Reading a budget document without a finance degree

Every year a government releases a budget document running to hundreds of pages, and every year most of us do the sensible thing: we read the news coverage. Someone else did the work, boiled it down to five things that matter, and we move on.

I did that for years. But I also started, slowly and badly, reading the actual documents. Not all of them and not cover to cover. Just enough that I no longer felt like I was reading a press release about a press release. Here is what I learned, in case you want to try it too.

The headline numbers are the least interesting part

The finance minister stands up and announces a total, and that total becomes the story. “X billion for health care.” “Record defense spending.” They are real numbers, but they are also the most carefully chosen ones, picked to sound good. The total tells you almost nothing about what is actually changing, because spending is always rising somewhere.

What matters is the comparison. Up compared to what? Compared to last year’s cash figure, or last year’s figure adjusted for inflation? Per person, or in total? Over four years, or just in the first year? A number that looks enormous in a speech can be a real decline once you account for population growth and rising costs. And a number that looks modest can be a real increase if it replaces something that was quietly shrinking.

The documents usually tell you this, but you have to look for it. Find the table, not the paragraph. The paragraph was written by someone who wants you to feel a certain way. The table was written by someone who has to be accurate.

Learn to see the gaps

This is the trick that took me longest to grasp: budgets are as much about what is missing as about what is in them.

Look at the time horizon. If the document gives detailed figures for two years and then a vague sentence about “future expenditure frameworks,” that is a signal. It means the hard decisions have been pushed past the next election or the next awkward vote. It is not exactly a conspiracy. It is just how politics works. But it is worth noticing.

Then look at the so-called baseline. A lot of announcements are measured against a scenario in which spending was already going to rise. So a “new investment” of several billion may be an increase to a plan that was already on the books. It is not dishonest, but it is easy to read a headline and think the money is entirely new when it is partly already promised.

And check the fine print on things like efficiency savings. Every budget promises to find money by being more efficient. Sometimes it is real. Often it means “we assume departments will spend less than we told them to spend,” which is a polite way of saying the numbers may not add up. If the savings are large and unexplained, treat them with suspicion.

The assumptions do the work

Budgets are forecasts, and forecasts rest on assumptions. Growth will be such and such. Inflation will be so and so. Unemployment will stay low. Wages will rise by a certain amount.

You do not need to be an economist to ask whether those assumptions look believable. If the growth forecast is far rosier than what independent forecasters are saying, the budget arithmetic is probably too good to be true. If the document assumes wages will grow faster than they have in years, the tax receipts it is counting on may not arrive.

The good news is you do not have to judge this alone. Budget documents usually include an independent forecast alongside the government’s numbers, and the gap between them is often the most telling thing in the whole package. If the independent forecaster is more pessimistic, that is your story.

A five-minute routine that actually works

Here is a concrete way to do this on a real document, using a fictional but typical example: a government proposes 4.2 billion for a new child care program over five years.

Step one: open the summary table, not the minister’s foreword. Find the line for the program and check the yearly split. Suppose it reads 200 million, 400 million, 900 million, 1.3 billion, 1.4 billion. That back-loading is normal, but it means the first-year number in the headline is a fraction of the total, and most of the money arrives after the next election.

Step two: find the baseline. The table may show the program against a “previously planned” line of 1.5 billion. So the genuinely new money is 2.7 billion, not 4.2 billion. Still real, just smaller than the announcement suggests.

Step three: check the assumptions page. If the forecast assumes 3 percent growth while the independent forecaster says 1.5 percent, run the tax revenue line at the lower figure for a rough sense of the shortfall. A simple rule: each percentage point of missing growth costs roughly 0.5 percent of GDP in revenue, which on a 300 billion economy is about 1.5 billion a year. That single gap can swallow the entire new program.

Step four: look for the missing years. If detailed figures stop after year three, note it. The hard choices were deferred.

None of this makes you an expert. It makes you someone much harder to fool, and that is most of the fight.