When a unit comes back, its cost has to come back too. A refund credits the cost of goods as well as reversing the revenue, and there is a genuine question about which cost that is. The unit was bought at one price, sold on a day when a particular cost was effective, and returned on a day when the cost may have changed. Credit the return at today's cost and a store can manufacture margin out of nothing simply by having supplier prices move between the sale and the return. The rule that prevents it is short: the credit is valued at the cost that was effective on the original order date, so the credit exactly reverses the original charge.
- A refunded unit reverses its goods charge, because the stock came back and can be sold again.
- The credit uses the cost that was effective on the original order, never the cost effective today.
- That makes the reversal exact, so a price change between sale and return creates no phantom gain or loss.
- The credit lands on the refund day, because that is when the return happened.
The problem the rule solves
Imagine a unit sold in March at a cost of €18,50 and returned in June, by which time the same unit costs €21,00 to buy. Credit the return at June's cost and the statement records a €2,50 gain that no trade produced: the store simply benefited from an accounting choice, and it would benefit again on every return whenever costs rise.
Credit it at March's cost and nothing invented happens. The March charge and the June credit are the same number, the goods return to stock, and the only real effects are the ones that actually occurred: the margin handed back, the parcel already spent, and the processing cost of the return.
Why effective dating makes it possible
This only works if the system can answer what a cost was on a past date, which is exactly what dated cost rules provide: each rule carries the day it took effect, the previous rule is closed the day before, and any order can be priced by the rule that governed its own day.
A store holding one cost per product cannot do this at all. It has no March, only a present, so every historical question is answered with today's number and every return is credited at whatever the cost happens to be now.
| Rule used | Credit booked | Effect |
|---|---|---|
| Cost on the order date | €18,50 | exact reversal, nothing invented |
| Cost on the refund date | €21,00 | €2,50 of phantom gain |
| Average cost | somewhere between | a small error, in either direction |
Which day the credit lands on
The refund day, alongside the refund itself. The sale happened in its own month and keeps its charge; the return happened in another and carries the credit, which is the same recognition date discipline that keeps a closed month closed.
What this does to a month
The rule keeps two months honest at once. The month of the sale carries the full cost of the goods it shipped, because it did ship them, and the month of the return carries the credit for what came back, valued at the same number. Neither month reaches into the other, and the cost of goods line stays a description of what was actually consumed.
The alternative quietly makes the cost of goods depend on when returns happen rather than on what was sold, which is the sort of coupling that produces a margin nobody can explain and everybody stops trusting.
The cases at the edges
- Goods that come back unsaleable get the credit and a write-down, because the stock returned but its value did not.
- A goodwill refund with no goods returned gets no credit at all; nothing came back to credit.
- A chargeback gets no credit either: the customer kept the product, so the cost stays spent.
- An exchange credits the returned unit and charges the replacement, which is two events rather than one.
- If the original cost was wrong, fix it with a dated correction rather than by adjusting the credit.
What a spreadsheet usually does instead
A hand-built statement almost always credits returns at whatever cost is currently in the sheet, because there is only one cost in the sheet. In a year of rising supplier prices that produces a small favourable error on every single return, which accumulates quietly and always in the direction that makes the store look better.
It is not a large number per unit and it is entirely systematic, which is the combination that survives review. Anyone reconciling a spreadsheet P&L against a bank account will find a persistent unexplained gap, and this is one of the two or three usual suspects.
Why this matters more as costs move
In a stable year the rule is almost invisible. In a year of supplier increases, freight swings and currency movement it is the difference between a return being a straightforward reversal and being a small arbitrary gain or loss booked at random. Landed costs move for reasons that have nothing to do with any individual sale, and a statement that lets those movements leak into returns is a statement that cannot be trusted about either.