How bad are things in Indian banking?

Ever since US-64 and Indian Bank, there has been a heightened consciousness about the dangers of the guarantees which have been freely given out by the Indian state. It sounds easy for a government to announce that it will stand guarantee against some eventuality, which may seem fairly remote at the time. However, any insurance contract is not free. The likelihood of my dying tomorrow is fairly low, but an insurance contract which pays if I die tomorrow is emphatically not free.

We can calculate the price of a guarantee, and the recipient should normally pay the fair insurance premium. When a guarantee is given for free, it implies a subsidy. It may seem like an invisible subsidy, but it is no different from writing a check to the beneficiary of the guarantee

One of the most important guarantees that's operative in India today is the umbrella of protection on the banking industry, that's called "the safety net". How much are these guarantees worth?

In order to price a guarantee, we have to first accurately define the terms of the contract. It is unfortunately hard to define exactly what the terms of this deal are. Is it like the deposit insurance contract run by the DICGC, where each depositor stands to get upto Rs.100,000 when a bank fails? Or is it a complete guarantee covering all liabilities? Or is it a promise to endlessly continue to bring in fresh equity capital to replace the bleeding capital of a weak bank?

Suppose we agree on one definition of the safety net:

In this case, it is possible to apply modern option pricing theory to execute two remarkable pieces of calculation.

Notice that the shareholders of the bank have a one--sided contract. If the bank does well, they get all the upside, but if the bank is closed down, their liability is limited to the equity capital in the bank. This gives shareholders a "call option" on the value of assets of the bank. The equity market capitalisation that we observe is the value of a call option on the assets of the bank. We can turn this upside down, and infer the value of assets of the bank (at market prices) from the observed level and volatility of equity prices.

The second step involves pricing the guarantee. The "safety net" has to pay out the gap upon the event that the bank's assets drop below 90% of the liabilities. The price of this credit guarantee is exactly the price of a "put option" on the assets of the bank. Once we know the value of assets of the bank, we can apply option pricing principles to get an estimate of the price of the put option.

In short, we can take information for the level and volatility of the share price, and infer the value of assets of the bank. This is interesting data in its own right, since it tells us something about the capital shortfall in banks. In addition, we can use this to calculate the fair price of the safety net.

This approach is entirely based on exploiting the information embedded in the level and volatility of stock prices. How reliable are these? When the shares of a certain bank are highly liquid, the stock market works as a remarkable information processing device, pooling information from thousands of individuals, ranging from professional fund managers, to individuals reading this paper, to insiders of the bank who are illegally profiting from special information that they possess.

The stock market appeals to the self-interest of speculators by giving them profits when they contribute high quality information into the market. For example, it is widely believed that information rapidly leaks from interim income tax filings to stock market speculators. The stock market is interesting because it harnesses a strong pool of information producers, and reacts rapidly to new information. In addition, the stock market is good at taking a holistic view of the bank, reflecting a combination of asset quality, profitability, franchise value, etc. It does not narrowly penalise NPAs, or get easily fooled by `evergreening', etc.

In a recent working paper, Measuring systemic fragility in Indian banking: Harnessing information from the equity market, Susan Thomas and I apply these ideas to Indian banking. We isolate 19 banks with highly liquid shares, and work out the market value of assets, and the value of the safety net. Our findings may be summarised as follows:

  1. In our sample, the value of the safety net as of 31/3/2000 ranges from 0.5% per year for HDFC Bank to 8.4% for Dena Bank. Most of this is a subsidy, when compared with the fee of 0.05% that DICGC charges.
  2. The average value of the subsidy for the 19 banks is around 6-6.5%. This works out to a subsidy of between Rs.30,000 to Rs.40,000 crore per year.
  3. Our calculations imply that the stock market believes that 16 of these 19 banks have a market value of assets which is below the market value of their liabilities. It would require Rs.40,000 crore of fresh equity to top off these assets. All but one bank (HDFC Bank) has a buffer between assets and liabilities which is smaller than 10%. It would require Rs.103,000 crore of fresh equity to bring all 19 banks up to this point, i.e. to cap leverage at 10 times.
  4. Our sample of 19 banks which are actively traded on the stock market are likely to be the better run banks in India. They account for 59% of the liabilities of the Indian banking sector. Hence, we would obtain a conservative set of estimates for Indian banking as a whole if we multiplied the above numbers by 1.69. This gives us the following estimates: the subsidy of the safety net is worth an annual payment of roughly Rs.70,000 crore, and the equity capital required to have assets which are 10% ahead of liabilities works out to Rs.173,000 crore or 9% of GDP.

These numbers are conservative estimates for the difficulties of Indian banking. If the Indian government tried to reinsure the safety net from international insurance companies, it would quite probably find itself with a price-tag which is larger than our subsidy estimate of Rs.70,000 crore a year.

Are we, as a country, doing this right? It is hard to see how helping Indian banks is worth such an expenditure. I suspect there are more votes to be found by spending this money on health and education. Further, this subsidy constitutes a large 600 basis point bias in favour of one kind of financial intermediation (i.e., banking) over other forms (e.g. securities markets). This is exactly as inefficient and inappropriate as a large government subsidy which helped scooters at the expense of motorcycles.

Back up to Ajay Shah's media page