Government Revenue: Where the Money Comes From

Government revenue comes from taxes on what is earned, taxes on what is bought and sold, and receipts that are not taxes at all. What actually arrives is a rate applied to a base, and the base moves with the economy whether or not anybody changes the rate. That is why the receipt side of a budget is an estimate in a way the spending side is not.

Two things already established make that answer readable, and both are covered separately. The first is that direct tax and indirect tax are two different kinds of tax, one taken on income and one taken on a transaction. The second is the single line every deficit in a government account is measured against: revenue receipts of Rs 3,00,000 crore in the Republic of Sankhya, the invented government used for teaching throughout these notes. Revenue receipts have so far stood as one number. The three sources underneath that number are what follows.

The reason it is worth opening is not tidiness. A total gives the size of something. A total gives no sign at all of how it will behave next year, and the receipt side of a government account behaves in a way that surprises people who have only ever seen the total. Two governments can collect the same Rs 3,00,000 crore and one of them can lose a fifth of it in a bad year while the other loses a twentieth, for reasons that were fixed long before the bad year arrived.

Where does a government's money actually come from?

From three places. Most people can name two, so all three are worth naming before any one of them is looked at.

The first is tax on what is earned. Somebody has an income, a share of it is taken, and it does not matter whether that person bought anything that year. The second is tax on what is bought and sold. Somebody buys a thing, a share of the price goes to the government, and it does not matter what that person earns. The third is every other rupee that comes in without any of it being a tax: the fee charged for a licence, the price of a service it provides, the dividendA share of a company's profit paid out to whoever holds its shares. A government holding shares in a company receives them like any other holder does. on shares it holds, the royaltyA payment made for the right to take or use something, most often a natural resource. It is usually set as an amount per unit taken rather than as a share of income. paid for the right to take something out of the ground, the interest on money it has lent.

The third source is the one readers forget, and forgetting it produces a picture of government in which the only thing a state does is take, when in fact a government also charges for things and earns on what it holds.

Here is the Sankhya receipt side, built from its three parts. Direct tax brings in Rs 1,20,000 crore. Indirect tax brings in Rs 1,50,000 crore. Non-tax receipts bring in Rs 30,000 crore. The three add to Rs 3,00,000 crore, exactly the revenue receipts figure the deficits are measured against. Nothing was added or removed on the way: the total is the three sources and there is no fourth line hiding in it.

What separates a tax receipt from a receipt that is not a tax?

One thing separates them, and it is not the size of the amount or the department that collects it. The separator is whether the payer got something named in exchange.

Think about two payments a household makes in the same week. The first is the tax inside the price of a bag of rice. Nobody at the shop offers the household anything in return for that particular rupee. The tax is taken because a rule says it is taken, and the household would have preferred to keep it. The second is the fee paid to renew a driving licence. The licence fee buys a specific thing: a licence, issued, with the household's name on it. Refuse to pay it and the household simply does not get the licence. Refuse to pay the tax on the rice and something quite different happens.

A tax is taken under compulsion with nothing specific given back, and a non-tax receipt is received in exchange for something named, and that single difference is why the two respond to completely different things.

The consequence is where the distinction earns its keep. A tax receipt moves with the size of the thing it sits on. If incomes in Sankhya grow, the direct tax collection grows without anybody doing anything. A non-tax receipt does not work that way. A non-tax receipt moves with what the government has decided to charge for, how much of that service people actually take, and what the government happens to hold and what those holdings earn. A government that has lent more money receives more interest. A government that holds shares in more companies receives more dividends. Lending and shareholding are decisions, not the size of an economy.

So the third line is not simply a small version of the first two. In a year when incomes and transactions both fall, the two tax lines fall together and the non-tax line may not move at all. The licences still get renewed and the loans still pay interest. The non-tax line is worth watching on its own before the whole receipt side is assumed to move together.

What is the difference between a rate and a collection?

A rate is not a collection. A collection is a rate applied to a base, and the difference between the two decides how much of a receipt side a government actually controls.

Take the vegetable seller outside a railway station. She sets her price: Rs 40/- a kilo. Rs 40/- a kilo is her rate, and the rate is entirely her decision. The rate is not her income. Her income for the day is Rs 40/- multiplied by however many kilos the passengers actually bought, and she does not decide that number. She can put the price up and watch the kilos fall. She can leave the price alone and find that a festival brought twice the crowd. Her price is a decision. Her takings are a decision multiplied by something that merely happened.

A government is in exactly that position, on a much larger scale. A government sets the rate. The base is set by everybody else.

A government sets the rate and does not set the base. Announcing a rate therefore says far less about collections than a reader expects it to.

Worked on the Sankhya indirect tax: in the year of the account, transactions carrying that tax came to Rs 12,00,000 crore, and the rate applied to them was 12.50 per cent. The two multiplied give Rs 1,50,000 crore, the published indirect tax collection exactly. The same decomposition works on the other two lines. Sankhya's assessed incomes came to Rs 8,00,000 crore at a rate of 15.00 per cent, giving Rs 1,20,000 crore of direct tax. Its chargeable activity and its holdings came to Rs 3,00,000 crore returning 10.00 per cent, giving Rs 30,000 crore of non-tax receipts.

Now notice what the government actually controls in that multiplication. The government controls the 12.50 and it controls the 15.00. The government does not control the Rs 12,00,000 crore of transactions. The transactions figure is the sum of what millions of people chose to buy. The Rs 8,00,000 crore of assessed incomes is out of its hands as well. Businesses earned that money and employers paid it. Both bases are outcomes, not settings.

Why do receipts move with the cycle when no rate has changed?

Because the base moves, and the base is made of the two things the economy is made of.

Hold the rate in the multiplication perfectly still and change nothing else. When output rises, more people are working and more businesses are profitable, so assessed incomes rise, so the direct tax base rises. At the same time more is being bought and sold, so the indirect tax base rises. Both bases have grown, both collections grow with them, and no minister decided anything. Run the same reasoning backwards for a downturnA stretch in which the economy grows more slowly than usual or shrinks outright. What causes one and how it plays out is covered separately. and both collections shrink, again with nobody deciding anything.

Receipts move with the cycle whether or not anybody decides anything. The movement is the automatic part of fiscal policy, seen from the receipt side rather than the spending side.

Work the Sankhya indirect tax again with the rate held at exactly 12.50 per cent. Suppose the base falls by a tenth, from Rs 12,00,000 crore of taxed transactions to Rs 10,80,000 crore. The collection is now 12.50 per cent of Rs 10,80,000 crore, or Rs 1,35,000 crore. The collection is Rs 15,000 crore lower than the published figure, and the fall is exactly a tenth of the line. The rate did not move, so the collection had to move in step with the base. Revenue receipts for the year fall from Rs 3,00,000 crore to Rs 2,85,000 crore.

Here is the part worth sitting with. The same fall can be reported in two ways and both are true. One report says collections came in Rs 15,000 crore short of the receipt side. The other says the rate was left exactly where it was. Only the first sounds like news, and the second sounds like nothing happened, and yet the second is the whole explanation of the first. The reassuring report is the one that should send a reader straight to the base.

Why is the receipt side of a budget the harder side to estimate?

Because the two sides of a budget are not the same kind of statement, even though they are printed in the same font on facing pages.

The spending side is largely a set of decisions. A government decides what to spend on salaries, what to spend on interest, what to spend on capital spendingSpending that leaves an asset standing behind it, such as a road or a building, rather than spending consumed within the same twelve months. The difference between the two kinds is covered separately.. The spending numbers can be wrong, and they routinely are, but they start life as intentions somebody actually formed. Sankhya's total spending of Rs 4,00,000 crore is that kind of number.

The receipt side is not a set of decisions but a set of forecasts. Sankhya's Rs 3,00,000 crore is the answer to a question nobody in the government can settle: how much will be earned, and how much will be bought and sold, in a year that has not happened yet. The rates are decided. The bases are guessed at, using last year's bases, an assumption about how fast nominal outputThe size of everything produced, counted in the money of the day, with no adjustment for price changes. Nominal output rises when prices rise, when quantities rise, or when both do. will grow, and an assumption about complianceWhether the people who owe a tax actually file and pay it. A base only turns into a collection to the extent it is reported and paid..

The asymmetry between the two sides of a budget is structural rather than a failure of forecasting, and a reader who treats both sides as equally firm has misread the document.

None of this is a criticism of anybody's arithmetic. Even a perfect forecaster faces the asymmetry. One side of the document records what a government has chosen. The other side records what it expects other people to do, and no technique turns the second into the first. The gap between what a receipt side estimated and what actually arrived is a live question in its own right, and it is covered separately.

Is borrowing a source of government revenue?

No. Borrowing is the confusion most worth clearing, and counting it as revenue changes what a whole account means.

Money that a government borrows arrives in exactly the same form as money it collects. Borrowed money is rupees, it turns up in the same account, and it pays for the same salaries and the same roads. A household knows this feeling well. Money from a salary and money from a loan both arrive in the bank, both buy groceries, and only one of them is income. Nobody adds this month's borrowing to this month's pay and calls the sum earnings. One of the two has to go back out again, with something extra on top.

Borrowing arrives as money and leaves as a liabilityAn amount that has to be paid back to somebody. It sits on the side of an account recording what is owed, not the side recording what was earned.. A liability is not revenue, and it is never counted on the receipt side.

For Sankhya the consequence is plain: revenue receipts are Rs 3,00,000 crore, made of the three sources set out above. Total spending is Rs 4,00,000 crore. The Rs 1,00,000 crore between them, the fiscal deficitThe gap between everything a government spends and everything it receives that is not borrowed. How it is measured, and how it compares with the other two deficits, is covered separately., is not on the receipt side and never appears there. The whole of it is funded by something that is not revenue at all.

Hold that firmly. A reader who quietly counts borrowing as a receipt closes the Sankhya account with Rs 4,00,000 crore coming in and Rs 4,00,000 crore going out, concludes that the account balances, and has lost the single most important fact in it. The account does not balance. The account is short by Rs 1,00,000 crore, and that shortfall is met by taking on an obligation to pay the money back with interest. The effect of borrowing on the price of money, and the way the obligations pile up over years into a stock of public debt, are both taken up separately and neither belongs on the receipt side.

What does an analyst actually watch on the receipt side?

Not the total. The composition. The composition says how far the total will move when the year goes badly.

Start with the household version, the same idea in ten seconds. Two households both bring in Rs 60,000/- a month. The first gets all of it from one salary at one employer. The second gets Rs 20,000/- from a salary, Rs 20,000/- from a small tailoring business and Rs 20,000/- from rent on a room. The totals are identical. The exposure is not remotely identical, and anybody lending to either household would want to know which one they were looking at before anything else.

A receipt side works the same way, with one extra twist that is worth getting right. A receipt side is not only about how many sources there are. The width of each source's base matters just as much.

A narrow base moves further in a downturn than a broad one, so two receipt sides collecting exactly the same rupees can lose very different amounts in the same bad year.

Work it with two versions of Sankhya's indirect tax, both collecting exactly Rs 1,50,000 crore, and hold the rate perfectly still in both. The broad version taxes Rs 12,00,000 crore of transactions at 12.50 per cent: it reaches nearly everything that gets bought, including the things people buy in every kind of year. The narrow version taxes Rs 4,00,000 crore at 37.50 per cent: it reaches only a slice, and the slice it reaches is the kind of spending a household cuts first when money is tight.

Now put both through the same bad year, in which output falls by 4.00 per cent. The broad base falls roughly in step with output, by 4.00 per cent, to Rs 11,52,000 crore, so the collection falls to Rs 1,44,000 crore, a fall of Rs 6,000 crore. The narrow base falls three times as far, by 12.00 per cent, to Rs 3,52,000 crore, so the collection falls to Rs 1,32,000 crore, a fall of Rs 18,000 crore. Same starting collection, same rate untouched, same downturn, and one of the two loses three times as much as the other. The difference was decided by the shape of the base long before the bad year arrived.

The habit worth building is short. Any receipt side raises two questions before the total is looked at: how many sources is this resting on, and how wide is the base under each one. A lender assessing a borrower whose earnings depend on government contracts is asking the same question about the same government, one step removed.

The Republic of Sankhya, the vegetable seller outside the railway station and the two households bringing in Rs 60,000/- a month are invented.
Educational material. Not advice on any investment, tax, budget or market position.