Skip to main content

Harpur v Brazel

This case recently went to the Supreme Court (the highest court in the United Kingdom for civil claims) and, understandably, it has generated a lot of interest as it has implications for many zero hours employees as well as term-time only employees.  This is a complicated area of law as the holiday pay calculation methods for workers with no fixed hours changed in April 2020 under the Good Work Plan, causing a further administrative burden for employers.  Here, we will look at what the changes mean.

Ms Brazel was a music teacher who was engaged by the Harpur Trust in a term-time only basis.  This means that the employee only works for part of the year, but there is an expectation that they will recommence employment each year (these contracts are common for teachers, as they are not contracted to work during the school holidays).   Her employer based her holiday pay on the weeks she actually worked, and pro-rated the 5.6 weeks of annual leave entitlement down to 4.7 weeks, as she was only engaged to work for 39 weeks per year.  However, she argued that this was not correct, and that all workers were entitled to 5.6 weeks of leave per annum, irrespective of the amount of weeks actually worked.  Initially, she lost her claim, but the Employment Appeal Tribunal allowed it to progress.  After going through the Court of Appeal, and then the Supreme Court, the ruling is that all workers are entitled to 5.6 weeks of leave.

The Supreme Court agreed with the Court of Appeal and the Employment Appeal Tribunal that an Employment Tribunal (the initial court) had been wrong to read words into the Working Time Regulations 1998 to cap those workers’ annual leave entitlement at 12.07% of annualised hours.   ACAS used to advise that the 12.07% method was acceptable, and this was used across the board by many employers, but ACAS changed their guidance and removed this calculation method.

The Working Time Regulations


The Working Time Regulations 1998 allow for holiday to be pro-rata for part-time employees, for example, and the government’s own holiday calculator online (at also allows for these calculations to be carried out.  The issue comes because there is no way to give any worker less than 5.6 weeks’ holiday a year.  As an example, if somebody works 3 days per week then this can be calculated as 5.6 x 3 = 16.8 days of annual leave – as they are still receiving 5.6 weeks of their “normal” working week.  But with atypical workers, this is not so easy to calculate and you are still required to give them 5.6 weeks of leave, based on weeks worked.  5.6 is the magic number to remember in all situations.

In theory, this means somebody could be a seasonal worker, working 4 months a year, with a continuing zero hours contract, and return each year (and be expected to return) and still receive 5.6 weeks of holiday.  This would be the same as someone who worked all year round.  There are laws to prevent part-time workers being treated less favourably than their full-time counterparts, but there are no laws to prevent them from being treated more favourably, as unfair as this may seem.

There should not be any practical impact for term-time workers whose salary is annualised and paid in 12 equal monthly instalments, because they already receive the same pay during weeks of holiday as during working time, usually dictated in the employment contract. 

The Impact

The biggest impact in practice is on hourly paid workers who have a permanent contract but, for whatever reason, have a number of unpaid non-working weeks during the leave year. The greater the number of non-working weeks, the greater their holiday entitlement (and therefore holiday pay) will be as a percentage of annual working time and earnings.

For truly casual workers, this isn’t an issue as there won’t be a continuing contract, they will work for an assignment and then in practice be paid their holiday entitlement and pay at the end of the assignment, and receive their P45.

Calculation Method

The only reliable calculation method is now to use the Calendar Week Method. If the worker takes a week’s holiday, they should be paid a week’s pay calculated according to the statutory formula. Employers will need to go back at least 52 weeks to establish all the weeks worked in that period, and go further back, if there are non-working weeks, to establish 52 paid weeks on which to base holiday pay.  If there aren’t 52 weeks of pay as somebody hasn’t worked long enough, you would use all the weeks worked to establish the “normal” rate of pay for them.  This may produce a different rate of pay each time holiday is taken, and will be very time-consuming.  

However, there is still a difficulty in expressing the holiday entitlement in terms of days or hours, which is not covered in the WTR 1998 (as it probably didn’t anticipate the boom of the gig economy when it was brought into law).  One possible solution is to base it on the number of days in an average week over a representative period.

Practically, a way to manage this would be ensuring that zero-hours workers are given at least some work every week when they are not taking annual leave, to reduce the value of holiday pay. In other cases involving term-time workers, it may be possible to engage the worker at the start of term on a fixed-term contract for the term, give a week’s paid holiday during half term, and a payment in lieu of any outstanding accrued holiday at the end of term, before re-engaging again after the school holidays. However, this could cause a number of issues around re-administrating the employee joining the company again, whether continuity is preserved between contracts (as there is the expectation to return to work again, so arguably an umbrella contract may exist) not to mention a successive series of fixed-term contracts could extend beyond 4 years, and qualify for additional statutory protection.

Claims for unpaid annual leave have a backstop of 2 years, and are subject to the same time limits as other claims.

Tell us about your HR needs