Skip to content

← Accounting

Checking the Sums

Chapter 11 · The Books

In this chapter 5 sections
  1. A sheet folded down the middle
  2. Yasmin’s first week, on one sheet
  3. What the totals catch
  4. What slips through
  5. Putting it right

Every entry in Yasmin’s books has now been written at least three times. It began on a bill or a receipt, went into the journal or one of the busy-day books, and was copied from there onto two pages of the ledger. A shop that sells a few hundred things a week does that copying thousands of times a year, by tired people at the end of long days, and every copy is a fresh chance to write 51 for 15, to miss a line, or to put an amount on the wrong side of a page.

So sooner or later every bookkeeper asks the same question: is any of this right? Reading the whole ledger again won’t answer it, because the person who made a slip once will often read straight past it a second time. What is needed is a test that doesn’t depend on anyone noticing anything, a sum that comes out one way when the books are in order and another way when they aren’t. Double entry provides one.

A sheet folded down the middle

Pacioli’s instructions for it were exact. When a merchant closed his books for the year, or moved to a new ledger because the old one was full, he was to take a sheet of paper and write on its left the debit totals of every account in the ledger and on its right every credit total. Each column was added up into what Pacioli called the grand total, the summa summarum. If the two grand totals came out equal, “your Ledger was very well kept and closed”. If one was bigger than the other, there was a mistake somewhere, “which mistake you will have to look for diligently with the industry and the intelligence God gave you”. A merchant who couldn’t keep his books well, he added, would “go on groping like a blind man and may meet great losses.”

In the summary of rules at the end of the treatise he put it more plainly still. A trial balance of the ledger is “a sheet of paper folded lengthwise in the middle”, the debtors on the left and the creditors on the right, and he was careful about the words: what must be equal is the total of the debits and the total of the credits, not the number of people on each side. The sheet still has that fold today, though it is now ruled into columns.

The reason it works is the one the whole book is built on: each entry adds exactly as much to the debit side of the books as to the credit side, so if the totals differ, somewhere an amount went onto one side without its partner.

In Nepali this list is सन्तुलन परीक्षण, a test of balance, and it is worth being careful about what it is not. It isn’t an account, and nothing is posted to it. It is only a summary of the accounts, drawn up on a particular date, to check the arithmetic of the books. The Cambridge syllabus says the same in its own words: a statement of the ledger’s balances on a particular date.

Yasmin’s first week, on one sheet

At the end of her first week Yasmin’s ledger has seven pages: cash, capital, the bank loan, purchases, Sujan, sales and rent, balanced on साउन 7 with each balance carried down. Her trial balance copies those balances onto one sheet, in the usual form: a serial number, the account’s name, its ledger page in the L.F. column, and its balance in the debit or the credit column.

S.N. Particulars L.F. Debit (Rs.) Credit (Rs.)
1 Cash a/c 1 355,000
2 Capital a/c 2 300,000
3 Bank loan a/c 3 200,000
4 Purchases a/c 4 230,000
5 Sujan’s a/c 5 30,000
6 Sales a/c 6 70,000
7 Rent a/c 7 15,000
Total 600,000 600,000

The L.F. column means anyone checking the sheet can go straight to the page a figure came from. The side each balance takes follows the rules from earlier chapters. The debit column holds what the shop has (cash), the goods it bought and what it spent (purchases, rent). The credit column holds where the money came from (her capital and the loan), what the shop owes (Sujan) and what it earned (sales). Every account sorts the same way: assets, expenses and losses on the debit side; capital, liabilities and income on the credit side.

That is the balance method, शेष विधि, and it is the one almost everyone uses, because each account takes one line. There is an older way, the total method, जम्मा विधि, which lists the full total of each side of every page instead. For Yasmin’s week it gives a longer list: the cash page’s Rs. 570,000 received and Rs. 215,000 paid out; Sujan’s Rs. 50,000 paid to him and Rs. 80,000 owed; and so on through the pages. Its grand totals come to Rs. 865,000 on each side, which is exactly the total of her journal for the week, since every amount journalized has been posted once on each side. That is the method Pacioli described. It is rarely used now, because it makes a long sheet and says less: a list of balances shows at a glance what each account comes to. There is a third, the mixed method, which sets down both the totals and the balance of every account, and it too is hardly used.

A real business at the end of a year has dozens of lines, sometimes hundreds, but the sheet is the same. Its totals have to agree; and if they do, the next step, drawing up the statements of profit and of what the business owns and owes, can start from this one sheet instead of from the whole ledger. That is one of the reasons for drawing it up, along with testing the arithmetic, helping to find errors, and giving an auditor a place to start.

What the totals catch

Here is the same sheet with slips put into the week’s books, one at a time.

Try it

The Books: Yasmin’s trial balance, with one slip

Each slip replays the week with that one mistake in it; the lines it changes are marked. A made-up shop and made-up figures.

The totals
agreediffer by Rs. 0

The week as it was written: every entry put the same amount on each side, so both columns come to Rs. 600,000.

Some slips upset the totals. They are called one-sided errors, एकपक्षीय गल्ती, because they touch only one side of the books.

Suppose the rent was posted to the rent page as Rs. 51,000 instead of Rs. 15,000, two figures swapped. The debit total rises to Rs. 636,000 while the credit total stays at Rs. 600,000. Bookkeepers have a trick for this kind of slip. The difference, Rs. 36,000, divides exactly by 9, and it always will when two figures have been swapped. Write a two-figure number with tens a and ones b; swapped, it becomes ten b plus a, and the gap between the two is nine times the gap between a and b. A difference that divides by 9 is no proof of a swap, but it is a good reason to look for one. The same is true of a figure with a nought added or lost: Rs. 15,000 written as Rs. 1,500 leaves a gap of Rs. 13,500, nine times Rs. 1,500.

Suppose Sujan’s Rs. 50,000 was posted to his page but never to the cash page. The cash page then shows Rs. 405,000, and the debits come to Rs. 650,000: the totals are apart by exactly the amount that was missed.

Suppose the week’s sales went onto the wrong side of the sales page, the debit side instead of the credit. Now one side has lost Rs. 70,000 and the other has gained it, so the totals, Rs. 670,000 against Rs. 530,000, are Rs. 140,000 apart: twice the amount. Halve a difference like that and you have the figure to hunt for, an entry of Rs. 70,000 somewhere.

Suppose a page was added up wrongly, the cash received totaled as Rs. 580,000 instead of Rs. 570,000. The cash balance comes out Rs. 10,000 too high, and so do the debits, at Rs. 610,000. And the trial balance can go wrong all by itself: if the rent was left off the sheet, the debits come to Rs. 585,000; if a column of the sheet was added wrongly, the totals disagree although the ledger is fine. These are all errors the trial balance reveals: a wrong amount on one side, one half of an entry not posted, a posting on the wrong side, a page or a subsidiary book added up wrongly, a balance carried down wrongly or left out of the sheet.

What slips through

The widget’s last four slips are the more useful thing to know about a trial balance, because the totals agree there and nothing on the sheet looks wrong. These are called two-sided errors, दुईपक्षीय गल्ती: each changes the debit side and the credit side by the same amount, so the test can’t see it. Accountants give them names, and the Cambridge syllabus asks for six of them.

The plainest is an error of omission. The rent was never written down at all: no journal entry, nothing on either page. Cash looks Rs. 15,000 too high and rent shows nothing, and the totals still agree at Rs. 600,000, because cash kept exactly what rent should have taken. A sale never entered, the kind of slip a shopkeeper most fears, behaves the same way. Its opposite is an error of duplication: Sujan’s bill entered twice, Rs. 80,000 debited to purchases twice and credited to him twice. Now the books say she bought goods worth Rs. 310,000 and owes him Rs. 110,000, and the totals agree at Rs. 680,000.

An error of commission puts the right amount on the right side of the wrong page, a page of the right kind. If the Rs. 50,000 paid to Sujan was posted to the bank loan page instead of his, both are accounts of what the shop owes, and both have been debited. The books say she owes the bank Rs. 150,000 and Sujan Rs. 80,000, when she owes the bank Rs. 200,000 and Sujan Rs. 30,000. Her total debts are right; only who they are owed to is wrong. Nothing on the sheet shows it, but Sujan will, the next time he sends a statement of account and it doesn’t match his page in her ledger.

An error of principle puts an amount in an account of the wrong kind. Suppose the week’s rent were debited to an account for the shop’s premises, as if it had bought something that would last, instead of to rent. The debit is still a debit, the totals still agree, but an expense now sits among the things the shop owns. The arithmetic is untouched; the meaning isn’t. The week’s profit was Rs. 3,000: sales of Rs. 70,000, less the Rs. 52,000 that the goods sold had cost, less the rent. Take away the rent, and the books show a profit of Rs. 18,000 that never happened.

An error of original entry goes wrong before the journal, or in it, and is then copied faithfully everywhere. If Sujan’s bill of Rs. 80,000 was written in the journal as Rs. 8,000, both pages get Rs. 8,000, and his page now says that she has paid him more than she owed: he seems to owe her Rs. 42,000, when she owes him Rs. 30,000. The totals agree at Rs. 570,000. A complete reversal uses the right accounts and the right amount but swaps the sides: the cash sale entered as a debit to sales and a credit to cash. Both totals drop by the same Rs. 70,000, to Rs. 530,000, and agree.

Last come compensating errors, two separate slips that happen to cancel. If the rent page was added up Rs. 1,000 too much and the sales page Rs. 1,000 too much too, one on each side, the totals agree at Rs. 601,000, both wrong.

So the agreement of the totals says one thing only, though it is a useful thing: that every amount in the books has a partner of the same size on the other side. Whether the amounts were the right ones, on the right pages, or written down at all, it can’t say. Other checks catch some of what it misses. A supplier’s statement of account catches a payment on the wrong page, the bank statement catches a cheque entered wrongly in the cash book, and counting the cash in the till catches a cash sale never written. Thinking about what an entry means catches the rent written as premises. The books can be wrong in ways that only looking at the world outside them can show.

Accounting software is built to refuse an entry whose two sides differ, and it posts both halves itself. That makes slips like the widget’s first three much rarer, and it changes nothing about the last four. A program can no more post a sale nobody entered than a clerk can.

Putting it right

An error, once found, is never scratched out or written over. The rule is firm: no केरमेट, no crossing out and rewriting. A wrong entry is put right by another entry, so that what was written and what was changed can both be read later. Pacioli asked for the same: a good bookkeeper should always say why such a difference arose, “so that the books are above suspicion.” Deliberate errors, छलकपट or fraud, are set apart from honest ones made by carelessness or not knowing; only the honest ones are this chapter’s business.

How the correction is made depends on when the slip is found. A one-sided error found while the books are still being written up, before the trial balance is drawn up, is put right on the ledger page itself, with a line on the side that needs it: the rent page posted as Rs. 51,000 gets a credit line of Rs. 36,000, saying why.

A two-sided error needs a journal entry, because two pages are wrong. There is a simple way to find it: write the entry that should have been made, write the entry that was made, and the correction is whatever turns the second into the first. Sujan’s payment should have been Dr Sujan, Cr Cash; it was made as Dr Bank loan, Cr Cash. The cash is right, so the correction is Sujan’s a/c Dr. Rs. 50,000, To Bank loan a/c Rs. 50,000. The bank loan page loses the debit it never should have had, and Sujan’s page gets it.

Then there is the awkward case: the trial balance doesn’t agree, the year’s statements are due, and the slip can’t be found yet. The difference goes into a temporary account, the suspense account, उज्राती खाता, which has other names too, among them भुलचुक खाता. It goes on whichever side of the trial balance is short, and the sheet then agrees. Suppose Yasmin’s sheet came out at Rs. 636,000 against Rs. 600,000 and she couldn’t see why. She would add a line to the credit column, Suspense a/c Rs. 36,000, and carry on.

When the slip turns up, the rent posted as Rs. 51,000, it is put right by a journal entry through the suspense account: Suspense a/c Dr. Rs. 36,000, To Rent a/c Rs. 36,000. The rent page comes down to Rs. 15,000, the suspense account is left with nothing in it, and it closes. The suspense account is only ever for one-sided errors; a two-sided error leaves the totals equal and so never puts anything there. While it stays open, a debit balance in it is shown with the assets on the balance sheet and a credit balance with the liabilities, which is a confession more than a figure: something in the books is wrong, by this much, and nobody yet knows where.

Some corrections change the profit, and the Cambridge syllabus makes a point of this. Had Yasmin worked out her week’s profit while the rent page still said Rs. 51,000, she would have shown a loss of Rs. 33,000 instead of a profit of Rs. 3,000. Correcting the rent raises the profit by Rs. 36,000. Correcting the error of principle the other way, moving the Rs. 15,000 from premises back to rent, takes the profit from Rs. 18,000 down to Rs. 3,000 and the premises off the list of what the shop owns. An error found is not only a figure fixed on a page; it can change what the business thought it earned and what it owns.

Pacioli’s way of finding a mistake, before the folded sheet, was to sit down with a helper. The helper took the journal, the merchant the ledger, and they went through every entry together: the helper called out the ledger pages where each entry should be, the merchant found it, and if it matched, they both marked it, two dots in the journal for an entry’s two halves and one in the ledger beside each. Nothing was to be marked “either by you without him, or by him without you.” Anything left unmarked at the end was a mistake to be corrected, with its own entry and its own date. Checking line by line, two people and two books, still works on a stubborn difference, though it can take an afternoon.

The trial balance, then, is a list of every balance in the ledger, debits in one column and credits in the other, and the two totals must agree because every entry put the same amount on each side. When they don’t, some amount went onto one side without its partner, and the size of the gap is a clue to what it was. When they do, the books are in balance, which isn’t the same as being right: a sale never entered, an entry made twice, an amount on the wrong page of the right kind, or written wrongly from the start, all pass the test. Errors are put right by entries that leave the old ones readable, through a suspense account if the sheet had to be made to agree before the slip was found.

Found a mistake? Please tell us.