Accounting

Five unresolved, none wrong

Bookkeeping is supposed to be precise. When the evidence runs out, the honest estimate is a better friend than the tidy number.

LPLP
Accounting
In this article

The short version

  • When the document that would settle a question does not exist, estimate, label the estimate and say what would replace it.
  • Proportional allocation gives a tidy, constant rate that proves nothing about the loan. Pick a basis that follows how the thing really behaves.
  • Leave open items open. Five unresolved items are better than one wrong one.
  • A visible, reasoned estimate is a professional judgment. A silent one is a small misrepresentation.

There is a popular image of accounting as the discipline of exactness: every cent traced, every balance tied, nothing left to judgment. Practitioners know better. A surprising share of the work is deciding what to do when the document that would settle a question does not exist, cannot be obtained or arrives too late. How an accountant behaves in that gap says more about the quality of the books than any number of perfectly reconciled months.

Consider a common case. A growing company borrows from a revenue-based lender and, for most of a year, books each monthly repayment wholly against the loan balance. No interest is recorded. When someone notices, the lender’s statements for the later months are available, and the interest in them can be booked precisely. For four earlier months they are not. What is known is the total: the difference between what was paid and what the principal fell by comes to roughly $17,000. What is unknown is how that total divides between February, March, April and May.

How do you split interest you cannot see?

The easiest method is to allocate the total in proportion to the loan balance each month. It is quick, it ties to the known figure and it produces a pleasingly consistent result: an implied interest rate that is exactly the same in every month. People take that consistency as confirmation. A constant rate from proportional allocation says something about the method and nothing about the loan. The lender’s actual statements for the later months, as it happens, show rates that are far from constant.

My basis was days elapsed between payments, on the view that interest accrues with time. An unscheduled wire counted as a payment on its own date, and a refund counted from its own date. The monthly figures came out uneven, less elegant and more defensible. If the loan terms say interest works differently, the basis changes, and that is one more reason to write the basis down. Both methods are estimates. Only one is built on the mechanics of the thing being estimated.

Why should the memo say it is an estimate?

The choice of method matters less than what is written beside it. The journal entries in this case carry memos stating that the figures are an estimate reconstructed from payment history and not taken from lender statements, describing the basis, and noting that they will be replaced if the statements turn up. A reader who opens the ledger in three years, an auditor or a buyer’s adviser, can see at once which numbers are evidenced and which are judged.

The same habit shows up elsewhere in my work. When four months of credit-card statements could not be obtained from an issuer, I reconciled the account across the gap to a balance derived from the surrounding statements, and the working paper says that this was done and why. When a run of equipment purchases had been posted to a fixed-asset account, I raised whether they should have been capitalized as a policy question for the client and did not correct it as an error. Until someone with authority decides, consistency of treatment governs. When a check could not be settled, I left it open and labeled it as open. My working rule: better five unresolved items than one wrong one.

Why is it so hard to leave items open?

The incentives point the other way. A reconciliation that closes at zero looks like competence. A working paper with open items looks unfinished. The accounting packages I use reinforce the bias by offering, at the end of every reconciliation, to post any remaining difference automatically, with no explanation required. Multiplied across thousands of small businesses, the result is a large volume of books that balance perfectly and cannot be relied on.

What works against that is mostly procedure:

  • Investigate before adjusting. A difference is a clue. Bisect the period, find the item, fix it where it lives. (The search problems in “The search that found nothing” are a good reason to do this carefully.)
  • Set a threshold for stopping. Above a stated amount, an unexplained difference halts the work and goes to someone with authority instead of into an adjusting entry.
  • Write memos that answer questions before they are asked. The amount, the period, what was tested, what was ruled out, the probable cause and what would resolve it.
  • Say what the working paper did not test. Limitations stated plainly protect the reader, and you.

Corrections go in as new entries and never as edits, so the error and its fix both stay visible. “The customer who wasn’t there” shows what that looks like over a full year.

What does a good estimate look like to the reader?

Owners, lenders, investors and tax authorities do not expect perfection. They expect that the people preparing the numbers know which ones are firm and which are soft, and will say so. An estimate that is disclosed, reasoned and replaceable is a professional judgment. The same estimate posted silently as if it were a fact is a small misrepresentation, however close to the truth it turns out to be.

The habit of mind is easy to state and hard to keep. When you do not know, say so, show your working, and leave the door open for the evidence to arrive. Books kept that way will sometimes look messier than their neighbors’. They will also be the ones still standing when somebody checks.

This is my own practice and not formal accounting guidance. Which treatment applies to a given estimate depends on the facts and the framework your books follow, so confirm it with your accountant or auditor.