How to budget on an irregular income
By Luigi PooleUpdated
Budget from a baseline month — your lowest realistic income — and pay yourself that amount as a monthly salary from a buffer account where all income lands. Good months fill the buffer, lean months draw on it, and windfalls get split by a rule you set in advance.
Most budgeting advice assumes a number you do not have: next month's income. A salaried employee can split a known paycheque into percentages; a freelancer, contractor, or commission earner cannot, because the paycheque is the thing that varies. When deposits swing between $1,500 and $7,200, a budget built on the average is wrong in most months — too loose in the strong ones, broken in the lean ones — and a budget rebuilt from scratch every month is not a budget, it is a running improvisation.
The fix is not better forecasting. It is separating what you earn from what you spend, so that your spending stops tracking your invoices. Three moves accomplish that: set the budget from a baseline month, pay yourself a fixed salary out of a buffer account, and decide in advance — by rule, not by mood — what happens to the money that stacks up when work is good.
Set the budget from a baseline month
The baseline is your lowest realistic monthly income — not your worst month ever, and certainly not your average. Pull the last year of deposits and sort them. Discard the true outliers in both directions — the month a client went dark, the month two projects happened to pay at once — and find the figure you cleared in ten or eleven months out of twelve. That is the number your essential spending has to fit inside.
The average fails for a structural reason: by construction, income lands below it in roughly half of all months. A budget set at the average is a plan for going backwards every second month and repairing the damage in between. The baseline flips this. Size the essentials — housing, groceries, insurance, minimum debt payments — to fit within the baseline, and a lean month becomes an on-plan month rather than a crisis. Everything above the baseline is real money, but it is not spending money by default; it has standing jobs, covered below.
If you are new to variable income and have no year of history to sort, run the logic in reverse: total your essential costs, set the baseline there, and treat every dollar above it as overflow until enough months accumulate to tell you what your floor actually is.
Percentage frameworks still have a place in all this — just not applied to income. The 50/30/20 rule works fine on an irregular income once it is pointed at the salary you pay yourself rather than at whatever happened to land this month.
Pay yourself a salary from a buffer account
Two accounts make the system run. Every dollar you earn — every invoice, payout, and commission cheque — lands in a buffer account: a separate chequing or savings account you never spend from directly. Then, on a fixed date each month, you transfer a fixed salary to your personal account, and live on that.
The salary is the baseline you just set, or slightly below it while the system is new. The date matters less than its fixedness — the first of the month, the day before rent, whatever maps onto your bills. What you have built is an employer of one: the buffer absorbs the chaos on the earning side, while the personal account sees a payroll as steady as any staff job. Every ordinary budgeting technique — categories, percentages, automatic transfers — starts working again, because the number they all depend on finally holds still.
Run your numbers50/30/20 budgetThis is also where discipline becomes cheap. You are not resisting spending in a good month through willpower; the good month's money simply never arrives in the account you spend from.
Six months through the buffer
Here is the system running. The freelancer below pays themselves $4,000 a month, starts with $6,000 in the buffer, and has a genuinely lumpy half-year — the worst month brings in about a fifth of the best.
| Month | Income into buffer | Salary drawn | Buffer at month end |
|---|---|---|---|
| 1 | $5,500 | $4,000 | $7,500 |
| 2 | $2,800 | $4,000 | $6,300 |
| 3 | $7,200 | $4,000 | $9,500 |
| 4 | $1,500 | $4,000 | $7,000 |
| 5 | $3,900 | $4,000 | $6,900 |
| 6 | $6,100 | $4,000 | $9,000 |
Income over the six months totals $27,000 — an average of $4,500 — while the salary drew $24,000, so the buffer grew by $3,000, from $6,000 to $9,000. That $500 monthly gap between average income and salary is the engine of the whole system: it is what refills the buffer after months like the fourth, when $1,500 came in and $4,000 went out. And notice what the fourth month felt like from the personal account's side: nothing. Rent cleared, groceries were bought, the salary landed on schedule.
| Income into the buffer | Salary drawn | |
|---|---|---|
| M1 | 5500 | 4000 |
| M2 | 2800 | 4000 |
| M3 | 7200 | 4000 |
| M4 | 1500 | 4000 |
| M5 | 3900 | 4000 |
| M6 | 6100 | 4000 |
Same six months as the table: deposits range from $1,500 to $7,200 while the salary stays at $4,000.
The trend line is also the honest test of the salary level. A buffer drifting downward across six months means the salary is set above what the work actually earns, and no buffer size fixes that — cut the salary or raise the income. A buffer drifting upward, as here, hands the surplus to the overflow rule below.
Size the buffer
One month of salary is the minimum at which the system functions: the salary you draw this month was earned last month, and a single slow stretch no longer reaches your rent. Two to three months of salary is the comfortable range for typical freelance variance — enough that two weak months back to back are absorbed without drama. Strongly seasonal income needs more than that, because a wedding photographer or a tax-season accountant is not smoothing noise; they are bridging a predictable off-season, and the buffer has to carry most of it.
The buffer is not your emergency fund, and keeping them in separate accounts is worth the small extra admin. The buffer absorbs normal variance — the ordinary difference between a strong invoicing month and a slow one. The emergency fund covers abnormal events: a major client disappearing, illness, the laptop dying the week a project is due. Self-employment argues for a larger emergency fund than a staff job does, since the income and the emergency can arrive together — how big it should be is its own question, and the emergency fund calculator will put a number on yours. The two funds also back each other up in a defined order: if the buffer runs dry, the emergency fund is the next line of defence; if the emergency fund is being tapped for ordinary slow months, the baseline was set too high.
Split the good months by rule
Money above the salary needs standing orders, written once, in a calm month — because the alternative is renegotiating with yourself every time a large invoice clears, and the large-invoice version of you is not the careful one.
Two skims come off the top of every deposit before anything else. First, tax. Nobody withholds for you now, and the CRA will expect instalments once your bills are large enough — so skim a fixed percentage of every deposit into a separate tax account the day it arrives. Your most recent return tells you what share of gross income the year's total tax took, which is the honest starting rate; round it up, never down. Second, the lumpy annual costs — insurance premiums, software renewals, professional dues, equipment replacement — which are exactly what sinking funds exist for. A sinking-fund calculator converts each one into a small fixed monthly amount, which the salary then carries like any other bill instead of meeting as an ambush.
After the skims, deposits fill the buffer to its target. Above a ceiling — the target plus one month of salary is a sensible line — the excess gets split by a fixed rule: perhaps half to long-term investing, a third to named goals, the remainder spent freely as the reward for a good month. The exact percentages matter far less than the fact that they were chosen in advance. A windfall with no standing instructions defaults to lifestyle; a windfall with a rule builds something.
Raise the salary slowly
The system has one moving part — the salary — and it should move rarely and asymmetrically. Raise it only after the buffer has sat above its ceiling for three or four consecutive months, because that is evidence the income genuinely stepped up rather than merely had a streak. Cut it early and without ceremony whenever the buffer trends down, since a small cut taken soon beats a large one taken late. The asymmetry is deliberate: irregular income punishes optimism and forgives caution, and a salary you raise slowly is a salary you almost never have to cut.
Common follow-ups
How big should the income buffer be?
+
One month of your self-paid salary is the working minimum; two to three months is comfortable for typical freelance variance. Strongly seasonal income needs more, because the buffer is bridging a predictable off-season rather than smoothing noise. Keep it separate from your emergency fund, which covers abnormal events.
Should the baseline be my average income?
+
No. Income falls below the average in roughly half of all months, so a budget set there fails half the time by design. Use your lowest realistic month — the figure you cleared in ten or eleven of the last twelve — and treat everything above it as buffer and goals.
What happens if the buffer runs out?
+
Cut the salary immediately and treat the situation like a lost job — pause the extras, and lean on the emergency fund only if essentials are at risk. A buffer that empties during ordinary slow months is a sign the salary was set above what the income actually supports.
How do I handle taxes with no employer withholding?
+
Skim a fixed percentage off every deposit into a separate tax account the day it arrives, before the money counts as income you can pay yourself. Your most recent return shows what share of gross income the total tax bill took, which is an honest starting rate.
Does this work for commission and gig income, not just freelancing?
+
Yes — the system only assumes income varies, not why it varies. Commission earners and gig workers often have a steadier floor than project freelancers, which means a smaller buffer does the same job. The steadier your worst month, the closer the salary can sit to your average.