Working days in payroll and project planning
7 min read · Updated 14 August 2026
A month can hold anywhere from 18 to 23 working days. For fixed monthly salary that is invisible. For hourly payroll, capacity forecasts, contractual deadlines and invoicing, it is a spread of nearly 30 per cent between the lightest month and the heaviest — and it is entirely predictable in advance.
The monthly spread
Take Netherlands in 2026. The year holds 254 working days, an average of 21.2 per month. The actual distribution is not close to flat.
| Month | Working days | vs average |
|---|---|---|
| January | 21 | -0.2 |
| February | 20 | -1.2 |
| March | 22 | +0.8 |
| April | 20 | -1.2 |
| May | 19 | -2.2 |
| June | 22 | +0.8 |
| July | 23 | +1.8 |
| August | 21 | -0.2 |
| September | 22 | +0.8 |
| October | 22 | +0.8 |
| November | 21 | -0.2 |
| December | 21 | -0.2 |
July carries 23 working days against 19 in May. That is 4 days of difference, or 21% more available capacity in one month than the other. At quarter level the variation smooths but does not disappear: Q3 holds 66 working days against 61 in Q2.
Where this actually causes problems
Hourly payroll against a fixed monthly figure
If hourly staff are budgeted at a flat monthly cost, the budget is right twice a year and wrong the rest of the time. The error is not random — it is systematic, concentrated in the months with holiday clusters, and it compounds across a large workforce. Budgeting from the actual working-day count per month removes it entirely, and the count is known years ahead.
Capacity and delivery forecasting
A team forecasting at “20 working days per month” will overcommit in May and undercommit in July. Because the light months cluster around Easter and Christmas, the shortfall tends to arrive exactly when quarter-end deadlines do.
Contractual deadlines
“Within 30 working days” lands on a genuinely different calendar date depending on where in the year it starts, and on which country’s holidays apply. A period starting in late March in a Catholic-calendar country can run a full week longer in wall-clock time than the same period starting in July. Where a contract spans jurisdictions, specify whose working days govern.
Invoicing and revenue recognition
Time-and-materials billing tracks working days directly. A month with 19 billable days against a forecast of 21 produces a revenue gap of several per cent that has nothing to do with performance. Flagging the working-day count alongside the forecast makes the variance explainable before it is questioned.
The year-over-year trap
The most common planning error is carrying this year’s working-day count into next year’s model. Annual totals move, sometimes by several days, because holidays rotate through the week and weekend placement changes.
| Country | 2026 working days | Change from 2025 |
|---|---|---|
| Slovakia | 254 | +4 |
| Italy | 254 | +3 |
| Luxembourg | 254 | +3 |
| Slovenia | 254 | +3 |
| Croatia | 252 | +2 |
A three-day swing on a workforce of 200 people is roughly 4,800 hours of capacity appearing or vanishing between one year and the next. It is fully knowable in advance, and it is routinely missed because the previous year’s figure gets reused.
A practical method
- Pull the per-month figure, not the annual average. Build the monthly series for each country you operate in, for the year you are planning — not the year you are in.
- Decide your holiday definition once. Statutory public holidays only, or also customary days off and regional holidays? Write it down. Different teams silently using different definitions is a common source of reconciliation pain.
- Handle regional variation explicitly. For Germany, Spain and the United Kingdom, a national figure understates the holidays in some regions. If your staff are concentrated in Bavaria or Scotland, the national number is the wrong baseline.
- Layer leave on top separately. Working days are the gross figure. Annual leave, bridge days and sickness all reduce it further, and bridge days in particular cluster predictably.
- Re-check the year before you commit. Holiday legislation does change. Croatia restructured its calendar in 2020, Slovenia restored 2 January in 2017, Luxembourg added Europe Day in 2019.
Country-specific things that will surprise you
- Finland. Christmas Eve and Midsummer Eve are not statutory holidays but are almost universally non-working. Budgeting them as working days overstates capacity.
- Estonia and Lithuania. The working day is legally shortened — by three hours and one hour respectively — before certain holidays. The day counts as a working day but does not deliver full hours.
- Hungary and Latvia. Working days are officially relocated to Saturdays in some weeks. The annual total holds; the monthly distribution does not.
- Germany, Spain, United Kingdom. Regional holidays can add three or four days beyond the national baseline depending on location.
- Malta and Ireland. Weekend holidays convert into extra leave entitlement. The calendar does not change, but the leave liability does.
Every country page here carries the monthly breakdown, the quarterly totals and the year-over-year delta for the market and year you select. Start from the country comparison, and see how working days are calculated for the definitions behind the figures.
Related guides
- How working days are calculatedThe method behind a working-day count: which days are excluded, why two calculators disagree, and the edge cases that trip up payroll.
- Bank holiday vs public holiday"Bank holiday" is a British legal term that most of Europe does not use. What each country actually calls its days off, and why the distinction matters for payroll.
- Bridge days and long weekendsA bridge day is a single working day between a public holiday and a weekend. How to find them, what each country calls them, and how much leave they actually buy.