Gender Pay Gap Review Panel - Matthew Harrison - Submission - 2 October 2018

Submission - Matthew Harris on

Dear Deputy Doublet and other panel members,

You have called for evidence regarding the existence of a gender pay gap' in Jersey. I am hopeful that your review will have a positive effect in assisting Jersey in moving towards equality of opportunity for all. It is important to note, however that the terms of reference can tilt a study in one direction or another, and I hope that you will consider both the assumptions inherent in the title and the underlying reasons for any differences in pay between genders which you may find in Jersey.

I should note to start with that I am a very strong believer in equality of opportunity, and that I abhor equality of outcome. Someone who wants to be a doctor and has the capacity to be a good doctor and the diligence to work hard to become and remain a good doctor should be able to be a doctor regardless of race, gender or anything else. But that does not mean that there should be (for instance) as many male doctors as female doctors or vice versa. If (and the evidence from scientific studies strongly suggests that there is) there is a difference between men and women in their inclination to become doctors, then the ratio will not be 50:50. (In fact as I understand it across the western world the number of female doctors is becoming far greater than the number of male doctors over time.) The same regarding inclination is true for any profession, trade or occupation. Bricklayers, for instance, are overwhelmingly male. It is interesting to note that people rarely if ever bring up such outdoor, physical, sometimes dangerous professions in discussions of job distributions and pay levels between men and women. I hope that you will include the amounts that unskilled farm and construction labourers, for instance, are paid if you do any sort of overall figures for the' pay gap.

I noted the bias inherent in even a title. "Gender pay gap" brings with it a number of assumptions, some of them liable to blind one to the more nuanced reality of anything in the actual world rather than a person's ideology. Implicit in the idea is that there is one thing and one thing only which constitutes a gap in pay between men and women. That is certainly not the case. The dynamics which exist in different occupations, trades and professions may well be very different. The term also implies that if one were to take the average salary of all men and the average salary of all women and compare them, it is desirable that they should be the same, or even that there is meaning in such a figure. It is not obvious that their being the same is a bad goal in and of itself, but it is indeed a bad goal, for reasons which I hope will be clear from what I say below. The key one is that it conceptualises men and women as groups to be seen as separate from each other rather than lots of individuals who should be judged individually on their merits and have the same moral value as individual human beings, not as members of any particular group.

I suspect that there is a degree of inequality in the pay levels of men and women in similar roles in Jersey. A degree of this may relate to bias. There may also be a degree of bias with regard to promotions and choice of candidates for roles. There is, however, strong evidence from other countries that although a degree of gender bias exists, there are significant factors which explain the variances in the population as a whole, leaving gender bias as a relatively minor element. One factor which correlates negatively with pay, for instance, is agreeableness (in the technical sense- one of the big five personality traits). The extent to which the correlation differs between the genders can differ itself. However over the population women are (on average) more agreeable than men. The difference is not huge (about half a standard deviation), but it is large enough that approximately speaking if one were to choose one man and one woman at random, the chance that the woman would be more agreeable than the man is about 60%. If it is indeed the case that agreeableness is negatively correlated with pay, then the average pay of men and women will differ, and men will earn more on average if everyone is treated equally. In such a case then, a difference

between average pay (all other factors being equal) would actually show fair treatment of individuals not a gender bias. It seems far better to me to have a situation where people's behavioural characteristics rather than their gender drive their success or otherwise. If we move to the idea that everyone regardless of ability or effort should have the same outcome we will return swiftly to the dark ages, as the desirability of competence will evaporate.

Another key difference between men and women lies in interest. On average (as ever averages say nothing about individuals so in a comparison of two individuals this may not be the case) women are more interested in people and men are more interested in things. A recent study showed that this is evidenced from the first day after birth. Elements such as the lack of scalability of caring professions make it likely that they will attract lower pay than less person-focussed work. Because women have, on average, more of an inclination towards nursing, for instance, the lack of scalability may have an overall negative impact on relative pay levels. It is key, therefore to separate occupations and not aggregate them if at all possible because combination of dissimilar professions gives a false view with regard to the actual fairness and militates against the freedom of choice and equal treatment we are all seeking.

Even more dangerous ground, but important nonetheless, is a separation between mothers, women who are not mothers, fathers, and men who are not fathers. The choices made by individuals and couples as to the share of childcare have an impact on pay. In a case where a mother chooses to work fewer hours, or even just work a full-time job rather than a job which requires an 80-hour week, in order to spend time with her children, her pay will be lower than a father who chooses to continue to work 80 hours in order to ensure that the family's standard of living is high. The roles can of course be reversed, and it should be possible for all to choose without discrimination. But across the population more women choose to reduce their hours than men. And the relationship between pay and working hours is not linear. A 10% increase in time spent working on average generates far more than a 10% increase in pay. Fair treatment results in unequal average outcomes because of fair treatment on an individual level. So inequality of outcome is actually a desirable goal, as long as it results from fair treatment. Inequality is not necessarily unfair.

The figures shown on your website for this review regarding the proportions of directors, senior managers and the overall workforce in the States could to at least some extent be driven by the above tendencies. There may also be an element of the profile (that is, a higher proportion of men at higher than at lower levels) which is driven by the tendency of men to be more interested in status than women are. The extremes of any normal distribution (which is the type of distribution which tends to exist in trait differences) are where the differences are most obvious. To return to agreeableness, just half a standard deviation of difference at the mean gives a huge difference at the tail of the curves. If you take the far left of the distribution (low agreeableness) there are vastly more men than women. Assuming the same standard deviation on both curves, if you go out to 3 standard deviations from the mean on the female curve (leaving about 0.1% of women to the left) you're only 2.5 standard deviations from the mean on the male curve (leaving 0.6% of men to the left). That goes a long way to explaining why most people in prison are male (leaving aside the slight bias towards leniency which studies show in the sentencing of women) - people who commit crimes bad enough to cause people to go to prison are the most disagreeable people in general. Equally, it could be an element of why more senior people are men. Other traits may also have an effect. Women are on average higher in neuroticism (in its technical sense roughly equivalent to susceptibility to negative emotion), so on average they may be less inclined to the extremes of stress relevant in positions with higher levels of responsibility and prominence. Add to that the working hours required and the differential in desire for status. I present the above merely as a hypothesis as to why the figures presented may not be a sign of as much dysfunction as they appear to show. I do not present it as a hard claim, although there are studies from other countries which show correlations in all the directions noted above. In aggregate they may well explain a significant proportion of the effects seen. That is not to say that there is no residual effect of gender bias. I am not a statistician, but I wonder if the size of the data set, even if personality assessments and inclination studies were carried out for all States staff, would allow effective evaluation statistically in any event. Our intuitions about sample size are very often wrong (as has also been robustly demonstrated), so perhaps I am wrong about that, but I suspect that your whole project is on shaky ground from a sample size point of view in Jersey unless you zoom out so far that you bring in the sorts of confusions I hope I am convincing you to avoid.

As a society, by focusing on the simple dichotomy of male and female, in a world where the fundamental battle for equality of treatment and opportunity has been won (the very concept of comparing male and female rates of pay would have been nonsensical in 1850 or far more recently), risks turning a functional society into one at war within itself, as indeed appears to be happening in some other western countries. The hell that is generated by seeing these complex issues in terms of group identity such as male/female, worker/capitalist or the like are to be seen in the twentieth century. Hundreds of millions of people died because people tried (on numerous occasions) to generate equality of outcomes, including by pitting groups against each other. That is true both where fascism was tried, and where the equally reprehensible and arguably even more murderous communism was tried. Communism has been tried far more often and resulted in more deaths, perhaps because it is easier to see why fascism is terrible from a moral point of view than it is to see why communism is reprehensible. We need both left and right of the political spectrum to steer our way, but the extremes of ideology generated by focussing on group identity are the road to ruin. Our society's view of the pre-eminence of the individual is sacrosanct. It is the reason for our affluence, our progress, our freedom, and our stability as a society. I hope that your review will support this view and find to what extent there is discrimination, and to what extent other factors are the reason for differences. I suspect that discrimination is not as great an element as some would have us believe, even though it exists and we should aim to eliminate it.

Yours sincerely, Matthew Harris on