Global menu

Our global pages


Education briefing – Court of Appeal decision on calculating holiday pay for term-time only staff

  • United Kingdom
  • Education - Briefings


The Court of Appeal has confirmed that an employer was wrong to cap holiday pay at 12.07% of annualised hours for a zero hour contract employee working on a term-time only basis. Instead, it should have based it on average earnings over the 12 week period immediately before the relevant leave was taken, even if doing so provided more favourable results for term-time workers when compared to full-timers.


The calculation of holiday pay for workers who have no normal working hours can present logistical difficulties for institutions. Strictly speaking, the statutory requirement is that payment for each period of holiday taken has to be calculated based on the average pay which the worker has received in the previous 12 weeks of work.

This is because section 224 of the Employment Rights Act 1996 (ERA) provides that, for employees with no normal working hours, the amount of a week's pay is the amount of the employee's average weekly remuneration in the preceding twelve weeks they have worked. In arriving at this average, any weeks in which no remuneration was payable by the employer to the worker are ignored and replaced by earlier paid weeks, to ensure that the calculation is always based on 12 paid weeks.

For convenience, however, many employers have historically calculated the holiday pay as an additional 12.07% of the relevant hourly rate for every hour worked (an approach previously endorsed in ACAS guidance).

However, in our briefing of 16 March 2018 we reported the Employment Appeal Tribunal (EAT) decision in the case of Brazel v Harpur Trust which demonstrated that this calculation is not always reliable and may result in underpayments to workers who are employed on a casual or zero hours basis and do not work the full year, such as term-time or seasonal workers. We pointed out that this could create exposure to claims for unlawful deductions and historic underpayments, and recommended that institutions should review their approach in light of the decision.

The Trust appealed the EAT decision but that appeal has been unsuccessful with the Court of Appeal confirming that using the figure of 12.07% in such circumstances may well result in an under-payment.

Facts of the case

Mrs Brazel was one of a number of visiting music teachers employed by the Trust. She worked at Bedford Girl’s School (the School) under a zero hours term-time contract. Her contract of employment stated “As a visiting teacher, requirements for your services will depend upon a varying level of demand for individual personal tuition in your subject/instrument. Demand may vary from term to term. There are no minimum hours of work guaranteed to you and you have no normal hours of work.” She was paid an hourly rate of £28.77.

Mrs Brazel was entitled to 5.6 weeks paid holiday a year, which she had to take during the normal School holidays or at such other times as were convenient for the School. A payment for accrued holiday was made to Mrs Brazel 3 times a year - at the end of April, August and December – and calculated at 12.07% of her hourly rate for the hours she had worked in each term.

Mrs Brazel argued that her holiday should have been calculated on the average weekly hours she had worked in the 12 weeks immediately preceding the end of each of the three terms – given that her holidays could only be taken outside term time.

This calculation equated to a rate for holiday pay of 17.5% for each hour. On this basis Mrs Brazel claimed that she had been underpaid by £1,360.72 since 2011. The School contended that its calculations were correct because it had used the 12.07% figure suggested at the time in the ACAS guidance.

Employment Tribunal decision

Mrs Brazel brought a claim of unlawful deduction from wages. The Employment Tribunal dismissed her claim holding that as she worked for only part of the year her entitlement should be pro-rated and this could be done by applying the 12.07% calculation. The Tribunal agreed with the Trust that to do otherwise would unfairly reward part-time workers, as a worker working 32 hours a week would receive a higher rate of holiday pay than a worker who worked the same irregular hours pattern over a “standard” 46.4 week working year (i.e. 52 weeks less 5.6 weeks statutory leave).

The Tribunal thought that words should be read into to the WTR to state “where a worker has no normal hours and works fewer than 46.4 weeks per year any such payment should be capped at 12.07 per cent of annualised hours.”

EAT decision

Mrs Brazel appealed to the EAT, which agreed with her and overturned the Tribunal decision. In reaching its decision the EAT concluded that:

• The exercise which the Tribunal had to carry out was a relatively simple one as Mrs Brazel’s entitlement to holiday pay was set out clearly in her contract.

• As Mrs Brazel worked irregular hours the application of section 224 of the ERA enabled a week's pay to be computed in a simple and straightforward manner using the 12 week averaging method.

• Whilst that approach could favour someone who does not work throughout the year, the EAT could not see how that justified words being read into the WTR.

• The Tribunal had overlooked the fundamental fact that, whilst part-time workers are not to be treated less favourably than full-time workers, there is no principle to the opposite effect.

• To impose a limitation which altered the ERA to the disadvantage of part-time workers, ostensibly to redress a potential grievance that might be brought by full-time workers, stood the logic and purpose of the legislation on its head.

Court of Appeal decision

At the Court of Appeal, the Trust again argued that to calculate holiday pay in the way put forward by Mrs Brazel was inequitable as it meant that she would be entitled to a much higher proportion of her actual earnings (17.5%) than if she worked full-time (12.07%) - with the result that someone in her position, who worked only 32 weeks of the year, was entitled to holiday pay calculated on the same basis as if she had worked 46.4 weeks.

The Trust submitted that this “inequity” could be avoided by pro-rating the calculation to reflect the number of weeks in the year that Mrs Brazel actually worked.

In support of this argument, the Trust claimed that to interpret the legislation as argued by Mrs Brazel would produce even greater anomalies in the case of employees with, like her, permanent but zero-hours contracts who worked for even smaller proportions of the year - for example a school cricket coach, who would only work for one term, or invigilators, who worked only during the exam season. In principle, the Trust said, a permanent employee who worked only one week of the year, for which he or she earned, say, £1,000, would then be entitled to 5.6 weeks (notional) annual leave, for which they would receive £5,600. The Trust argued that this could not have been the intention of the WTR.

The Court of Appeal rejected this argument on the basis that “on any natural construction” the WTR makes no provision for pro-rating and that the calculation required by the legislation “is straightforward and should be followed”. Whilst it accepted that applying the terms of the WTR without a pro-rata reduction for part-year workers would produce odd results in extreme cases, it noted that “general rules sometimes produce such anomalies when applied in untypical cases”. In any event, the Court said it would expect it to be unusual for workers whose services are required for only a few weeks a year to be employed on permanent contracts, as opposed to being engaged on a freelance basis.


It is worth noting that the ACAS guidance on calculating holiday pay now states “If you do casual work with no normal hours, for example on a zero-hours contract, your holiday pay will be based on the average pay you got over the previous 12 weeks. These should be weeks in which you were paid. If you were not paid in one of those 12 weeks (because you did not work), the last paid week before that should be used to calculate your holiday pay”. In addition, the BEIS guidance, published in February 2019, on calculating holiday pay for workers without fixed hours or pay, refers to using a 12 week holiday pay reference period in calculating holiday pay for term-time workers.

Using the 12.07% accrual rate has been an attractive option for institutions where individuals work irregular hours. However, as this case demonstrates, that leaves the institution open to a potential claim because the method does not necessarily comply with section 224 of the ERA.

Whilst, on the face of it, the figure of 12.07% is a logical way of calculating the average pay over a period of time where there are no normal working hours, it is based upon an assumption that the person is potentially working 46.4 weeks a year. If, like Mrs Brazel, they work term-time only, then their working weeks are compressed into a shorter period of time (in her case 32 weeks) and the ratio between working weeks and weeks of holiday is different. As a result, the 12.07% accrual method is unlikely to reflect average pay over the preceding 12 week period.

This decision will only have implications for those workers who have no normal working hours and particularly (but not exclusively) for those who work on a term time basis or intermittently throughout the year. Institutions should consider whether they have exposure to underpayments for any workers in this category and, if so, whether to take any corrective action in respect of past and/or future payments.

One practical issue will be the difficulty of calculating holiday pay on an individual basis every time a zero hours or casual worker takes holiday. In the Brazel case, holiday was deemed to be taken outside term-time, so the statutory calculation could be made 3 times a year based on average pay over the previous 12 weeks. So if institutions engage casual term-time only staff on contracts which similarly control when holiday is deemed to be taken, the calculation may be more straightforward.

If the contract, however, permits casual staff without normal working hours to take holidays during term-time or at times of their choosing, then a separate calculation would need to be made on each occasion holiday is taken.

It should also be noted that, from April 2020, the reference period for calculating holiday pay for workers with no normal working hours will be extended to 52 weeks or (if the worker has been employed for a shorter period, the number of complete weeks for which the worker has been employed). Under the extended reference period, weeks in which no remuneration is payable by the employer to the worker are still ignored, save that no account is taken of remuneration more than 104 weeks before the calculation date. Given that this longer reference period also requires weeks for which no remuneration is paid to be discounted, the same risk of underpayment will arise if holiday pay is calculated at 12.07% for each hour worked.

We’d be very happy to help with the assessment of exposure to claims and the options for dealing with the impact of this decision.

For more information contact

< Go back

Print Friendly and PDF
Subscribe to e-briefings