The three deficits

One gap, measured three ways, and the three answer genuinely different questions.

The definitions

Fiscal deficit = total expenditure − total receipts.

The headline. How much the government must borrow this year. It is the number in every news report and the one the Budget sets a target for.

Revenue deficit = revenue expenditure − revenue receipts.

How much of the recurring spending was not covered by recurring income. This is the borrowing that funded salaries, pensions, subsidies and interest rather than anything that will exist next year.

Primary deficit = fiscal deficit − interest payments.

The deficit excluding the cost of past borrowing. It is what this year's government did, stripped of what previous governments committed it to.

India's three, to the rupee

April to September 2026, in ₹ crore, from the Controller General of Accounts:

	Budget Estimate	Actual to September	% of BE
Fiscal deficit	16,95,768	7,14,413	42.1%
Revenue deficit	5,92,344	1,38,468	23.4%
Primary deficit	2,91,796	68,403	23.4%

Every one of those is derivable from chapter 1's table, and each derivation lands exactly:

fiscal   = 25,49,835 − 18,35,422 = 7,14,413
revenue  = 18,95,288 − 17,56,820 = 1,38,468
primary  =  7,14,413 −  6,46,010 =   68,403

Interest payments were ₹6,46,010 crore, which is where the third line comes from and which chapter 4 is largely about.

The same three identities hold on the Budget Estimates too, which is worth checking once because it confirms you are reading the right rows rather than numbers that happen to look plausible.

What each one tells you

The fiscal deficit tells you the borrowing requirement. That is genuinely useful: it is how much new government paper the bond market must absorb this year, which chapter 6 and the fixed income subject both care about. It tells you almost nothing about whether the borrowing is sensible.

The revenue deficit tells you whether the borrowing bought anything. This is the one the Indian framework has historically cared most about, and the aspiration behind it is sometimes called the golden rule: borrow for capital, pay for current spending out of current income. A revenue deficit of zero would mean every rupee borrowed went into something that lasts.

The primary deficit tells you what this year's decisions cost. A government can inherit an enormous interest bill and run a prudent year; the fiscal deficit will not show that and the primary deficit will. It is the fairest measure of current policy and the least quoted of the three.

The question the three answer together

Now use them. Of every hundred rupees borrowed in the first half of 2026-27, how much funded current spending?

revenue deficit ÷ fiscal deficit = 1,38,468 ÷ 7,14,413 = 19.4%

About a fifth. Four-fifths of the borrowing went to capital purposes rather than to running costs.

Put the other way, capital expenditure of ₹6,54,547 crore was 91.6% of the fiscal deficit — the government borrowed ₹7.14 lakh crore and spent ₹6.55 lakh crore on capital account. Strip out the ₹1,51,220 crore of that which is loans disbursed rather than assets built, and direct capital spending still accounts for 70.5% of the borrowing.

That is a far better composition than the headline deficit alone suggests, and it is invisible unless you take the three measures together. Anyone who read only "₹7.14 lakh crore of borrowing" has the number and none of the information.

The direction, which is the other half

The same table carries last year's figures for the same point in the year, and they are less comfortable:

% of Budget Estimate, at end-September	2026	A year earlier
Fiscal deficit	42.1%	36.5%
Revenue deficit	23.4%	5.2%
Primary deficit	23.4%	−1.7%

The revenue deficit was at 5.2% of its full-year allowance a year ago and is at 23.4% now. The primary deficit was negative — a primary surplus at this point in the year — and is now 23.4%.

So the composition is good and the trend is worse. Both statements are true, both come from the same two columns, and a commentary offering only one of them is offering half the table. Chapter 5 is about reading these percentages properly, including why the 42.1% headline is less alarming than it was reported to be.

Three cautions about deficit numbers

A deficit is a flow; debt is the stock. This year's borrowing adds to an accumulated total that chapter 6 takes up. A small deficit on a large debt and a large deficit on a small debt are different situations entirely.

"Percentage of GDP" is the usual framing and introduces a second moving part. A deficit-to-GDP ratio can fall because borrowing fell or because nominal GDP rose, and chapter 3 of the growth subject showed nominal GDP rising 10.3% in a quarter. The rupee figures above have one moving part; the ratio has two.

The monthly figure is not the annual figure. The CGA says so itself, in a note under the table, and chapter 5 quotes it in full because it is the single most important sentence in the document.

The point

Three deficits measure one gap differently: fiscal is the borrowing requirement, revenue is how much of it funded recurring spending, and primary is the deficit excluding the cost of past borrowing. For April to September 2026 they were ₹7,14,413 crore, ₹1,38,468 crore and ₹68,403 crore, each deriving exactly from the receipts and expenditure lines. Taken together they say that about a fifth of the borrowing funded current spending and the rest went to capital — a better composition than the headline implies — while the revenue and primary deficits have both deteriorated sharply against the same point last year.