Algo trade: Why Sebi wants a clamp down

First things first—what is algo trading?

Algo trading refers to the use of computer programs (or algorithms) to buy and sell shares at lightning speeds, often in milliseconds. This type of trading relies on complex mathematical formulae to trigger buy/sell orders without human intervention. Algorithmic trading was introduced on Wall Street in the 1970s. It arrived in India in 2008 and initially only institutional investors were allowed to deploy this powerful tool. The field was opened for retail investors in 2016. Currently, algo trading accounts for 50% of the volumes in equity derivatives and over 65% of cash market trading.

What does this trading chain look like?

There are three main players—the stockbroker, the client (proprietary or retail traders), and third-party algo service providers. Some traders know how to write their own algos but most investors use third-party algorithms. Clients subscribe to various plans offered by these service providers and can run their algos on the broker’s server through an application programming interface (API). The use of APIs allows clients to access brokers’ trading platforms without manually logging in. However, when clients execute orders through the API, neither exchanges nor brokers can identify if they are algo or non-algo trades.

What are the market regulator’s concerns?

Many algo trading service providers flout Sebi rules by promising ‘guaranteed’ returns to gullible retail investors. Some also missell basic algos as complex trading strategies or inflate their record. These developers are outside Sebi’s regulatory ambit, even though they give investment advice. Algorithmic trading strategies are offered on various unregulated platforms.

What are the latest proposals from Sebi?

Sebi reportedly held a meeting last month with market infrastructure institutions, stockbrokers and algo providers. The regulator proposed stopping all open APIs where the broker is unaware of how it is being used by the client. It also suggested two models. Under the first, stockbrokers will have to get approval for their algo platforms and take the responsibility for data security. As per the second, algo platforms would be regulated and can share their past performance only after it is verified by a performance validation agency.

Are there any systemic risks from algo trades?

Since algorithms are based on specific market conditions, any unexpectedly volatile or ‘Black Swan’ event can cause algos to misfire in unison, triggering a market meltdown. Two freak trades in domestic derivatives last year were blamed on algos. Even institutional investors’ algos can go haywire, but these are vetted by the exchanges, so there’s a layer of supervision. That’s not the case with algos used by retail clients, who mostly use APIs, preventing exchanges from knowing if the trades are algorithmic or manual.