How to count working days without getting it wrong

Working days sound like the least interesting number on this site. They are also the one people most often get wrong, usually by a margin big enough to matter: day rates, notice periods, project plans, holiday entitlement and pro-rata salaries all sit on top of it.

2026 has 261 weekdays. That is the ceiling, before any country's holidays are subtracted, and it is already not 260, which is the figure most people reach for by habit.

Trap one: holidays that fall on a weekend

A country with twelve public holidays does not automatically lose twelve working days. Some of them land on a Saturday or Sunday, where they cost nothing in working time.

This is why the working-day calculators on this site count holidays that actually fall on a weekday rather than simply subtracting the length of the holiday list. You can check the year for any country on its working days page, which lists the individual dates as well as the total.

A related complication: some countries move a holiday to the following Monday when it falls at the weekend, so the working day is lost after all. The United Kingdom and Ireland do this routinely; most of continental Europe does not. Where a substitute-day rule exists, a naive weekend check undercounts.

Trap two: movable feasts

Most public holidays sit on a fixed calendar date. A minority move every year because they are pinned to Easter, which is itself calculated rather than fixed. Good Friday, Easter Monday, Ascension, Pentecost and Corpus Christi all shift by up to five weeks between years.

Easter Sunday in 2026 falls on 5 April. The standard way to compute it is the anonymous Gregorian algorithm, and every holiday tied to Easter is then expressed as an offset in days: Good Friday is Easter minus two, Easter Monday is plus one, Pentecost is plus forty-nine.

Trap three: there are two Easters

This one caught us, and it is worth stating plainly because it is the kind of error that produces confidently wrong dates.

Countries with a predominantly Orthodox Christian calendar celebrate Easter on a date derived from the Julian calendar, which usually differs from the Western date, sometimes by more than a month. Running the Gregorian algorithm for Greece, Moldova, Albania or North Macedonia produces dates that look plausible and are simply not their holidays.

Rather than publish wrong dates, our dataset omits the Easter-derived holidays for those countries and counts only their fixed-date ones. Their working-day totals are therefore slightly high, and the country pages say so. An acknowledged gap is more useful than a confident error, and this is the sort of thing we would rather have on the record than quietly approximated. The reasoning behind that policy is on the methodology page.

Trap four: regional holidays

Germany is the standard example. Its public holidays are largely set at state level, so Bavaria observes several days that Berlin does not. Spain layers national, regional and municipal holidays. Switzerland varies by canton. Any single national figure for these countries is a simplification, and for payroll purposes the relevant calendar is the one for the specific location of employment.

Putting it to use

The two calculations people actually need this for:

• Salary to day rate. Divide annual gross by the working days in the country, not by 260 and not by 365. In Estonia in 2026 that divisor is 252.
• Day rate to annual income. Multiply by the days you expect to bill, not by the working days available. Holiday, illness and gaps between contracts are yours to absorb, which is why the freelance floor in the employer cost guide is built on billable days rather than calendar ones.

And if you are converting between monthly salary and an hourly rate, the length of the standard week matters as much as the number of days, which varies from 35 to 48 hours depending on where you are. That is covered in the guide to standard working hours.