Buying a State's 10-Year SDL at 68 bps
A public sector bank's investment committee is evaluating a 10-year state development loan (SDL) at 7.86%, when the same-tenor G-sec yields 7.18%. The state's numbers: debt-GSDP 31.8%; own tax revenue growing 6.2% against nominal GSDP growth of 9.5%; fiscal deficit 3.6% of GSDP against an FRBM ceiling of 3.0%, with a 0.5% cushion permitted only if power-sector reform milestones are met; and ₹9,400 crore of borrowings routed through three state-owned SPVs carrying state guarantees, of which ₹2,100 crore has already fallen due and was serviced by taking a fresh SPV loan. Power discom dues to generators stand at ₹12,000 crore. The state's SDL spread over G-sec has widened from 28 bps to 68 bps in a year. The mandate caps the bank's exposure to any single state, and this state is at 55% of its cap. Treasury argues the 68 bps compensates for the risk; the risk head wants the state placed on negative watch. SDLs carry nil credit risk weight and are SLR-eligible.
1. What is this state's fiscal position signalling, and is 68 bps adequate compensation for a 10-year exposure? Show your reasoning.
Read the worked answer
Reading the signals. Debt-GSDP of 31.8% is unremarkable; the trajectory is not. Own tax revenue growing at 6.2% against 9.5% nominal GSDP means the revenue base is shrinking relative to the economy, so a 3.6% deficit is being financed without commensurate revenue effort. The 0.5% reform-linked cushion is conditional; the extra 0.1% above the 3.5% ceiling has no identified source. And the ₹9,400 crore of guaranteed SPV borrowing - 1.5–2% of GSDP if GSDP is around ₹6 lakh crore - is debt in substance: ₹2,100 crore already due was rolled into a fresh SPV loan, which is a classic sign of a hidden liability being refinanced rather than repaid. Adjusted for this, effective debt and deficit are materially worse than reported, and the spread widening from 28 to 68 bps says the market has begun to price it.
Is 68 bps enough? Decompose it:
- Carry: 68 bps per year × 10 years = 6.8% of face if the spread holds and the bond is held to maturity.
- Mark-to-market: modified duration at 7.86% for 10 years ≈ 6.7. A 50 bps further widening costs ≈ 3.4% of value; a 100 bps widening ≈ 6.6%.
So one bad quarter at 100 bps wipes out the entire ten-year compensation. The bank must survive roughly five years of carry to recover a single 50 bps shock. For a state with unaudited off-budget debt and a rollover already in progress, 68 bps is not adequate compensation at size - it is only adequate in a small, held-to-maturity, matched-liability position.
2. What should the investment committee decide, and what conditions should attach to any purchase?
Read the worked answer
Decision: do not veto the state, but do not add at the current size and spread. Approve a small, ring-fenced position with hard conditions, and put the state on watch.
Conditions to attach:
- Size cap. Restrict fresh exposure to, say, 10% of the state cap (moving from 55% to 65% at most), funding it from maturing G-sec rather than new deposits.
- Held-to-maturity classification. Book it in HTM within the bank's HTM ceiling and match it with liabilities of similar tenor, so a spread blow-out does not hit the P&L or the CRAR through mark-to-market.
- Spread trigger. Stop buying if the spread crosses 90–100 bps; that is the market repricing the very risk the analysis flags, not a bargain.
- Disclosure condition. Require the state's annual budget and Finance Accounts to disclose SPV/guarantee borrowings; absence of disclosure is itself a reason to stay small.
- Monitoring calendar. Review after the state's next borrowing programme, after the discom reform deadline, and on any downgrade of the state's guarantee redemptions.
Reasoning on the mandate. The bank's exposure cap exists to stop exactly this: a nil-risk-weight asset that generates a tempting spread. Because SDLs are SLR-eligible and carry zero credit risk weight, regulatory demand keeps spreads structurally tighter than the credit deserves - meaning the market price understates risk and the bank must do its own credit work. Pricing at 68 bps is a signal to reduce, not to add.
3. How should the bank treat the ₹9,400 crore of guaranteed SPV borrowings in its internal assessment, and what does the policy framework say about them?
Read the worked answer
Treat them as state debt. Three tests all point the same way: the SPVs are state-owned, the borrowings carry an explicit state guarantee, and ₹2,100 crore of maturing principal was repaid by fresh SPV borrowing - a rollover of state obligations by another name.
Policy framework. The Fifteenth Finance Commission recommended that off-budget borrowings be treated as debt of the state for all purposes, including FRBM compliance, and that states legislate a ceiling on guarantees; the Centre has also linked a part of additional borrowing space to power-sector reform milestones. A bank's own credit assessment should therefore recast the state's debt-GSDP and fiscal deficit to include guaranteed SPV liabilities, irrespective of how the state presents them in its budget documents.
Practical treatment in the appraisal:
- Add ₹9,400 crore to outstanding debt and recompute debt-GSDP - the ratio moves from 31.8% to roughly 33.4% on these assumptions.
- Treat the guarantee as a contingent liability likely to crystallise, given the rollover.
- Note that SPV debt servicing competes with the state's own debt service, which is why the spread widened before any formal default.
- Distinguish a guarantee from a comfort letter: only the former is legally enforceable, and even a guarantee is worthless if the guarantee ceiling is already exhausted.
The exam-relevant point is that the headline fiscal deficit of a state is not an adequate risk measure once SPVs, guarantees and discom dues are in the picture, and the SDL spread is a lagging, regulatory-distorted indicator of that risk.