Fifteen-month treasury deployment under a hard mark-to-market loss limit
A corporate treasury has ₹10 crore to deploy for about 15 months. It is choosing between three instruments: the 8.24% GOI 2033, quoted at a clean price of ₹104.60 with modified duration 6.4; a one-year bank certificate of deposit at 7.35%; and a AAA-rated 10-year corporate bond at 7.60%, 55 basis points over the comparable G-sec, with modified duration 6.1. The G-sec pays coupons on 8 January and 8 August, and settlement is 12 July. Consensus expects a 25 basis point repo cut at the review six weeks away, while the treasury believes credit spreads may widen by 30 basis points over the year because of stress in NBFC funding. The investment policy requires that no quarter show a mark-to-market loss greater than 2% of the portfolio, and all investments are fair-valued and reported quarterly.
1. Compute the accrued interest and the amount actually payable per ₹100 face value of the G-sec, and the current yield. Why will the yield to maturity be lower than the current yield?
Read the worked answer
Accrued interest (30/360 day count, semi-annual coupon 8.24 / 2 = ₹4.12 per half-year, 180-day half-year):
- Days from 8 January to 12 July on 30/360 = 6 months (180 days) plus 4 days = 184 days.
- Accrued = 4.12 × 184 / 180 = ₹4.21 per ₹100 face value.
Dirty price payable = clean price + accrued = 104.60 + 4.21 = ₹108.81. On ₹10 crore face value the outflow is about ₹10.881 crore; the treasury should budget for the dirty price, not the screen price.
Current yield = coupon / clean price = 8.24 / 104.60 = 7.88%.
Why YTM is lower: the bond is bought at ₹104.60 against a redemption value of ₹100, so the investor loses about ₹4.60 of capital over the remaining life (roughly 8.5 years). That pull-to-par loss is amortised into the YTM, which works out to roughly 7.5%. Current yield ignores it. A bond bought above par always has YTM below current yield; below par, the reverse. On settlement the seller also receives the ₹4.21 accrued interest, which is why the clean price is quoted separately.
2. If the 25 basis point cut happens and the curve falls 25 bp, estimate the price gain on each bond. Then assess whether the G-sec is compatible with the 2% quarterly mark-to-market loss limit over a 15-month holding period.
Read the worked answer
Estimated price change for a parallel 25 bp fall:
- G-sec: 6.4 × 0.25% = 1.6%, about ₹16 lakh on ₹10 crore, plus carry of roughly 7.5% annualised.
- AAA bond (spreads unchanged): 6.1 × 0.25% = 1.525%, plus carry of 7.60%.
Against the policy limit. Modified duration cuts both ways. A 30 bp rise in yields produces 6.4 × 0.30% = 1.92% on the G-sec and 6.1 × 0.30% = 1.83% on the corporate bond, both inside 2% but uncomfortably close. A 35 bp rise gives 2.24% and 2.135%, which breach the limit. Over 15 months, a 35 bp adverse move is an ordinary event, not a tail, so the 10-year paper is not compliant with a hard quarterly loss cap unless the position is hedged or run at a much smaller size.
Judgement: the expected return of the G-sec if the cut lands is roughly 7.5% carry + 1.6% = about 9.1%, which beats the CD at 7.35%. But the CD has no mark-to-market risk and locks the return. Given the policy limit, the CD or a short-duration instrument is the defensible choice; taking the 10-year would mean the treasury is betting the policy limit will not be tested.
3. The treasury expects G-sec yields to fall 25 bp but credit spreads to widen 30 bp. What is the net yield change on the AAA bond and its approximate price effect, and what does this imply for buying credit to play a rate view?
Read the worked answer
Net change in the corporate bond's yield = fall in the benchmark + widening of the spread = -25 bp + 30 bp = +5 bp, i.e. its yield rises slightly even though the government curve rallies.
Approximate price effect = -modified duration × change in yield = -6.1 × 0.05% = -0.305%, a small mark-to-market loss, while the G-sec gains 1.6% on the same view.
Implication. The AAA bond's return is dominated by two separate bets: the rate view and the credit view. A 55 bp spread buys about 0.55% a year of extra carry over the G-sec, but a 30 bp widening gives back more than half of it plus the mark-to-market loss. If the treasury holds both views, it should not bundle them in one 10-year instrument: express the rate view in the G-sec, where duration is unencumbered by spread risk, and take credit exposure only where the carry adequately compensates for the spread volatility. Note also that wider spreads raise the bond's yield, so the loss is temporary if held to maturity, but the policy is marked to market quarterly, which makes temporary losses real for reporting.