Daily data from electronic payments systems: a new treasurehouse of data on the economy

John W. Galbraith and Greg Tkacz have a paper which, to me, implies that RBI should initiate release of daily data on traffic on payments networks such as credit cards, debit cards, etc.

The coolest internship program I can think of

DEA has an internship program. I hope it will help build the next generation of reformers. Update: DEA has a press release about their recent retreat.

SEBI's move on short selling

SEBI has moved on short selling and securities lending. See the two PDF files at the bottom of this page. For the motivation, about why this is important, and the political economy of the constraints, see this blog post by Jayanth Varma. A quick summary of where we stand on this question:

• Instititional investors are prohibited from selling a security they do not own, but some of them (mutual funds, FIIs) have fairly good flexibility on using single-stock futures where shorting is easy and convenient.
• Non-institutional investors - who make up the bulk of the market - are able to go short, but lack a mechanism for borrowing shares. Once again, shorting is easy and convenient using the single-stock futures. In addition, many day traders short-sell within the day but are forced to buy back within the same day since borrowing shares is infeasible.

My sense of the situation is that India does not suffer from a significant problem in the ability to express negative views about stocks, given the strength of the individual stock futures market. However:

• Reverse cash and carry arbitrage requires access to borrowed shares.
• Many securities lack individual stock futures. Obtaining market efficiency on these critically requires that speculators have a mechanism for selling borrowed shares.

SEBI's announcement envisages short selling for the same stocks where individual stock futures are available. The impact there will, then, be limited to enabling the activities of arbitrageurs. While that is beneficial, much more remains to be done.

The really important issue is the mechanism for borrowing shares. Will this work frictionlessly? In my intuition, demand for borrowing is small and the supply (with institutional investors) is quasi-infinite, so access to borrowed shares should become possible at very low prices. I wonder what the charges for borrowing securities are in the UK and the US.

As emphasised in the MIFC report, the need of the hour is an integrated securities lending mechanism covering shares, corporate bonds and government bonds, so that short selling can flourish on all three markets. SEBI needs to urgently solve the problems of the borrowing mechanism, so as to then move forward on implementing this larger agenda. Short selling is more urgently needed on the bond market, where the minimum semblance of market efficiency is presently lacking.

Mobis Philipose has a good article in Mint reviewing the SEBI announcements and their consequences, and Business Standard had a good editorial on this subject today:

Economic liberalisation is, ultimately, about the idea that resource allocation driven by markets works better than that driven by a Planning Commission. This requires markets that work well. This calls for ample information disclosure, checks against market power, and a free play of both optimistic and pessimistic views. Short selling - expressing a negative view about a stock that one does not own - is thus an integral part of any well-functioning financial system. Sebi's recent moves on short selling are, hence, based on the correct vision for financial sector policy.

When derivatives trading on individual stocks began, a "technology" for expressing speculative views about individual stocks became available. An optimist can buy futures or call options, or sell put options. A pessimist can sell futures or call options, or buy put options. Thus, with derivatives trading, a level playing field between positive and negative views is assured; short selling is no longer a big issue. Unfortunately, Sebi says that short selling will only be available for these stocks. Sebi would do well to permit derivatives trading and short selling on more stocks. Short selling for stocks on the derivatives list matters only insofar as it supports reverse cash and carry arbitrage. When the futures price is `too low', arbitrageurs borrow shares, sell them, and buy the futures. The impact of this announcement will, then, be indirect: it will help cure the persistent underpricing of futures that has been found in India.

Some features of Sebi's announcements are unfortunate. Institutional investors are prohibited from squaring off positions within the day. They are required to disclose that an order is a short sale at the time the order is placed. Retail investors are being asked to make the same disclosure at the end of day. Brokers have to supply this data to exchanges who will then release these to the public. These notions are not grounded in serious policy analysis. For the spot market, short selling is invisible: on T+2, when deliveries have to be made, the short seller supplies shares just like any other seller. A thorough policy analysis effort on Sebi's part would have led to the simple removal of all restrictions on short selling.

The real challenge is not in short selling but in effecting borrowing of shares. Sebi proposes to set up an exchange-traded mechanism for borrowing shares. This involves one key rigidity that will hamper its success: borrowing and lending can be done only for seven days. This is inconsistent with the needs of futures arbitrage. When an arbitrage opportunity surfaces (say) three days from futures expiration, the futures arbitrageur needs a way to borrow shares immediately for a maturity of three days. While an exchange traded mechanism sounds sophisticated, the mainstream solution found internationally - that of merely borrowing securities OTC from institutional investors - appears to be a superior solution.

Update: Andy Mukherjee disagrees, saying that India might well be onto something very important by emphasising an anonymous order-matching mechanism for borrowing shares.

Paper on PNs

Manmohan Singh has written a good IMF working paper on participatory notes in India. I have checked and I'm sure this URL works.

Who needs weather derivatives?

Once I was chatting with a kabab vendor on the streetside. I asked him "What problems do you face?". He said that it was Id that day, and he faces a massive risk. If the crescent moon is sighted, he gets flooded with business, and runs out of chicken. If it's cloudy, and the moon is not sighted, then he is holding too much inventory of chicken and it gets wasted. I thought to myself: Wouldn't it be neat if he could get a contract that pays if it's cloudy.

The New York Times has a fascinating story on a new place where weather derivatives are useful. They describe applications in the apparel industry :

And the manufacturer Weatherproof, which supplies coats to major department stores, has bought what amounts to a $10 million insurance policy against unusually warm weather, apparently a first in the clothing business.

Fredric Stollmack, the president of Weatherproof, said that unseasonable weather, once a widely mocked excuse for poor performance in the industry, is the new norm, forcing companies to make sweeping changes in how they manufacture and sell clothing.

...

For Weatherproof, forecasts and climatologists are not enough. The majority of the companys business is done in November and December and if the weather is unusually warm, as it was during those months last year, sales plunge. (The last several months were not much better, with August, September and October combined the warmest ever recorded for six states, according to Planalytics, a weather research firm.)

So in a closely watched experiment, Weatherproof signed a contract that guarantees it would be paid as much as $10 million if daily temperatures in New York City are warmer than the historical average for December, 37 degrees. The higher the temperature this month, the more money Weatherproof will be paid.

Weatherproof bought its coverage from a 1-year-old company called Storm Exchange, which also sells such contracts to oil and electricity companies.

The Internet changes the music business

This blog entry is written jointly with Kaushik Krishnan.

The crisis of the old music business

With computer technology, music has gone from LPs and CDs into computer files. Crashing prices of networks and storage has meant that transferring music files from one friend to another is effortless. Through these changes, music has acquired characteristics of a public good: it is non-rival (my consumption of music doesn't come in the way of yours) and non-excludable (it has become impossible to stop piracy in anything but a police state). Hence, the traditional business model of the music industry is in deep trouble. As an example of the difficulties that music companies are facing, see Robert Sandall writing in Prospect magazine. David Byrne has also written a very nice piece in Wired magazine. (You might also want to look at a conversation between Thom Yorke and him on the future of music.) He shows that there has been a drop in music sales in general and a steady increase in the sales of music electronically:

Many experiments are afoot on rethinking business models in this age of the Internet. The essential opportunity lies in utilising the very computer technology - which has obsoleted the record / CD business - by linking up the ultimate artist to the ultimate consumer, so as to eliminate the overhead of various middlemen. Byrne writes that a large portion of the cost of a CD is in overheads; the payments by the buyer of the CD mostly don't reach the artist:

The sarodist Suresh Vyas pointed out to us that the 15% of the overhead that's spent on marketing/promotion is in the interest of the artist, insofar as it is about raising publicity and awareness. And yet, a key change of the electronic world is that friends pass on music to friends, giving a powerful word-of-mouth phenomenon through which awareness can be increased. It is different from the marketing blitzkrieg of pop music of old, but that doesn't mean it's non-existent.

The Apple Music Store does not solve many of these problems

The leader of the pack, in terms of revenues, is Apple Music Store. However, I would personally never buy a single minute of music from Apple Music Store, given their closed standards: you can only use their music files on Apple hardware. Even though I have an ipod and my main machine is an Apple Macbook Pro, I wouldn't dream of being tied down to them like this. In addition, they have digital rights management (DRM) of a sort that I find offensive. Electronic distribution should help by lowering overheads. As yet, the situation is one where Apple makes money, but the musicians still get very little (as shown in the above diagram).

Today, I saw magnatune, an alternative way of organising the music business, that I think has a bright future. Magnatude has been around since sometime in 2003. Here is how it works:

1. For starters, experimentation in their catalogue is convenient and free. They have pages sorted by genre, such as this page of `world music', which work as a free radio station. This helps you to sample their material.
2. They make it easy to shift from listening to the radio to buying. While something is playing, the album cover is displayed. Click on it and you are looking at the material produced by this artist. Here is an example.
3. When you click to `buy', it gives you a choice between download or physical disc.
4. When you go to the download page (here is an example), you are asked to pick a number for what you will like to pay - between $5 (Rs.200) and $18 (Rs.720).
5. They make it very convenient - all you have to do is type in a credit card number and the CVV.
6. This takes you to a download page (here's an example) which offers various file formats. All the files are free of Digital Rights Management (DRM), and both low-res and high-res files are on offer.

The download password works for 60 days, so if something goes wrong in the download, it's easy to restart it. Every time you buy an album, they give you three gift coupons using which three of your friends can download the same album for free.

I find it to be quite a transformation when compared with the traditional music business - whether it's the old record companies or the Apple Music Store -

1. It is easy to evaluate material on the website without paying for it.
2. Customers have flexibility to pay between $5 and $18 for an album;
3. Half the revenues goes to the musician, which is a lot better than the traditional business;
4. Downloading files is, of course, nicer than buying CDs;
5. Yet, this is done without bringing any DRM into the picture;
6. High-resolution FLAC or WAV files are on offer, as are low-res files for those who prefer them;
7. Non-commercial use of the purchased material is free.
8. The website is extremely well thought out and easy to use.

I think they will go far, and are a far more impressive model for what the music business can be in the Internet age when compared with the market leader, Apple Music Store. Here are some links:

• Wikipedia entry on them.
• Interview with the founder, John Buckman, on creativecommons.org.
• An article by the founder on openDemocracy.
• More information from under the hood.

Other efforts

Magnatune is very impressive, but it's only one of a new breed of `open source record labels'. One example worth examining is onclassical.com.

Some other music sites worth exploring are pandora which is the offspring of the Music Genome Project. Pandora used to be available everywhere but it is now restricted to US users only due to legal issues. Suresh Vyas pointed us to http://www.underscorerecords.com which pays 70% to the artist.

Ruminating on what is happening

If I may ruminate on what is going on, the free software movement has shown the way in shifting from products to services. In this world, products have public goods characteristics (non-rival and non-excludable) and are free. Associated services are not public goods (they are rival and excludable) and are not free. So it is possible to earn money from consulting, configuring, adapting and modifying free software - but not from selling it. A good programmer will never starve, but in this world there is no possibility of scoring another Bill Gates.

In similar fashion, when music has acquired public goods characteristics, musicians will have to shift from revenues based on products (sale of CDs) to revenues based on services (concerts, custom compositions / performances, etc). The Byrne article shown above lists six strategies that musicians can use in this changing environment to still chalk out a living for themselves. A good musician will never starve, but revenues like those obtained by music companies of old are not feasible in this world. Open source record labels fit well into this emerging ecosystem, while many traditional firms do not.

Rajappa Iyer asks a deeper question. The `old deal' offered to musicians was: With a low probability, you will get Led Zeppelin payoffs. This helped attract a certain kind of person, with low risk aversion, to take the plunge into the tens of thousands of hours of effort that is required to try to become a good musician. Most didn't make it, but a tiny few became fabulously rich. In the new world, where this low-probability massive-payoff is not in the picture, will there be a reduction of supply of individuals who are willing to undergo such penance? My first answer would be that the risk aversion of people who choose to make music will be higher than that found in the old world. :-) But that is surely only a part of the story.

Watching markets work - IFCI

There was an announcement on 19th that the IFCI control transaction would not go through. The CMIE website shows this at 6:11 PM, so this would be after the market closed at 3:30. The stock price graph from Friday (14th) to Thursday (20th), from Yahoo Finance, is fascinating (click on it to see it more clearly). At the close of business on 17th, IFCI and Nifty were together. A gap showed up on 18th morning and widened slightly by 19th. On 20th morning, there was a massive gap and the opening price was over 30% below that found on Friday.

But then, IFCI is defunct. The only reason to buy IFCI is to get a license to be an Indian financial firm - a reflection of the entry barriers in Indian finance. The 20th closing price (Rs.76.75) is still a hefty P/E of 7.89 and a P/B of 2.81.

I often meet people who are repelled by sharp price fluctuations. Episodes like this are a vivid demonstration of markets in action, and of the importance of sharp movements of the price when there is a sharp change in the situation. The valuation of IFCI on Friday was conditional on the sense that a management transformation was around the corner. On Tuesday and Wednesday it was clear there were problems, and on Wednesday evening the announcement came out that the deal was not going through as envisaged. It is an efficient market that responds swiftly and clearly.

Look closely at the intra-day chart for 20th (Thursday) only. I was fascinated to see how the opening price of Thursday morning was roughly true - there was no significant undershooting or overshooting. (Once again, click on the above graph to see it more clearly). There was a lot of turnover, but right in the first few minutes, the market basically got the closing price right. This IFCI story is a clean experiment because the firm is defunct; it has no expectations of cashflows that vibrate intra-day owing to news about the economy and the firm. The only news affecting the firm is the impending transaction (or lack thereof).

Through all this, the IFCI stock futures market was quite resilient. (As an aside, once a stock has futures trading on it, there are no price limits, which was essential to allowing these large price movements to take place unhindered). At closing time, the `market by price' of IFCI futures at NSE was:

I worked out the `liquidity supply schedule' (LSS) that's implied by this MBP:

The LSS shows the impact cost associated with all possible transactions. Negative values on the x axis pertain to selling while positive values pertain to buying. It looks quite liquid. At the touch, the spread is ~ 0.26%. The one-way impact cost for buying Rs.5.6 million is 0.23% and the one-way impact cost for selling Rs.6.2 million is 0.26%. The `market-by-price' display (of the top five prices) alone gets you to buying/selling over Rs.10 million, where impact cost of a tad beyond 0.3% is seen. All in all, it's a good display of market resiliency - a big price move took place, but the futures market was not spooked, atleast by NSE closing time. On the subject of resilience, you might like to see this.