If, for instance, the charges of collusion and unfair access are true, the market structure changes proposed by Sebi alone will not help
The Securities and Exchange Board of India (Sebi) said it is examining various options to allay concerns about unfair and inequitable access to stock exchanges. It has invited public feedback on seven proposals to curb algorithmic trading.
In March this year, a Sebi-appointed committee had concluded that some traders had unfair access to National Stock Exchange’s market data and trading systems (click here to read more). This was about a year after a whistleblower made allegations about collusion between NSE officials and certain traders. It is likely Sebi’s fears and concerns about algorithmic trading have increased because of the charges and the committee’s adverse findings.
Mint reported on 27 May that NSE has refuted charges that it provided unfair access (click here to read more). But here’s the unfortunate part—the NSE case hasn’t been brought to a closure yet. The regulator hasn’t disclosed whether it has found wrongdoing or not. This is critical for the issue at hand.
If, for instance, the charges of collusion and unfair access are true, the market structure changes proposed by Sebi alone will not help. If any governance issues are left unaddressed, no amount of rule changes will curb the menace of unfair access. In fact, wrongdoers may gain more if market structure is made complex.
The regulator must first penalise those involved in the wrongdoing, if any, as well as in any attempted cover-up (click here for more). Else, it would be akin to punishing the entire market for the sins of a few.
Be that as it may, Sebi’s discussion paper on algorithmic trading cites other reasons why rules need to be revisited. The paper states, “Academic literature also indicate that algorithmic trading may have accentuated the issues of adverse selection costs for non-algorithmic traders and (may have) increased probability of ‘flash crashes’ vis-a-vis the situation in the pre-algo era." Unfortunately, it doesn’t list the literature it is referring to.
Besides, some of Sebi’s proposals are not well thought out. Market participants had assumed its bizarre two-queue proposal from 2013 had been given a quiet burial; but it has resurfaced in the discussion paper.
Under this system, stock exchanges would implement a new order-handling architecture, comprising two separate queues for co-located and non co-located orders. The order-matching engine would pick an order from each queue alternatively for execution. According to a CEO of a leading trading software provider, not only will this lead to unnecessary complications and hurt the market, but high frequency trading firms will also find ways to be at the top of both queues. As such, the intent of providing equitable access to non-co-located traders will not be achieved at all. Besides, the move can lead to a bureaucratic nightmare on what exactly qualifies as non-co-located order flow. Sebi’s earlier paper suggested servers situated in close proximity to exchanges would be treated as co-located order flow.
Some other proposals such as minimum resting time have been considered by other regulators such as the Australian Securities and Investments Commission and the European Commission, but have been dropped. See for a detailed discussion.
Interestingly, however, the Toronto-based TMX group has introduced a new order type called Long Life, which has a minimum resting time of 1 second. The intent is to make its markets attractive for natural investors who do not compete on speed. But the exchange group has made this feature optional, with those closing the Long Life order type getting priority with execution.
In the US equity markets, too, solutions to address concerns about adverse selection costs for non-algorithmic traders have come from the market participants. The new exchange by the IEX Group, the hero in Michael Lewis’s Flash Boys, is a prime example. Its trading architecture includes a speed bump—one of Sebi’s proposals—to deter participants who compete primarily on speed.
ICAP’s EBS, a forex trading platform, aggregates orders received in a 3 millisecond time-frame and then randomises their place in the queue. In other words, the first order to arrive at the trading platform may not necessarily be the first to get executed, hence minimising the need for ultra-fast speed. Venkatesh Panchapagesan, adjunct professor, finance & control area, Indian Institute of Management-Bangalore, says randomisation of orders is the least disruptive of Sebi’s seven proposals. Some of the other proposals can result in unintended consequences where the costs to the markets far outweigh the benefits.
But note that in the examples given above, individual exchanges made changes to attract natural investors rather than high-frequency traders. Is it best then to leave it to market participants to find solutions to problems associated with algorithmic trading?
While that may be the Chicago tradition of leaving it to the markets to find a solution, it may not always be the best approach. With high-frequency trading, regulators need to answer the question of whether they can be up-to-speed with market participants when it comes to market surveillance. While this column has argued earlier that Sebi should use the same tools as high-frequency traders, that now looks like a distant dream.
Besides, as some academicians have highlighted, the current market structure incentivises a technological arms race, which can be socially wasteful expenditure. Panchapagesan adds there may be a case for slowing down the markets using the least disruptive method.
Sebi would have done itself a favour by presenting data from the Indian markets to support this view. Hopefully, it can redeem itself by including such data before it tinkers with market structure, as well as clear the air on the NSE matter.
Subscribe to Mint Newsletters
* Enter a valid email
* Thank you for subscribing to our newsletter.
Never miss a story! Stay connected and informed with Mint.
our App Now!!