How budgeting works: plan, split, saving rate

A budget is a plan for money you have not spent yet. That is the whole difference between budgeting and tracking. Tracking tells you where the money went; a budget decides where it goes. Both are worth doing and only one of them can change an outcome, because by the time a tracker has an answer the money has already left the account.

The thing being planned is cash flow: money in against money out, arranged in time. Income usually arrives in lumps on fixed dates. Outflow arrives in three shapes at once, a steady drip of daily spending, a handful of large payments on their own dates, and a scatter of bills that turn up once or twice a year. A budget holds those two patterns against each other in advance, so a large payment lands in a month you already knew about.

What comes out at the bottom is a residual: income minus everything planned. That residual is the only part of the month that becomes savings or extra debt repayment. Most other lines are transfers out of your account and into somebody else's, with one exception worth naming: the principal buried inside a required loan payment leaves the account too, and still raises net worth, because it reduces what you owe. A plan measured in cash and a plan measured in net worth do not agree line for line. A budget that produces a residual by accident produces a different one every month, and a budget that assigns the residual first produces the same one on purpose.

None of this needs an app, a spreadsheet or a named system. It needs a number decided in advance and a way of noticing when the month stops matching it. Software adds speed and legibility. The decision is the part that does the work.

Costs behave differently, and the difference decides what you can do about them inside a given month.

[Fixed costs](/definitions/fixed-costs) are the same amount every period and arrive whether or not you change your behaviour: rent, loan payments, insurance premiums, subscriptions. Fixed does not mean permanent. It means unresponsive to this month's decisions, because changing one takes a move, a refinance, a switch of insurer or a cancellation. Utilities sit awkwardly and are filed as fixed in the table below for convenience: the contract is fixed, the bill still moves with the season and with how much you use.

[Variable costs](/definitions/variable-costs) answer to decisions made inside the month: groceries, fuel, eating out, discretionary spending of every kind. This is where a mid-month correction is possible, and where a plan that is too tight fails first, because these are the only lines with give in them.

Periodic costs break plans built from twelve identical months: car maintenance, an annual renewal, birthdays. Predictable in total, irregular in timing. Divide the yearly figure by twelve and set it aside monthly, which is what a sinking fund is.

Fixed lines total $2,340, or 52 percent of $4,500 of take-home pay, leaving $2,160 that a decision made this month can still reach. Not all of that is spending money: $900 of it is the planned saving and $130 the periodic set-aside, so the genuinely discretionary part is smaller than the headline. The share is still the most useful number here: a household at 52 percent has steering room, and one at 80 percent lives inside contracts signed months ago.

The rule splits take-home pay three ways: 50 percent to needs, 30 percent to wants, 20 percent to saving and to debt repayment beyond the minimums. Minimum payments sit inside the 50. Elizabeth Warren and Amelia Warren Tyagi popularised it in a 2005 book, and its appeal is that it is one line of arithmetic with no category system to maintain.

Run the household above through it. Needs, meaning rent, utilities, groceries, transport, insurance, the two loan payments and the periodic set-aside, come to $3,135. That is 69.67 percent of take-home pay against a ceiling of $2,250. Wants are whatever is left once saving is out: eating out, hobbies and the gym, and the $45 of subscriptions, which the table files as fixed because the amount never moves and the rule files as a want because nothing depends on them. Fixed and needed are different questions. Those lines together come to just over 10 percent of pay against an allowance of 30, and saving lands on 20 percent exactly. The household hits the saving line and misses the other two by about twenty points in opposite directions.

Now raise the rent to $2,100 and change nothing else. Needs become $3,785, or 84.11 percent. Under 16 percent of pay is left for wants and saving together, so even a household that spends nothing on wants cannot reach the 20 percent line without cutting a need. The rule has not diagnosed anything. It has reported that housing is expensive, which the household already knew.

That is the general failure, and it is arithmetic rather than bad luck. The 50 is a claim about what needs cost, and what needs cost is set by local rents and by income. As income falls, essentials take a larger share by necessity, so the rule gets hardest to meet exactly where the pressure is worst. As income rises, 30 percent to wants stops being a limit worth having. And because the split runs on after-tax income, the same gross pay produces a different starting figure in different countries. Of the three numbers the saving line is the one that survives the criticism best, and even that is a convention rather than a figure anyone derived. Read the 50 and the 30 as description rather than instruction. The budget split calculator runs the three percents on a take-home figure.

Zero-based budgeting works from the opposite end. Rather than applying a split from above, you start with the income actually in hand and assign it line by line until nothing is unassigned. The name comes from corporate budgeting, where a department rebuilds its request from zero each cycle instead of adjusting last year's figure.

The zero at the bottom means fully assigned, not fully spent. Saving is a funded category like any other, and in the household above it is the $900 line, decided before the groceries figure rather than after it.

What the method really does is force the trade-off to happen in advance. A percentage rule sets the buckets from above and never prices the lines inside them against one another. A zero-based plan does: adding $45 to one line means taking $45 off another, because the total is pinned to income. It also closes the most common leak in a loose plan, which is that money assigned to nothing gets spent on everything.

The costs are real. It wants a session at the start of each month and a short check each week, and it wants re-planning whenever the month disagrees with the plan, which is most months. Moving money between categories mid-month is the method working rather than a breach of it: the rule is that the total holds, not that the first draft was right.

For irregular income there is a variant worth knowing. Spend this month from last month's income. Once a full month of expenses is banked, the amount to allocate is a known figure already sitting in the account rather than a forecast, which is what makes commission or freelance pay budgetable at all. Getting there takes one month of expenses saved first, a near relation of the starter buffer an emergency fund begins with, though sized on the whole plan rather than on essentials alone.

The saving rate is the residual as a share of income:

$$\text{saving rate} = \frac{\text{income} - \text{spending}}{\text{income}}$$

For the household above, $900 out of $4,500 is 20 percent.

Notice what is absent from that formula: any category. Two households on the same pay with the same saving rate build cash at the same speed, whether one spends heavily on rent and lightly on everything else or the reverse. The split describes how a household lives; the gap describes what it banks. Small lines matter only through the residual, and what decides how far any one of them moves it is how often it repeats: a change to a monthly line counts twelve times a year, a one-off counts once.

The rate does two jobs at once: it puts more in, and lowers the spending that has to be funded later. Ignoring investment returns, the years of work needed to bank one year of spending is $(1 - s) / s$:

Moving from 10 percent to 20 percent more than halves the wait, nine years of work down to four: the money going in rises while the target it chases falls. Returns shorten every row without changing their order, which is what compound interest does to a stream of deposits.

Read the table for what it measures, which is time against your own spending rather than the size of the pile. A high rate on a small income reaches the years sooner while still accumulating less money than a low rate on a large one, so how fast and how much are separate questions with separate answers, and the rate answers only the first.

Two qualifications. Measured on take-home pay, the rate misses a retirement contribution deducted before the money lands, so a payroll saver beats the figure shown. And repaying debt raises net worth without creating a liquid balance, which is as true of the principal inside a required payment as of anything paid above the minimum, so it is saving in one sense and not in the other. Pick which you count and stay with it.

Budgets fail in a few predictable ways, and each has a dull fix.

• Too many categories to maintain. Eight lines you keep up with beat forty abandoned in March.
• No line for periodic bills. The $130 set-aside exists so a tyre and a renewal notice are not emergencies.
• Planning from income that has not arrived. Allocate from the balance actually in the account, which is what the last-month-first variant above is for.
• No buffer. A plan with nothing spare in it fails on the first surprise, and a first surprise always turns up.
• One bad week read as proof the whole thing was pointless. Ten minutes of review a week catches a category running hot while there is still time to move money, which makes a bad week an adjustment rather than a verdict.

Two habits do most of the work on top of that. Move the saving on payday, before it is available to be spent, which is zero-based budgeting with the saving category funded at the top. And re-price the fixed lines once a year, because insurance, subscriptions and phone contracts drift upward quietly and one afternoon there changes twelve months of outflow.

One floor deserves measuring on its own: required debt payments, because they cannot be re-planned, and missing them costs money and, where credit reporting exists, marks the file lenders read later. How credit scores work covers the United States version. The debt-to-income calculator at the top of this page measures that floor the way a United States lender does, on gross pay rather than take-home, and on the numbers here the ratio is 32.33 percent. The 43 percent it is tested against in the worked example below is a widely quoted United States benchmark rather than a live rule: it came out of one version of the mortgage regulations, which no longer turn on a single ratio, and it has outlived them as a habit. Automated underwriting on mainstream loans goes above it where the rest of the file is strong, government-backed programmes higher again, and the only ceiling that decides anything is the one the lender in front of you applies. Other countries test affordability by their own rules, often by stressing the payment at a higher rate instead of ratioing it against gross pay.

Everything here is educational material rather than advice about your own circumstances, and every figure is illustrative. The arithmetic is general. The right split for your household is not.