Here is a new case which you think initially might be quite helpful on the calculation of holiday pay, but which then suddenly veers off into the contractual undergrowth, and actually isn’t. However, what it does do is administer a sharp lesson about the wisdom of trying to incorporate broad principles into individual employment contracts.

In Flowers and Others –v – East of England Ambulance Trust the EAT had to consider whether voluntary overtime should be included in the calculation of holiday pay, i.e. overtime which the employer was not obliged to offer and the employee was under no obligation to do. It had to consider this from two perspectives, the underlying Working Time Directive requirement and the terms of the relevant employment contract. These stated

Pay during annual leave will include regularly paid supplements including….payments for work outside normal working hours….Pay is calculated on the basis of what the individual would have received had he/she been at work. This would be based on the previous three months at work…

Unsurprisingly, the EAT found as a point of principle that voluntary overtime should be included in the calculation of holiday pay if it had been paid over a period of time sufficient for it to become “normal” for Working Time Directive purposes or “regularly paid” as referred to in the employment contract of the staff concerned. This was “a question of fact and degree”, said the EAT, little assisted by the obvious but not immediately helpful comment from Bear Scotland in 2015 that “normal pay is that which is normally received”.

This is hopeless stuff for employers. Except at either extreme of voluntary overtime working (all the time or almost never), who knows whether they are in or out, whether that position varies with each employee or with varying working patterns over the course of a year, or even whether two Tribunals considering the same patterns on the margins could reach different decisions, each equally unappealable?

So now, that principle established, the question will go back to the Employment Tribunal for an assessment of whether the voluntary overtime of each of Mr Flowers and his colleagues was actually “regularly paid” within the meaning of that contract clause. Even then, there may still be a tension between the remaining parts of that clause – what the individual would have received had he/she been at work is not necessarily the same figure as his/her average over the preceding three months. It may be so where there is little fluctuation in demand for overtime working (perhaps in an Ambulance Trust, I don’t know), but in workplaces affected by seasonal considerations or peaks and troughs in customer demands, certainly not.

If you apply the three months backwards average in those cases you can come up with some very odd results. Someone who takes holiday when a busy period slackens off will receive credit for overtime which he would not have carried out if he had been at work over the period of his leave. By contrast, someone who takes times off just as it gets busy will see no credit for the voluntary overtime he/she undoubtedly lost out on by going on holiday.

In Flowers the EAT decided that the appropriate test was the three month back-average and not the more legally appropriate but possibly harder-to-determine what the employee would have earned if he/she had not been on leave. The effect of that last sentence in the contract term was therefore to take away from the relevant employees the burden of showing that they had actually suffered any loss by going on leave. Instead the assessment of their pay was converted into a simple formula. This makes for easier administration, yes, but since you can’t contract out of the Working Time Directive, it remains to be seen if that arrangement would withstand a challenge from our second class of employee above, i.e. someone who would have earned more if at work than that formula would provide. Other than by rolling over altogether and including allowance for all overtime, however sporadic, any attempt to convert principles into formulae will always carry that risk.