The Employment Appeal Tribunal, as everyone now knows, has decided that certain types of overtime earnings should be included in the calculation of holiday pay. However, the blanket press coverage of that decision has to some extent obscured the progress of another holiday pay case, Lock –v- British Gas, which is dealing with the same basic issue (should holiday pay be limited to basic salary only?) but in relation to commissions rather than overtime. In Lock the ECJ decided in effect that an employee should not earn less during a period of leave than he would have done had he not taken that holiday. The Leicester Employment Tribunal (to which Lock has now returned from its jaunt to Europe) is to decide how that basic principle is to be implemented, presumably an average of some sort over a given reference period. But will that really work? Here are my thoughts on that point:
In considering the calculation of holiday pay (and the extent to which an amount in respect of commissions needs to be included) the Tribunal needs to determine whether the holiday absence has had any impact on earnings, and if so, by how much.
In Lock it is pretty clear that the employee did suffer loss – by the nature of his work he made lots of small deals which were mechanically linked to payments received shortly thereafter, and it would be relatively easy to show from the history of his deals before and after the holiday period that he would have done approximately X deals with £Y financial consequences over the holiday period. In such circumstances, where there is a regular flow of very small transactions, the precise length of the proposed reference period is not very important – it would always produce approximately the same answer.
However this will not be the case in businesses subject to seasonal fluctuation or those which operate commission schemes which are “lumpier” and/or less mechanical in their outputs, for example where the figure is driven part by sales and part by a discretionary assessment of a broader contribution by the employee. I think that it would be inappropriate to include in a Lock-type calculation any figure which could not be seen as a direct ratchet off the notional sales not made in the absence period.
However, even this is a more difficult question than it looks. What certainty could the employer have, for example, that the employee had not benefited from the transactions which would have fallen due over the holiday period? Perhaps he worked extra-hard in advance of the holiday to clear his desk of them, or caught up with them after he returned (in each case then receiving the commission that he would have earned had he carried them out over the holiday period).
That introduces a further complexity, i.e. that the “loss” to the employee will also be a function of the length of the holiday. An absence of one day is in the majority of cases most unlikely to dent an employee’s commission earnings in any way, suggesting that it should be paid at basic salary only. However, an absence of, say, a week or more may create a meaningful impact on what he could achieve during the relevant month. This creates the extremely unattractive prospect that the daily rate for short holidays should be less than that for longer absences.
It is therefore my view that even the relatively simple circumstances of the British Gas commission scheme in Lock create questions which would need to be answered before it would be appropriate to apply a reference period to assess average commission earnings. In particular, it seems to me that it would be incumbent on the employee to establish first that he had suffered some financial loss through not being in the office over the relevant period. That loss should have to be proved empirically, not statistically. Only such a system would permit those who had genuinely lost out to be paid appropriately but also prevent those who could not point to any loss arising from their leave being rewarded without justification.
The obligation to show actual loss would greatly reduce the risk of manipulation of recent prior earnings by the employee prior to his absence on leave. It would also minimise the risk of employers being required to take into account as “average” figures which were one-offs or irregular only. An employee who does a handful of deals each year, perhaps selling very high-value items like art, building designs, aeroplanes, ships, etc., might receive no commission for many months but then a very substantial lump sum when a deal closes. Taking any account of that lump sum during leave in the following months (whether three or six or twelve) pays no heed to the more or less total certainty that he would not have received any commission specifically in respect of the holiday period anyway.
You could argue that the employer could always refuse consent to the taking of leave at a time which would artificially inflate the employee’s holiday pay. Employees might contend equally that employers would only allow holidays at times of minimum commission earnings. The potential employment relations issues arising from such disputes are obvious and the Tribunal should do its best to avoid an outcome which creates space for such complaints either way.
Although I believe that loss should be proved empirically rather than statistically, this may be difficult for commission schemes where individual deals being worked on are very transient (been and gone in a day or so, like Lock) or hard to forecast. Statistical evidence of what has been earned in the period around the leave will then certainly be relevant, but it should not be determinative, and it should not displace any finding of fact by the Tribunal on evidence from the employer that the actual likelihood of any actual loss arising during that specific period of absence was minimal.
Therefore it is my view that the ECJ Judgment in Lock could be complied with at a domestic level by a Tribunal decision that holiday pay claimants have to establish first on a balance of probabilities that they have suffered a particular loss of income as the result of a particular period of absence. While the Leicester Tribunal is superficially making a decision only in respect of Lock’s specific facts, it seems clear that the ruling will be pored over for principles and precedents of relevance to future cases even where the relevant facts are very different. If the Tribunal limits itself to deciding merely that the reference period should be 12 or 17 or 26 weeks, I would say that it is missing both the underlying thrust of the Lock decision (compensation for loss) and the substantial impossibility of applying any given reference period equally equitably to the enormous variety of commission arrangements on which people can be employed.