Friday, January 02, 2015

Are Indian banks systematically mispricing risk?

by Harsh Vardhan.

The primary objective of a financial system is to efficiently allocate capital. The key step in allocating capital efficiently is to assess and price risk correctly. Correct pricing of risk ensures that it is properly distributed - capital providers get to hold assets that reflect their risk appetite and present them with an efficient risk-reward trade off.

Banks are an important vehicle for linking up the savings and investment of India. It is critical that banks price risk correctly. There are concerns about how Indian banks are pricing risk. The chart below shows the average risk premium charged by the banking system on commercial loans in comparison with the risk spreads on various ratings of bonds.

 
This shows that private sector banks priced their loan book as if it was rated between AA and A whereas public sector banks priced loans as if they were rated between AAA and A. The risk spreads spiked in FY 2009, the year of global financial crisis. The financial years 2006 to 2008 appear to be `carefree' lending years: there was the biggest ever boom in bank credit, and risk premia were the lowest.

This analysis may be challenged on several points:

  1. Indian corporate bond markets are not deep enough to provide reliable risk spreads
  2. Loan risk spreads cannot be compared with bond risk spreads as loan covenants are generally tighter than bond covenants
  3. Bank loan books carry a significant retail lending portfolio, which is generally of much lower risk, and hence the risk of the whole lending book comes down


The first criticism is valid - the bond market is indeed very shallow, and that hampers our ability to read information from it. However, there are two key data points that will throw some more light on this issue. Firstly, in the chart below we show the rating distribution of the top 3 rating agencies in India for the financial year 2009 and 2013. These three agencies account for over 95% of all bonds rated in India. The data shows that the median and mean rating of the bonds was BBB in both the years. The overall rating profile deteriorated over these 4 years. Also, the companies that issue rated bonds are typically larger in size and more established. The firms that banks lend to are, to a greater extent, Small and Medium Enterprises (SME) which are riskier than the large companies that access the bond market. Thus, it is fair to expect that the average rating of the portfolio of banks should be worse than the average bond rating.


Is this an issue of underestimating the risk, or bad pricing? Unfortunately the analysis does not throw much light on the cause of mispricing. But here is another piece of data that may be insightful. In FY 2012, the Indian banking system had ~ 76 Mn commercial loan accounts (i.e. loan accounts for commercial loans excluding Individual loans). Of these about 72,000 were loan accounts of value above 5 crore, and this subset accounted for 80% of the loans of banks. Currently, around 40% of these accounts are rated by rating agencies. (Under Basel II, banks are allowed to use external rating for their loans). This means that about a 3rd of the loan book by value of the banking system is rated by the same rating agencies that rate bonds. It's unlikely that the ratings agencies use radically different (and lower) standards for rating bank loans. This implies that the cause of underpricing may not be misunderstanding risk but mispricing it.

The second criticism, of loan covenants assigning better rights to the lender (i.e. banks) than bond covenants is plausible. It is hard to test empirically unless we get the data on recovery rates of loans vs. bonds. I feel that while this is a factor at work, it cannot explain the significant difference between the loan spreads and bond spreads.

Finally, the criticism about retail loans in the portfolio. It is true that the retail secured lending (eg housing, car loans) has demonstrated a much better risk profile over the last decade or so. However, it is important to note that retail lending constituted about 22% of the overall loan book of Indian banks, and hence the lower risk on retail is unlikely to significantly lower the overall risk.

I hope to have persuaded you that there is evidence in favour of systematic mispricing. This takes us to the next question: Why would there be such systematic mispricing? There are two possible reasons. First, excessive competition, especially in the larger and relatively better rated lending, that drives pricing down.

The other reason is more nuanced. Historically, a significant part of the banks' deposits base (~ 33%) was under interest rate regulation - these are the "CASA" (Current Accounts, Savings Account) resources that all banks chase. The chart below presents some interesting analysis. Consider this: if the banking system as whole did nothing but took CASA deposits and invested in the risk free 10 year government security (we call this Risk Free Margin on CASA) then the margin it would make is more or less the same as the pre-tax profits of the banking system! Over the last decade, the risk free margin on CASA was more than the pretax profits of the system except in a few years when the yields on government securities had fallen precipitously - FY 03, FY04 and FY09. This "lazy profit" has been created by regulations. It is this profit pool that allows banks to underprice loans, effectively creating a system wide wealth transfer from (CASA) depositors to commercial borrowers. Most banks use a cost plus approach to loan pricing where the loans are priced at a margin over the cost of funds, which are kept low by interest rate regulation. Despite lifting the SA interest regulation a couple of years ago, only 3 relatively small banks have changed their SA pricing.


Using sources of funds whose cost has been kept low, because of regulations, to underprice risk in lending, effectively engineers a large scale wealth transfer from depositors to borrowers. This wealth transfer can be interpreted as being an integral part of the system of financial repression which is in operation in India.

Systematically mispriced risk will result in misallocated capital. Is this one of the reasons that the system regularly runs into NPA crises? Is this systematic transfer of wealth the reason why households avoid bank deposits and prefer to hold other kinds of assets?

As the Indian economy grows, our financial system should not just grow in size, but also evolve to become more sophisticated and allocate capital more efficiently. With a monolithic system dominated by public sector banks, we have not seen such evolution. Banks today look and behave the same way they did 10 years ago, they are only a lot bigger. Fundamental change in the framework for financial economic policy will give a banking system that allocates resources better, and is safer.

1 comment:

  1. The second criticism cannot be brushed away that easily. The portion of the corporate bond market that caters to secured debt is the one directly comparable to loans. Unless there is a comparison of ratings/spreads in that particular area, the "mispricing" hypothesis cannot be a persuasive one.

    ReplyDelete

Please note: Comments are moderated. Only civilised conversation is permitted on this blog. Criticism is perfectly okay; uncivilised language is not. We delete any comment which is spam, has personal attacks against anyone, or uses foul language. We delete any comment which does not contribute to the intellectual discussion about the blog article in question.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.