightmare. According to a Mumbai-based exchange official, as far as possible, exchanges must act as the main safety net, with its own pre-trade filters and pre-set throttle rates etc to ensure that suspicious orders don’t find their way into the order book. Anant Jatia, director, Forefront Capital Management Pvt. Ltd. says that there are always people out there who want to game the system, and so it makes added sense for the exchange to provide the additional layer of safety. Sebi’s algorithmic trading rules make the exchange primarily responsible for market integrity, but also have guidelines for trading firms.

One can argue that a large number of safety filters can slow things down for trading firms and this will work against the interests of overall market liquidity. But it’s far better to have an orderly market, rather than rogue algorithms disrupting the market every now and then.

Of course, despite Sebi’s seemingly robust guidelines, India has had its share of algorithmic trading accidents. As Misra says, “Some consistency across trading venues is required. This, coupled with best practice controls upstream at broker level will help create a safety net. However, this will never be completely foolproof to totally prevent episodes such as those on Muhurat day trading, as algorithms change frequently based on market conditions. It merely reduces the probability of such issues occurring. Until someone can come up with a detection system that is very predictive and can anticipate problems at the market-wide level and the firm level, ahead of the risk becoming a reality, rather than reactive as most systems are today, the chance of things going wrong will always remain."

Proponents of algorithmic trading say that critics should cut the segment some slack since all automated systems go through problems. This argument, however, doesn’t cut ice because problems in electronic trading systems lead to large financial losses for a large number of market participants. Asic says in its consultation paper, “An aberrant algorithm generates not only costs that are borne by the firms using the algorithms, but also negative impacts for all market participants by impairing the fairness and orderliness of the market." India has had its fair share of mishaps and has learnt this truth from example. Sebi must now ensure that all exchanges are following its guidelines in a consistent manner, and errant trading firms are penalized appropriately, so that adequate safety filters are in place.

Your comments are welcome at