Fiscal, Revenue and Primary Deficit Compared

The three deficits are three cuts of the same gap. The fiscal deficit is everything the government must borrow. The revenue deficit is the part of that borrowing which funded day to day spending rather than assets. The primary deficit is the fiscal deficit with interest on past borrowing taken out. The remainder shows what this year's decisions cost as against what earlier years already committed.

Two things have to be in place before the arithmetic starts. The first is an account for the Republic of Sankhya, an invented country used here for teaching. The Sankhya account was set out earlier with one gap named in it: Rs 1,00,000 crore, the amount that had to be borrowed. Nothing was taken apart at that point. The second is the output measure, the single total that everything the economy produced in the year adds up to. For Sankhya that total is Rs 17,47,200 crore. Every percentage in this guide sits on that denominator, so a figure of 5.72 or 0.57 has a known total below the line.

Now comes the taking apart. One gap, three ways of cutting it, and the three cuts answer different questions. The same set of books yields three statements: how much was borrowed, what the borrowing bought, and how much of it this year actually chose.

Why does one account give three different deficits?

Here is the picture to hold on to before any arithmetic. There is one account. The account has money coming in and money going out, and going out is bigger. The difference between the two is the gap, and there is only one gap.

Consider a household. A salary comes in. Money goes out on rent, groceries, school fees, the electricity bill, the interest on a loan taken three years ago to build an extra room, and, this month, on a new water tank for the roof. If the household spent more than it earned, it had to find the difference somewhere. The shortfall is one number. But different people would ask different things about it. The bank lends the difference, so the bank wants to know how much had to be found. A parent wants to know whether the extra went on the water tank still standing on the roof or on groceries long since eaten. And the household itself wants to know how much of the shortfall came from choices made this month. The interest bill was fixed the day the loan was taken, and nothing anybody did this month could have changed it.

Three different questions are being asked about one shortfall, and the three deficits are exactly those three questions asked of a government account. The three deficits are not three separate facts about the economy. All three are readings of a single gap, so a movement in any one of them can only be read against the other two.

Sankhya's account for the year runs as follows.

Three rows of spending, one total of receipts. Taking the receipts away from different combinations of those three spending rows is what produces the three measures. One subtraction, done three ways, produces all three measures.

What is the fiscal deficit, and what does it oblige?

Take everything the government spent and take away everything that came in without creating a liability. On the Sankhya account that is Rs 4,00,000 crore out against Rs 3,00,000 crore in, so the answer is Rs 1,00,000 crore. Set beside Sankhya's output of Rs 17,47,200 crore, that gap is 5.72 per cent.

Notice the phrase doing the work: came in without creating a liability. Taxes do that. Fees and dividends do that. A non-debt capital receiptMoney arriving as a capital item without creating a liability, such as the sale proceeds of something the government held, or the repayment of a loan it made earlier. Borrowing brings money in too, but borrowing leaves an obligation behind. does that too, though the Sankhya account has none, and the absence of such receipts is what makes the identity between the three deficits exact. Borrowing does not. Borrowing brings money in and leaves an obligation behind it. So the fiscal deficit is not really a measure of anything abstract. The fiscal deficit is a bill, and a bill is an amount somebody has to be persuaded to lend.

The fiscal deficit is the borrowing the year requires, so it is the number that adds to what the government already owes. The borrowing requirement is why the fiscal deficit is the figure a lender, a rating body or anyone pricing government paperThe instruments a government issues when it borrows, held by banks, insurers, pension funds and others, and traded between them afterwards. How such an instrument is priced belongs to fixed income and is not worked out here. looks at first. The fiscal deficit is a flow: one year's addition. The debt stockThe whole amount of borrowing outstanding at a point in time, as against the fresh borrowing of a single year. One is a level and the other is a flow, and adding this year's flow is how the level moves. it adds to is a level, and how that level is measured and when it starts to matter belongs to the outstanding stock of government borrowing.

What does the revenue deficit show that the fiscal deficit hides?

The fiscal deficit gives how much was borrowed. The fiscal deficit says nothing whatever about what the borrowing went to. Two governments can borrow the identical Rs 1,00,000 crore, one of them to build power lines and one of them to pay pensions, and the fiscal deficit will not tell them apart. The revenue deficit is the cut that does.

Split the spending into two kinds. Revenue spending gets consumed inside the same year it is paid: salaries, pensions, subsidies, the running cost of an office, and the interest bill. Capital spending ends in something that outlasts the year: a road, a bridge, a school. On the Sankhya account, revenue spending is the interest of Rs 90,000 crore plus other revenue spending of Rs 2,40,000 crore, so Rs 3,30,000 crore. Set that against revenue receipts of Rs 3,00,000 crore and the revenue deficit is Rs 30,000 crore. Against output that is 1.72 per cent.

Back to the household for a moment. Borrowing to buy a water tank leaves, at the end of the year, a debt and a water tank. Borrowing to buy groceries leaves a debt, and the groceries are gone. In both cases the debt is identical and the interest is identical. The difference is what stands on the other side of the debt. Borrowing that funded consumption spendingSpending used up in the year it happens: wages, pensions, subsidies, fuel, the electricity bill of an office. Consumption spending buys something real and leaves no item behind that can be pointed at afterwards. leaves nothing behind, and borrowing that funded an asset leaves the asset, and the revenue deficit makes precisely that distinction where the fiscal deficit cannot.

Now the hard part, and it is where most readers overreach. The revenue deficit does not say the asset was worth building. A road that nobody drives on is capital spending. A bridge to a place with no reason to go there is capital spending. The measure sorts spending by the accounting head it sits under, not by whether it was any use. The revenue deficit answers what kind of spending, and only what kind. Whether the thing was worth doing is a judgement that needs evidence the account does not contain.

Why is interest stripped out to get the primary deficit?

The third cut asks a question the first two cannot. Of everything the government had to borrow this year, how much of it followed from what this year's ministers actually decided?

Look at the interest line. Rs 90,000 crore is due on borrowings raised in earlier years. Nobody sitting in the finance ministry this year chose that figure. The figure was fixed when the money was raised, by the amount raised and the terms it was raised on. The interest line is debt serviceThe yearly cost of carrying borrowings already taken: the interest falling due on them, and where a repayment is due, that as well. The obligation was fixed when the money was raised, so debt service is met before anything else is decided., and it goes out before a single new decision gets made. A government could cancel every road, freeze every salary and abolish every subsidy, and the Rs 90,000 crore would still be due.

So take it out. Fiscal deficit of Rs 1,00,000 crore less interest of Rs 90,000 crore leaves Rs 10,000 crore. Against output that is 0.57 per cent, and the residue is the primary deficit.

Interest is the price of decisions already taken, and no government can unmake it in the year it is sitting in, so what remains once it is removed is the part of the borrowing that this year's own choices produced. Separating the inherited from the chosen is the entire reason the measure exists, and the primary deficit is the only one of the three that even attempts the separation.

Two cautions, both worth carrying. Stripping interest does not make it stop existing. The money leaves. Nobody is pretending it does not. And the primary deficit answers a different question, so it cannot be quoted as a nicer version of the fiscal deficit when the fiscal deficit is uncomfortable. The borrowing still comes to Rs 1,00,000 crore.

What does a primary deficit of 0.57 beside a fiscal deficit of 5.72 establish?

The whole comparison turns on one point, and the point comes down to a single sentence.

Sankhya's fiscal deficit is 5.72 per cent of output. Its primary deficit is 0.57 per cent. The whole distance between those two numbers, 5.15 percentage points of output, is interest on borrowing taken in earlier years. Put in rupees rather than shares: of the Rs 1,00,000 crore Sankhya must borrow this year, Rs 90,000 crore of it is going straight back out as interest on what it borrowed before. Ninety per cent of the new borrowing is servicing the old.

Almost the whole of the borrowing is to pay interest on past borrowing, and that is a single arithmetic fact with no verdict attached to it.

Read that sentence carefully. There is a strong pull to add something to it, and the sentence is complete without the addition. The arithmetic fact does not say Sankhya is in trouble, and it does not say Sankhya is fine. It does not say the borrowing is too high, or that the interest bill is unsustainable, or that anything must now be done. Whether the interest bill is too large is a different question altogether. The answer needs things the account cannot supply: where the money goes and whether it will produce anything, how fast the economy is likely to grow, at what cost the government can keep borrowing, who is lending and for how long, and what people in the country want their government to be doing. Every one of those is a separate enquiry, and several are questions about preferences rather than arithmetic.

The contrast redirects attention. If almost all of the borrowing is interest, then a change in the interest bill matters far more to next year's fiscal deficit than a change in any single spending programme, simply because of the sizes involved. And a reader who quotes the 5.72 alone has described the borrowing without once mentioning that nine tenths of it was committed before the year began. Naming the interest share is not a verdict either, only a more complete description.

What does the fiscal deficit less capital spending actually give?

One more move on the Sankhya numbers. The fiscal deficit of Rs 1,00,000 crore less the capital spending of Rs 70,000 crore comes to Rs 30,000 crore, exactly the revenue deficit.

A reader is very tempted to treat the match as a confirmation. Two different routes, the same answer, so the account must be right. Stop there. The reading is wrong, and wrong in a way that keeps reappearing.

The relationship is an identity: it holds by construction whenever there are no non-debt capital receipts, so it confirms nothing at all and is not a cross-check. The reason, in one line of algebra written out in words. The fiscal deficit is total spending less receipts. Total spending is revenue spending plus capital spending. So the fiscal deficit is revenue spending plus capital spending less receipts. Subtracting capital spending from that leaves revenue spending less receipts, the definition of the revenue deficit. The second route is the first route rearranged. The second route could not have come out differently.

The identity can be seen rather than taken on trust. Whatever figure in the Sankhya account is changed, both sides move together, every single time. Raising capital spending raises the fiscal deficit by the same amount that is then subtracted. Cutting the receipts raises both the fiscal deficit and the revenue deficit by exactly what was cut. Halving the interest lowers both by exactly the same amount. There is no setting of the account, however wrong, in which the two sides disagree. The test is not looking at the account at all, so an account with a transposed digit in it, an account with a whole programme left out and an account made up outright would all pass it.

So what is it good for? Something narrower, and worth saying precisely. The rearrangement establishes that the borrowing exceeded capital spending by Rs 30,000 crore. Rs 30,000 crore of what Sankhya borrowed this year funded day to day running rather than anything that lasts. The statement is a real one about the account, and it happens to be the statement the revenue deficit already made. The rearrangement is a way of seeing the revenue deficit from the borrowing side rather than from the spending side. The rearrangement is a different view of one fact, not a second fact.

How do all three come off the same account in one pass?

All three cuts sit in one table, so every subtraction stands beside the other two. The middle column is the sum being done and the right column is what it comes to.

Every share in that last column is taken against the same output figure of Rs 17,47,200 crore, and only a common denominator lets the three be set beside each other at all. Down the rupee column stand the three sentences that matter: Sankhya had to borrow Rs 1,00,000 crore, Rs 30,000 crore of that borrowing funded day to day running, and only Rs 10,000 crore of it followed from this year's own decisions. The three sentences are the whole of it.

Which of the three should be read first?

All three, in a fixed order. The answer sounds like a dodge until it becomes clear that each of the three answers a question the others physically cannot.

The fiscal deficit comes first, and it settles a question of fact that everything else depends on: how much money has to be raised from lenders this year. For a bank treasuryThe desk inside a bank that manages its own funding and its own holdings of instruments, as distinct from the parts of the bank that lend to households and businesses. working out how much government paper is coming to market, or a lender looking at how much competition there will be for savings, this is the number that matters and the other two are commentary. The fiscal deficit is also the flow that adds to what is already outstanding.

The revenue deficit comes next, and it establishes what the borrowing bought. Rs 30,000 crore of Sankhya's Rs 1,00,000 crore went on running costs, so a tenth of a year's borrowing has left nothing standing. The capital half is the part that turns into orders for somebody, so an equity analyst looking at whether a government will keep ordering steel, cement and equipment cares about the split rather than the total.

The primary deficit comes last, and it establishes how much of the borrowing this year's decisions produced. Sankhya's Rs 10,000 crore against a fiscal deficit of Rs 1,00,000 crore says that almost everything being borrowed was committed before anyone sat down to decide anything.

The household version is the same three questions. How much had to be arranged this month. Whether the extra went on the tank or on the groceries. And how much of the shortfall was the loan instalment that could not have been touched anyway. A reader who quotes one of the three has described a third of an account and will be confidently wrong the moment somebody asks a question the other two answer.

The Republic of Sankhya and the household with the water tank are invented.
Educational material. Not advice on any investment, tax, budget or market position.