Investigating inventory shrinkage and count variances

Investigating inventory shrinkage means working out whether a difference between book and physical quantities is a genuine loss of goods — theft, evaporation, breakage, or process yield loss — or an error in the counting or the underlying data, before accepting management's proposed write-off. Investigation should be triggered by value-based thresholds rather than quantity, and the resulting write-off needs a specific authorization trail, not a single blanket journal entry passed at year end.

Genuine shrinkage versus an error masquerading as shrinkage

A count-to-book variance has exactly one of two underlying causes: stock that genuinely no longer exists in the quantity the books assume, or a record somewhere that is simply wrong. These need entirely different responses, and the investigation has to establish which one it is before any write-off is proposed.

The investigation sequence

1. 01Rule out data errors firstUnit of measure, posting cut-off, and duplicate or missing count sheets are the cheapest explanations to check and the most common cause of a variance that looks alarming but is not real. Clear these before assuming the loss is genuine.
2. 02Check the transaction history for the specific item around the count dateA gap in recorded movement right before or after the count date often explains an apparent shortage without any stock actually being missing.
3. 03Ask for the client's own explanation, in writing, item by itemNot a blanket explanation for the total variance. A specific explanation per significant item, or per category of similar items, that can be corroborated.
4. 04Corroborate the explanation against an independent sourceA damage report created at the time of the damage, not reconstructed afterward; a security incident log for a theft claim; an industry-typical yield-loss percentage for an evaporation or process-loss claim.
5. 05Look for a pattern across periodsA shrinkage rate rising steadily year over year at a specific location, with no corresponding change in process or product, is a different conversation from a one-off variance — it points toward a control weakness that needs addressing, not just an accounting entry.

Setting a materiality threshold for investigation

Investigating every single-unit variance across a large stock-keeping unit count is not a sensible use of time, and no audit does it. The threshold needs to be set deliberately, by value, and applied consistently rather than chosen to avoid uncomfortable items.

• Set a value threshold, not a quantity or percentage threshold alone — a thousand-unit variance on a low-value fastener and a five-unit variance on a finished product can carry the same rupee exposure.
• Lower the threshold, or investigate regardless of value, for items that are easily resold, cash-equivalent, or previously flagged as a control concern.
• Aggregate small variances across similar items before dismissing them individually — a pattern of small negative variances across an entire product line sums to something material even if no single line crosses the threshold.
• Document the threshold chosen and the basis for it in the working paper, so a reviewer can assess whether it was set appropriately rather than reverse-engineered from the results.

Write-off authorization

Once a variance is accepted as genuine shrinkage rather than a data error, the write-off itself needs a documented approval trail appropriate to its size — this is a control test as much as a substantive one.

1. 01Confirm the client has a stated approval limit for inventory write-offsWho can approve what value, and whether larger write-offs escalate to a more senior authority or the board — many mid-market clients have no written policy at all, which is itself worth noting.
2. 02Trace the write-off journal to its approvalA signed or system-logged approval at the appropriate level, dated before or at the time of the write-off, not backdated afterward to support an entry already passed.
3. 03Check for split write-offs designed to stay under an approval thresholdSeveral write-offs for the same cause and around the same date, each individually below the approval limit, is a pattern worth asking about directly.
4. 04Confirm the write-off is reflected consistently in both the financial books and the stock/quantity recordsA write-off passed only in value, with the quantity still showing in the perpetual stock system, leaves a reconciling item that will resurface at the next count.