Why Iceland’s wage gap policy is dumb

Iceland has introduced legislation that requires firms with more than 25 employees to prove that they pay their female employees as much as their male employees. This shifts the burden of proof from the person who wants to claim sexism to the firm, which now has to demonstrate annually that it is not sexist. The intention is explicitly to reduce the gender wage gap, which is 7% in Iceland (which, I believe, is the lowest anywhere on earth). This is a rather dumb policy for achieving that. This policy would only work if prejudice towards women was widespread in Iceland, which I find very hard to believe. 

The wage gap is not a function of mal-intent, of prejudice. Econometric evidence has suggested for many years now that when you compare apples with apples (which, I must add, is very hard to do) women are paid pretty much the same as men. What's interesting is to what extent we tolerate the non-prejudicial factors that do lead to wage-inequality overall, and to what extent we consider these sexist. In short, the question you need to ask is: if we had perfect equality of opportunity, would men and women on average make different life choices? If the answer is yes, then you will have a wage gap (not necessarily in men's favour). I am inclined to believe that some of the difference in choices is biological in origin rather than environmental, but you don't need to engage with that question to see that this policy is an expensive distraction. Iceland's policy is unlikely to address any of the factors that encourage men and women to make different choices (Iceland has already done a lot on this front that I think should be imitated elsewhere, especially viz. child rearing). I can only hope that a silver lining will emerge when the government’s auditing offices realise how complex an administrative task this will be and put together some tools for helping people to quickly analyse a firm’s payroll data to assess its wage equity. 

Before I start, some quick links to stuff I’ve written before that shows I’m not writing this because I’m some kind of MRA, but rather because I’m ally who hates dumb policy: here on issues in Japan, and here on women deserving equal pay in tennis.

Let’s get some basics out of the way first. The gender wage gap refers to the average wage of all women compared to the average wage of all men. It is not a comparison of a woman and a man working identical jobs with identical experience, ability, education etc. That is to say, it is not a comparison of an apple with an apple. There is no accounting for differences in age, education, experience, industry, career interruptions, ability etc. When many of these things are accounted for, the gender wage gap shrinks precipitously. One of my PhD colleagues recently did this accounting for Australia using fixed effects modelling and the HILDA panel, and the gap that remained after control was 3 per cent. The raw figure is 17.7 per cent. And I stress that the unexplained residual is unlikely to be made up entirely of prejudice (though of course people like Weinstein do exist). Most prominently, our ability to proxy experience, ability and education quality is poor.

Let me unpack a few of these controls. If a disproportionate number of men work in finance and a disproportionate number of women work in childcare, then you will get a large wage gap at the aggregate level even if men and women are paid the same for the same work in finance and childcare simply by virtue of the fact that finance pays more than childcare. We could certainly have an argument about why finance is paid more than childcare, but that’s much more complex than the wage gap (the short answer is demand, supply and productivity rather than patriarchy). If career progression is substantially a function of time on the job (because firms want to be sure you will stick around before they invest time, money and training in you), then career interruptions will reduce your wage in the long run. If women then do the majority of child-rearing because of the need to recover from childbirth, the demands of breast-feeding, and various cultural and biological systems that encourage them to do so, then women will have more career interruptions than men, and so women will earn less, even if they are paid the same amount for the same work. This is why paternity leave policy is so critical for reducing the wage gap—it encourages men to interrupt their careers as much as women. It is also why it is questionable to refer to the gender wage gap as a matter of 'justice' as this implies some degree of evil: birth isn't even a structural factor--it's a biological one.    

The key point is that the gender wage gap does not refer to equal pay for equal work. I can’t believe I still have to state that explicitly. If you complain about the wage gap and don’t understand this basic fact, you’re contributing to ignorance not justice. I was gobsmacked to discover exactly this error of understanding while reading bell hook’s (she doesn’t use capitals—I am not performing a microaggression) Feminism is for Everyone. This is one of the most prominent feminists on the planet and she hasn’t even taken the time to understand the most basic indicator of economic equality between the sexes.

So the first dumb thing about the Icelandic policy is that it does nothing to get at the root causes of the wage gap, and seems to perpetuate the view that it exists because firms are sexist (in a prejudicial sense), which is not supported by the evidence I’ve seen, though I would certainly expect some pockets of sexism to exist in all OECD countries, even  in Scandinavia.

The second dumb thing about the policy is that it fails to realise how hard it would be to ‘prove’ that you aren’t discriminating. I had this conversation at length with economists and management-consultant types yesterday, and everyone agreed that the accounting involved is basically impossible unless your burden of proof is very low (in which case the legislation serves no purpose). The reason is that you have almost no way to compare apples with apples. Unless you’re running a factory or firm with very clear job descriptions and pay grades, you can’t put two people side by side in the firm who are identical except for their genitals. Let me give some examples to illustrate how complex this gets, especially if you want to tackle structural patriarchy (and perhaps even biology) and not just administrative practices.

First, US Senator Elizabeth Warren, prominent feminist and critic of the wage gap, was recently attacked by the right-wing press in America after it emerged that the wage gap in her office was worse than that in the rest of the country at 30 per cent. The dumbest thing about both the original story and the reply from The Huffington Post (both were pretty solid though) is that they talked about salary gaps instead of wage gaps. Then the comparison is even more ridiculous, because you are potentially comparing people who work 60 hours a week (chief of staff) to people who only work 35 hours a week (receptionist), or even to people who work part time if you don’t look at full time equivalent hours. In any case, it turns out that the gap was a function of Warren’s chief of staff, Mindy Myers, who was earning $170 000 p.a., leaving to work for The Democratic Senatorial Campaign Committee.

The Huffpost article concludes triumphantly from the adjusted payroll records that Warren actually pays women more than men. NO SHE DOESN’T! It’s not equal work. If Warren had two chiefs of staff, one male and one female, working exactly the same job, and the woman got more, then we could say that Warren pays women more.

Stepping it up an academic gear, we have Goldin and Katz’s (G&K) articles following MBA graduates for 16 years. You can find a summary of the pieces here, and the actual articles here (for the initial analysis) and here (for the stuff focussing on career interruptions). These are some of the most prominent papers in the wage gap literature because they very nearly manage to compare an apple with an apple. G&K follow all Chicago MBA graduates for 10–16 years (depending on cohort) and track their pay differences. They are able to control for a very rich number of observable differences between people (like courses chosen, GPA, sector, occupational title etc.) and thereby use statistics to construct very close comparisons of graduates who differ mostly in terms of sex rather than other factors. They find that newly minted MBAs have ‘nearly identical labour market outcomes’ at the beginning of their careers, but that the paths of men and women rapidly diverge. Fifteen years after graduation, the gap is nearly 40 per cent!

G&K identify three main causal channels driving this divergence: ‘difference in training prior to MBA graduation, differences in career interruptions, and differences in weekly hours’. Children are the main cause of all 3.

There are two broad classes of ‘solutions’ to these dynamics: schemes that minimise the disruption of parenting and encourage men to parent as much as women, and policies that aim to give women exactly the same wages post-children that they would have had had they not had children.

I’ve written about the first set of schemes before, here. There are three main things. The first is paternity leave that encourages men to parent. This is typically packaged as a ‘daddy bonus’, where the mother gets more leave if the father also takes leave. This makes it hard for firms to assume that the mother will drop out of the workforce after kids and therefore preemptively start investing in men. The second is maternity leave schemes that reserve the mothers position in a firm’s career track for a while, usually 3 years, until she is ready to return to work after child-rearing. The main effect of this is to preserve the trajectory of a woman’s wage growth over time. There is a lag compared to men and it takes a bit of time to get back on the earlier trajectory, but she gets there in the end (as we are seeing now through a different channel as baby-boomer mothers finally reach executive positions after returning to work 15 years ago).The final policy is cheap and accessible child care, which makes it possible to continue your career with minimal disruption. This is mostly a question of supply; something Australian governments have failed to appreciate since ABC went bust.

The second set of proposals I honestly find bizarre. They basically expect firms, especially government organisations, to promote women faster than men and pay them higher to make up for the fact that women take time out for kids. This is reasonable if you think that wage equality is the principle dimension of social justice, but it seems to me a sure-fire way to encourage firms to not hire women. You’re basically exploding the implicit costs of hiring women. More broadly, this seems to me an incredibly entitled attitude. Parenting is important for society, so I think we should support it somewhat e.g. by the policies mentioned above. But parenting is also a lifestyle choice, and it is only in very rare cases that we subsidise someone’s lifestyle choices. It seems almost egregious to me to suggest that mothers should be paid by everyone else such that the financial costs of motherhood become nonexistent. What stuns me about most such proposals is that they're made with only an implicit assumption that equity along gender lines is the foremost if not only 'justice' consideration in society--the authors are rarely explicit about their strong value-judgement. To put this into context, the enormous cost of such schemes would leave a lot less money leftover to fund other equity programs, like early childhood interventions for inner city youth, housing for the homeless, domestic violence shelters etc. There is such a staggering unawareness of the fact that money doesn't grow on trees in these articles. 

The final complexity in wage-gap policy that has to be discussed is the differences in how men and women approach promotion, competition and interviews. Basically, women on average put their hands up for promotion less regularly than men, and often only when they are overqualified. Women are not as competitive in temperament, which means they tend to shy away from rat-races and other dynamics that characterise many major firms’ human resources systems. And finally, women tend to be more modest and reserved in interviews, which can see them underplay their strengths. Here is a slightly longer summary written by a journalist trying to be funny and failing--the links are good though. If you want links to the academic literature, hit me up in the comments. In the context of Iceland’s legislation, if you’re required to demonstrate that your HR practices are not fragile to these psychological biases, then every firm required to report is going to be done for discrimination, including the ones with female managers and HR departments.

I think there is a very interesting conversation to be had about whether these gender psychology dynamics are a function of patriarchy and capitalism. That said, that conversation is rarely scientific in my experience, and this legislation is not a way to unstick the structures of social oppression or whatever. That takes time mostly, and all the trends are going in that direction already. Notably, as women are entering the workforce in greater numbers, hanging around longer and coming back in more, firms are adapting HR and management practices to make better use of female talent. One positive side-effect of the legislation might be a injection of scientific rigor into attempts to measure patriarchy and things like that.   

Think about all this (the ignorance problem, the measurement problem, the evolutionary psychology dimensions) for a second in the context of Iceland’s legislation. If it’s just a matter of having all your ‘level 5—manager’ staff paid the same, then it shouldn’t be too crushing an administrative burden. If women disproportionately leave early to pick up the kids, use flexitime and work from home allowances, and work less overtime than the men, then you might run into some trouble when comparing bonus payments, but I don’t think it’s going to be a disaster. But this is so obvious a thing that anyone who wasn’t getting paid what they were administratively meant to could easily just move to sue under the old ‘burden to prove wrongdoing’ legislation. There is no need to move to a burden to prove equality.

But if the legislation starts to stray into things like why most upper and middle managers are men, if the auditing requirements can’t control for a person who is 5 years into a level 5—manager position compared to someone who is only 1 year into it because they took 3 years out to have kids and consequently isn’t paid as much because they aren’t as experienced, etc. then it’s going to be a mess. The admin burden will be enormous, and enormously wasteful. If the legislation is trying to get into the psychological dynamics of capitalism then it’s going to be a disaster.  

In some ways, I am happy this is happening in Iceland, because it’s a country the size of Canberra (~300 000 people) and if disaster strikes it won’t hurt too many people. I mean, how many firms with more than 25 people even are there in Iceland? The administrative burden will be rough, but as the government has noted, bureaucracy in a welfare state is already rough and Iceland is doing fine. It’s a nice experimental environment. At the same time, the legislation is injecting more stupidity into a debate already clogged with passionate ignorance, righteous laziness and misguided policies, so I’m far from happy about it.