How different cultures actually use vacation time

Long haul versus short bursts

When I compared the duration of vacation requests among the U.K., U.S., Germany, and France, distinct cultural patterns emerged. Germany and France are the kings of the “long holiday.” After excluding one-day leaves (the majority of leave requests), 1.3% of French and 1.8% of German approved leaves were for 16 or more days, compared to 0.6% for the U.K. and 0.5% for the U.S. American employees are much more likely to request two-day leaves but less than half as likely to request ten-day ones as the U.K.

The North American allergy to sick days

When it comes to sick leave, North Americans are seemingly "healthier"—or at least, they take less time off for illness. 39% of European workers in the sample took sick time in 2025, compared to only 29% of U.S. and Canadian workers (p < 0.0001). When restricting the sample to only those with fixed vacation days to ensure that we exclude workers taking sick leave under their flexible vacation policy, Europeans are still much more likely to take sick leave (30%) than the U.S. and Canada (25%, p < 0.05)

However, the duration of sick leave varies wildly by country. France stands out for its high percentage of longer-term (11 or more days) sick leave requests. One potential explanation: France is the only country of the four to require certification from doctors for any sick leave request (the U.K. and Germany allow employees to self-certify up to 7 and 3 calendar days, respectively, while the U.S. leaves policies up to the employer), so people may be less likely to request sick leave for less serious illnesses.

The British and Irish take their Christmas leave very seriously

Finally, since we’re only weeks away from Christmas, I examined leave patterns around the holiday. Christmas is a public holiday in every country we analyzed, but most of the dates around it are not. I looked at the percentage of workers taking at least three annual leave days around Christmas this year—after December 22 and before January 2 (inclusive). Because most European countries also recognize Boxing Day as a holiday, an extra three leave days will result in a week’s holiday.

The U.K. and Ireland love a proper Christmas break. Despite their relatively lower leave entitlements, 28.8% of Irish workers and 27.0% of British ones already have their three-day holiday booked. Since Christmas and Boxing Day are automatic bank holidays for most, that means they’ll have a full working week away!

Cooks, process operatives, and office managers were the most likely to have to use annual leave for Christmas Day, rather than receiving it as an automatic holiday, highlighting that even on “universal” holidays, some jobs don’t automatically get the ability to disconnect.

This analysis complements our economist brief on the unlimited vacation paradox and the transatlantic vacation gap.

Methodological details

I analyzed 160,000+ approved and used time off requests for vacation and sick leave from over 18,000 employees for thousands of companies across the European Economic Area, Switzerland, the U.K., the U.S., Canada, and Mexico in 2025 (leave is grouped into a year based on the first day of the leave). All data is current as of December 4, 2025. Leave that has already been approved for December is also included. Only workers who have been present on Deel’s time off platform for the entirety of 2025 were included. Workers are assigned to countries based on the country of their contract.

Vacation leave includes paid leave policies such as annual leave, RTT days, vacation days, paid time off, marriage/wedding leave, moving leave, and personal leave. Sick leave is paid and includes disability leave, critical illness leave, paid sick days, blood donation leave, and doctor’s appointment leave. All home-working, remote work, working abroad, TOIL, unpaid leave, and overtime requests were excluded from the analysis (as they don’t represent paid time away from work). Note that employees could take days that they had rolled over, or could take less time than they were entitled to in order to roll over vacation days.

Leave could be taken in increments of hours, weeks, months, calendar days, or business days. The majority of leave was taken in business days, but all other leave was converted to business days according to the following formula:

• Weeks: multiplied by 5 (number of business days in a week)
• Years: multiplied by 261 (there are between 260 and 262 working days in a year, but it’s usually 261)
• Months: multiplied by 21.75 (261 working days/12 months)
• Hours: multiplied by between 7 and 8.4, depending on the standard for the country (e.g., French working days are 7 hours, but U.S. working days are 8)
• Calendar days: multiplied by 1 (most calendar day requests appeared to be taken in a manner that was similar to business days, apart from in Estonia)

I excluded all known part-time employees and independent contractors from the analysis. Sometimes days were taken in increments (e.g. a half day or a part day for sick leave); these were added to full-day increments. Most calculations were median days off: this was calculated by summing 2025 leave by policy type (vacation or sick leave) and worker, then taking the median. Statistical significance for medians was calculated using a Mann-Whitney U test and a t-test of proportions for frequency differences.

Flexible leave was defined as having access to an unlimited number of days off on Deel’s platform. Fixed leave had both an accrued number of days and the ability to take only a fixed number of days beyond that in a given month (since time off often accrues each month, some employers gave their employees the ability to take a certain number of days off beyond their accrued amount). Measures of entitled time off don’t include days rolled over.

When comparing U.S. and Canadian leave, I ran a panel regression of U.S. & Canadian workers with company-level fixed effects. I set the number of days taken off as the dependent variable and a U.S. worker dummy, broad job group dummies, a dummy for non-EOR/PEO contracts, and “normalized” salary as the independent variables. I defined normalized salary as individual salary in USD divided by the 2024 median salary for full-time employees in the country (found here for Canada as $65,520 CAD or $47,401 USD and here for the U.S. as $61,620 USD), converted to USD. The use of this normalized metric was necessary to account for baseline differences in compensation between the two countries. The resulting coefficient for the U.S. dummy variable was strongly positive and statistically significant (p < 0.001), indicating that U.S. employees take significantly more time off, even after controlling for organizational factors, contract type, job function, and relative compensation level.