How easy is it to manipulate a financial market?

A lot of people in India believe (a) that market manipulation is easy and (b) that it happens all the time.

An excellent blog entry by Michael Stastny offers some good reasoning on the issues. The key insight is to distinguish between the ease with which a price can be distorted by a manipulative cartel and the ease with which the cartel is able to walk away holding profits. In the past, I have called this `the Abhimanyu problem' (if this literary allusion means nothing to you, that's okay). It's easy to walk into a large long position, with a huge MTM profit. It's enormously more difficult to walk out with no position and clean cash.

Market manipulation is a business, and the basic rules about entry continue to hold. If market manipulation were an easy business offering supernormal rates of return, there would be a flood of entry, and a lot of people would plunge into doing it!

The thought process about position limits needs to be illuminated by such reasoning.

Bond trading stopped in Japan for two hours because of a computer systems glitch

Jayanth Varma and I have previously written about the systemic risk aspects of computer security [link] [link]. Today, there is story in the New York Times about a computer systems problem owing to which bond trading in Japan stopped for two hours. Add this one to the list of computer disasters that have affected finance.

Estimation of the equity premium in India

Jayanth Varma and Samir Barua have written a mini-paper on estimation of the historical equity premium in India. They wage a brave war against difficult data problems, and come up with an estimate of 8.75 percent.

A while ago, in October 2005, I had written some notes about my views about the equity premium in India. In that note, I had argued that looking forward, one might estimate an Indian equity premium like 8%. With a short rate of roughly 5% and an equity premium like 8%, this means nominal equity index (total) returns might work out to around 13%.

An equity premium of 800 bps is huge! Just to be safe, in all our discussions about the New Pension System, much more conservative numerical values have been utilised. As an example, in my paper A sustainable and scalable approach in Indian pension reform, which is forthcoming in the LKYSPP India/China book, the three scenarios (page 28) have values of the equity premium of 3.5%, 4% and 4%.

Update: Kaushik Gala says:

Thanks for pointing to the IIM-A equity premium paper. You must be familiar with other estimates as well:

1. 5.2% by Damodaran,
2. 11% by Mehra,
3. 6% by JM Morgan Stanley,
4. 7% or 4.6%, by Haribhakti.

It was in Bill Gates' self interest to leave

Look at the intra-day time-series of the Microsoft share price superposed with the S&P 500: So it looks like the market feels it was worth giving a golden handshake to Bill Gates worth around 4% of his ownership for relinquishing control of the embattled giant. (How much is that?) I was fascinated by how slow the information processing was. I seem to remember that the announcement came out after hours on Friday? I would have thought that by monday morning, everyone would have made up their mind on how this changes things. But actually, the trading on monday seems to have gone on till evening discovering the new price. The monday evening gap, of around 4%, between MSFT and the S&P 500 was then stable on Tuesday. In the period after Tuesday, the gap rose a bit.

Central bank transparency

There is a traditional belief that monetary authories have to be enigmatic, and that market participants have to then zealously watch the central bank to pick up crumbs of information and decipher cryptic clues. The transition from Greenspan to Bernanke, with a market that has to learn a whole new style of communication, has highlighted the poor institutional structure of US monetary policy. In a Bloomberg column a while ago, Andy Mukherjee had pointed out that the phrase "unfolding constellation of uncertainties" used by RBI had never before been seen by google. The recent RBI rate hike was a surprise to markets because there was no scheduled monetary policy meeting.

In an article Plain english versus mumbo jumbo in Business Standard today, I argue that modern monetary economics involves a powerful move towards central banks that speak in plain english and are transparent. In a well functioning monetary regime, market participants would only zealously watch the economy, not the central bank. Sound institutions involve a nuanced relationship between data releases, rules, and rate changes.

The enforcement process at SEBI

The debate on retail quotas for IPOs appears to be continuing, while in the meantime, the SEBI "IPO Scam" order has been having a rough time in the courts and the SAT. An editorial in Business Standard yesterday looks beyond the specifics of this order, and has suggestions for improving SEBI's processes:

• The infamous "ex parte order", which is supposed to only be applied in a grave emergency, is being misused; each SEBI chairman should have only one go at doing this in his three years.
• The first step of an enforcement action should be a well-drafted show-cause notice sent - in private - to the accused. This should come up to UK/US standards of drafting quality.
• The release of this show cause notice to the public constitutes libel, for it is a mere accusation.
• The next step should be a quasi-judicial hearing, in private within SEBI, where the investigators argue as the "prosecution", and a dedicated "bench" of two board members listen to the defence, and award a penalty. There should be a full separation between this "bench" within SEBI and the board member(s) who handle investigations. This will reduce the mistakes caused by prosecutorial zeal.
• It is better to take on one entity at a time, with "five pages of top quality order", instead of writing one jumbo order about 24 entities. It is better to not have policy mistakes like the retail IPO quota, for that sets up an insuperable task of enforcement when millions of households are given an incentive to break the law.

I am, personally, an optimist on how SEBI is faring and where SEBI is going. I think that SEBI is on the right track on the core issue of : rule of law. SEBI has full clarity through SC(R)A and the SEBI Act on what it does. Both these acts are philosophically sound, and have been amended repeatedly so as to solve problems. Regulated entities regularly challenge SEBI at the level of the letter of the law, and SAT has proved to be a successful specialised court hearing financial cases. So while it looks messy, I think the basic framework "rule of the law in the public eye" is in place, and there are self-correcting forces - such as the public failure on the "IPO Scam" order - which will keep pushing SEBI in the right direction. It's messy, like democracy, but that's about okay in my book.

If all of finance operated in such a fashion - with repeated legal challenges to the regulator, and a rule of law - then many of the problems of Indian finance would be solved.

Trying to understand what happened on the stock market in May

I saw some fascinating data, from CMIE, on the behaviour of foreign investors in the recent days of enhanced market volatility. There are small difficulties in measurement in this data - it includes options notional values, which aren't that clearly interpreted. But options trading is small so it doesn't contaminate the results that much.

The data shows the net purchase by all foreign investors on the equity derivatives market, and on the equity spot market, both measured in rupees crore. I also show the official closing price of the Nifty spot, and of the May Nifty futures, in the table. A useful piece of background is that the last date in the table (25 May) was expiration date for the May contracts.

	Buy on F&O	Buy on spot	Nifty	May Nifty
May 10	-156	322	3754	3745
May 11	-852	-1199	3701	3693
May 12	-1124	18	3650	3633
May 15	-1084	-728	3503	3462
May 16	537	-533	3523	3520
May 17	390	-423	3635	3641
May 18	430	-810	3389	3364
May 19	1877	-1361	3247	3224
May 22	1914	-929	3081	3021
May 23	1266	-1243	3199	3191
May 24	883	-1935	3116	3087
May 25	1017	-1632	3178	3180

As an example, if I may read out the last row in the table to you, on 25th May, FIIs purchased Rs.1,017 crore of equity derivatives in notional value, and sold Rs.1,632 crore on the equity spot market. The official Nifty close was 3177.7, and the May futures nicely converged to the spot with an expiration-date closing price of 3180.15.

The story that I think this data tells is like this.

• I know, it seems like a long time ago, on 10 May, Nifty closed at 3754.25. On the three following days, foreigners appear to have been successful speculators, selling on both the spot and the derivatives market. Nifty dropped to 3502.95. By 15 May, a big negative basis (-1.17%) had opened up.
• So from this point onwards, foreign investors were steadily buying on the derivatives while selling on the spot market. This sounds to me like reverse cash and carry arbitrage.
• Particularly, from 19th to 25th, local speculators were selling the Nifty futures, while the FIIs were doing the work of buying on the futures and selling on the spot. The basis showed up values as large as -2% on 22nd may, when FII purchase on the derivatives peaked. I find it reassuring for my story that on the days with the largest FII purchases on derivatives (19-25), the basis was at it's biggest negative values (-0.7, -2, -0.3 and -0.9). Of course, this "basis" is computed by comparing the official closing price on the futures and the official closing price on the spot, so it surely misses out the intra-day story.

These are highly aggregated numbers summing up the behaviour of over 1000 distinct entities., and it's always dangerous to anthropomorphise across aggregation. So one should not interpret this as "the behaviour of all FIIs" - instead it is the behaviour of the aggregate FIIs, and I'm sure there is huge heterogeneity within that class of investors.

The other remarkable thing in this data is how small these values are. The equity market (NSE + BSE, spot + derivatives) typically does over Rs.45,000 crore of turnover a day. The values for FII turnover seen here are all small, much smaller than what you might think if you read the newspapers shouting about FII exit (here's an example) leading to a drop of Indian equities.

The numbers in this table sum up to: A sale of Rs.10,456 crore of equities on the spot market, coupled with a purchase of Rs.5,100 crore on the derivatives market. Once again, the average per-day net buy/sell is small change compared with the size of the equity market. And, this reasoning and evidence suggests that a full picture of what FIIs are doing on both spot and derivatives is important to understanding what is going on.

In an ideal world, near-infinite arbitrage capital should be in play through sophisticated IT systems, so that the fair value of the index futures is circumscribed between a very tight "no-arbitrage band" at all times. In India, there are two big problems holding this back. First, SEBI and NSE have banned the IT systems - they insist that arbitrage be done in a labour intensive way. Second, the big institutional investors who are the natural players in this fixed income game - banks, pension funds, insurance companies - are prohibited from doing it.

India is doing some pioneering stuff by world standards in overcoming this problem by selling arbitrage funds to retail customers through mutual funds. I disagree slightly with the sales pitch that some of these funds make - they should clearly say it's a fixed income investment, but they don't always say that. For the rest, it's a great way to make progress delivering risk/return profiles to customers that existing institutional investors are unable to, and helping the country achieve "near-infinite capital in play for arbitrage".

FIIs have found a niche in derivatives arbitrage, with a comparative advantage owing to their regulators being better than ours. They have near-infinite capital and could thus do a great job in bringing about market efficiency on the derivatives. What holds them back is (a) the ban on IT systems, and (b) limits on FII ownership of individual stocks.