At the end of July, the Supreme Court issued its decision in a case concerning the entitlement to annual leave, and pay for that leave, under the Working Time Regulations 1998 for workers who do not work a full year, otherwise known as “part-year workers”. The decision has considerable ramifications to organisations which employ term-time staff, or zero hours workers, casual or bank staff such as schools, universities, colleges, the care sector, retail and hospitality.
The facts in Harpur Trust v Brazel  UKSC 21 was that Mrs Brazel was a visiting music tutor who although engaged under a contract for the full year, only actually worked during term-time i.e. when there were pupils to teach. She was paid only for the hours she worked during term-time, and was not allowed to take holidays during term-time. As is fairly common in the sector she was treated as taking her holidays entitlement of 5.6 weeks in 3 equal amounts of 1.87 weeks at the start of each of the summer, winter and spring holidays.
Prior to 2011, her holidays were paid by looking at the hours she had worked in the 12 weeks prior to her holidays (but ignoring any weeks she didn’t work) and then working out what her average weeks’ pay was. The Trust however changed the method of calculation in 2011 and started by following the method recommended by ACAS, that is by saying that holiday entitlement accrued as hours are worked i.e. that she would be paid 12.07% of her total pay in any term in respect of the hours she worked and was paid for.
Mrs Brazel raised a claim for unlawful deduction of wages, claiming that the later method was not in accordance with the Working Time Regulations (“WTR”) and resulted in her receiving less holiday pay, which it did!
For a number of years, it has been common for organisations who engage zero hours workers, casual staff or individuals who only work part of a year (but are engaged on a contract for the full year) to pay holidays they are entitled to under the Working Time Regulations 1998 (WTR) by paying 12.07% in respect of the hours they are worked.
Confused? Well, let us explain.
Holidays under the Working Time Regulations 1998
Under the WTR workers are entitled to 5.6 weeks annual leave. For someone who works 5 days a week this is the same as 28 days. If someone works 3 days a week they are still entitled to 5.6 weeks but as a working week for them is 3 days, it equates to 16.8 days. However, what about someone who only works on some weeks and not others, as was the case with Mrs Brazel. The practice has grown up of calculating holiday pay as 12.07% of the hours they worked. This followed both ACAS and BEIS guidance at the time. Why 12.07%? Well, that’s because if someone gets 5.6 weeks holidays a year it means they in fact work for 46.4 weeks a year, and therefore 5.6 weeks as a proportion of those working weeks – 5.6/46.4 – is 12.07%.
Lots of organisations with people who work variable hours, or as in the case the Supreme Court heard, only worked during term-time (and even then the hours varied from week to week during the term) worked out the entitlement to annual leave (and pay) for leave on that basis.
For example if someone only worked 33 weeks then rather than get 5.6 weeks entitlement they would get 3.98 weeks (33/46.4x5.6).
The “conformity principle”
The principle underlying this was that if someone only worked for part of a year (because other than holidays they did not work every week) then they would not be entitled to 5.6 weeks but an entitlement based on the number of weeks they worked, and therefore this had to exclude any weeks they did not work. This is called the “conformity principle”. In essence their holiday leave entitlement was a pro rata proportion of the full time entitlement. Most people thought this was a fair way of apportioning holiday entitlement. Although as this case shows the law is not always about fairness!
The question for the Supreme Court was whether this is what the WTR provides, and if not, should it be interpreted as providing for that. The answer is a resounding NO.
The Supreme Court found that although the Working Time Regulations (WTR) were passed to implement the Working Time Directive (WTD), there was nothing to prevent the WTR from introducing a regime which was more favourable than the WTD, and that is what the WTR did.
Whilst the case law which has developed concerning the WTD has (with some limited exceptions) essentially confirmed the “pro rata” or “conformity” principle for holidays under it, the wording of the WTR provides (in some cases) for a more favourable entitlement.
Nothing in the WTR provided that the entitlement to the 5.6 weeks leave must be prorated based on the number of weeks they actually work in any leave year. That is different when someone starts or leaves partway through a leave year, and there is specific provision in the WTR that the holiday entitlement is prorated.
In addition, and most importantly for this case, the WTRs are clear on how a week’s holiday pay should be calculated for people whose hours vary from week to week, and that is by looking back at the last 52 weeks in which they carried out work (previously it was 12 weeks but this increased in 2020) and averaging what a week’s pay is. It is not possible to apply another method of calculation which, albeit may seem fairer, is not what is provided for in the WTRs themselves.
What does it all mean?
Well, the Supreme Court recognised that this could give rise to some strange results. For example an exam invigilator who is engaged on a permanent contract but only works 3 weeks a year would be still be entitled to 5.6 weeks holiday each holiday year, and their pay would be worked out by averaging out what they were paid. However, there is a limitation as in calculating the average over 52 weeks you only look back a maximum of 104 weeks i.e. 2 years. That means for the invigilator they would have received 6 weeks pay over the 2 years and this six weeks pay would be used to work out the average weekly wage, which would be multiplied by 5.6 weeks to give them the holiday pay they are entitled to.
Mrs Brazel was successful in her claim for unlawful deduction of wages. Such claims for underpayment of holiday pay can be brought, although legislation was introduced some time ago to limit backdated claims to 2 years. Of course, that can still be a pretty big liability.
Ultimately it means that unless Parliament changes the method of calculating entitlement to leave and how pay is calculated for that leave, employers have to apply the method in the WTR.
One important point to note that this case is significant for workers who are only entitled to the minimum annual leave of 5.6 weeks under the WTR. If an employer gives more holidays then, if they use a pro rata method of applying a percentage, it may well be there isn’t in reality an issue. For example, if someone gets 35 days holiday a year (7 weeks) then the percentage of pay will be higher as it will be 15.5% (45 working weeks i.e. 7/45 = 15.5%).
In some cases, such as term time working, where it is known how many weeks a person will work it will be relatively easy to work out what the correct percentage to apply will be to the 5.6 weeks entitlement. For example, in universities and colleges where a person works only the weeks students are in, then it is easy to work out how many working weeks there are for those people in a holiday year. It will inevitably be more difficult where there isn’t a set pattern say for relief workers or those on zero hours contracts where there is no predictable pattern.
This has some pretty big consequences for employers who engage staff and there are some immediate steps to take:
- Identify any workers who do not work every week, but who are engaged under a contract for the full year.
- Review the basis on which annual leave entitlement and pay was calculated.
- Undertake an assessment, if the 12.07% method was used, whether they have been underpaid.
- Ensure you change the basis of calculating going forward.
- Consider whether you will offer to reimburse any underpayments already made.
- Look at your systems to see how you can implement any required changes. This will not be straightforward as one of the reasons the 12.07% method was used was it could be incorporated into systems.