MC Explains | What is algo trading, why SEBI is seeking to regulate the segment

Algo trading uses algorithms, or a predefined set of commands, to dictate the exact criteria for buying and selling stocks and other assets such as futures and options (F&O), commodities and currency derivatives.

A simple example: Buy 100 Infosys shares every time the rupee depreciates 5 percent against the US dollar. The trades take place faster and manual monitoring is done away with.

It is similar to employing a broker to buy and/or sell shares based on specific instructions—just much, much quicker. Computers analyse the data, the scope for error is low. Such trades step-step significant price changes as the orders are executed within seconds.

Who uses algo trading?

Generally, algo trading is used by mutual funds, hedge funds, insurance companies, banks, and other institutions to execute a large number of high-volume trades that are otherwise impossible for humans to carry out.

Over the past decade, the rise of fintech firms has led to an increase in retail participation. Some of these platforms are Zerodha, 5Paisa, Alice Blue Algo Trading Platform, Fox Trader Algo Trading Platform and Mastertrust Algo Trading Platform.

Any regulation by SEBI could impact such low-cost, fintech-based brokerages that have been adding millions of clients.

What is SEBI’s concern?

Algo trading has a 50 percent share of the Indian financial market (2018), according to the National Institute of Financial Management. In the United States, around 80 percent of all trading are algo-based.

The concern is that while exchanges approve of algo trades submitted by brokers, those by retail investors using APIs cannot be identified as algo or non-algo —by either the broker or the exchange.

SEBI’s consultation paper pointed out that it is these unregulated or unapproved algo trades that pose a risk to the market. It added that these can be misused for systematic market manipulation and to lure the retail investors by guaranteeing them higher returns.

What is SEBI proposing?

SEBI suggests that all orders from an API should be treated as algo orders and be subject to control by stock brokers. APIs must be tagged with a unique algo ID provided by the stock exchange granting approval for the algo, the regulator said.

The regulator said there should be technological tools to ensure appropriate checks “to prevent unauthorised altering or tweaking of algos” and suggested that brokers could either provide in-house algo strategies developed by an approved vendor or outsource the services of third-party vendors.

Recent developments

The National Stock Exchange (NSE) is exploring an "unsupervised machine learning model" to plug anomalies in algorithmic orders, PTI reported this month. In its annual report for 2021-22, the exchange said that it had developed well-curated market surveillance mechanisms backed by a robust technology architecture over the years.

NSE’s surveillance systems identify malpractices and ensure timely management of identified breaches. In order to upgrade and strengthen its surveillance capabilities, the exchange rolled out key initiatives in 2021-22 including deployment of alerts to detect market abuses in equity stock options, OTM (over-the market) contracts, capture multi-leg reversal cases as well as abusess wherein option contracts are traded at away prices without a change in the underlying assets. "Exchange is exploring with an 'Unsupervised Machine Learning model' to detect anomalies in algo orders," the annual report noted.

Algo trading‘s history in India

Algo trading isn’t really new in India’s financial markets. It was introduced and allowed by SEBI in 2008; initially, it started with Direct Market Access and was restricted to institutional investors. After stock exchanges started leasing co-location servers to brokers and fintech firms, retail participation started growing.

In 2012, the regulator put in place broad guidelines for algo trading in the Indian securities market.

“Any order that is generated using automated execution logic shall be known as algorithmic trading,” it had said. It also specified minimum “order-level risk controls” that would include price checks, quality limit checks, a system to identify dysfunctional algos, and the requirement of a monthly report on algo trading by stock exchanges.