Payroll budgets drift for predictable reasons, and almost none of them are pay rises. They are timing, employer costs, the shape of joiners and leavers through the year, and variable pay landing in a different period from the one it was earned in. Building the forecast from employment cost rather than salary removes the largest error at the outset.
Budget employment cost, not salary
Gross pay plus employer contributions, levies, pension and the cost of benefits. That figure is meaningfully higher than salary and is the number that hits the accounts. Budgets built from salary alone understate by a consistent margin and produce a variance every month that somebody has to explain in the same way each time.
Model the timing of headcount
A role approved for the year that is filled in month seven costs less than half a year. Modelling joiners by expected start month rather than spreading the annual cost evenly is the single change that most improves a payroll forecast, and it also makes the recruitment plan visible in the budget.
Treat variable pay and leavers explicitly
Bonus and commission are earned in one period and paid in another, and accruals have to reflect that or the year end will move. Leavers carry final pay, untaken holiday and sometimes notice, which arrive as a lump rather than as a saving. Both are forecastable and both are commonly left out.
Questions people ask about payroll budgeting
Why does payroll always exceed budget?
Most often because employer costs were excluded and joiners were spread evenly rather than modelled by start month.
How should untaken holiday be handled?
As an accruing liability that crystallises when somebody leaves. Ignoring it produces a surprise in the month of a departure.
Where does the data come from?
Payroll for actuals and the HR system for the establishment. Forecasting from either alone leaves out half the picture.