Prof. Jayanth R. Varma's Financial Markets Blog

About me       Latest Posts       Posts by Year       Posts by Categories

Why market surveillance can no longer be left to the exchanges

I wrote a column in the Financial Express today arguing that the financial market regulators need to get directly involved in real time market surveillance.

Traditionally, securities regulators globally have regarded the exchanges as the front line regulators with primary responsibility for market surveillance. As a result, regulators have traditionally not invested in the computing resources and the human capital required to perform real time surveillance themselves. A number of developments are making this model unviable in the developed markets and the same factors are at work, a little more slowly, in India as well.

I think it is time for Indian regulators like Sebi, FMC and RBI to develop in-house real time market surveillance capabilities rather than rely on the capabilities that may currently exist at the exchanges or exchange-like entities that they supervise (NSE, BSE, MCX, NCDEX, NDS).

I believe there are two key factors that make this regulatory shift necessary. First is the dramatic change in the nature of exchanges themselves. In the past, exchanges were regarded as ‘utilities’ providing key financial infrastructure and regulatory services. In recent years, they have evolved into businesses just like any other financial services business. Many observers in India (including some of the exchanges themselves) have been concerned about this transformation, but this is a global phenomenon and it is delusional to deny this reality. Concomitantly, there has been a blurring of the line between exchanges and brokers. Globally, alternative trading systems and dark pools have gained market share in recent years, and the operators of these systems are half way between traditional exchanges and large broker dealers, in terms of their business models and regulatory incentives.

In India, too, we have seen the blurring of the line between exchanges and non-exchanges. Examples include the subsidiaries of regional stock exchanges that trade on national exchanges; the exchanges in the commodity space whose promoters had or have large trading arms; and RBI regulated entities that perform many functions of an exchange but are not legally classified as exchanges.

The second and even more important factor is the rise of algorithmic and high frequency trading that links different exchanges together at much shorter time scales than in the past. Each exchange looking only at the trading in its own system has only a very limited view of what is happening in the market as a whole. It becomes very much like the story of the six blind men and the elephant.

The best example of this is the flash crash in the US on May 6, 2010. The US SEC, which like other regulators had never dirtied its hands with real time surveillance, found itself struggling to figure out what happened in those few turbulent minutes on that day. In an interim report, the SEC stated: “To conduct this analysis, we are undertaking a detailed market reconstruction, so that cross-market patterns can be detected and the behaviour of stocks or traders can be analysed in detail. Reconstructing the market on May 6 from dozens of different sources and calibrating the time stamps from each source to ensure consistency across all the data is consuming a significant amount of SEC staff resources. The data are voluminous and include hundreds of millions of records comprising an estimated five to ten terabytes of information.”

This is what happens when a regulator leaves it to others to do its job, but is forced one day to do the job itself. Is it not scandalous that a systemically important institution like an exchange or a depository is not required to synchronise its clocks to a standard time (say GPS time) with an error of not more than a few microseconds at worst? Exchanges are willing to spend a fortune to bring down the latency of their trading engine to a millisecond or so to attract trading volume, but are unwilling to spend a modest amount to synchronise their clocks because nobody asked them to.

There is another important hidden message in this. Modern finance is increasingly high frequency finance and those who do not dirty their hands with it become increasingly out of touch with the reality of financial markets. Doctoral students in finance today, for example, have to learn the econometrics of high frequency data and grapple first hand with the challenges of handling this data.

Unless regulators collect this high frequency data and encourage their staff to explore it, they risk becoming progressively disconnected with the reality that they are supposed to regulate. Interestingly, the US derivatives regulator, CFTC, is moving rapidly to develop this capability. They already collect all trade data on a T+1 basis and run their own surveillance software on that data. Over the next year, they hope to enhance this to receive the entire order book data from the exchanges that they regulate. All regulators worldwide need to move in that direction.

It is true that this will be difficult, expensive and time-consuming for Indian regulators. That is all the more reason to start immediately.

Posted at 12:19 pm IST on Wed, 16 Jun 2010         permanent link