The Union Budget: Structure, Timeline and What to Read First

The Union Budget is the government's statement of what it expects to receive and what it proposes to spend, put before the elected house so that the spending can be authorised. The document is two things at once: a set of proposals, and a request for permission. Most of what a reader hears about it comes from the announcements, and most of what actually binds the government sits in the authorisation.

Two things already built sit underneath that sentence, and neither is extended here. The first is that a government acts by spending and by taxing, and that when the spending is larger than what comes in, the difference has to be borrowed. The second is that the gap can be split several ways, each split answering a different question. Both of those are arithmetic, and the arithmetic does not change below. The shape does. The same material stops being an account and becomes a document: a thing with parts, an order it moves through, and words on it that do not all carry the same weight.

Every rupee figure below belongs to the Republic of Sankhya, an invented country built for teaching. The Union Budget and the Ministry of Finance are the real document and the real office that prepares it. Anything a real government has actually proposed in a particular period is the sort of detail somebody has to keep current: accurate in the week it is written down, out of true some while afterwards, and identical in appearance either way. The offices that hold those details and keep them current are named below.

What kind of document is the Union Budget?

Start at a kitchen table. The shape is identical and there is nothing intimidating about it. A household sits down before the month starts and writes two columns on a sheet of paper. On one side, what it expects to come in: a salary, a little rent from a room let out, the interest on a deposit. On the other, what it plans to pay out: school fees, the loan instalment, groceries, and this month a new cylinder regulator. The sheet is a plan. The sheet is not the passbook.

The passbook is a different object entirely. The passbook says what actually landed and what actually left, and it is written after the fact by an institution that had no opinion about any of it. No member of that house would mix up the two. The plan and the passbook are physically different things kept in different places. Printed in the same document, in the same typeface, under the same heading, the confusion becomes not just possible but likely.

A budget states what is expected and what is proposed. A record of what actually happened is a different document produced later, and reading an estimate as an outcome is the commonest error a first-time reader of any budget makes. The Union Budget is prepared and presented by the Ministry of Finance, and the whole of it, receipts and spending alike, is written in the forward tense. A receipt figure in it is what the government expects to collect. A spending figure in it is what the government proposes to spend and is asking to be allowed to spend. Neither is a measurement.

There is a second half to the description that matters just as much. The document is not only an estimate, it is a request. A government cannot spend from public money on its own authority: it has to ask, and the asking is done through this document, put before the legislatureThe elected body that makes law and grants permission to spend public money. In India the Union Budget goes before Parliament. The word covers the institution rather than any particular sitting of it.. So a reader holding a budget document is holding an estimate and a petition bound together, and almost every misreading of one comes from forgetting one of those two halves.

What are the main parts of a budget document, and what does each part carry?

A budget is easier to hold once the question stops being what the parts are called and becomes what each part carries. Four kinds of thing go into the pile, and the useful separation is not by name, it is by how much weight each carries.

The first is a statement of receipts: everything the government expects to come in, sorted into heads so the reader can see where it is expected to come from. The second is a statement of spending: everything the government proposes to pay out, sorted by who is doing the spending and what it is for. The third is a set of tax proposals: changes to what is charged and on what. The fourth is a stack of supporting statements. The supporting statements explain and break down the first three, and they are by a wide margin the bulkiest part of the pile.

The four parts are not equally binding, and knowing which is which is most of a reader's job. The statement of spending is the part that turns into permission, once it has been through the stages described further down. The tax proposals are proposals until they are agreed and given legal effect. The statement of receipts is a forecast, and no forecast binds anybody to anything. A receipt figure is an expectation about the behaviour of millions of taxpayers, not a commitment by the government. The supporting statements bind nobody at all. The supporting statements exist to explain, and explanation is exactly what a reader most needs and least often reads.

Think of a wedding, on the scale most households actually run one. There is the guest list with the caterer's rate against it, and the guest list is where the money is really committed. There is the list of what the household hopes cousins will contribute, a forecast and nothing more. There is the running note of who has been asked for what, and the note explains the first two. Everybody in that house knows the guest list is the binding document and the contribution list is a hope. A budget has exactly the same asymmetry, and the reason it gets missed is that all four parts of the budget arrive printed to the same standard on the same day.

What does the same account look like when it is laid out as a document?

Here is the Republic of Sankhya again, with figures built for teaching and standing for no real country's public finances. The arithmetic below is the same arithmetic built earlier. The arrangement is what differs: the receipts sit in one statement, the spending sits in another, and the difference between them appears at the foot as an amount that has to be borrowed, rather than as a result somebody has calculated.

Look at what the last row is doing. The arrangement is the teaching here. In a document, the gap is not presented as an answer to a question about the government's health; it is presented as an operating requirement. A spending statement larger than a receipts statement can only be made to work by borrowing the difference. Sankhya proposes to pay out Rs 4,00,000 crore and expects to take in Rs 3,00,000 crore, so Rs 1,00,000 crore has to come from somewhere, and borrowing is where.

The same Rs 1,00,000 crore for Sankhya carries a name, the fiscal deficitThe amount a government has to borrow in a period because what it pays out is larger than what it takes in. It can be split several ways, and how it splits is set out separately., and the deficit can be split several ways, each split answering a different question about the borrowing. How the deficit splits is covered separately. The structural point is narrower: the number appears in the document as a requirement to be met, and every other figure in the pile is a part of one of the two statements it came from.

What is the order of stages a budget passes through?

The word timeline means one thing here: the order, and nothing else. The order is the durable part. Any calendar attached to it is a maintained record that goes quietly wrong, so no date, month, season or frequency appears below. The sequence below has five positions in it, and the positions do not move.

First the document is presented, and presentation means it is placed before the legislature and becomes public. Second it is debated, and debate means it is argued over, examined, questioned and put through whatever scrutiny the house applies to it. Third it is authorised: the house grants permission for the spending. Authorisation is the stage the whole document was built to reach. Fourth the money is actually spent, by the offices that were given permission, for the purposes permission was given for. Fifth there is a later account of what was actually spent, prepared after the period is over and put through an independent auditAn examination of accounts by somebody who did not prepare them and does not report to whoever did, so that the report on what happened does not come only from the office being reported on..

Money cannot be spent at the presentation stage, and that is precisely where most readers assume it moves. Presentation is the only stage most people ever see. The stage that carries the news is the first one. The stage that carries the power is the third. Between them sits the second stage. Debate is where a proposal can be argued with, and almost nobody outside the house watches it.

The fifth stage closes the loop, and it is the reason a budget is not a record. The account of what was actually spent is a separate document, produced later, by a different process, and checked by somebody who did not prepare it. In India the constitutional auditor of the Union's accounts is the Comptroller and Auditor General.

What is the difference between what is announced and what is authorised?

The distinction between an announcement and an authorisation is the one everything else rests on, and it is worth slowing down over.

An announcement is a statement of intention. The announcement says a thing is going to be done. An announcement can be broad, it can be aspirational, it need not carry an amount, it need not name who will spend it, and it commits nobody to anything by the mere fact of having been said. An authorisation is a different object entirely: it is permission granted by the legislature to a named office to spend up to a stated amount on a stated purpose. Three things are fixed in it and all three matter. There is a ceiling. There is a purpose. There is somebody who has to answer for both.

An announcement is an intention and an authorisation is permission to spend a stated amount on a stated purpose, and a reader who cannot tell the two apart will read a headline as a commitment. The headline is not lying. The headline is reporting, accurately, that an intention was stated. At the moment of the announcement no permission has been granted, so the headline cannot report whether permission will be granted, in what amount, and to whom.

Here is the household version, and it is exact rather than approximate. A parent says at dinner that the house will get a proper water tank this year. The dinner statement is an announcement, and it is honestly meant. Nothing has been permitted, no amount has been set, and no shop has been told anything. Some weeks later the household sits down, agrees that Rs 40,000/- may be spent on a tank and on nothing else, and tells the son to go and buy one within that figure. The agreement at that sitting is an authorisation. The tank may still not arrive: the son may find nothing suitable, the money may be needed elsewhere, the shop may close. A ceiling, a purpose and a person now exist, and none of the three existed at dinner.

The gap between the announcement and the authorisation is not deception; it is the process working exactly as it was designed to work. The design puts a stage of argument between what a government would like to do and what it is permitted to do. If announcement and authorisation were the same act, that stage would not exist, and the whole apparatus of debate and scrutiny would be decorative. A reader who is disappointed to find that an announcement did not bind anybody has, without realising it, been hoping for a system with the scrutiny removed.

Why does the same line appear more than once with different labels?

Open a budget document, find a single line itemOne named row in a statement of accounts, carrying its own figure. A statement is built out of line items, and a total is the sum of the ones beneath it. such as capital spending, and it will very often carry more than one figure. To a first-time reader that looks like an error or a contradiction. The extra figures are neither an error nor a contradiction, but three different kinds of number sitting next to each other, and the document is being careful rather than careless.

An early estimate is made before the period starts, out of what the office expects. A revised estimate is made partway through, when some of the period has actually happened and the earlier expectation can be corrected. An actual is what was finally paid out, reported after the period is over and put through the audit described above. Three numbers against the same line are three different kinds of number rather than a contradiction, and each answers a different question: what was expected, what was expected once better information arrived, and what happened.

The reason the three end up printed side by side is worth naming, and it is the direct cause of the mistake below. A budget document does not carry one period, it carries three at once. The document carries the coming period as an early estimate, and that is the period permission is being asked for. The document carries the running period as a revised estimate, and the office now knows more than it did when the period opened. And it carries the last closed period as an actual, the only period for which an actual can exist. Three adjacent columns, three different kinds of figure, and nothing in the layout stopping anybody from subtracting one from another.

Follow Sankhya's capital spending line through the whole grid.

The closed year read across shows what a revision actually looks like. Sankhya expected to spend Rs 62,000 crore on capital, revised that to Rs 59,000 crore partway through, lower by Rs 3,000 crore, and finally spent Rs 57,000 crore, lower by a further Rs 2,000 crore. Across the whole year the line came in lower than the early estimate by Rs 5,000 crore, ten per cent below where it started. Nothing in that sequence is a policy change. The line is one number being restated as the period stopped being a forecast and started being a fact.

Read the grid down instead and notice the emptiness in the lower right. There is no actual for the running year because the year has not finished, and there are two blanks against the year ahead because it has not started. The one comparison a reader can always make between two years is early estimate against early estimate. An early estimate is the only kind of figure that exists for every period the document covers.

Telling a change from a revision

What should a first-time reader look at before anything else?

Three totals, in this order: the total of spending proposed, the total of receipts expected, and the gap between them. Not a scheme, not the head that happens to be in the news, and not the part of the document that has been written about most. The reason is structural rather than a matter of taste. Every other figure in the document is a share of one of those three totals, so a reader who starts with a single scheme is holding a numerator with no denominator anywhere in sight.

Watch what that does with Sankhya's own figures. Suppose the document carries a scheme of Rs 12,000 crore. Read alone, Rs 12,000 crore is an enormous number and there is nothing whatever a reader can do with it. Read against the three totals, it lands immediately: it is three per cent of the Rs 4,00,000 crore proposed to be spent, four per cent of the Rs 3,00,000 crore expected to come in, and twelve per cent of the Rs 1,00,000 crore that has to be borrowed. Same figure, three sentences, and the reader now knows roughly what size of thing they are holding.

The habit generalises past budgets entirely, and that is why it is worth building. A grocery bill of Rs 3,000/- means one thing in a household earning Rs 20,000/- a month and something quite different in one earning Rs 2,00,000/-, and nobody would try to interpret the bill without asking about the income first. A budget is the same object at a larger scale. Take the denominators first and every later figure has somewhere to sit.

What does somebody who reads budgets for a living do differently?

One thing above all: they open the supporting statements before they open the speech.

The reasoning is not about scepticism, and it is certainly not about anybody being untruthful. A speech is built to be listened to. A speech carries what is being said, and what is being said is chosen, ordered and given emphasis by whoever is saying it. Choosing and ordering that way is a perfectly ordinary property of speeches. The supporting statements are built to be filed. The supporting statements carry what is being asked for, and what is asked for is bounded: an amount, against a purpose, against an office. Only the second of those two is a thing a plan can rest on, and only the second has a ceiling in it.

An analyst reading a budget is not looking for what is new, they are looking for what is bounded, and the bounded part is almost never the part that was announced. The practical shape of that habit is short. Take the three totals first, then the statement of spending, then the head in question read against those totals. The speech comes last, and only then, to show which parts of the document somebody chose to talk about, genuinely useful information of a different kind.

A lender does the same thing for a narrower reason. Somebody assessing a contractor who builds water systems wants to know whether the work the contractor is bidding for sits inside a permitted amount for a stated purpose, or inside a stated intention with no ceiling attached to it yet. The two situations look almost identical in a newspaper and behave nothing alike on a credit file. The first is a bounded thing with an office answerable for it. The second is an intention that may or may not reach the third stage of the sequence, and a lender who cannot tell them apart is lending against a headline.

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