SEBI’s proposal to regulate Algo/Hi frequency trades

This has resulted in a vigorous debate in media, amongst stock brokers/investors and the public. There have been some views that SEBI should not regulate such matters at all since, amongst others, this creates hurdles in the development of technology . The suggested methods have also been critically analysed. On other hand, there have been other views that SEBI should indeed regulate such matters on ground such that some parties obtain  certain special and unfair advantages through such trading. There are also concerns that these are being abused in a manner that the public and perhaps even SEBI does not realize such abuse, considering the sheer complexity involved.

Algo trading has increased exponentially. Indeed, the volumes are so large that just two figures should highlight it. As per SEBI, 80% of all orders and 40% of all trades are now generated through computer algorithms.

However, algo/hi-frequency trading have a dark side too. There has been a history in the United States of it being abused by certain traders to make huge profits at the cost of investors. There has been a huge debate over this in India too when SEBI is said to be investigating the alleged role of National Stock Exchange in a similar context.

Algorithmic/hi-frequency trading (“Algo trading”) is also said to have resulted in market crises (notable amongst these is the so-called Flash Crash of 2010 in the USA).

On other hand, there are obvious advantages of Algo trading including that of higher liquidity, lower spreads, etc.

These types of trades are also not understood well by investors and the public generally. Hence, the recent SEBI consultation paper can be a good opportunity to consider the background of the subject and some related matters.

Some concepts
Algo trading is conceptually simple to understand though, in practice, the manner in which such trades are carried out can be quite complex. The SEBI paper has explained some basic terms that are worth a review. This will also help one understand the various measures suggested by SEBI for regulating them.

Algorithmic Trading
The paper describes it as – “Algorithmic trading (for brevity, Algo), in simple words, is a step-by-step instruction for trading actions taken by computers (automated systems). Typically, trading algorithms enable the traders to automate the process of taking trading decisions based on the preset rules / strategies.”.

To put it simply, in Algo trading, the process of placing trades is automated using computers. Software is developed incorporating detailed instructions when to buy/sell, etc. and it monitors market data and places trades accordingly. There is nil or minimal human intervention. There are several advantages. The first, obviously, is very high speed. The time taken by a human operator to press a few keys is in computing time astronomically higher than the time the algo trading software takes to place/execute the order. Secondly, in case of repetitive situations, where the decision making follows standard parameters, it does not make sense using human intermediaries. Further, this also enables traders to carry out large trades usually at microscopic margins.

Hi-frequency trading (HFT)
Hi-frequency trading is really a type of Algorithmic trading. Algo trading as explained earlier is software-based trading with nil or minimal human intervention. HFT involves carrying out of extremely fast trades in very small fraction of seconds often taking advantage of the edge in information over others. The paper explains HFT as:-

“High Frequency Trading (HFT) is a subset of algorithmic trading that comprises latency-sensitive trading strategies and deploys technology including high speed networks, colocation, etc. to connect and trade on the trading platform. The growth and success of the high frequency trading (latency sensitive version of algorithmic trading) is largely attributed to their ability to react to trading opportunities that may last only for a very small fraction of a second.”

Co-location
Co-location (“Colo”) is considered to be a contentious issue. It basically means providing stock market intermediaries/hi-frequency traders’ servers a physical location that is very near stock market servers. Often, the servers are in the same building that the servers of the exchange are located in. Physical nearness to the exchange servers that receive and process trade data is critical since nearer the physical location to such servers, the faster can a intermediary/hi-frequency trader can receive and send back data. And thus act and profit on it, particularly if one is a hi-frequency trader.

High order-to-trade ratio
This means that the ratio of orders placed over actual trades executed is very high. The rest of orders are cancelled. This again is a common feature of HFT.

Issues faced
SEBI has identified the following issues that arise out of Algo trading and related aspects:-

(i) Unfair access or denial of faster access to persons not having co-location facility. To take a simple example, a person from New Delhi is physically quite far from the stock exchange servers in Mumbai and thus suffers a time disadvantage (even if of fraction of seconds) as compared to a person in Mumbai.

(ii) There is more price volatility.

(iii) HFT imposes costs on other market users

(iv) Algo trading results in a technological arms race.

(v) In times of high volatility, SEBI would get limited opportunities to intervene etc.

Solutions suggested by SEBI
SEBI has placed for discussion certain solutions. These are explained below with their advantages/disadvantages  including experience in regard to these solutions in other countries.

(i) Minimum resting time for orders:- Under this method, each order is not allowed to be modified/cancelled till a minimum resting time elapses. This will ensure that the order will be available for some time for execution and thus fleeting orders would be reduced. It is interesting to note that the resting time proposed is 500 milliseconds (1 second = 1000 milliseconds). Thus, this would affect only those parties whose orders undergo change in fractions of seconds.

(ii) Frequent batch auctions:- Orders for a specified period of time of 100 milliseconds will be grouped together and matched, instead of the continuous order matching mechanism. Thus, the advantage of time that a person may have over others owing to co-location, better technological equipment, etc. would be neutralised to an extent.

(iii) Random speed bumps:- This involves delaying orders randomly by a few milliseconds. The result is that this neutralises to some extent the speed advantages.

(iv) Randomization of orders received during a specified period of say 1-2 seconds:- Thus, the orders received during this period would be shuffled randomly and their time sequence altered. All orders within a specified period would have an equal chance and once again the speed advantage is neutralised.

(v) Maximum order to trade ratio:- This will ensure lesser fleeting orders and also that orders are entered into the system with a greater opportunity of their being converted into trades.

(vi) Separate queues for co-location and non-co-location orders:- One order from each queue would be taken alternatingly. Once again, the objective of neutralising speed advantage may be achieved to an extent.

(vii) Providing tick-by-tick feeds to all market participants:- Tick-by-tick data feed, as SEBI describes, “provides details relating to orders (addition+ modification + cancellation) and trades on a real-time basis”. This data is provided by exchanges for a fee. SEBI has suggested that data of top 20/30/50 bids/asks, market depth, etc. be provided to all. This would create a level playing field to all participants irrespective of their technological or financial strength.

Consideration of solutions
The opposition to regulating Algo trading is on various grounds. The first, of course, is that SEBI would be putting hurdles to technological developments and this would not be a wise thing to do. Further, each of the methods suggested create their own inequities. There would also be software and other changes required to provide for such solutions. There would need to be regulatory check to ensure that these solutions are in place. Interfering with such trading would also result in higher ask-put price differences, lower liquidity, etc. Some of the solutions offered, as SEBI itself points out in the paper, have been rejected in some places where they were originally proposed or adopted.

Having said that, there are abuses that need to be considered. While SEBI has already mandated fair, transparent and equitable rules in granting nearness to exchange servers, there have been concerns about this in one way or the other. Further, the sheer complexity of algo trading may result in a group of insiders abusing it to their advantage by prior arrangement particularly if the exchange or its staff plays truant. Thus, the measures suggested may, even if indirectly, help control such abuses. Further, SEBI may also need to regulate such trading to prevent such abuses.

Abuse of hi-frequency trading
Serious abuses have been pointed out from HFT arising out nexus between HFT traders on one hand, and brokers/exchanges on the other. Through a complicated mechanism including giving preferential treatment to HFT traders, it has been found internationally (and allegedly in India too to an extent) that HFT traders hugely profited at the cost of investors. By monitoring quotes on multiple stock exchanges, they came to know in advance impending orders. Effectively, they thus bought (or sold) cheap and sold high (or low) to investors who were not only slower but were also duped by alleged unfair underlying understanding. This has been described in lucid detail in the bestselling book Flashboys by Michael Lewis.

In India too, there is a shadow of this. A whistle blower wrote to SEBI and Moneylife (a financial magazine) about alleged irregularities by National Stock Exchange. It appears that SEBI is looking into this matter.

Thus, abuse of HFT trading can be a serious issue. The HFT traders, as described in the book Flashboys, do not profit in large amounts per trade. Their skimming is small amounts. But on a cumulative basis, they would make large amount of profits. There is a cascading effect of this. Investors end up paying higher price. In turn, this raises the cost of capital for companies seeking to raise capital from the markets. Generally, this would harm the crediblity of markets too.

Conclusion
In the author’s view, SEBI is wrong in proposing measures to slow down the speed of trades/data exchange. This would be restraining developments in technology. Indeed, it is submitted, this is not the real issue at all. The real issue is alleged inequitable access to speedy information and alleged abuse of algo trading through irregular means. For this purpose, SEBI would have to understand and keep pace with the technical developments in algo trading and closely monitor such trading and provide for mechanism to monitor trades and uncover abuses. While the existing Regulations of SEBI relating to frauds and unfair trade practices are general and perhaps broad enough to cover such abuses, SEBI may consider providing for certain matters specifically, describe them in detail and provide for punishment.