A V Rajwade: Issues in bond and derivatives markets

WORLD MONEY

Given the share of floating rate bond issues in government borrowings, the need for reforms is all the more urgent.

The government has accepted, in principle, the recommendations of the Patil Committee on the development of the corporate bond market in India. An efficient and long-term corporate bond market, for both primary issues and secondary trading, is necessary if the needs of the corporate world in general, and the infrastructure sector in particular, are to be satisfied. At present, the primary market is limited to a handful of well-known AAA-rated issuers and the maturity rarely extends beyond five years. The secondary market is illiquid.

While the Patil Committee recommendations are to be welcomed generally, I would like to comment on one disagreement and one omission. The committee has recommended that "the standardised practice of 30/360 day count convention, followed for dated government securities, may be made mandatory for all new issues of corporate bonds". Firstly, the day-count convention in the secondary market has little to do with issue terms. Second, to my mind, this would be a retrograde step: on the contrary, there is a strong case for moving the G-sec market to the actual/actual convention. (In the corporate bond market, there is no uniform practice either in terms of coupon periodicity, or the day-count convention.) The present market practice is artificial and a relic of an era when computing power was neither cheap nor readily available, and compromises on calculation accuracies had to be made. The practice has long been abandoned in the US and the UK government bond markets in favour of the actual/actual day count. The reason is simple: the latter correctly reflects the actual accrual of interest payments: the half-yearly coupons are due on dates six calendar months apart (say April 24 and October 24). Thus, the same half-yearly coupon represents interest for the actual number of days, which can vary from, say, 181 to 184 days. The actual/actual day-count convention reflects this reality. To be sure, the committee has gone on to recommend that "a suitable road map may be finalised to migrate interest payment conventions across all fixed income instruments, including government securities, to an actual/actual basis". The two-stage approach, first corporate bonds moving to 30/360 day count, and then both migrating to actual/actual, seems to be uncalled for, and probably exaggerates the difficulties involved. To my mind, only a minor, and not very difficult, change in the IT systems should allow the move from 30/360 to actual/actual in the G-sec market even now. The Fixed Income Money Market and Derivatives Association of India (FIMMDA) needs to move in the matter on a priority basis.

One point not covered in the recommendations is the need for a three/six months floating-rate benchmark in the derivatives market. At present, the most liquid floating rate benchmark is the NSE Mumbai inter bank offer rate (Mibor) on which the overnight index swap (OIS) is based. With the use of Mumbai Interbank Forward Offered Rates (Mifor) as a benchmark in pure rupee transactions banned, the other floating rate benchmark getting some popularity is the one-year INBMK "" the yield on one-year (balance) maturity government bonds. A three/six months floating rate benchmark would be extremely useful for the banking system to manage the interest rate risk on long-term bonds and, otherwise, also better manage the asset liability maturity mismatches. A three/six month floating rate benchmark would also be useful for the securitisation market. Ideally, the benchmark should be the term interbank rate "" but the market is still not very liquid. In the meantime, we probably need to look at T-bill yields. The NSE Mibor and one-year INBMK benchmarks are somewhat artificial with few actual transactions taking place at the benchmark rates. The NSE Mibor is based on a poll at 9.30 am and the polled rates are not necessarily indicative of the willingness of the quoting bank to actually lend money at that rate. In fact, most of the days, NSE-Mibor has been above the call rate. As for the other benchmark, there are few transactions in one-year maturity bonds.

As for the overnight rate, there is a case for the market to consider switching to the overnight CBLO rate based on dealt and bid and offer prices on the CCIL's trading platform. This is an anonymous order matching system and the participants are committed to the prices they are quoting. A weighted average of the CBLO deals actually done and the outstanding bids and offers at, say, 10 am in the morning would be a far more realistic benchmark for the overnight rate. To be sure, it is a rate for secured loans, unlike the unsecured interbank money, but does not suffer from the other weaknesses of the existing benchmarks.

While on the subject of floating rates, FIMMDA's valuation method for floating rate bonds seems to be faulty. Given the fact that in the current fiscal year, floating rate bond issues are expected to be about 10 to 15 per cent of the total government borrowings, the need for reform is all the more urgent.

Email: avrco@vsnl.com