Robin Allen QC and Anna Beale consider the implications of the EAT’s decision in McNeil v Commissioners for HM Revenue and Customs , one of the first appellate cases to consider what is required to show “particular disadvantage” in an equal pay claim based on indirect discrimination under the Equality Act 2010. Previous formulations of the test were interpreted as requiring proof of something often described as ‘sex taint’.
The statutory provisions
In an equal pay claim, once the claimant has pointed to a better paid comparator of the opposite sex, who performs equal work, the burden shifts to the employer to show that the difference in pay is due to a material factor that is not directly or indirectly discriminatory.
Section 69(1) and (2) of the Equality Act 2010 set out the tests to be applied by a court in determining whether a particular factor is indirectly discriminatory. The claimant must show that, as a result of the factor, she and persons of the same sex doing work equal to her are put at a particular disadvantage when compared with persons of the opposite sex doing work equal to hers. If she is able to do so, the respondent must show that the factor is a proportionate means of achieving a legitimate aim. Of course these provisions have to be interpreted so as to conform to European Law and the right to equal pay in Article 157 TFEU.
In Chief Constable of West Yorkshire Police v Homer , Baroness Hale made a succinct comparison of the meaning of the ‘particular disadvantage’ test contained in section 19 and section 69 of the Equality Act 2010 with the tests contained in earlier anti-discrimination provisions:
“Previous formulations relied upon disparate impact—so that if there was a significant disparity in the proportion of men affected by a requirement who could comply with it and the proportion of women who could do so, then that constituted indirect discrimination … [T]he new formulation was not intended to make it more difficult to establish indirect discrimination: quite the reverse. It was intended to do away with the need for statistical comparisons where no statistics might exist. It was intended to do away with the complexities involved in identifying those who could comply and those who could not and how great the disparity had to be. Now all that is needed is a particular disadvantage when compared with other people who do not share the characteristic in question.”
Thus far, there has been little by way of judicial interrogation of how ‘particular disadvantage’ might be shown. Simler P’s judgment in McNeil provides some much-needed and useful guidance on this issue in an equal pay context.
The claim was brought by women in grades 7 and 6 of HMRC, who were paid less than identified male comparators in the same grade. Their argument was that the pay differential resulted at least in part from the application of a length of service criterion in determining pay within the grades, and that this criterion placed women at a particular disadvantage.
By the time the case reached the employment tribunal, the claimants no longer relied on pleaded allegations that the length of service criterion disadvantaged women owing to later career starts and/or career breaks to have children. Significantly, the claimants also accepted that median average basic pay statistics did not show significant and persistent disparity favouring men in either grade (the differentials ranged, over a seven year period, between 2.3% and 1.2% in grade 7, and 1.9% to 1.5% in grade 6). The claimant’s argument on ‘particular disadvantage’ rested on a contention that there was differential distribution of men and women within the pay scales, showing clustering of women towards the lower end of the pay scales, and men towards the upper end.
The expert evidence
The claimants relied on an analysis conducted by an expert instructed on their behalf which showed that the proportion of women in the lowest quartile of the pay scale for each grade was higher than expected, and in the upper quartile was lower than expected. This led the expert to conclude that the uneven distribution of men and women within each quartile was statistically significant.
The respondent’s expert disputed the validity of this approach. His position was that this type of analysis was not appropriate where differences in basic pay, as opposed to differences in access to particular pay elements, were in issue. Grouping ‘continuous data’ such as basic pay data into quartiles within each pay scale, without providing the associated numbers, was meaningless. The appropriate test to determine whether there was disparity of pay between men and women within each pay scale was to compare average basic pay.
‘Particular disadvantage’ shown through uneven distribution?
Employment Judge Snelson accepted the respondent’s expert’s evidence. The relevant comparison in this case was between the basic pay of the two groups, that being the term challenged. Whilst the distribution of men and women within a segment of the overall sample might provide a partial picture of apparent advantage, in circumstances where there was no significant long-term difference between the basic pay of men and women in the two grades, it could not substantiate the assertion that women were collectively disadvantaged.
Simler P essentially agreed with this analysis. In the appeal, the claimants sought to emphasise an aspect of their submissions below, to the effect that a simple comparison of basic pay did not properly reflect the factor causing disadvantage (i.e. length of service) which operated only on a proportion of the overall basic pay. The disadvantage caused was better reflected in the distribution analysis carried out by their expert. Simpler P rejected this argument. Like EJ Snelson, she pointed out that the ‘less favourable term’ relied upon by the claimants was the term as to basic pay. Just as respondents could not be permitted to aggregate terms to show that an overall package was equally favourable for men and women despite differentials in individual terms, claimants could not be permitted to disaggregate a single term as to basic pay.
Both tribunals therefore reached the perhaps unsurprising conclusion that, whilst differential distribution of men and women within a pay range may indicate a problem, or require further investigation, without evidence that it reflects a sufficiently significant or patterned pay inequality, particular disadvantage is unlikely to be established.
The end of Armstrong – the end of ‘sex taint’?
Although it was strictly unnecessary in view of his finding on the main issue, EJ Snelson went on to hold, relying on the case of Naeem v Secretary of State for Justice in the Court of Appeal (which in turn relied on a line of authority commencing with Armstrong v Newcastle upon Tyne NHS Trust ) that it was open to the respondent to avoid a finding of particular disadvantage by showing that the underlying reason for any disparate impact was not related to sex – i.e that there was no ‘sex taint’ . He concluded that such an argument would have succeeded in the present case, as the reason for any disparity was that women had only more recently been recruited in greater numbers to these historically male-dominated grades.
On appeal, Simler P accepted the claimants’ argument that this aspect of the decision could no longer stand following the Supreme Court’s decision in Naeem and Essop v Home Office . To the extent that Armstrong decided that a finding made of particular disadvantage could be rebutted by showing that the underlying reason for the disadvantage was not related to the protected characteristic in issue, it was not to be followed. However, it remained open to a respondent to dispute a case of apparent disadvantage based on statistics by advancing an explanation or evidence to show that the statistics were not significant, or the result indicated by them was not significant.
Simler P also confirmed that it was not necessary for individual claimants to show that they shared an underlying causal factor with other women which led the material factor to place them at a disadvantage, such as short service as a result of childcare responsibilities or maternity leave. It would be sufficient to show that the criterion of length of service caused women as a group disproportionately to be paid less than men, and caused each individual claimant to be at the same disadvantage (in pay terms).
As the Supreme Court in Essop/Naeem did not expressly comment on the Armstrong line of authority, this decision brings welcome clarity at appellate level. Armstrong cannot be relied upon to give employers a second bite at the cherry once significant and reliable statistics have shown a particular disadvantage. Of course, as Simler P points out, if it can be shown by the respondent that the statistics are not in fact significant and reliable, particular disadvantage will not be made out and no objective justification will be required.
Armstrong has been around for thirteen years and from the start it has been argued by some that the Court of Appeal erred in suggesting that there was some further step that had to be crossed before the obligation to justify arose. The issue came up in Gibson and others v Sheffield City Council  where Smith LJ held, without departing from Armstrong, that the significance of the need to show ‘sex taint’ was not great (see  in particular). Permission to appeal to the SC in that case was granted and Robin Allen QC and Dee Masters of Cloisters were briefed by the EHRC to argue that the ‘sex taint’ argument was wrong. Regrettably (though understandably) Sheffield settled the case rather than face the SC. It has taken another 7 years for the Armstrong error to be rectified. Better late than never!
 – Robin Allen QC and others have long argued this in lectures.
 –  EWCA Civ 63,  I.C.R. 708