Fixed interest portfolios and risk

In the annual policy statement for 2004-05, issued by the Governor of the Reserve Bank of India on May 18, 2004, while discussing the government securities portfolios of the banking system, he stated that "such holdings above the statutory liquidity ratio amounted to Rs 2,69,777 crore (Rs 2697.77 billion) … Such large holdings of government securities by banks entail significant interest rate risk as the yields on government securities are already at their historically low levels."

This was of course not the first time that the central bank had cautioned commercial banks about the significant interest rate risk integral to fixed interest securities portfolios.

To be sure, on this subject, there is an obvious conflict of interests involved for the central bank itself: as the supervisor for the banking system, it would like to limit the interest rate risk banks assume; on the other hand, as merchant banker to the government of India, it has been deliberately lengthening the maturity structure of government debt. Banks being the single largest buyers of government debt, longer maturities add to their interest rate risk.

But this apart, did the banking system take heed of the cautions from the central bank? While no system-wide hard data are available, first quarter results do evidence a sharp fall in treasury income of the banks. (Not surprisingly, 16 of the 17 primary dealers are reported to have registered losses in the first quarter.)

Rumours in banking circles indicate that, for some of the banks, the problem of the provisioning needed for marking to market could make a significant dent in the current year's net profit.

One does get an impression that an eight-year long stretch of almost continuously falling yields and rising prices (since the bad old days of 1995-96), had led to a sense of complacency; perhaps there was also an element of wishful thinking involved in the expectation that the party would not end soon, if only because of the existence of surplus liquidity in the system.

The positive side of the falling yields and rising security prices was the significant profit on the sale of investments booked by many banks, much of it used to increase provisions against NPAs and clean up the balance sheets.

However, the question is how many banks simultaneously de-risked the portfolio. For those who missed the opportunity, the current year could be a tough one, particularly after the surprise inflation number that came out last Friday.

To be sure, separately from the cautions sounded by the Reserve Bank, there were enough pointers to the possibility of the interest rate cycle turning in the opposite direction -- in other words, of prices of fixed-interest securities falling.

Quite apart from the general principle that nothing can defy Mr Newton forever, the benchmark 10-year security was giving negative real yields since March last year.

It is elementary of course that negative real yields are not sustainable. Interestingly, both Dr Rangarajan and Dr Reddy, than whom there are no more authoritative commentators on monetary policy, made the point last Saturday in a function in Pune.

This apart, one wonders how conscious bankers were of the basic risk-reward relationship in fixed interest securities. In a positive yield curve scenario, which is normal, higher yields come only at the cost of a higher price risk.

Interestingly, even after real yields turned negative, there were enough opportunities to de-risk the portfolios by selling longer-term securities and buying short-dated ones. For example, in October 2003 the yield difference between one-year and ten-year securities was less than .5 per cent -- the price risk on the ten-year security is anywhere 7–8 times higher than on the one-year security!

At least to my mind, the negligible extra yield hardly justified the significantly higher price risk. Even as recently as April, just before the yields started going up, the difference between one- and ten-year yields was barely 0.7 per cent.

Those who missed out on these opportunities may well have to pay the price of large provisioning in the current year. While, as many analysts have pointed out, the inflation number could go up from the past week's level in the near term, it could well fall before the fiscal ends.

The question, however, is whether there would be a corresponding fall in bond yields. Personally speaking, despite the existence of surplus liquidity in the market, I am pessimistic on the subject:

For one thing, the inflation shock has probably changed market psychology from wishful thinking to worries about rising yields;

The demand-supply situation in the money market is also likely to undergo a change. Bank credit went up by Rs 42,000 crore (Rs 420 billion) in the first quarter of the current year. By all estimates, the fiscal deficit could be higher than budgeted (the first quarter has already eaten up as much as 60 per cent of the year's budgeted revenue deficit). And, the recent fall of the rupee and higher forward margins reduce the attraction of leading foreign currency receipts and lagging payments, thereby increasing demand for bank credit.

In the last annual policy statement, the RBI governor also stated that capital charge for market risk would be introduced for the trading book by March 31, 2005, and for the available for sale category by the end of March 31, 2006.

One advantage of the capital charge for market risk is that, since most banks are facing a shortage of equity capital, it opens the eyes of the board and top management to the interest rate risk in a far more pointed fashion than the duration number.

Incidentally, the investment policy guidelines in the RBI's master circular (July 17, 2004) are silent on risk measurement, whether through duration or VaR.

Some guidance on prudential limits on the interest rate risk would be useful at the current stage of evolution of the system. (The contrast with the RBI's control on FX exposures/gaps is remarkable: a relic of the "precious" foreign exchange, and the "worthless" rupee mindset?)

One expects of course that, as a corollary to the capital charge, the requirement of investment fluctuation reserve and the present ad hoc risk weight for securities portfolios would be eliminated.

But capital charge apart, the supervisor needs to consider improving balance sheet disclosures relating to the fixed interest securities portfolio, at least in two areas:

To facilitate comparisons, the book value of the AFS portfolio needs to be at market value. (This need not imply that surpluses are automatically carried to profit and loss.) Under extant regulations, the same security can be carried at different prices by different banks depending on when it was purchased, thus rendering yield comparisons meaningless.

In the maturity pattern, there is only one classification for investments over five years. The risk profiles of a six-year security and a 25-year one are radically different and the disclosure should be meaningful enough for analysts to estimate the risk exposure with some degree of reliability.

Powered by