Methodology & evidence base
How the validity-gap estimate is built · what each coefficient rests on
why it survives the hardest challenge a finance leader can put to it
Four operating principles
A sizing number is only worth presenting if the finance leader on the other side of the table can check it, argue with it, and find it conservative. Four principles enforce that.
Each coefficient sits at the 25th percentile across its evidence range — deliberately low. A defensible-low number a controller can argue against on the low side is more useful than a defensible-high number an analyst discounts on sight. Actual cost is almost always higher.
Every figure is citation-traceable to a named primary source: an audited filing, a government circular, a rating-agency note, or an identified industry study. Where sources of different strength conflict, the harder source wins.
The most-challenged coefficient — dispute carrying cost — is not validated against a single balance-sheet line, because that line is systematically misleading at mature companies. It is validated across four independent signals that together survive the distortion.
A method that hides its weaknesses is not one you should trust. The known limits are listed below, in writing, because each one bears on how a careful reader should weight the estimate against their own situation.
Source discipline · hardest first
Sources are graded by how directly they can be verified. A claim resting on an audited filing outranks one resting on industry research; where they disagree, the audited filing decides.
The cost anchors, derived
Two cost families are tracked, and they are never summed together. The first is the Cost of Invalidity (COI) — four leakage lines that net to a conservative reference rate of ≈2.38% of revenue. The second is reconciliation labour, a separate operating-leverage line. Each carries a published floor coefficient, an explicit formula, and a named source. Below, X is your annual revenue and Y your annual procurement spend (both in ₹ crore).
| Cost anchor | Floor coefficient | Formula | Primary source of the rate |
|---|---|---|---|
| Procurement leakage Cost of Invalidity |
2.55% of Y | Y × 0.0255 | 25th-percentile interpolation across material-price, quantity, scope-change, duplicate-invoice and timing leakage. Anchored Xelix 0.35% (best-in-class) to WorldCC ≈8.6% full-lifecycle erosion (Deloitte/WorldCC 2023); SC&H recovery-audit data. |
| Dispute carrying cost Cost of Invalidity |
1.0% of Y | Y × 0.01 | Conservative end of SC&H 1–3% disputed-value-at-any-moment range; corroborated by listed-company multi-metric evidence (see below) and IOFM / CreditPulse aging benchmarks. |
| GST input-credit lapse Cost of Invalidity |
0.5% of input credit | (Y × 0.18) × 0.005 | Input-credit lapse rates only — ClearTax 1% monthly floor, IRIS GST 3–5% typical first-pass mismatch. Sourced to lapse-rate studies, not to any disputed-order amount. |
| TDS reconciliation loss Cost of Invalidity |
0.5% of TDS receivable | (X × 0.025) × 0.005 | First-pass mismatch 5–15% (ClearTax); structural three-party mismatch mechanics (Terra Insight); quarterly volume thresholds (BDO India). |
| Reconciliation labour Separate operating-leverage line · never summed into COI |
3 FTE per ₹50 Cr @ ₹40k/mo | X × 28,800 | Productivity benchmarks APQC / Stampli / Auxis / IFOL (3 FTE is below the industry median of 4); Indian SMB loaded-cost baseline. Reported alongside the COI reference rate, never added to it. |
A further question — what share of your spend runs through formal written agreements — is asked, but it is an eligibility gate, not a cost line. It separates the addressable zone — the enforcement-eligible, contract-governed surface Axiom acts on — from full company-wide COI, which is carried as context. Below roughly 40% contract-governed spend, recognition enforcement has too little surface to act on, and the estimate is withheld rather than inflated. It routes the conversation; it never multiplies the number.
The evidence base · twenty-five companies, audited
The dispute coefficient does not rest on assertion. It rests on the publicly disclosed dispute exposure of twenty-five listed Indian companies — contingent liabilities, disclosed adjudication orders, and arbitration amounts, drawn from audited financials covering recent years. A representative sample:
Multiple concurrent tax and customs disputes across states in a single year — a ₹374.68 Cr demand alongside several smaller orders, plus historical arbitration of ₹188.53 Cr. The pattern: many disputes running at once across jurisdictions.
A ₹1,979.09 Cr arbitration-awards portfolio assigned to a wholly-owned claims subsidiary; total disclosed dispute exposure near ₹3,153 Cr. A mid-cap that built an entire subsidiary to manage disputes as a permanent function.
On a single thermal-plant contract, a ₹1,557 Cr claim returned only ₹101 Cr — 93.5% rejected as invalid — while a counter-claim allowed ₹700 Cr against it. Rating commentary tied margin moderation to legacy claim write-offs.
Seven-plus separate input-tax adjudication orders across five states, each from ₹32 lakh to ₹4.81 Cr — the many-small-disputes-many-jurisdictions pattern that contract-governed transaction structure is built to prevent.
A competition-penalty dispute outstanding nine-plus years — a live example of the "more than three years" aging bucket — sits alongside concurrent tax demands totalling roughly ₹2,679 Cr open at once.
A state mining-tax dispute of roughly ₹17,300 Cr compounded over nearly two decades. Rating analysis notes the leverage effect of whether it is classed as contingent or operating.
A multi-year, multi-issue tax cycle on a distributor of this scale: income-tax demands of ₹233.66 Cr, ₹175.10 Cr and ₹136.25 Cr across consecutive years; a ₹91.74 Cr indirect-tax demand later quashed at appeal. The receivables-flow case in full.
A largest single dispute of ₹67.47 Cr over input-credit treatment, plus a long tail of lakh-to-crore-scale orders across states — the pharma pattern of many concurrent small disputes, cumulatively significant.
A regulator demand of roughly ₹9,779 Cr carried as contingent liability for years — among the most-disclosed regulatory disputes in Indian corporate history, still live.
This is the publicly admitted residue — what has reached audited disclosure. The unrecognised and unreconciled portion of the gap, the part that has not yet escalated into a formal dispute, sits upstream of this and is structurally larger. The point of the evidence is not any one company's number; it is that disputes are an architectural feature of contract-governed execution under existing tools, not an occasional incident.
The strongest challenge — answered
"If disputes are structural, show me the disputed-payables line in their audited filings."
It is a fair challenge — and on the single line item, the challenger is right. Mature large-caps routinely show zero or near-zero disputed trade payables across multiple years, even with documented active disputes. That is not evidence the disputes are absent. It is the operational signature of a mature finance function, produced by three mechanisms:
Disputed amounts are moved out of trade payables into contingent liabilities, legal provisions, or contract liabilities before they are ever formally tagged "disputed." The amount is still there — just not on that line.
"Disputed dues" requires positive evidence of an acknowledged-but-contested liability. Many live disputes — including multi-year arbitrations — never qualify, because the company's position is that nothing is owed at all.
Quarterly-close discipline settles, writes off, or reclassifies aged payables faster than disputes resolve. Items that would technically qualify are off the line before the audited statement is produced.
So validating against that one line validates against the wrong number — true on the line item, wrong about the gap. The defensible method combines four metrics, each capturing a facet the single line misses:
Where reclassified disputes actually surface. Ratios run a few per cent at large-caps with institutional capacity to absorb them, and materially higher at mid-caps carrying dispute residue directly.
Payables beyond one year are a near-proxy for de-facto disputes regardless of label — if it has not settled in a year, something in the transaction is contested.
Doubtful-debt provisioning, exceptional losses on disputed contracts, rating actions citing working-capital stress — disproportionately concentrated where absorption capacity has been overrun.
Dedicated claims subsidiaries, routine multi-state arbitration, promoter pledging linked to dispute-driven working-capital crises — the structural tells.
Applied across the evidence base and read at the conservative 25th percentile, the four metrics recover the dispute signal the single line misses and land at the 1.0% floor. The coefficient is multiply-anchored: by audited filings, by independent industry benchmarks, and by the structural admission of major firms that have built permanent machinery to manage disputes.
Two institutional patterns · both out of reach for mid-market
The clearest proof that disputes are structural is that the most sophisticated companies have built dedicated institutional machinery to handle them — through two distinct routes. Neither is available to a mid-sized business, which carries the residue directly.
One major construction firm runs a dedicated claims-management subsidiary, incorporated in 2009 and operating for seventeen years as a permanent function.
A second major firm takes the opposite route — realising disputed claims actively while carrying the residual directly on its own books.
Either route requires institutional capacity — a subsidiary, or capital-markets and government-scheme access — that a mid-sized firm does not have. The existence of either pattern, at scale, is the validation: disputes are a feature of the architecture, not a run of bad luck.
How self-reported figures are read
When a company's own reported figures — its year-end disputed line, its statutory late-payment disclosure — are used as evidence, they are read through a single discipline tied to the audit opinion. This governs how the evidence is weighed; it is never something asked of you as an input.
When the latest statutory audit opinion is clean and the internal-controls opinion names no weakness on reconciliation, vendor records, balance confirmation or records integrity, a near-zero disputed line genuinely tells you the book is in order. The self-reported figure can be read at face value.
Under any modified opinion — qualified, adverse or disclaimer — or a named control weakness on that axis, the same near-zero line is not low; it is uninformative. The disclosure is impugned. The reading falls back to the operating cycle, never to the self-reported figure.
The logic mirrors the challenge above. The same institutional mechanisms that drive a healthy company's disputed line toward zero "because the book is clean" are exactly inverted under a modified opinion, where a near-zero line reads "because the disclosure is void." What matters is the type of opinion, not its severity — a disclaimer is the most severe class yet still cannot affirm that a self-reported figure is reliable. A statutory late-payment disclosure can corroborate the residue where the book is clean; under a modified opinion it carries no weight either.
A January 2026 regulatory inflection
In January 2026 the highways ministry issued a circular removing disputes of ₹10 crore and above from arbitration, routing them through tiered administrative settlement with civil litigation as the fallback. It applies with immediate effect across toll, hybrid-annuity and engineering-procurement contracts.
The effect on the evidence base is direct: disputes that previously cleared via arbitration in two-to-four years will now sit longer in the oldest aging bucket as they migrate to civil courts. The arbitration route that some firms had institutionalised becomes less effective. This strengthens the case for prevention over resolution — whatever the forum, preventing a dispute is structurally superior to resolving one, and recognition enforcement acts before a dispute can form.
What the method does not yet know
Each limit below is surfaced because it bears on how a careful reader should weight the estimate. None of them is hidden in a footnote.
Every coefficient is a 25th-percentile floor. Actual exposure in many mid-market environments runs two-to-three times higher. The floor exists for cross-company defensibility — to be hard to argue down — not to claim what a given business actually carries.
The most institutionally sophisticated firms reclassify and actively realise disputes, so their quantifiable commercial-dispute line collapses toward zero. The exposure does not disappear; it relocates into harder-to-quantify places — regulatory and tax contingents, unbilled-recognition residue, intra-group concentration. The floor is best read as a conservative proxy for that larger, less-quantifiable recognition-integrity exposure, not as the whole of it.
The formula applies single floor coefficients uniformly. Real leakage and dispute-aging patterns differ by industry — cement, steel, fashion, pharma, distribution and real estate do not behave alike. Per-sector adjustment factors are work in progress; until they land, the estimate is a sector-agnostic floor.
The listed-company evidence clusters among larger firms, so the estimate is most confident there. For smaller businesses — broadly under ₹100 crore in revenue — the coefficients are extrapolated from theory and market data rather than anchored to audited filings of directly comparable firms.
Several leakage and reconciliation benchmarks are predominantly US and UK data; the dispute base and the government and tax layers are India-specific. The three layers carry different citation weight, and the India-specific layers govern where they apply.
The number estimates the cost you carry today from the validity gap. It does not forecast how much of that would be recovered by deploying recognition enforcement — recovery depends on how much contract-governed surface comes into scope, your existing controls, and how much of the gap is structurally addressable. None of the public data here is a substitute for your own books.
In summary
A Cost of Invalidity at a conservative ≈2.38%-of-revenue floor, with reconciliation labour reported as a separate operating-leverage line — never summed into it. An evidence base of twenty-five audited filings. A four-metric defence against the hardest challenge. A reading discipline for what self-reported figures can and cannot mean. And the limits, in writing. The estimate is built to be checked.