Algorithmic trading refers to orders generated using automated execution logic. Co-location refers to the positioning of servers on the exchange premises so that a broker renting a co-location rack is able to gain in terms of trading costs from the speed at which information about a stock or its derivative flows between the exchange and the servers to which the broker’s trading terminal is connected.
At present, stock exchanges do not allow sharing of co-location racks among brokers. While BSE offers some basic co-location facilities free of cost, at NSE, the cost of taking one co-location rack on rent is Rs35-40 lakh annually. Due to the high fixed costs, only large brokers are able to offer algo-trading facilities to their clients.
The market regulator wants to create a level playing field by directing stock exchanges to allow sharing of co-location racks and limiting the charges exchanges can levy on brokers for availing algo-trading facilities.
To do this, Sebi may ask exchanges to ensure that at least five brokers share each co-location rack at exchanges, by paying the rent to the exchange directly or to the vendor if such co-location racks are sold to a third-party vendor, who is neither an associate of the exchange nor a party related to the broker.
The move, if implemented, will lead to sharing of rent for each co-location rack among brokers, and if Sebi stipulates capping the amount that can be charged from each broker, then small and mid-size brokers too will be able to gain access to algo-trading facilities for a monthly rent of Rs40,000-50,000.
“Sebi’s aim is to allow even small and mid-size brokers to get the benefits of co-location facilities so that they too are able to offer algo-trading services to their clients at reasonable rates. This will reduce the disparity between algo and non-algo users existing today," said the first person mentioned above, on the condition of anonymity.
Kunal Nandwani, founder and chief executive of uTrade Solutions, a trading technology provider, said the move will make algo-trading access for brokers fairer and more equitable.
“Sebi’s move will reduce the entry barrier for brokers into algo-trades. The small brokers and traders minimum cost to co-location for faster algo-trading shall drop by at least 1/5th compared to now if Sebi’s move is implemented. This in turn should lead to deeper liquidity in the market, and more trading volumes and opportunities for all stakeholders," added Nandwani.
In August 2016, Sebi had in a discussion paper proposed seven new ways to level the playing field between those using high frequency trading (HFT) systems and regular market users.
The proposals included introduction of a minimum resting time between HFT orders, matching orders under a batch system, introduction of random delays of a few milliseconds in order processing, revision of the order queue randomly every 1-2 seconds to discourage time-sensitive trading strategies, capping of the order-to-trade ratio, employing separate queues for orders coming from co-located servers as well as those from other servers, and reviewing the tick-by-tick data feed for all market participants.
“The norms that were proposed in the discussion paper may take some time, because before finalizing them Sebi wants to study observations and recommendations made by the IIT Bombay professor who has been given the assignment to suggest rules for algo.
Brokers have shown a high demand to Sebi for changing the co-location rules first.
Sebi has asked exchanges two months back to recommend ways to rationalize algo-trading costs so that gradually the gap between algo and non-algo users reduces.
Sebi is still examining the best possible options but the new co-location rules may come in the next three-four months," said the first person cited above.
Experts feel Sebi’s move may help in increasing the proportion of algorithmic trades in India and matching global standards.
K. Suresh, president of the Association of National Exchanges Members of India (ANMI) said Sebi’s co-location norms may increase algo-trading volumes by at least 10% in the near term and eventually help in reducing algo-trading costs for all clients.
“The cost is a deterrent for small and mid-size brokers. Such brokers may not need the full rack. Their volumes are not so big and hence they may need only a small share of the rack to provide algo facilities only to some of its clients. Sebi’s move will therefore help. ANMI is also told that exchanges may rent out their co-location racks in small portions at one-tenth of the cost," he said.
“In India, algo-trade constitutes around 40% for the cash market, in the US it is 93%, in the UK it is around 80% and the global average is around 75%. Sebi’s move will not only result in more algo-based trades but also create a level playing field between algo and non-algo traders. Some of the norms suggested by Sebi in the discussion paper are bit harsh. They should be tweaked and this move on co-location should help in finding a balance," added Suresh.
The main reason for using algorithms is to cut down the average cost of trading, which is why the concept is growing fast globally. According to a 9 January report by Business Insider, the global algorithm trading market is expected to grow from $8.79 billion in 2016 to $18.16 billion by 2025 at a compound annual growth rate CAGR of 8.7% between 2017 and 2025.
In May 2015, Sebi announced some norms to ensure fair and equitable access to the co-location facility and to ensure that such a facility does not compromise integrity and security of the data and trading systems.
In its discussion paper, Sebi had mentioned that as some market participants across the globe have highlighted the concern of unfair access and inequity for non-co-location, non-HFT participants vis-à-vis participants using trading algorithms and co-location to trade, securities market regulators are examining various proposals to address such concern.