Sharp Blue: Liberty and egality

Ashley and I have been having a slow-motion argument about such issues as feminism and equal opportunities for a while now. I don’t seem to be able to articulate my view to her adequately using IRC, so I thought I’d try to explain myself in more detail in the hope that we will then be able to argue more about the real issues. I think that it all comes down to this: good intentions are not sufficient to produce good policies, and in fact policies that at first glance seem to favour good outcomes often turn out on more detailed examination to actually be counterproductive. For example, consider this part of her comment attached to my earlier article, “Feminism and Equality”:

I am not calling for specific special treatment of women per se, I am calling for the addition of programs that would benefit any man who was a stay at home parent, who was affected by any of these issues… I was saying that home makers should be recognized for their contribution to society. I believe I even mentioned within my writing that these are not exclusively women, but statistically more women are home makers or provide secondary incomes… The society I live in has no respect for the woman’s traditional role, it gives that role no value.

At first glance, this might seem unobjectionable, but it is not, and I object strenuously. Let’s look at the proposal in more detail:

1. I don’t think it’s too controversial to say that most homemakers are women, so that policies that affect homemakers primarily affect women. In what follows I will thus use “homemaker” and “female homemaker” interchangeably.
2. It’s not possible to just value women’s traditional role more highly: it’s only possible to change our system of values so that traditional female roles become relatively more highly valued and less traditional roles for women become relatively less valued. Valuation of roles is actually the allocation of scarce resources - money, praise, time, whatever - and no system of allocation can magically make more of these resources appear in a direct way. Any such change would not benefit women in general, because whatever some gain others would lose (or else men would lose, which seems rather sexist).
3. Changing the “landscape of valuations” would distort the choices of women relative to their current choices. If we have a social policy to value homemaking more than working, then more women will choice to become homemakers than do now, and less will choose to work. A policy to value homemaking is thus the same as a policy to discourage women from working.

So the net result of this policy is to tax men and working women to pay for a scheme that gets women out of work and into homes. It is a policy that invalidates the choices of many women, effectively reducing them closer to the moral status of children who require some authority (in this case government) to choose for them. It is a social policy that implicitly supports making women more homogenenous as a group, because women who were previously office workers and hairdressers and data analysts will afterwards all be homemakers. It is, furthermore, a social policy that reduces the degree to which women are engaged with society, that reduces the number of women in positions of power and influence. It seems to me to be a policy that is not so much anti-feminist as actively misogynist, and yet it was decided upon with the best of intentions.

Another example of this counterproductivity is the attempt to equalise sex ratios in various careers, such as the various schemes to get more women involved in science or engineering. Let’s consider a toy model of such a scheme. For the purposes of this model, let’s assume that all careers are sex-balanced except for mathematics and nursing: most mathematicians are men and most nurses are women. Furthermore, let’s suppose that all the careers have reached a state of equal opportunities, by which I mean that candidates are only given jobs on the basis of ability and not of irrelevant factors such as sex. Given this (which seems to me to be close to true), the difference in sex ratios in mathematics and nursing must come down to a combination of two effects:

1. A statistical difference in relevant abilities between the two sexes. Perhaps, for example, men have greater spatial skills on average and women have higher empathy, and mathematics requires spatial awareness whereas nursing requires empathy.
2. A statistical difference in inclinations between the two sexes: men just prefer mathematics and women nursing.

There will be a correlation between these two factors, because there will be a feedback between inclinations and ability: people who like something practice at it and get better, people who are better at something find it more satisfying. This means that men will be both more likely to apply for mathematics jobs and more likely to get them, whereas women will be more likely to apply for nursing jobs and more likely to get them. This seems to me all well and good. Mathematicians and nurses are paid about the same, and I’m not sexist so I don’t care whether men or women prove the theorems I use or look after me when I’m ill. However, some people might object to this state of things…

Let’s suppose that someone starts up a scheme to get more female mathematicians. This can’t happen in isolation, because a woman can only work as a mathematician at the cost of her not working as something else. Furthermore, we don’t want to change the total number of mathematicians, so another net result of this scheme is that some men who would have been mathematicians will now be forced to take up some other career. Futhermore, the sex ratios are balanced everywhere else except in nursing, so the only way to make our scheme work is to not just have more female mathematicians, but fewer female nurses and male mathematicians, and more male nurses. (Despite this net result, the detailed mechanism might be more complicated. For example, suppose there are also some interior designers and interior design requires a combination of spatial skills and empathy. Then the detailed mechanism could involve some women choosing to become interior designers and not nurses and some choosing to become mathematicians rather than interior designers; and similarly some men choosing interior design rather than mathematics and some choosing nursing rather than interior design.)

Now, if this were the only effect then I wouldn’t mind such schemes. As I said, I don’t care the sex of the people who prove my theorems or tend my ailments, and perhaps both mathematicians and nurses would be happier in more sexually balanced workplaces (I know I’d be happier if there were more female software engineers around!). Unfortunately, that would be the only effect. Remember that despite the inequality of outcomes we have (in this example, and probably in reality) a system with equal opportunities. To distort the outcomes, we have to in some way discard these equal opportunities:

1. We can choose to accept candidates who are less able to prove theorems or care for the sick solely on the basis that the former happen to be female and the latter male.
2. We can distort the landscape of choices by, for example, paying female mathematicians more than male mathematicians and male nurses more than female nurses.

Or we adopt some combination of the two strategies. The net results will be not only the equalisation of sex ratios but some combination of two other things:

1. Lower quality or fewer theorems and worse healthcare
2. More expensive theorems and healthcare.

Even worse, we’d probably have to pay for a whole bunch of government or industry bureaucrats to administer and monitor the scheme, so we’d be taking lots of people out of productive work at the expense of higher taxes or more expensive services to make services worse and more expensive, and at the same time as adopting explicitly sexist policies, throwing away equality of opportunity, and taking responsibility for people’s choices out of their own hands. This seems to me to be an extremely stupid, insulting and counterproductive thing to do, no matter how good the intentions behind the scheme.

Similar arguments apply to a vast number of other situations in which governments override people’s freedom of choice in the interests of “equality” (which is actually nothing of the sort!) or job creation or whatever. So why does this keep happening? I think it’s for one simple reason: the few benefits are more visible than the vast costs. After all, it’s easier to see the more sexually balanced workplaces than it is to see the more general increase in cost, decrease in quality and diversion of labour into less productive pursuits, just as it’s easier to see people working in job-creation schemes than it is to see the compensating loss of jobs elsewhere. It will only be when we learn to think more deeply and widely about such issues that good intentions will lead to good outcomes.

Update (15/7/2003): My friend Cerstin objected to my statement that “mathematicians and nurses are paid about the same”. In the US, at least, this seems to be true. There, the average starting salary of a nurse is $32,930, rising to $50,000 after five years, whereas the average starting salary of a mathematician is $36,800 rising to $55,000 after five years. These aren’t quite the same, but they’re in the same ballpark, and both woefully underpaid. A more valid flaw in my toy model is the relative numbers of people in each profession: there are 2,170,000 American nurses but only 13,500 mathematicians. This will change things a lot. Firstly, it will presumably make being a mathematician more prestigious and hence a more attractive career. Secondly, it means that it takes a smaller number of additional female mathematicians to equalise the sex ratio amongst mathematicians than it would take additional male nurses to equalise the sex ratio of nurses. But, as I said, it’s a toy model.

Update (19/7/2003):The toy model described in this article flows naturally from a very few observations and postulates. I’d appreciate it if my (supposedly) feminist critics would tell me which are incorrect:

• I have noticed no difference in the average competence of men and women performing any given job, and assume that this applies to all or at least most jobs.
• I have noticed sometimes quite large differences in the numbers of men and women performing any given type of job, and assume that this reflects the true numbers of each sex performing those jobs rather than being a statistical fluctuation in my sampling.
• I take it as axiomatic that men and women are adult human beings who are capable of deciding on the basis of their personal preferences between a range of options of varying attractiveness to each individual.
• Technical assumption I: any men selected for some class of posts are on average the most competent men who apply for that class of posts, and similarly any women selected are on average the most competent applying women. (I’ll call this an assumption of normality in future. An abnormal selection would be one in which the most competent are not chosen. A biased selection would be one in which factors other than ability are significant. Selections may be any combination of biased, unbiased, normal and abnormal.)
• Technical assumption II: selections are not pathological in the sense that some fraction of applications by women and/or men are thrown away at random. (It seems to me highly likely that this is true, but if not it can make biased selections seem like equal opportunity ones.)
• Technical assumption III: selections are “local” - each selection, whether biased or unbiased, is independent of the outcomes of other selections. (I’ve not thought through the implications of this being false, but again it seems to me highly likely to be true. If not, then our society is permeated by a vast conspiracy about which I’ve remained entirely unaware.)
• Technical assumption IV: we may use inductive and deductive reasoning.
• A moral axiom, which is not necessary for any of the arguments I’ve presented by which I expose here in the interests of transparency: each individual man or woman is on average more competent to make decisions about his or her life than anyone else is (and, especially, is more competent than governments or self-selected “elite” groups are). (Actually, this might be an observation rather than an axiom!)

Update (21/7/2003): Ashley has posted her reply to this. You should all read it.