All Eyes On SEBI’s Proposed Algo Trading Regulations

With the rampant misuse of API by unregulated market participants leading to fraudulent activities resulting in losses to the investors, the SEBI proposed that any trade executed through API should be classified as algo trade

The nascent form of algo trading was formally introduced in India in 2008 through Direct Market Access (DMA) facility. Algo trading (short for algorithmic) includes any type of automated rule-based trading where decision making is delegated to a computer model. High Frequency Trading (HFT) is a type of algo trading, which is latency sensitive and is characterized by a high daily portfolio turnover and high order-to-trade ratio.

The exchanges provide additional advantage to trading members and data vendors where they could locate their trading or data vending systems within or in the proximity of stock exchanges, called co-location or proximity hosting facility. This is to minimize or reduce the time (latency) for sending orders to the trading system of the exchange.

By automating, the users or traders need not worry about being online when the transactions are done or when manually entered.

This is done through API (Application Programming Interfaces) given by the broker to develop their own trading algorithm. The rudimentary form of such a system is the provision of an order limit set by the broker, indicating the price at which an order could be placed by the trader, provided in advance so that the transaction is completed when the price of the security reaches that level.

Over the years, the regulator through various circulars issued in ’13, ’15 and ’19 has enforced broader framework on this subject by facilitating algo trading with audit requirements, fair and equitable access to their co-location and even allowed direct connectivity between two colocation facilities of the stock exchanges. With the rampant misuse of API by unregulated market participants leading to fraudulent activities resulting in losses to the investors, the SEBI proposed that any trade executed through API should be classified as algo trade and must obtain approval from the exchange.

While this brings transparency to the system, the approval process is time-consuming and at times not opportunistic as the market is dynamic. Any delayed approval could turn redundant if the market trends change not favouring a particular algorithm.

Also, issues like unequal access to algo trades due to lower latency and availability of Tick-by-Tick (TBT) data, contribution to price volatility, market noise (excessive order entry and cancellation), denial of profit opportunities to other investors, clogging the pipe that carries the orders thus slowing down the order messages of other investors have drawn regulatory attention.

To benefit the retail investors, the recent consultation paper has categorised algos into two types: White box/Execution algos and black box Algos. The former is transparent and can be replicated by the users, while the latter remains undisclosed. To provide such algos, the providers should register as research analysts (RA) and maintain detailed reports on the logic and functionality. This makes all third party API providers to come under the regulatory framework and thereby protect the investors.

The regulator also proposed certain thresholds on order size and speed to be considered as algo trades. Accordingly, the exchange may additionally publish minimum, maximum and mean latencies at the 50th and 99th percentile.

It was publish reference latency a regards the time taken for the order message to travel between a reference rack in the colocation facility and the core router. This allows the trading members to compare the actual latencies in the colocation facility vis-à-vis the reference latency published by the exchange.

Another amendment is the review of testing requirements for software and algos. Now, the exchanges may be allowed to provide a simulated market environment for testing of software, including algos. Such a facility may be made available over and beyond the current framework of mock trading. After assessing the robustness of the facility, the decision to phase out with monthly mock trading may be taken in consultation with SEBI’s appropriate technical committee.

These are part of the proposals made by the consultation paper that SEBI has recently floated. We will have to wait for the actual law once the feedback is received by the Board.

(The writer is a partner in Wealocity Analytics, a SEBI registered Research Analyst)