Introduction

Certainly, some people think so. Lara Techekirian, writing for the Erevan report [1], cites a study from 2020 by UN Women [2], arguing that there is a roughly 10% adjusted (i.e. taking into account differences in occupation sector, job level, working hours, etc) wage gap, and follows up with recommendations of the study on how to close it. The same study is cited by Seda Hergnyan in hetq.am [3], who puts it bluntly: “The authors of the study concluded that all this is evidence of gender discrimination.” But does the study actually demonstrate that?

First, we have to understand how discrimination is actually estimated in the labour markets. Discrimination is an elusive thing, as it is not only about attitudes in one’s head, but how those attitudes translate into actual actions. One cannot (usually) straight up ask the employer whether he pays the female employees less because he hates women, and expect to get an honest answer. What is done instead is that other variables are controlled for. This means that we don’t just compare women and men on average (which would be the unadjusted wage gap), but we compare women and men with characteristics as similar as we can get. Do women who work the same amount of hours get paid less than men? What if we add job seniority? And then occupation sector, educational background, and so on. Once a reasonable number of factors is controlled for, the residual gap (the difference that is still left) is theoretically attributable to discrimination. I say “theoretically” because there could still be other factors that are not accounted for, and proper studies usually avoid giving a definitive conclusion on what the effect of discrimination is.

Analysis of UN Women study

Now one of the key factors to take into account is work experience. Any employer out there knows that good work experience can be more valuable than a university degree. Indeed, reading the UN study’s excellent literature review on this topic, one can see that “the omission of experience […] may significantly impinge on the calculated gender pay gap” (p. 10), that “pay gap may be significantly influenced […] by missing key variables (e.g. experience)” (p. 10). It notes how “equally important as education is experience. In particular, women tend to have more interruptions in the workplace than men, especially related to childbirth and child-rearing.” (p. 11), and that education and experience can explain “a large portion of the gender pay gap” (p. 12). At some point, it covers the important nuance I stated earlier, saying on page 12 that

The unexplained gender pay gap is often thought to represent discrimination. Yet, this is often a naïve approach to the discussion and understanding of the gender pay gap, for a few reasons: (a) the estimation of the adjusted pay gap may still be missing important personal or labour-market characteristics that may significantly impact the gender pay gap; (b) unobservables – notably ability, motivation, devotion, attentiveness, risk aversion, attitude to work, ties and social networks, among others – may all affect the wages of men and women distinctively and yet cannot be captured by observed variables; and (c) women with particular characteristics (e.g. more-educated, career-minded women) may tend to self-select into the labour market.

Again, this is an excellent review of the literature and issues in analyzing gender wage gap. The factor of experience is also included in the mathematical model that the study uses. However, the impression one gets is that everything up to the actual data calculations was written by one person, and the rest (especially the conclusion) by another. Why?

Because by the time we get to the actual data, the importance of experience has dwindled down to a single sentence: “Work experience is not available in our survey” (p. 30). Now this is not the study’s fault, as this data is simply not collected by the government body, Armstat, in collaboration with which the study was conducted. Still, one would expect at least some kind of elaboration on the fact that such an important metric is missing from the analysis. But surely this would be raised in the conclusion, cautioning the reader from making definitive deductions, right?

No, not a word. And while the authors do say that the residual gender pay gap of 10% “could be ascribed to labour-market discrimination and the work of unobservable factors” (p. 44), – not quite the way hetq.am presented it, – they move on to recommending policy interventions without suggesting to first conduct a more detailed analysis [4]. But then again, do such studies ever conclude by saying that the markets function reasonably well, and that government intervention could lead to more harm that good?

Concluding remarks

I won’t discuss the merits of their recommendations in detail here. Suffice to say that a binding minimum wage would help current workers, but will deter future job growth and make it harder for low-skilled workers (many of whom are incidentally women) to get into the market, as their productivity simply might not be enough to justify the minimum wage. The “hallmark way” of prescribing gender quotas arguably benefits only women who are already ahead, all the while enforcing the stereotype that women can’t make it on their own and need government force to propel them upwards. Many of these suggestions are also assuming that women and men are inherently equally interested in all kinds of activities and occupational sectors, for which no evidence is provided, and the contrary can be suggested based on large-scale findings in the literature [5-8].

There are legitimate ways to help individuals, which would incidentally reduce the wage gap, if that’s the statistic we are so worried about. Individuals who are high in agreeableness (many of them women) have a harder time fighting for salary raises and promotions. Psychologists can to some extent train such individuals to be less agreeable when it comes to negotiating salary. Such an approach would be helpful for all individuals who are high in agreeableness, whether they are men or women, and would focus on the outcomes in their individual lives, rather than removing a statistical disparity on the papers.

Last thing to note is that UN Women’s paper is not the only study on gender wage gap in Armenia. World Bank had a report published in 2018 [9]. Of course, they don’t have access to data on experience either, but they completely butcher it by introducing “potential experience” instead, which is constructed by “subtracting from age the years of schooling” (p. 6). This assumes that a woman and a man who are of the same age, and studied for the same number of years, also must have the same experience, which completely ignores the whole point of child-rearing and different marital responsibilities that can cause a gap in experience in the first place.

So is there enough evidence to conclude that there is a gender wage gap in Armenia, and that women are paid less than men with the same characteristics? If one has to go on the basis of these studies, the data is simply inconclusive, notwithstanding the rush conclusions to which some journalists may arrive.

[1] https://evnreport.com/economy/closing-the-gender-gap/

[2] ANALYSIS OF THE GENDER PAY GAP AND GENDER INEQUALITY IN THE LABOUR MARKET IN ARMENIA, UN Women, 2020

[3] https://hetq.am/en/article/120224

[4] One of their technical recommendations is to “Wisely choose a referent survey, or choose between survey and administrative data” (p. 45), but it has to do with the supplier of data, not it contents (i.e. work experience data), and is not a requirement for implementing the other, more significant, government interventions.

[5] Stoet and Geary, 2018, The Gender-Equality Paradox in Science, Technology, Engineering, and Mathematics Education Available at: https://journals.sagepub.com/doi/pdf/10.1177/0956797617741719

[6] Lippa et al, 2014, Women’s Representation in 60 Occupations from 1972 to 2010: More Women in High-Status Jobs, Few Women in Things-Oriented Jobs Available at: https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0095960

[7] Schmitt et al, 2016, Personality and gender differences in global perspective, Available at: https://onlinelibrary.wiley.com/doi/full/10.1002/ijop.12265

[8] Charles and Bradley, 2009, Indulging Our Gendered Selves? Sex Segregation by Field of Study in 44 Countries, Available at: https://www.jstor.org/stable/10.1086/595942?seq=1

[9] Rodriguez-Chamussy et al, 2018, The Economics of the Gender Wage Gap in Armenia. World Bank