Statistics, used well, can play a vital part in exposing inequalities and persuading people of unjustness. However, used less well, they risk undermining the argument even if they are initially superficially attractive. Earlier this month, the Chartered Institute of Personnel Development (CIPD) and the High Pay Centre published their survey of FTSE100 CEO pay packages for 2016. It casts a sharp light on the complex topic of the pay of chief executives of the leading companies listed on the London Stock Exchange, providing a breakdown of the composition of CEO pay and highlighting the differences between CEO and staff pay packages by sector and by company size. It also includes a short comment on Brexit (the number of companies reporting Brexit as a risk has increased more than 23-fold in the year). It also has a page on gender pay differences. The CIPD’s press release is titled ‘Reality bites: Average FTSE 100 CEO pay packet drops by 17% in the past year’ and has the sub-title ‘Slowdown in CEO pay welcome, but huge gap remains between the top and the rest’. While the gender pay disparity is mentioned once or twice it does not gain top billing. The Independent, by contrast, headlined one of their stories on the report with ‘Male FTSE 100 chief executives earn 77% more than female counterparts, finds new research.’ This is undoubtedly a catchy headline but its prominence obscures a key finding (treated by the article as a subsidiary point): that out of the 99 firms reviewed only six had a female chief executive.
For there to be only six leading, listed firms with a female chief executive is clearly worthy of note as is the fact that the number increased only by one in the year under review and is set to inch up in the following year. The report – and to be fair the article – also shows an 8% increase in the number of women on the remuneration committees but also that there are nine companies with no women on those committees. Furthermore, over three quarters of the companies have no female executive directors (the increase number of women on boards, driven by the Davies Review, being predominantly through non-executive directorships). The pipeline of potential female chief executives is thus currently heavily restricted and anything that causes attention to be paid to this must be welcomed by those valuing diversity and equality. The 77% pay differential figure, however, is problematic.
The section on gender pay (on page 11) starts by saying ‘As a FTSE 100 CEO, it is more likely that your name is David than you being a female’, and therein hangs one of the problems. The 77% figure is derived by taking the average of the male CEO pay and comparing it with the female CEO pay and as such is arithmetically accurate. But comparing an average of 93 observations with one of, say, only six or seven creates inherent difficulties. It is just as accurate to say that those not called David (92) earn 51% more than those called David (7) or that the 93 whose name doesn’t start with an I earn 69% more than the six called Ian, Iain or Ivan (these figures are drawn from the complete list published in The Independent). An even more stark example of small numbers skewing averages can be seen when looking at the companies’ names: Those who run FTSE 100 companies which begin with a P (5), a C (7) and a W (5) earn 74%, 183% and 391% more than those who run companies beginning with an M (5). The gender pay differential may become more of a concern when the number of women CEOs increases but at the moment it is that increase that must surely be the primary concern.
The second problem with the 77% is the bold statement in the report that ‘… as a male, you are likely to earn 77% more on average than your female counterpart.’ The 77% figure may be accurate but misleading, but this phrasing in the report unfortunately verges on the spurious. It is true that average of male CEO earnings is 77% higher than the average female CEO earnings. But it is also true that only 30% of the male CEOs have earnings at or above that average. Furthermore, 39.8% of male CEOs have earnings below the female average. It is thus untrue to imply that male CEOs are likely to earn 77% more than their female counterparts. On these figures, one could expect a male CEO to earn more than his female counterpart but in around four out of ten cases that would not be the case, let alone earning 77% more.
Statistics can provide illumination and support important arguments but they need to be robust and carefully used or they may end up overshadowing significant points and, worse, undermining the credibility of the specific – and the wider – argument if they crumble under analysis.