Average DSCR vs Minimum DSCR: The Distinction Indian Appraisal Notes Require

Average DSCR is coverage across the whole repayment period; minimum DSCR is the worst single year in it. The average is what most appraisal notes lead with, and it is the number that cannot default. The minimum is the year the borrower actually misses an instalment. YuSight builds the year-wise schedule with 100% of figures cited to the source page, so both are auditable.

Key facts

• A missed instalment is dated from the day it falls due, not from the average. The RBI Master Circular on IRAC norms (RBI/2023-24/06, 1 April 2023) classifies a term loan as an NPA where "interest and/ or instalment of principal remains overdue for a period of more than 90 days," with SMA-0, SMA-1 and SMA-2 marking 30, 60 and 90 days (RBI). Classification tests one date in one year.
• The RBI prescribes neither an average nor a minimum DSCR. The Reserve Bank of India (Project Finance) Directions, 2025 (RBI/2025-26/59, 19 June 2025) require the repayment schedule to be "realistically designed to factor in the initial cash flows" and cap the tenor, including moratorium, at 85% of economic life — with no ratio attached (RBI). Both floors used below are bank policy.
• Asset quality is strong, which is when tail-year risk stops being priced. GNPA for scheduled commercial banks hit "a multi-decadal low of 2.2 per cent at end-March 2025 and 2.1 per cent at end-September 2025" (RBI, *Report on Trend and Progress of Banking in India 2024-25*, 29 December 2025).
• In the schedule below the average reads 1.51x and the minimum 1.10x — a 0.41x gap, with the worst year at Year 7 of 8, not the first repayment year.

What is the difference, precisely?

The minimum is unambiguous: the lowest annual DSCR in the schedule. The average is not. It can be a simple mean of the annual ratios or a ratio of summed numerators to summed denominators, each computed with or without the moratorium year — four methods, four answers, as set out in DSCR calculation for term loans in India.

The deeper difference is testability. A minimum DSCR can be tested every year on audited accounts. An average DSCR cannot be tested until the loan matures, by which point the covenant has no remedy left. A sanction letter that covenants an average covenants something monitoring cannot check.

A schedule where average passes and minimum fails

Kanhaiya Cold Chain Private Limited, Nashik — temperature-controlled warehousing, brownfield expansion of two chambers. Term loan ₹1,600 lakh at 9.25%, door-to-door 8 years: Year 1 moratorium with interest serviced, then a step-up principal ladder across Years 2 to 8, negotiated to keep early outgo low. Book depreciation ₹185 lakh a year, cash credit interest ₹22 lakh a year (₹14 lakh in the part-year), tax at 25%. A six-year cold-storage contract with a QSR chain runs to the end of Year 6; projections assume the space is re-let from Year 7 at lower tariffs.

Interest is on the opening balance. All figures in ₹ lakh.

Numerator is PAT + depreciation + term-loan interest; denominator is term-loan interest + principal. Year 7 in full:

PBT = 470 − 185 (dep) − 60 (TL interest) − 22 (WC interest) = 203 Tax at 25% = 51 PAT = 152 Numerator = 152 + 185 + 60 = 397 Denominator = 60 + 300 = 360 DSCR = 397 ÷ 360 = 1.10x

The averages, on the same eight years:

Against a policy of average ≥ 1.40x and minimum ≥ 1.20x, this proposal passes the average test comfortably at 1.51x on the simple mean of all eight years, fails it at 1.37x on a ratio-of-sums basis excluding the moratorium, and fails the minimum test at 1.10x either way. One schedule, three verdicts, decided by presentation.

What is the moratorium effect?

Year 1 shows 2.32x, the strongest year in the loan, and it is coverage of nothing — the denominator contains no principal. Including it lifts the simple-mean average from 1.39x to 1.51x: 0.12x bought by a year in which the borrower repays nothing.

A moratorium year is an interest-cover year. Show it so the sanctioning authority can see the structure, and exclude it from every average. Where interest is capitalised rather than serviced the denominator is genuinely zero and the year should read "NA" — averaging a division by zero is how a note ends up quoting a figure nobody can reproduce.

What is the tail-year effect?

The usual heuristic — the worst year is the first repayment year, because principal is flat while profits grow — holds for a straight amortisation. It fails here, for two reasons both visible at sanction:

1. The step-up ladder loads the largest instalments last. Principal rises from ₹100 lakh to ₹350 lakh while interest falls from ₹148 lakh to ₹32 lakh, so the denominator is at its highest of the whole tenor in Years 7 and 8.
2. A contracted cash flow ends inside the tenor. The QSR contract expires at the end of Year 6. EBITDA drops ₹125 lakh in Year 7 and only partly recovers. This is the borrower's own projection, not a stress case.

Anything in this list should send an analyst to the tail: a step-up or ballooning ladder, an anchor tenancy or offtake agreement shorter than the loan, an interest subvention that lapses, a tax holiday that expires, a major overhaul capex, or an existing facility whose amortisation stacks on in a later year. Averaging does not merely hide a tail-year dip — it hides it best, because a late year is diluted by every earlier one.

What should a credit committee do with a 1.51x average and a 1.10x minimum?

Not approve on the average, and not decline on the minimum either. This is a structuring problem, and the arithmetic says which fixes work.

Flattening the ladder does not fix it. Replace the step-up with equal instalments of about ₹229 lakh from Year 2:

Year 7 revised: interest 42, principal 228 → denominator 270 PBT 470 − 185 − 42 − 22 = 221; tax 55; PAT 166 numerator 166 + 185 + 42 = 393 DSCR = 393 ÷ 270 = 1.46x (was 1.10x) Year 2 revised: interest 148, principal 229 → denominator 377 numerator unchanged at 419 DSCR = 419 ÷ 377 = 1.11x (was 1.69x)

The minimum simply moves from Year 7 to Year 2 and stays at 1.11x. Flattening relocates the failure to the year carrying the full interest load; it creates no coverage. Say so in the note, because "we will flatten the repayment" is the reflex proposal in committee.

What does move the minimum above 1.20x:

• Lengthen the tenor, subject to the 85%-of-economic-life cap, which cuts annual principal in every year at once.
• Reduce the loan. Additional promoter contribution cuts interest and principal together, lifting the whole curve rather than tilting it.
• Size the DSRA to the actual gap. Year 7 needs 1.20 × 360 = 432 against 397 available, a shortfall of ₹35 lakh; Year 8 needs 1.20 × 382 = 458 against 435, a further ₹23 lakh. ₹58 lakh covers the projected tail gap. The standard "one quarter's debt service" reserve would be ₹90 lakh and would bear no relation to where this file is thin.
• Covenant the driver, not just the ratio. The Year 7 dip is a contract expiry, so require evidence of renewal or re-letting twelve months ahead, with a cash sweep if it is not produced.

How should the appraisal note present both?

Most Indian notes state both numbers. Where they fail is in linking them to the sanction conditions and to what monitoring will test. Four lines fix it:

1. Minimum first, with the year named — "Minimum DSCR 1.10x in Year 7 (FY2033)", not "Average DSCR 1.51x".
2. The averaging method, spelled out — "1.51x, simple mean of annual ratios including the moratorium year; 1.37x on a ratio-of-sums basis excluding it."
3. The driver of the minimum year, in one sentence — here, contract expiry plus step-up ladder.
4. What the covenant tests, and how often. A condition reading "DSCR of not less than 1.20x tested annually on audited accounts" covenants a level the note itself projects will be missed in Year 7 — a breach written into the file on day one.

That mismatch is the practical cost of conflating the two: the proposal is approved on a number that is never tested, and monitored on a number that was never approved. How the covenant is defined is covered in covenant testing for DSCR and leverage; how the test then runs year after year in covenant monitoring in commercial lending.

FAQ

What is the difference between average DSCR and minimum DSCR?

Average DSCR is coverage across the whole repayment period; minimum DSCR is the lowest single year in it. The average tells you whether the facility is affordable in aggregate. The minimum tells you which year the borrower misses an instalment, and that is the year asset classification cares about.

Which DSCR does the sanction note quote?

Indian appraisal notes normally quote both, with the average on the summary page and the year-wise schedule in an annexure. The trouble is that the summary leads with the average, so the committee's attention lands on the number that cannot default.

What average DSCR do banks look for on a term loan?

Between 1.40x and 1.75x is the common range for a brownfield facility, with a minimum-year floor around 1.20x to 1.25x. None of it comes from the RBI; it varies by sector, tenor and security.

Can a proposal pass on average and still be declined?

Yes, and it should be, if the minimum year has no cushion and no fix. A 1.10x minimum means a ten per cent slip in that one year produces a missed instalment, and no amount of coverage in Year 3 helps in Year 7.

Should the moratorium year go into the average?

No. There is no principal in the denominator, so it is an interest-cover year rather than a coverage year. Show it in the schedule, exclude it from the average, and say in the note that you have.

Why does the worst year sometimes fall at the end of the tenor?

Because something changes late — a step-up ladder, a balloon, an offtake or lease contract expiring inside the tenor, a tax holiday ending, or a second loan's amortisation stacking on. Straight-amortisation intuition does not survive any of those.

Can you covenant an average DSCR?

You can draft it, but you cannot test it until the facility matures, which is when you have no remedy left. Covenant the annual minimum and use the average for sizing.

Does the RBI prescribe a minimum DSCR?

No. The Project Finance Directions, 2025 require the repayment schedule to be realistically designed against initial cash flows and cap the tenor at 85% of economic life, but attach no ratio. Every DSCR floor you are held to is your own policy.

Key takeaways

• The average is a sizing number; the minimum is a credit number. Quote the minimum first, with the year it falls in.
• "Average DSCR" has four definitions. Name the one you used, and say whether the moratorium year is in it.
• The moratorium year flatters the average — here by 0.12x — and repays nothing.
• The worst year is not always the first repayment year. Step-up ladders, expiring contracts and stacked amortisation push it into the tail, where averaging hides it best.
• Flattening the ladder often just relocates the failing year. Test the fix in the schedule before proposing it in committee.
• Set the covenant at a level the note's own projections support, or change the structure.

Building this schedule means pulling PAT and depreciation from audited statements, projections from CMA data, the ladder from the sanction terms and the existing debt schedule from the borrower's other facilities — the reconciliation the credit appraisal note format expects. A single-year check is what a DSCR calculator is for; a tenor needs the whole schedule.

Watch YuSight spread a real balance sheet and see the year-wise DSCR schedule build itself, with every figure traced to its source page.