Accounting Fundamentals: The Ground Floor of Record-to-Report

2026-10-10

These notes follow the section in the curriculum's own order. Part A covers the accounting equation and the different ways accounts are classified. Part B covers the core accounting concepts, from going concern to the rules for changing policies, estimates and errors. Part C covers the architecture of the financial statements and how they connect. The reference framework is Ind AS, with IFRS mentioned where it matters. The examples use small rupee amounts so the logic stays visible, and every journal entry is checked against the accounting equation, because that check is the quickest way I know to catch a wrong entry.

A1. The accounting equation

Everything a business owns has to have come from somewhere, and there are only two possible sources. Someone lent it to the business and expects to be repaid, which makes it a liability. Or the owners put it in, or the business earned it and kept it, which makes it equity. There is no third source, so the equation is not merely usually true. It holds after every single transaction, without exception.

Two transactions show it working. An owner invests ₹1,00,000 in cash, and the business then borrows ₹50,000 and uses it to buy a laptop.

Step 2 moved several things at once: cash came in from the lender, cash went out for the laptop, and a liability appeared. Both sides still balance, because every transaction has two sides, and double entry is simply the practice of recording both.

Once I began checking entries against the equation, they sorted themselves into two kinds. A pure balance sheet swap moves value between assets and liabilities, or between two assets, and leaves equity untouched. A loan-funded purchase, a collection from a customer and a payment to a vendor are all of this kind. A P&L-linked entry involves income or an expense, so equity moves as well: income increases retained earnings and expenses reduce them.

Rent of ₹5,000 shows the second kind clearly. The debit is always rent expense. Only the credit changes.

The equity effect is the same in all three. It comes from the expense, not from the credit side, which only decides which other part of the equation absorbs the entry. This is also the thread that joins the P&L to the balance sheet: an expense reaches the balance sheet through equity.

A2. Current and non-current classification

Both assets and liabilities are split by one test. Is the item expected to be realised or settled within twelve months of the reporting date, or within the normal operating cycle where that is longer? If so it is current, and otherwise it is non-current. (The operating cycle only matters for businesses such as real estate, where a normal cycle runs longer than a year.) For liabilities there is a second route to current: the entity does not have the right, at the reporting date, to defer settlement for at least twelve months.

A single loan therefore appears in two places on the balance sheet. A ₹1,00,000 loan repayable in four equal annual instalments shows ₹25,000 as current and ₹75,000 as non-current. A year later the next ₹25,000 moves across. For a ₹3,00,000 loan over five years the split is ₹60,000 and ₹2,40,000, and the same reclassification repeats at every close.

The more interesting case is that classification follows the right to defer, not the repayment schedule on paper. Suppose a loan carries a covenant that debt to equity must not exceed 2, and at the reporting date the ratio is 2.5. The lender can now demand repayment, so the whole loan, not just the instalment due within the year, becomes current. The only way out is a grace period of at least twelve months that the lender agreed to by the reporting date. A waiver granted in April does not rescue a 31 March balance sheet. This is why covenant compliance is a real step in the close and not a formality.

A3. Capital and revenue expenditure

The test is how long the benefit lasts. If it extends beyond the current accounting period, the cost is capital expenditure: it becomes an asset and is depreciated over time. If the benefit is used up within the period, it is revenue expenditure and goes straight to the P&L. A machine bought for ₹5,00,000 is capital, while the electricity it consumes, say ₹8,000 a month, is revenue.

The judgement lives at the boundary. A repair that restores something to its existing condition is revenue expenditure. An upgrade that extends the asset's life or raises its capacity is capital expenditure. (Under Ind AS 16, later expenditure is capitalised when future economic benefits are probable and the cost can be measured reliably, and day-to-day servicing is expensed.)

The antivirus renewal covers a year, but it buys a service and not a lasting asset, so it is revenue in nature. Any unexpired portion at the reporting date sits as a prepayment, which A5 covers.

A4. Operating, investing and financing activities

These three buckets are the backbone of the cash flow statement, and they connect directly to the previous section. Capital expenditure appears as an investing cash flow and revenue expenditure as an operating one, while borrowing and raising equity are financing.

Five events, classified:

The last row is the important one. The cash flow statement sees only actual cash movement. An accrual moves the equation (a liability rises and equity falls) but leaves the cash flow statement untouched until the payment is made. One point of choice: under Ind AS 7 a non-financial entity may classify interest paid as operating or financing, interest and dividends received as operating or investing, and dividends paid as financing or operating, provided the choice is applied consistently. I stay with principal flows here.

A5. Accrual and cash accounting

Cash accounting records a transaction when cash moves. Accrual accounting records it when the economic event happens, meaning when goods or services are delivered or the benefit is received, and treats the timing of cash as a separate matter. Company books are kept on the accrual basis (Ind AS 1 requires it for every statement except the cash flow statement), and that is why accruals, prepayments and depreciation exist at all. They stop the timing of cash from distorting the P&L.

Goods worth ₹1,00,000 delivered in March, invoiced at once and paid in April, are March revenue on the accrual basis (debit receivable, credit revenue) and April revenue on the cash basis. March's cash flow statement shows nothing for them either way. The same logic applies to a payment. Insurance of ₹12,000 covering January to December is paid in January:

On payment the entry is a balance sheet swap, debit prepaid insurance and credit bank. Each month then moves ₹1,000 from prepaid insurance into expense, which touches equity. The cash left once, but the expense spreads across twelve months. The purpose of accrual accounting is that each month's P&L tells the truth about that month and not about when a payment happened to be made.

A6. Real, nominal and personal accounts

The older classification is still taught, and still useful for working out which side of an entry to debit. It has three kinds of account, each with its own rule.

Textbooks are not unanimous about the bank account. Some treat it as personal because the bank is an artificial person, and others as real because the balance is the company's own asset. The debit and credit come out the same either way, and the modern classification in A7 removes the question, since a bank balance is simply an asset.

Mapped to the equation, real accounts are mostly assets. Personal accounts split three ways, since a vendor payable is a liability, capital is equity and a customer receivable is an asset. Nominal accounts are the income and expense accounts, which are closed into retained earnings at the end of the period. That last step is why expenses reduce equity, as A1 showed.

A7. The modern classification

Enterprise systems and financial statements use five categories instead, and this is the working classification for the rest of the curriculum.

Assets and expenses sit on the debit side, and liabilities, equity and income on the credit side. The category matters more than the balance, because a debit balance could be an asset or an expense, and a credit balance could be a liability, equity or income. The category decides whether an account lands on the balance sheet or in the P&L. Applied to a few accounts, trade receivables are an asset, accrued rent payable is a liability, share capital is equity, sales revenue is income, depreciation is an expense, and prepaid insurance is a current asset.

The modern scheme improves on the traditional one because each account has exactly one home, whereas "personal account" could mean a liability, equity or an asset. One quick test separates assets from liabilities: is the money coming to the company or going out from it? A loan puts cash into the business, so cash rises, but the money must be returned, so the loan itself is a liability. That is the two-sided logic of A1 again.

With income and expenses included, the equation takes its fuller form:

At the end of a period income and expenses are closed into retained earnings, which is why the simple form always holds on the balance sheet.

The concepts in this part are rules of judgement rather than mechanics, and I find it easiest to read each one as the answer to a specific problem. Going concern settles the basis on which everything is measured. Accrual and matching settle timing. Prudence, consistency and materiality discipline the judgements involved. Substance over form settles what a transaction really is, historical cost and fair value settle how it is measured, and revenue recognition applies all of this to the top line. The last topic covers what happens when a policy, an estimate or a previously reported figure turns out to need changing.

B1. Going concern

Financial statements are prepared on the assumption that the business will continue for the foreseeable future, unless management intends to liquidate it or stop trading, or has no realistic alternative but to do so (Ind AS 1, paragraphs 25 and 26). "Foreseeable" has a floor: management must look at least twelve months beyond the reporting date. The assumption is not that the company will exist forever, and bankruptcy is not the only way to lose it. A parent group deciding to wind up a subsidiary is enough.

The assumption quietly holds up much of Part A. Prepaid insurance of ₹11,000 is an asset only because the company is expected to be around to use the remaining months. Depreciating a machine over its life assumes the machine will keep being used. The split between current and non-current assumes loans will run to their schedules.

When the assumption fails, the balance sheet changes in three ways at once. Take a company with a machine carried at ₹3,00,000, inventory at ₹2,00,000, cash of ₹1,00,000 and a ₹2,00,000 long-term loan. Management decides to liquidate, and expects the machine to fetch ₹1,00,000 and the inventory ₹80,000.

Assets fall by ₹3,20,000 and equity falls by exactly the same amount, because the loan keeps its amount and changes only its classification. The equation makes the owners absorb the whole loss, which is why equity is called the residual claim. Ind AS 1 does not prescribe a liquidation basis. It requires disclosure of the fact that the statements are not on a going concern basis, the basis used, and the reason. The treatment above, assets at expected realisable value and liabilities treated as payable on demand, is a typical approach and ignores liquidation costs for simplicity.

Management's position falls into one of three situations.

In a month-end close the warning signs appear in routine work: vendor ageing with a growing tail beyond 90 days, bank balances that sit low or in overdraft, continuing losses with interest accruing but unpaid, and covenant tests that sit close to the line.

B2. Accrual and matching

Accrual, as A5 showed, decides when a transaction is recorded. Matching goes a step further and decides which period a cost belongs to: the one in which it helped to earn the revenue. Consider 100 units bought for ₹100 each, ₹10,000 in cash, of which 60 are sold at ₹150.

₹10,000 of cash went out, but only ₹6,000 reaches the P&L. The other ₹4,000 waits on the balance sheet as inventory and becomes an expense when those units are sold. Expensing the full ₹10,000 would show a loss of ₹1,000 in a period when the business actually earned ₹3,000 on what it sold. The purchase was a balance sheet swap, cash into inventory, and equity moved only on the sale.

Selling costs follow the same logic. If a salesperson earns a 5% commission on March sales of ₹9,600, the cost belongs to March even though it is paid in April: debit commission expense ₹480, credit accrued commission payable ₹480. Liabilities rise by ₹480, equity falls by ₹480, and the cash flow statement shows nothing until April. Booked on payment instead, March's profit would be overstated by ₹480, and April would carry a cost with no revenue behind it.

Costs match revenue in three different ways.

The three questions I use, in order, are these. Did the cost arise because of a specific sale? If not, does its benefit last beyond this period? If neither, it is immediate. A three-year maintenance contract paid upfront for ₹36,000 is debited to prepaid maintenance and released at ₹1,000 a month (₹36,000 over 36 months). Right after payment, ₹12,000 of it is current and ₹24,000 non-current, and the whole ₹36,000 is an operating cash outflow in the month of payment.

Matching has a boundary. It adjusts timing, but it does not let a cost be parked on the balance sheet because it might help future revenue. A cost becomes an asset only when it meets the definition of one, and the IASB's 2018 Conceptual Framework is explicit that matching is not an objective in itself.

B3. Prudence

Prudence is the exercise of caution when judgements are made under uncertainty. The traditional wording, never anticipate a profit but provide for probable losses, still describes how several standards behave. The IASB's 2018 Conceptual Framework is careful, though, to say that prudence does not require systematic asymmetry, and that it is not a licence to understate assets and income or overstate liabilities and expenses on purpose.

Receivables show the loss side. Customers owe ₹3,00,000 and the expected credit loss is estimated at 10%.

Receivables now stand at ₹2,70,000 net. Assets and equity both fall by ₹30,000, and cash is untouched because nothing has been collected or lost yet. Although the account is often called a provision, this is a loss allowance under Ind AS 109 and not a provision under Ind AS 37.

Inventory shows both sides of the asymmetry. Ind AS 2 measures inventory at the lower of cost and net realisable value (NRV), which is the estimated selling price less the estimated costs of completion and of making the sale. Take inventory that cost ₹10,000.

A fall is recognised at once and a rise waits for the sale. If a write-down was made earlier and NRV later recovers, it is reversed, but only up to the original cost. Provisions follow the same pattern: a provision is recognised when an outflow is probable, whereas a contingent asset is recognised only when the inflow is virtually certain (Ind AS 37).

One working rule helps in choosing the standard. If the question is the value of an asset already on the balance sheet, the asset's own standard governs: Ind AS 2 for inventory, Ind AS 109 for receivables. If the question is whether to create a new liability, it is usually Ind AS 37.

B4. Consistency

Once a policy is chosen it is applied the same way to similar transactions and from one period to the next, so that statements can be compared. A policy may change only when an Ind AS requires it, or when the new policy gives reliable and more relevant information (Ind AS 8, paragraph 14). Presentation follows the same discipline (Ind AS 1, paragraph 45).

The stakes are visible in a small example. A company buys 100 units on 1 January at ₹10 and another 100 on 1 February at ₹14, and later sells 100 units at ₹20 each.

Same goods, same sale, and profit differs by ₹200 purely because of the method. If a company could switch whenever it suited, comparisons across years would mean nothing and profit could be steered. (LIFO is not permitted under Ind AS 2.) A company in a period of rising prices that moves from weighted average to FIFO only because profit is below target meets neither condition for a change, and the timing and direction of the switch point to motive. The same discipline applies to small things, such as applying one capitalisation threshold to every asset. A justified change is applied retrospectively, which B10 explains.

B5. Materiality

Information is material if omitting, misstating or obscuring it could reasonably be expected to influence the decisions that the primary users of the financial statements make on the basis of them (Ind AS 1, paragraph 7). Two things follow. The test is about users' decisions and not simply the size of an amount, and it is specific to the entity: ₹20,000 is trivial for a company earning ₹10,00,000 and significant for one earning ₹50,000.

Auditors often use a benchmark such as 5% of profit before tax as a starting point. That is practice and not part of Ind AS, and it is only the quantitative lens. The qualitative lens asks whether a small amount flips something that matters: a profit into a loss, a covenant into a breach, a bonus trigger, a related party transaction, a fraud.

Take a company with profit before tax of ₹20,00,000 and an unrecorded expense of ₹30,000. That is 1.5% of profit, well below a 5% threshold of ₹1,00,000. Now add a loan covenant that sets a floor for profit before tax.

Same error, same company, same quantitative size. Only the covenant differs, and the answer changes. Two further limits apply. Small errors aggregate: four such errors total ₹1,20,000, which is 6% of profit. And even an immaterial error must not be left uncorrected on purpose to achieve a particular presentation (Ind AS 8, paragraph 41). In the first row the headroom is only ₹20,000, which is why at a close I look at the total of all unadjusted differences and not at each one alone. Materiality also supports sensible housekeeping: a ₹3,000 stapler is expensed and not capitalised, under a threshold applied uniformly to all assets.

B6. Substance over form

A transaction is recorded according to its economic reality and not the name on the paper. The IASB's 2018 Conceptual Framework treats this as part of faithful representation: showing a legal form that differs from the economic substance cannot faithfully represent anything. The idea is easiest to see in a personal setting. If I "sell" my bike to a friend for ₹12,000 but am bound to buy it back in three months for ₹12,300, I have not sold anything. I have borrowed ₹12,000 and agreed to pay ₹300 for the privilege.

Companies do the same thing, and there the label matters because it changes reported profit and reported debt. A company holds inventory that cost ₹10,00,000. On 31 March it "sells" the stock to a bank for ₹12,00,000 and commits to repurchase it on 30 June for ₹12,30,000.

The loan treatment is the correct one. The repurchase is certain and the price is higher, so the bank is effectively earning interest (₹30,000 over three months) and the company never gave up the risks and rewards of the goods. Ind AS 115 says so directly (paragraphs B64 to B76): an obligation to repurchase at an amount equal to or above the original selling price makes the arrangement a financing, and a repurchase below the selling price makes it a lease. Booking the sale would lift profit by ₹2,00,000 and take ₹12,00,000 of debt off the balance sheet, which is exactly what window dressing looks like. The loan entry is a pure balance sheet swap: cash up, borrowings up, equity untouched, the same pattern as the laptop in A1.

By the same test, goods sold for ₹6,00,000 with a compulsory buyback at ₹6,30,000 six months later are a loan, while an identical sale with no buyback obligation, only a standard warranty, is a sale.

The same logic appears elsewhere. Redeemable preference shares are legally share capital, but if they must be redeemed on a fixed date they are a liability under Ind AS 32. A machine "rented" for most of its life is in substance an asset financed by a loan, which is the idea behind Ind AS 116 (Section 16 of the curriculum).

Three questions help with any doubtful transaction. Who now controls the asset and carries its risks and rewards? Must the company pay the money back, whatever the document calls it? With the labels removed, is this a sale, a loan or a lease? Substance is established from the contract terms and objective evidence, such as buyback clauses and side letters. It is not a licence to reinterpret transactions freely. In RTR it tends to surface as large sales near year end followed by returns, or by repurchase from the same party, in the next period.

B7. Historical cost

Assets are recorded at what was paid for them and are not restated as market prices move. If I bought a plot in 2015 for ₹10 lakh and it would fetch ₹50 lakh today, the books still say ₹10 lakh, because the registry and the payment prove the first number while the second is only an estimate. Three reasons stand behind the rule. Cost is verifiable, since there is an invoice an auditor can inspect. It does not depend on anyone's opinion, whereas two valuers will give two numbers. And an unrealised gain has not been earned yet. The inventory example in B3, where goods with an NRV above cost stay at cost, is the same rule at work.

"Cost" means everything needed to bring the asset to the location and condition in which it can be used. A machine with an invoice price of ₹5,00,000, freight of ₹20,000 and installation of ₹30,000 costs ₹5,50,000 (ignoring recoverable GST). Depreciation is calculated from that cost, so a book value such as ₹3,00,000 is cost less accumulated depreciation and says nothing about market value. Land is the clearest case: bought for ₹8,00,000 and worth ₹30,00,000 today, it stays at ₹8,00,000 under the cost model, and it is not depreciated.

The rule has an obvious weakness. An old cost does not describe today's economic reality, and ₹10 lakh in 2015 is not ₹10 lakh now. That is why Ind AS 16 lets a company choose a revaluation model for a class of property, plant and equipment, and why financial instruments are often measured at fair value. Cost is the base, and those are the exceptions.

B8. Fair value

Fair value is the price that would be received to sell an asset, or paid to transfer a liability, in an orderly transaction between market participants at the measurement date (Ind AS 113). It answers a different question from cost. Cost asks what was paid. Fair value asks what the asset would fetch today, in a normal sale and not a fire sale. Cost is reliable but can be stale, and fair value is current but can rest on estimates.

A company buys listed shares for ₹5,00,000, held for trading, so changes in value go to the P&L.

The year 1 entry debits fair value loss (an expense) and credits investment in shares (an asset) for ₹40,000, and in year 2 the entry runs the other way for ₹30,000. Assets and equity move together, and cash is untouched because nothing has been sold. The point that is easy to miss is that each year's P&L effect is the closing fair value less the previous year-end carrying amount, not less the original cost. Measuring year 2 against the purchase price would show a loss of ₹10,000, but that figure is the two-year total, a loss of ₹40,000 followed by a gain of ₹30,000, and year 1 has already reported its share.

This sits in some tension with prudence. A gain is recognised before the shares are sold, which B3 would not allow for inventory. The difference is how objective the number is. A listed share has a public price every day, so the figure is observed and not argued over (the top of the fair value hierarchy in Ind AS 113, which Section 38 of the curriculum covers). Where the price is only an estimate, the quality of the estimate becomes the issue. Whether a fair value change goes to the P&L or to other comprehensive income depends on how the asset is classified, which comes up again in the sections on financial instruments.

B9. Revenue recognition

Revenue is recorded when the company has done what it promised, meaning the customer has obtained control of the goods or services. An order, an invoice and a payment are each tempting triggers, and none of them is the trigger. Ind AS 115, converged with IFRS 15, turns this into a five-step model in which the output of each step is the input of the next.

One running example carries through the steps. On 1 April ABC Ltd. signs a single contract with a solvent customer for 10 laptops and three years of support, for ₹4,50,000 plus 18% GST. Sold separately, the laptops would cost ₹4,50,000 and the support ₹50,000.

Step 1: Identify the contract

A contract is an agreement that creates enforceable rights and obligations, and it can be written, oral or implied by business practice. The model applies only when all five conditions hold (paragraph 9): the parties have approved the contract and are committed to it, each party's rights are identifiable, the payment terms are identifiable, the contract has commercial substance, and it is probable that the company will collect the consideration it is entitled to. "Probable" here means more likely than not.

Two of these conditions do most of the work in practice, and both echo earlier concepts. Commercial substance is substance over form: two companies selling the same goods back and forth to show revenue have changed nobody's cash flows. Collectability is prudence written as a rule: if the customer is insolvent, goods may ship but revenue does not follow. A phone order confirmed on WhatsApp, with agreed quantity and rate and 30-day terms, from a customer who has paid on time for years, is a contract. A signed and stamped 20-page agreement with a customer whose bank accounts have been attached is not, because collectability fails. Any cash received in that case is a liability, with a debit to bank and a credit to a liability, and equity is untouched. It stays a liability until the conditions are later met, or until the work is finished and the cash is non-refundable.

Two refinements belong here. Contracts entered into at or near the same time with the same customer are combined if they were negotiated as a package, if the price of one depends on the other, or if together they form a single performance obligation (paragraph 17). Take 10 laptops for ₹4,50,000 and three years of support for ₹10,000, signed on the same day, where support normally costs ₹50,000 and is cheap only because the customer is buying the laptops. Kept apart, ₹4,50,000 is recognised on delivery and ₹10,000 over three years. Combined, the discount is shared across both, ₹4,14,000 to the laptops and ₹46,000 to support, so ₹36,000 of revenue moves out of the delivery date and into the support period where it belongs.

The second refinement is contract modification (paragraphs 18 to 21). If a change adds distinct goods at stand-alone prices, it is a separate contract. If not, and what remains is distinct from what has been delivered, the old contract is treated as ended and a new one begins, with the leftover consideration plus the new price. Take 100 units at ₹100, with 60 delivered and ₹6,000 recognised, and 40 more units added at ₹70 each. The remaining consideration is 40 × ₹100 plus 40 × ₹70, which is ₹6,800 over 80 units, or ₹85 a unit, and the ₹6,000 already recognised is not touched. If the remaining work is instead part of a single, partly completed obligation, the change is a cumulative catch-up. A warehouse contracted at ₹10,00,000 with expected cost of ₹8,00,000 and ₹4,00,000 spent is 50% complete, so ₹5,00,000 is recognised. A design change then lifts the price to ₹12,00,000 and the expected cost to ₹10,00,000, which makes progress 40% and cumulative revenue ₹4,80,000, so ₹20,000 of revenue is reversed. A catch-up can go down as well as up.

Step 2: Identify the performance obligations

A performance obligation is a promise to transfer a distinct good or service, or a series of them. It matters because revenue timing follows the obligation and not the contract. The first task is to list every promise, explicit or implied. A promise can be implied by a company's customary practice or published policy, for instance if it always installs free of charge. Internal set-up work, such as creating the contract in the system, is not a promise to the customer.

Each promise is then tested for being distinct, and both tests must pass (paragraph 27): the customer can benefit from it on its own or with readily available resources, and it is separately identifiable from the other promises in the contract. A promise fails the second test, and merges into a single obligation, when the company integrates items into a combined output, when one item significantly modifies another, or when the items are highly interdependent. A machine plus routine installation that any technician could do is two obligations, whereas installation that heavily customises the machine makes one. Bricks, labour and design for a building are one integrated obligation. And a service delivered in the same way every period, such as 36 months of support or 12 months of cleaning, is a single obligation, a series, and not one obligation per month.

Three promises are easy to overlook. A warranty that only assures the product meets its specification is not an obligation of its own (it falls under Ind AS 37), but a warranty that gives extra service, or that the customer can buy separately, is. A discount option that exceeds what customers normally receive, for example 40% off the next order when 10% is typical, is a material right and a separate obligation, and its revenue waits until the option is used or lapses. And when a company arranges a sale for someone else, it must decide whether it is principal or agent. If it controls the specified good or service before the customer receives it, it is the principal and records the gross amount, the indicators being primary responsibility, inventory risk and freedom to set the price. If it only arranges the sale, it is an agent and records only its commission. A travel agency that bought seats in advance, bears the unsold risk and sets the price is a principal, whereas one that books for a fixed commission is an agent.

In the running example the laptops are distinct, since a customer can use them alone and the support does not modify them, and the support is distinct, since a laptop works without it and support is sold separately. That gives two obligations: the laptops, and a 36-month series of support.

Step 3: Determine the transaction price

The transaction price is the amount of consideration the company expects to be entitled to, excluding amounts collected for third parties. GST is collected on behalf of the government and is therefore not revenue. The other items are handled as follows.

Variable consideration is the part that needs judgement. The company estimates it using the expected value, a probability-weighted average that suits many similar contracts, or the most likely amount, which suits a binary outcome such as a bonus that is either earned or not. It then applies a constraint (paragraph 56): include only as much as it is highly probable will not later be reversed in a significant way. This is prudence written as a rule. Uncertainty is higher when the outcome depends on things outside the company's control, when it will take long to resolve, when the company has little experience, or when the range of outcomes is wide. The estimate is revisited at every reporting date, and any change flows through revenue.

Take a product sold at ₹100 a unit, where the price becomes ₹90 retrospectively if the customer buys more than 1,000 units in the year. The outcome is binary, so the most likely amount applies. In the first quarter the customer buys 200 units and is not expected to pass 1,000, so revenue is ₹20,000. In the second quarter the customer buys another 900 units, 1,100 in all, and the threshold will be crossed. Second quarter revenue is 900 × ₹90 = ₹81,000, less a catch-up of ₹2,000 (200 units × ₹10) on the first quarter's units, which gives ₹79,000. The total of ₹20,000 and ₹79,000 is ₹99,000, which equals 1,100 × ₹90.

Returns work the same way. If 100 units are sold at ₹1,000 and 10% are expected back, revenue is ₹90,000 and the other ₹10,000 is a refund liability. A financing component matters when payment is far from delivery. Goods delivered today and paid for in two years at ₹1,21,000, with a 10% discount rate, are worth ₹1,00,000 now. That is the revenue, and the remaining ₹21,000 is interest income, ₹10,000 in year 1 and ₹11,000 in year 2. If the gap between delivery and payment is a year or less, no adjustment is needed, so ordinary 30 to 60 day credit never triggers this.

In the running example the price is ₹4,50,000. The GST of ₹81,000 sits outside it, and payment is due in 30 days, so there is no financing component. The invoice totals ₹5,31,000.

Step 4: Allocate the transaction price

The total is one number, but each obligation has its own timing, so each needs its own amount. Allocation is in proportion to relative stand-alone selling prices (SSP). The best evidence of SSP is the price at which the company sells the item separately. If there is none, the company estimates it using an adjusted market assessment or expected cost plus a margin. A residual approach, which subtracts the other SSPs from the total, is allowed only when the price is highly variable or not yet established.

A discount is shared across all obligations in proportion, unless there is observable evidence that it belongs to only some of them. A variable amount is assigned wholly to one obligation only when it relates specifically to that obligation. If the transaction price changes later, the change is allocated on the same basis as at the start, and the SSPs are not refreshed.

The discount of ₹50,000 is shared, ₹45,000 against the laptops and ₹5,000 against support. Without allocation the laptops would carry the full ₹4,50,000 and the support nothing, which would bring ₹45,000 of revenue forward to the day of delivery even though it relates to three years of service.

Step 5: Recognise revenue

Revenue is recognised when an obligation is satisfied, which means the customer obtains control. Control is the ability to direct the use of the asset and obtain substantially all of its remaining benefits. The first question is whether this happens over time or at a point in time. It is over time if any one of three conditions holds (paragraph 35):

If none holds, revenue is recognised at a point in time. The indicators of control passing are a present right to payment, legal title, physical possession, transfer of risks and rewards, and customer acceptance, and they are weighed together and not run as a checklist. Dispatch and invoicing on their own are not triggers. For obligations satisfied over time, progress is measured by an output method (units delivered, milestones, time elapsed) or an input method (cost incurred, labour hours), chosen once for each obligation and applied consistently. If the outcome cannot yet be measured reasonably but the costs are expected to be recovered, revenue is recognised only up to the costs incurred.

Where revenue, billing and cash fall at different times, three balances appear on the balance sheet. A receivable is an unconditional right to payment, where only the passage of time is needed. A contract asset is revenue already recognised for which payment still depends on something other than time. A contract liability exists when the customer has paid, or payment is due, for work not yet done. Cash received in advance is the mirror image of prepaid insurance: there the company paid first and held an asset, and here the customer pays first and the company holds a liability, released into revenue as it performs.

Take the running example at the delivery of the laptops on 1 April, when control passes. Payment is due in 30 days and the contract is non-cancellable, so I treat the company as having an unconditional right to the whole invoice, including the support portion that has not yet been performed. (A stricter reading of the standard's own illustration of a non-cancellable contract books the receivable for an unperformed portion only when payment falls due. Billing systems usually post the full invoice on the invoice date, and the example follows that route.)

The equation holds: assets rise by ₹5,31,000, liabilities by ₹1,26,000 and equity by ₹4,05,000. The cost of the laptops goes to cost of goods sold on the same day, which is matching again. Support is delivered evenly and the customer consumes it as it happens, so it is recognised on elapsed time: ₹45,000 over 36 months is ₹1,250 a month, with a debit to the contract liability and a credit to support revenue. Of the ₹45,000, ₹15,000 is current and ₹30,000 non-current, the same split as in A2. (GST is shown on the whole invoice for simplicity, and the time-of-supply rules are ignored.)

In the close, revenue shows up in a few recurring places. Cut-off follows the transfer of control and not the dispatch date: goods dispatched on 30 March that reach the customer's warehouse on 3 April belong to March or April depending on the terms and any acceptance clause. The deferred revenue roll-forward is opening balance plus billings less revenue recognised, which equals the closing balance. The line between a receivable and a contract asset is whether payment depends only on time. And in bundled deals, the revenue journals should tie to the allocation file. Section 13 of the curriculum returns to revenue from the close side, with billing to revenue reconciliation, unbilled revenue, the deferred revenue roll-forward and disclosure support. Contract costs, licences, bill-and-hold and consignment arrangements also belong to Ind AS 115 and are left out of this note.

B10. Accounting policies, estimates and prior-period errors

Ind AS 8 deals with three things that sound alike and are treated very differently. A policy is the rule or basis a company adopts. An estimate is a number needed to run the policy that cannot be known for certain. A prior-period error is a mistake in earlier statements, where reliable information was available and was misused or ignored. A household version helps: a fridge bought for ₹50,000. The policy is to record assets at cost. The estimate is that it will last ten years, so ₹5,000 a year is written off. If it later looks like lasting six years, the arithmetic changes going forward and the old years stay as they were. If instead the register shows ₹5,000 where the bill said ₹50,000, the old numbers themselves were wrong and must be corrected.

To decide which one is in front of me, I ask in order. Was reliable information available when the earlier statements were approved, and was it misused or ignored? If so it is an error. If not, did the measurement basis or principle change, as in FIFO to weighted average? If so it is a policy change. Otherwise, did new information or developments change a number? Then it is an estimate change. Where it is hard to tell a policy change from an estimate change, the standard says to treat it as an estimate (paragraph 35).

Accounting policies

Policies are the principles and practices a company applies in preparing its statements, and they decide what is recorded and how. An estimate then supplies the number: the policy says receivables carry an allowance for expected credit losses, and the estimate says 3%. Where an Ind AS covers a transaction, that standard is followed, and choices exist only where the standard itself offers them, such as FIFO or weighted average, the cost or revaluation model for property, plant and equipment, or where interest paid is shown in the cash flow statement. Where no Ind AS applies, management uses judgement to develop a policy that is relevant and reliable, looking first to similar standards and then to the Conceptual Framework.

A policy may change only when an Ind AS requires it, or when the new policy gives reliable and more relevant information (paragraph 14). Profit targets and convenience do not qualify, which is why the switch to FIFO in B4 fails. Applying a policy to genuinely different transactions, or to transactions that did not occur before, is not a change at all.

The treatment is retrospective, as though the new policy had always been used. Comparatives are restated, and if the effect reaches back before the comparative year, the cumulative amount goes to the opening retained earnings of that year. If the effect on the opening balance sheet is material, a third balance sheet, at the start of the comparative year, is presented (Ind AS 1, paragraph 40A). The catch-up never goes through the current year's P&L. Disclosure covers what changed, why, and the effect on each line item and on earnings per share. If the effect cannot practicably be determined, the policy is applied from the earliest date possible.

Take a company that started in year 1 and changes from FIFO to weighted average for a valid reason, with these closing inventories.

Year 1's comparative profit is restated down by ₹20,000, and year 2 absorbs only the further ₹5,000. Retained earnings fall by ₹25,000 in total, which matches the inventory difference (tax ignored). Pushing the whole ₹25,000 through year 2 would understate that year's profit by ₹20,000 and leave year 1 overstated.

Accounting estimates

Policies often call for numbers that cannot be observed directly, and those are estimates. The earlier examples are all here: the credit loss allowance, NRV, fair value, depreciation lives and residual values, and warranty provisions. Revising one is not an error. New information, new developments or more experience make revision necessary, and a change in an input or a measurement technique is an estimate change unless it corrects a past error.

The treatment is prospective. The effect goes into the P&L of the period of change, and of future periods if they are affected. Earlier years are not restated. If the change alters the carrying amount of an asset or liability, that amount is adjusted in the period of change (paragraphs 36 and 37).

A machine costing ₹10,00,000 is depreciated over 10 years on the straight-line method with no residual value, which is ₹1,00,000 a year. After three years its carrying amount is ₹7,00,000. At the year-end review, new technology cuts the remaining life from 7 years to 4. Years 1 to 3 stay at ₹1,00,000. From year 4 the charge is ₹7,00,000 ÷ 4 = ₹1,75,000 a year. Total depreciation over the life is 3 × ₹1,00,000 + 4 × ₹1,75,000 = ₹10,00,000, never more than cost. Each year's charge reduces both assets and equity and has no cash effect.

Prior-period errors

A prior-period error is an omission or misstatement in earlier statements that came from failing to use, or misusing, reliable information that was available when those statements were approved and could reasonably have been obtained (paragraph 5). It includes arithmetic mistakes, misapplied policies, oversights and misreadings of facts, and fraud.

The test that separates an error from an estimate change is whether the information was available at the time. Take a customer who defaults in year 2, when year 1 carried a 3% allowance. If the customer's condition deteriorated after year 1, that is new information and an estimate change: year 2 absorbs the extra allowance and year 1 is left alone. If an insolvency notice was already in the file at year 1 and was ignored, that is an error, and year 1 is restated. Hindsight is not allowed. An earlier year is judged only on what could reasonably have been known then.

Materiality is the gate. A material error is corrected retrospectively: comparatives are restated, or if the error pre-dates the earliest comparative, the opening balances of that period are restated (paragraph 42). In practice an immaterial error is usually corrected in the current period, after checking that it has not become material to the current year, though leaving an immaterial error uncorrected on purpose to achieve a presentation is not acceptable (paragraph 41). The correction never passes through the current P&L.

An example: rent of ₹2,00,000 for year 1 was not recorded, though the invoice was on file. It was spotted in year 2 and is still unpaid. Year 1 reported profit was ₹10,00,000, and year 2 profit before the correction is ₹12,00,000.

The two-year total of ₹20,00,000 is the same, but the periods are wrong in the first row, which is exactly what matching exists to prevent. In year 2's books the correction debits opening retained earnings (equity) and credits accrued rent payable (a current liability) for ₹2,00,000, so liabilities rise and equity falls. The year 1 comparatives show both the rent expense and the liability. The disclosure states the nature of the error, the amount of the correction for each line item affected and for earnings per share, and the amount at the start of the earliest period presented.

Not every error changes profit. A finance cost of ₹5,00,000 wrongly shown within operating expenses is a misclassification and is restated in the comparatives as well. Profit before tax stays the same, but EBITDA and margins change, which can matter for covenants defined on EBITDA.

In a close, this topic appears in a few places. The year-end review of useful lives, residual values and depreciation methods is an estimate review. Expected credit loss and provision assumptions are refreshed at each close, and the documentation should show when the new information arrived. When an old mistake turns up in a close or an audit, the first question is whether it is an error or an estimate, because an error means restatement and a materiality check and not a catch-up in the current P&L. And a policy change needs a memo with the reason under paragraph 14, the effect on each line item and the approval.

The simplest way I have found to hold the statements together is to treat the balance sheet as a photograph and the others as film. The balance sheet is a picture of the business on one date. The profit and loss statement, the statement of changes in equity and the cash flow statement each cover a period, and between them they explain how the opening photograph turned into the closing one. The notes explain how the numbers were made. These are not five separate documents but one system, and once the links are clear, each statement becomes a check on the others.

Under Ind AS 1 the complete set is as follows.

Each statement carries at least the previous period's comparatives, and when a policy change or an error correction has a material effect, a third balance sheet at the start of the comparative period is added (B10). The concepts of Part B reappear here as presentation rules: going concern, the accrual basis (except for the cash flow statement), materiality and aggregation, no offsetting, comparatives, and consistency of presentation.

In practice none of these statements is built by hand. Every account in the adjusted trial balance is mapped to a line of the statements, a design decision that Section 2 of the curriculum covers and that SAP implements as a financial statement version, and the classification in A7 does the sorting. Assets, liabilities and equity go to the balance sheet, income and expenses go to the P&L, and the net result of the P&L closes into retained earnings at year end. I take the statements in the order they are built: trial balance, P&L, changes in equity, balance sheet, cash flow and notes. Cash flow comes late because it is derived from the others. Section 28 of the curriculum returns to preparing the full set in detail.

A single example, ABC Ltd., runs through all of it, in ₹ lakh. The opening balances sit in the comparative column of the balance sheet. The year's events are these.

The statement of profit and loss

This statement asks how much was earned in the period, as income less expenses.

The format above, with a gross profit line, is the management view. The statutory Statement of Profit and Loss under Schedule III (Division II) classifies expenses by nature and has no gross profit line. Purchases of 120, a change in inventories of (10), operating expenses of 40, depreciation of 10 and finance cost of 3 total 163, and 200 less 163 gives the same profit before tax of 37.

Other comprehensive income (OCI) exists because some gains and losses are either unrealised or arise from remeasurement, and putting them through profit would make earnings swing from year to year. They bypass profit but still change equity, so they have to appear somewhere. Some never return to the P&L, such as remeasurements of gratuity obligations, gains on equity investments held at FVOCI and revaluation surplus on property, plant and equipment. Others are reclassified later, such as translation differences on foreign operations, cash flow hedge reserves and FVOCI debt instruments. Profit plus OCI is total comprehensive income. In the Indian Schedule III format OCI is a section of the Statement of Profit and Loss, and earnings per share (Ind AS 33) appears on its face. Tax on OCI is ignored in the example.

The statement of changes in equity

This statement tracks each component of equity from opening to closing. The rule is that the opening balance plus total comprehensive income plus transactions with owners equals the closing balance.

A separate statement is needed because equity moves for reasons other than profit: OCI, and dealings with owners such as share issues, dividends and buybacks. Retained earnings is therefore not the sum of all past profits, because dividends and transfers have come out of it. The adjustments from Ind AS 8 (B10) also appear here, in separate rows for restated opening balances, and not in the current year's profit. Schedule III presents equity in two parts, equity share capital and other equity. Share-based payment reserves and transfers between reserves are covered in Sections 24 and 25 of the curriculum.

The balance sheet

The balance sheet shows what the company has, what it owes and what belongs to the owners, on one date: assets equal equity plus liabilities.

Classification uses the twelve-month test and the right-to-defer rule from A2. The borrowings show it: each year ₹5 lakh moves from non-current into current, so non-current falls from 25 to 20 while the current portion stays at 5. Indian presentation lists non-current items first. Receivables, cash and investments sit within financial assets, and borrowings and payables within financial liabilities, which the table simplifies. Offsetting is not allowed unless a standard requires or permits it (Ind AS 1, paragraph 32). A vendor with a credit balance of ₹5,00,000 is not netted against another vendor's debit balance of ₹2,00,000 to show ₹3,00,000 payable. The debit balance is really an advance paid, so it belongs on the asset side, which is why vendor debit balances are reclassified at the accounts payable close. Finally, each line follows its own measurement basis: property at cost less depreciation, inventory at the lower of cost and NRV, receivables after the credit loss allowance, investments at fair value. Historical cost, fair value and prudence meet on a single page.

The statement of cash flows

Profit is measured on an accrual basis, but bills are paid in cash, which is why a profitable company can still run short of it. This statement shows where cash came from and where it went.

Profit was ₹27 lakh and yet cash fell by ₹6 lakh. Operating activities produced 30, all of it went into the new machine, and financing took out a net 6. The buckets are those of A4. Tax goes under operating unless it relates specifically to investing or financing, and interest and dividends follow the consistent policy choice described in A4, here with interest paid and dividends in financing.

The table uses the indirect method, which starts from profit before tax, reverses non-cash and non-operating items, and then adjusts for working capital. The direct method lists receipts and payments instead, here 190 − 110 − 40 = 40 as cash generated from operations. Operating cash flow is the same either way. Cash and cash equivalents are cash, bank balances and very liquid investments with a maturity of about three months or less when acquired, and a bank overdraft that forms part of cash management is included as a negative component.

Non-cash transactions do not appear as cash flows. The ₹2 lakh fair value gain is nowhere in the statement, and a machine funded directly by a lender paying the vendor would not be a cash flow either, so it is disclosed in the notes. Ind AS 7 (paragraphs 44A to 44E) also requires disclosure of the changes in liabilities arising from financing activities, and the usual way to give it is a reconciliation of the opening balance, the cash flows, the non-cash changes and the closing balance. For borrowings here that is 30 less 5, giving 25. The statement closes with the check that opening cash, plus the three activity totals, plus any exchange rate effect, equals closing cash: 15 + 30 − 30 − 6 = 9. Section 27 of the curriculum takes this statement much further.

The notes

The face of the statements carries numbers only. The notes show how the numbers were made, and they are an integral part of the statements, with each face line cross-referenced to its note (Ind AS 1, paragraph 113).

The roll-forwards in the notes are the same schedules that Section 55 of the curriculum builds later, and each face line must tie to the total of its note.

How the statements connect

The whole structure rests on one idea. The opening balance sheet plus the movements equals the closing balance sheet, and every movement is explained in the P&L, the statement of changes in equity or the cash flow statement. The ₹27 lakh of profit closes into retained earnings, appears in the statement of changes in equity and arrives at the equity on the balance sheet.

A movement with no home in any statement is an error or a misclassification. That is the question "how does it move" from the reconciliation framework, applied to the whole set. At every close, four tie-outs prove the set:

1. The balance sheet balances: total assets of 136 equal equity of 91 plus liabilities of 45.
2. The profit in the P&L equals the profit in the statement of changes in equity (27), and the closing equity there equals the balance sheet equity (91).
3. The closing cash in the cash flow statement equals the cash on the balance sheet (9). A mismatch means a mapping or classification error.
4. Every face line equals the total of its note, and every opening balance equals the prior year's closing balance.

Looking back over the section, what stands out is how little the equation ever had to bend. Every example, from a ₹1,00,000 investment to a ₹5,31,000 invoice, balanced on both sides, and nothing in Part B changed that. What the concepts changed was when a transaction is allowed to move one side of the equation, and by how much: whether an expense belongs to this month or the next, whether a gain waits for a sale, whether a loan stays a loan when the document calls it a sale. Each of those is a judgement, and judgements are where two sets of reported numbers come apart.

The next section, General Ledger Architecture, is where these classifications become something a system can enforce: the chart of accounts, the hierarchy above it and the mapping that carries each ledger balance into a line of these statements. If the mapping is wrong, none of the statements in Part C can be right, however careful the accounting behind them has been.