Menu

GGA RESEARCH & METHOD

Trade-Channel Credit Governance

Published and reviewed: 2 October 2026

Deep governance of a dealer book is not a credit policy document. It is four decisions — who gets credit, who decides it has gone wrong, when it moves to recovery, and who carries the loss when it does not come back. On most Indian trade channel books, three of those four sit with the person whose incentive runs the other way. This paper measures what that costs.

481 of 481Accounts where assignment matched the aged balance exactly
35% → 2.4%Recovery yield, within one year against beyond five
41.4%Of the book already beyond two years when handed over
72.2%Of the residue classified as risk five years on
₹5.81 crOwed by dealers who were visited and could not be found

The argument

Revenue that has not been collected is not revenue. It is a loan — and the person who made it is never the person who has to live with it.

01

Executive summary

This paper is not a case study. It uses one Indian trade channel book, worked over five years, to measure something that applies to almost every company selling through dealers and distributors: the cost of letting the person who created a receivable decide when it becomes a recovery problem.

  1. Recovery yield collapses with age, and the curve is steep. On the same book, worked by the same team: 35.2% of value recovered where the balance was inside a year, 8.1% beyond two years, 2.4% beyond five. Recovery deteriorates sharply with age, from 35.2% inside one year to 2.4% beyond five.
  2. Assignment for recovery is mechanically late. On a 481-account tranche tested in full, the value assigned matched the balance already beyond 240 days on 481 of 481 accounts, to the rupee. The company retained everything inside 240 days and released everything past it. Nothing is ever handed over while it is still young enough to convert well.
  3. By the time the book was handed over it was already old. At the point of engagement the ageing report showed 61.4% of value beyond 350 days and 41.4% beyond two years. The condition was created before any recovery effort began.
  4. Five years on, the residue is terminal. 47.4% of what remains is beyond five years and 72.2% is classified as risk. ₹5.81 crore is owed by dealers who were physically visited and could not be found. The client's own dashboard annotates ₹21.96 crore as “very low” against ₹2.15 crore “effective”.
  5. Nobody reverses the revenue, and the reasons are structural. Reversal is an admission, it reduces a reported number, and the person who booked the sale is rarely still accountable for it. The cash loss has already happened; carrying it preserves only the appearance.
  6. The remedy is a governance change, not a collections one. Four decisions — credit granted, credit gone wrong, moved to recovery, loss carried — should not all sit with the sales organisation. Where the date of assignment is set by rule rather than negotiation, the book never reaches the part of the curve where recovery stops working.
02

The four decisions

Every trade channel receivable is governed by four decisions. They are rarely written down, and in most companies three of them sit in the same place.

Where each decision usually sits, and where it should
DecisionUsually made byWhy that is a problem
Who receives credit, and how muchSales, with finance approving a limit set on last year's volumeThe limit reflects selling ambition rather than payment behaviour. A dealer's capacity to pay is assessed once, at onboarding, and rarely again.
Whether a balance has gone wrongSales, informally, in a monthly reviewThe person answering is the person who booked it. “They will pay” is a sincere belief, not a forecast, and it costs nothing to hold.
When it moves to recoverySales, by consent — typically after a conversationThis is the decision that matters most and has the least structure. It is where the delay measured in this paper is created.
Who carries the lossFinance, years later, as a provisionBy the time it lands, the originating salesperson may have moved territory, role or employer. Nobody who made the decisions is present when the cost appears.
Read the second column down. Three of the four decisions are made by the function whose objective is volume, and the fourth is made by a function that had no say in the first three. That is not a collections weakness. It is a governance design.

The consequence is a systematic bias in timing. Nobody in the chain has an incentive to move a balance to recovery early. Sales loses a relationship and gains an admission; finance lacks the standing to overrule a commercial judgement; the customer certainly does not ask. So the default is delay, and delay is exactly what the yield curve punishes.

03

What delay costs, measured

The argument for acting early is usually made by assertion. On this book it can be measured, because the same customers were worked by the same team across five years and the collections are recorded by the age of the balance when it converted.

Recovery yield by age of balance
Age when collectedBalance outstandingRecoveredYieldShare now classified risk
0–90 days₹2.64 cr₹1.33 cr33.5%16.2%
90–180 days₹0.59 cr₹0.42 cr41.8%5.5%
1–2 years₹4.51 cr₹1.06 cr19.0%49.7%
2–3 years₹3.23 cr₹1.06 cr24.8%62.2%
3–4 years₹2.29 cr₹0.25 cr9.9%83.5%
4–5 years₹3.01 cr₹0.39 cr11.4%66.4%
Beyond 5 years₹14.83 cr₹0.36 cr2.4%93.7%

How to read the two left-hand columns. Receipts on this book are applied first-in, first-out, so money collected is set against the oldest open invoice rather than against the band it was earned in. Balance outstanding is a closing snapshot; recovered is cumulative over the engagement. The two are not a division pair — dividing one by the other does not reproduce the yield column, and should not be expected to. The yield is measured against the balance in each band at the point the recovery was worked.

Grouped, the curve is unambiguous. Balances inside one year returned 35.2%. Beyond two years, 8.1%. Beyond five years, 2.4%. The right-hand column moves with it almost exactly: 16% of the youngest balances are classified as risk, against 94% of the oldest.

Two readings of the same fact. A CFO can read it as the cost of delay — each year of waiting removes a large part of what was still recoverable. A board can read it as the price of the governance arrangement described above, since the delay is not accidental and nobody chose it deliberately either.

A rupee moved to recovery inside its first year returns about 35 paise. The same rupee moved after five years returns two and a half. Nothing about the dealer changed in between. What changed is that everyone who could have acted had a reason not to.
04

Who decides, and when

If delay is the cost, the question is what governs the timing. On this book the answer is unusually clear, because one assignment tranche can be tested line by line.

A list of 481 dealer accounts was prepared for handover to recovery, carrying total outstanding of ₹15.92 crore. Each row showed the balance inside 240 days, the balance beyond it, and a column marked as the assigned scope.

A 481-account assignment tranche, tested in full
ValueShareTreatment
Total outstanding on the list₹15.92 cr100%Prepared for handover
Balance inside 240 days₹6.14 cr38.6%Retained by the company
Balance beyond 240 days₹9.78 cr61.4%Released for recovery
Assigned scope₹9.78 cr61.4%Identical to the aged balance
Accounts where assigned scope equalled the aged balance exactly481 of 481100%No exceptions

On every account, without exception, the assigned value equals the balance already beyond 240 days. The rule is mechanical: the company keeps what is young and releases what is old.

Which means recovery never sees a balance while the yield curve is still favourable. By construction it receives nothing before 240 days — and because the book was already 41.4% beyond two years, most of what it receives sits where yield is 8.1% rather than the 35.2% available inside the first year. Whatever recovery achieves is then measured against a starting position that guaranteed a poor number.

An outside recovery firm assessed on a book that is only ever handed over after it has aged is being measured on a decision it did not make. So, more importantly, is the internal team — whose performance on the balances that were retained is generally not measured at all.

The selection underneath the rule

The formal rule is age. The discretion is over which accounts reach the list at all, and that discretion usually sits with the commercial team.

There is an incentive here worth naming plainly, because it is rarely hostile and almost always present. A salesperson asked which accounts should go for recovery has reason to nominate the ones already beyond hope: the difficult conversations leave their desk, and if recovery then performs poorly, the outcome confirms what they said about the outside firm in the first place. Meanwhile the accounts they believe are still winnable stay with them, unmeasured.

The test that would settle it is not run. Nobody compares what the internal team collected on the balances it retained against what recovery achieved on the balances it released, adjusted for the age of each. Until that comparison exists, the claim that the internal team would have done better is unfalsifiable — which is precisely why it survives.

05

What the end of the curve looks like

The book examined here is now five years past the handover. It is worth describing, because it is the destination that late assignment leads to and most companies never see it laid out.

The residue, five years on — 685 accounts, ₹31.25 crore
ClassificationAccountsValueShare
Classified risk465₹22.58 cr72.2%
Closed — settled, credit notes, deposits adjusted160₹3.83 cr12.2%
Under active recovery51₹3.45 cr11.0%
Awaiting client settlement decision8₹0.35 cr1.1%
Out of scope1₹1.04 cr3.3%

Beyond five years: 47.4% of the value. Beyond two years: 74.7%. Inside 90 days: 8.4%. And of the risk pool, 93.7% is more than five years old.

The reasons, gathered by field visit across nineteen states, are not commercial disputes. They are the consequences of time.

Why the risk pool cannot be recovered
ReasonValueWhat it means
Refuses to pay — legal recommended₹10.54 crThe dealer acknowledges the balance and declines. Recovery now requires process, not persuasion.
Visited, not traceable₹5.81 crField verification found no locatable counterparty. The business has gone.
Refuses to pay — legal already initiated₹4.17 crIn process, with the associated cost and duration.
Proprietor deceased₹0.52 crNo successor accepting the liability.
Dispute raised₹0.32 crA commercial claim, unresolved long enough to have become unresolvable.
False billing₹0.16 crIdentified only on physical verification.
₹5.81 crore is owed by dealers who no longer exist at any address the company holds. That is not a collections failure. It is the arithmetic of asking five years after the fact.

The client's own dashboard states the position more bluntly than any consultant would. Against the risk pool it carries a note reading ₹21.96 crore very low, against ₹2.15 crore effective. The assessment has been made internally. What has not been made is the accounting decision that follows from it.

06

Why nobody reverses the revenue

A balance identified as unrecoverable can sit on a balance sheet for years after everyone involved has privately concluded it will never convert. That is not negligence, and understanding why it happens is more useful than deploring it.

Three reasons it stays

  • Reversal is an admission. It restates a number that was reported, celebrated and in some cases incentivised. Nobody volunteers for that, and no single person has the standing to insist.
  • The originator is not present. By the time the write-off is proposed, the salesperson who booked it has often changed territory, role or employer. The decision falls to someone with no part in creating it and nothing to gain from surfacing it.
  • There is no trigger. Ageing thresholds prompt provisioning discussions, not reversal decisions. Absent a rule that forces the question at a date, the question is simply never formally asked.

What carrying it actually costs

The cash loss has already happened. Goods left the warehouse, working capital was consumed, and no payment arrived. Carrying the balance does not defer that loss; it defers only its recognition.

Tax is paid on revenue never received. Income tax is charged on revenue recognised at invoice, not on cash collected. Relief on a bad debt arrives only when the write-off is taken — so every year of delay is a year of tax paid on money that was never going to arrive.

And on most trade channel goods, so is GST. Indian GST is payable at the time of supply rather than on receipt, and there is no general mechanism to recover it on a receivable that fails. On agri inputs the rates vary by line: seeds for sowing are exempt at 0%, so a seed book carries no GST leakage; but fertilisers attract 5%, farm machinery and tractors 5%, and pesticides and agrochemicals between 5% and 18% depending on classification. On a taxed line, output GST on an uncollected invoice is a permanent cash cost, paid on a sale that never happened in economic terms.

Rates are stated as a general feature of the Indian GST framework and vary by product classification and over time. Readers should confirm their own position; this is not tax advice.

Nobody suffers from a carried bad receivable except the company and its shareholders. The dealer keeps the goods. The salesperson keeps the credited sale. The tax has been paid. What remains is a number on a balance sheet that everybody has quietly stopped believing.
07

The stock that was never going to be paid for

One commercial practice generates more of this than any other, and it is worth stating without euphemism.

A dealer pays for stock that has sold. Not for stock delivered, not for stock invoiced — for stock that has moved off their shelf and been converted to cash from the market beneath them. That is not a negotiating position; it is their cash reality, and no relationship alters it.

Which makes quarter-end dumping self-defeating. Material pushed in excess of order, or unordered altogether, to close a period number does three things at once. It creates an invoice the dealer will not pay until the goods sell, which may be never. It occupies warehouse space and shelf position the dealer needs for goods that turn, which they resent. And it contaminates every subsequent invoice to that account, because the dealer now has a standing grievance to point at whenever payment is requested.

The period target is met and a receivable with a five-year half-life is created. On the yield curve above, a balance that ages past two years returns 8 paise in the rupee. Revenue recognised at 100 and realised at 8 is not a sale; it is a transfer of inventory into somebody else's warehouse at the shareholder's expense.

Dumping does not create revenue. It converts inventory into a receivable, and the receivable into a write-off nobody will authorise.

The seasonal case is sharper still. On a book selling into a crop cycle, dealer payment capacity peaks in a narrow post-harvest window. Collections on this book ran at ₹0.82 crore in January against ₹0.09 crore the following August — a ninefold difference driven by the calendar rather than by effort. Stock pushed just after that window has an entire season to sit unsold before the dealer next holds cash.

08

What deep governance looks like

Six changes. None requires a new system, and none is expensive relative to what the curve above costs.

  1. Set the assignment date by rule, not by conversation. A balance crossing a defined age moves to recovery automatically, with an exception process that is written down and approved outside sales. The single most valuable change available, because it operates on the part of the curve where recovery still works.
  2. Assign a representative sample, not the residue. Where an outside firm is used, include a proportion of younger balances. Otherwise both the firm's performance and the internal team's are unmeasurable, and the comparison that would settle the argument is never available.
  3. Measure the internal team on what it retained. The claim that the internal team would have collected better is testable: compare recovery on retained balances against recovery on released balances, matched for age. Until that number exists, the claim is not a fact.
  4. Move the credit decision to a body that carries the loss. Limits set on selling ambition and reviewed only at onboarding are not credit decisions. Where the shareholder bears the risk, the shareholder's representatives should set the exposure.
  5. Force the reversal question on a date. A balance beyond a defined age comes to the audit or risk committee with a recommendation, and a decision is recorded either way. The point is not to write off faster; it is to stop the question going unasked for five years.
  6. Price dumping into the sales incentive. Where a period-end despatch is not paid for within a defined window, the commission or credit attaching to it reverses. Sales targets that count despatch rather than realisation will produce despatch rather than realisation.
09

Method and limitations

This paper draws on one Indian trade channel book in agri inputs, worked from 2021 onward, together with fifteen monthly dashboards covering April 2025 to July 2026 and two handover files from December 2021 and November 2022. Analysis is computed from the client's own line-item data.

The assignment finding is a full-population test, not a sample. All 481 accounts in the November 2022 tranche were checked; on every one, the assigned scope equalled the balance beyond 240 days exactly. It is one tranche on one book and is reported as such.

The yield figures compare collections recorded against the age of the balance when converted, with pendency measured at a single date. Collections span twenty months and the pendency is a snapshot, so the ratios are directional rather than exact. The shape of the curve — sharply declining with age — is robust to that limitation; the individual percentages should not be read as precise recovery rates.

The handover ageing describes the full ageing report shared at engagement — 11,284 customers and ₹2,764 crore of outstanding — of which the value actually assigned for recovery was a subset. Percentages describing the age profile relate to that full report; no performance claim is made against it.

The reason taxonomy is the client's own classification, gathered through field visits across nineteen states, and is reported as recorded rather than independently verified.

The account of how assignment decisions are made reflects practitioner observation across trade channel engagements. The mechanical assignment rule is measured; the selection behaviour underneath it is inference from that pattern and from repeated experience, and is presented as such rather than as measurement.

No client, dealer or individual is identified. GGA does not publish the payment behaviour of any named company.

10

Definitions

Terms as used in this paper
TermDefinition
Assignment dateThe point at which a balance moves from internal follow-up to recovery. The single most consequential governance variable on a trade channel book, and usually the least formalised.
Adverse selection in assignmentThe tendency for accounts released to recovery to be systematically worse than those retained, because the selection is made by a party with an interest in the outcome.
Recovery yieldValue converted to cash as a proportion of the balance in an age band. Declines sharply with age; on the book examined, from 35.2% inside a year to 2.4% beyond five.
Carried revenueRevenue recognised and invoiced but not collected, retained on the balance sheet after it has ceased to be realistically recoverable.
DumpingDespatch in excess of order, or unordered, to meet a period sales target. Creates a receivable contingent on the dealer selling stock they did not want.
Post-harvest windowThe concentrated period in which agricultural channel customers hold cash. Payment capacity outside it is limited regardless of follow-up intensity.
11

Frequently asked questions

When should an overdue receivable be moved to recovery?

Early, and by rule rather than by negotiation. On the book examined here, balances worked inside one year returned 35.2% of value; beyond two years 8.1%; beyond five years 2.4%. Recovery deteriorated sharply with age, though the band-to-band sequence is not smooth. Where the assignment date is set automatically at a defined age with a written exception process approved outside sales, the book never reaches the part of the curve where recovery stops working.

Who should decide that a receivable has gone bad?

Not the person who created it. Four decisions govern a trade channel receivable — who receives credit, whether a balance has gone wrong, when it moves to recovery, and who carries the loss. In most companies the first three sit with sales and the fourth with finance years later, by which time the originating salesperson may have changed territory, role or employer. Credit risk is borne by the shareholder, so the exposure decision belongs with the shareholder's representatives.

Why do collections firms appear to underperform on trade channel books?

Frequently because of what they are given. On a 481-account tranche tested in full, the value assigned for recovery equalled the balance already beyond 240 days on every single account, to the rupee — the company retained everything younger. A firm that only ever receives balances already weighted toward the oldest bands is being measured on a decision it did not make. The internal team's performance on the balances retained is generally not measured at all.

What does carrying an uncollectable receivable actually cost?

More than the balance. The cash loss has already occurred — goods shipped, working capital consumed, no payment received — so carrying defers recognition rather than loss. Income tax is charged on revenue recognised at invoice, with relief arriving only when the write-off is taken. And Indian GST is payable at time of supply with no general mechanism to recover it on a failed receivable, so on taxed lines the output GST is a permanent cash cost on a sale that never economically happened.

Does GST apply to uncollected agricultural input sales?

It depends on the product line. Seeds for sowing are exempt at 0%, so a seed receivable carries no GST leakage. Fertilisers attract 5%, farm machinery and tractors 5%, and pesticides and agrochemicals between 5% and 18% depending on classification. Because Indian GST is payable at time of supply rather than on receipt, output tax on a taxed line that is never collected is not recoverable through any general mechanism. Rates vary by classification and over time and readers should confirm their own position.

Why is quarter-end stock dumping counterproductive?

Because a dealer pays for stock that has sold, not for stock delivered. Material pushed in excess of order creates an invoice contingent on goods the dealer did not want moving off their shelf, occupies warehouse space they need for goods that turn, and gives them a standing grievance to raise whenever payment is requested. The period target is met and a receivable is created that, once past two years, returns around 8 paise in the rupee.

How does seasonality affect trade channel collections?

Payment capacity concentrates in a narrow window. On the agri book examined, collections ran at ₹0.82 crore in January against ₹0.09 crore the following August — a ninefold difference driven by the crop calendar rather than by collection effort. Stock despatched just after a post-harvest window has an entire season to sit unsold before the dealer next holds cash, which is why despatch timing matters as much as credit terms.

What happens to a dealer book left too long?

It becomes unrecoverable in ways that have nothing to do with willingness. Five years after handover, the book examined here held 47.4% of its value beyond five years and 72.2% classified as risk. Field visits across nineteen states found ₹5.81 crore owed by dealers who could not be located at any address held, ₹0.52 crore where the proprietor had died with no successor accepting the liability, and ₹10.54 crore where the dealer acknowledged the balance and simply declined to pay.

Evidence boundary

This publication separates measured findings, modelled illustrations and operating observations. Population evidence does not predict the outcome of an individual receivable, counterparty or company.

Public law, policy terms and procedures can change. Current primary material and the facts of the live case govern.

Download the publication

Download the complete publication

Email GGA gg@ggaindia.co.in