Efficiency of Indian markets – we do serve free lunch here

Years ago, the Nobel laureate Milton Friedman of the University of Chicago argued famously that there is nothing called free lunch. Why? Because, at least in the context of capital markets, market forces through arbitrage activity will ensure that pricing inefficiencies will not persist for too long.

Electronic trading powered by algorithms (or “algo trading”) now ensure that “too long” is measured in milli- or even micro-seconds.

We in India take great pride in the fact that we went electronic years ago; we are surprised that exchanges in New York and Chicago continue to operate floor based markets.

At a recent conference organized by the Futures Industry Association in Mumbai, a senior official from the National Stock Exchange of India (NSE) announced that over 60 per cent of the trading activity on that exchange is now driven by algorithmic trading.

This is clearly a very impressive number, and indicates that local and global market participants are actively seeking trading opportunities.

Algo trades

Traders love exploiting pricing inefficiencies. The simplest and least risky strategy is to buy low and simultaneously sell high. For example, if a stock is trading cheaper on the Bombay Stock Exchange (BSE), then a trader will buy on BSE and sell it on NSE.

Assuming the cost of executing such trades is not too high, then these easy pickings are essentially equivalent to the “free lunch” described by Friedman.

Simple algos can be written to keep sniffing continuously for such opportunities, and thus ensure that nobody is serving a free lunch in the Indian stock markets.

On May 31, our markets served free lunch for at least about 30 minutes — the last half hour before close. The trigger for it was a pre-announced decision by MSCI, a global index provider, to reconstitute its MSCI India Index, an index tracked by many of the largest institutions in the world. Billions of US dollars are benchmarked to this index.

The reconstitution required these firms to sell the stocks being excluded from the index, and to buy the ones being included. This is standard practice in the global markets, and well anticipated. Fund managers wait till the close of trading on the day prior to the reconstitution to execute their trades. This ensures that their funds' net asset values (“NAVs”) track market prices at close. The funds tend to be agnostic of the actual pricing of the stocks.

May 31, 2011

So what happened in India on May 31, 2011? There was a huge spike in volumes on the NSE in the last 30 minutes of trading. The volume jump on BSE was slightly muted. Clearly, most of the foreign institutional traders were being executed on NSE. Knowing what we know about our traders' expertise and skill in exploiting arbitrage opportunities, the pricing on both exchanges should have been more or less the same.

The ask prices on BSE were consistently lower on BSE than NSE; the difference spiked up to over Rs. 3 per share around 3:25 pm. This was the same story across all the six stocks (Dabur, Mundra, Titan, Bank of India, Shriram and Asian Paints).

The fact that the market was unable to arbitrage away the price inefficiency brings up the obvious question: Why did this not happen? Given that algo trading is such a predominant force in our markets, why did we not see the same firms deploying their algos to hit the BSE bid-price and simultaneously, hit the NSE ask?

There is a simple answer – prior to deployment of algos, firms are required to seek approval from the exchanges. While SEBI rules are not explicit about it, the exchanges are at a liberty to frame their own rules on top of that prescribed by SEBI.

An algo to exploit arbitrage opportunities across exchanges should be able to send orders automatically to the different platforms. As of date, no firm has been able to get any such multi-algo approved. While BSE's approval process is not too granular in seeking details of the algos, NSE's is.

While there are ongoing discussions about the lack of depth in our markets, it is always instructive to consider specific rules in our overall regulatory framework which might be hindering the growth of market liquidity. This restrictive policy towards multi-exchange algos is one such example. It also impacts ability of our member firms to service their retail investors.

Clearing mechanism

There are other factors that contribute to the inefficiency of our markets including the fact that if an investor has purchased a stock on BSE and later sells it on NSE the same day, the delivery obligations cannot be netted. Being able to do so would reduce the cost associated with having to fund both delivery obligations – carry inventory of the stock, fund the purchase, and pay transactions tax on both trades. In the absence of a single clearing corporation, the best way forward is to build inter-operability among clearing houses, similar to what we have in place for the depositories.

In the context of clearing houses, interoperability would mean that an investor (operating via a member firm) can choose to have all his/her trades cleared and settled through a single clearing house; all margins and collateral will be held with a single clearing house, which will also handle all the payment and delivery settlement obligations on behalf of that investor. Connectivity among the clearing houses will allow the investor to trade on an exchange or platform of his choice, including buying on one exchange and selling on the other without having to fund/deliver on both obligations, as the connectivity will net off his positions.

In summary, while the reforms of the early 1990s have resulted in robust trading, settlement and risk management systems, inefficiencies still persist.

Implementing inter-operability among clearing houses, allowing netting of delivery obligations among different segments, along with a less restrictive approach towards multi-exchange algos will go a long way in ensuring that the pricing inefficiencies we observed on May 31 will not be repeated in the Indian markets. Free lunches will then disappear overnight.

(The author is Head — Product Strategy, Bombay Stock Exchange Ltd)

COMMENT NOW