The answer to Spot the plug: report B, the line called Balancing item.
Why B
Apply the test: could this line be computed from something that happened, without knowing the closing balance?
- "Claims paid" — yes, from the cash actually paid.
- "Interest accretion" — yes, from the opening balance and the rate.
- "Balancing item" — no. There is no event it corresponds to. It is 26.0 because 26.0 is what was needed; had any other line moved, it would have moved to match.
That is the structural definition of a plug, and it is why the name it is given never matters. B's plug is 26.0 against a 44.0 movement: 59% of the change is unexplained, and the report says nothing about it.
Why C's "Rounding" is not a plug
C carries a −0.2 line labelled Rounding, and it is legitimate — for a reason worth being precise about.
A rounding line is bounded by the presentation, not by the arithmetic. C shows five lines at one decimal; the most that rounding can contribute is a few hundredths per line. −0.2 is inside that. A "rounding" line of −26.0 on the same report would be a plug wearing a different hat.
So the test has two parts, and both matter:
- Is the line independently determined? A rounding difference is — it follows from the chosen precision.
- Is its magnitude consistent with its claimed cause? If a line's size can only be explained by "whatever was left over", the label is decoration.
Why A and D are clean
Every line in both corresponds to something that happened, and each is computable without reference to the closing balance. That is exactly what makes their reconciliation informative: because nothing was free to move, the fact that they tie is evidence.
The habit this is training
The reconciliation check is worth having only if it can fail. A model where some line always absorbs the difference has a check that cannot fail, and therefore tells you nothing — while looking, on every report, exactly like a model that is working.
Live and green
Press Run — the same battery as the exercise, with the answer in place. Every check ✓.