Do Women Make Less Because They Prefer Fixed Salaries?


There has been plenty of debate over the pay gap between women and men over the years. It definitely exists, but different studies show different levels of disparity. Some studies "prove" that it is the result of bias, while others "prove" that it isn't. The latter typically say that when controlling for education, hours worked, etc., pay is almost identical. I haven't looked at it enough to know what's true, but some new research out of Germany implies that women may make less because they prefer the sure thing.

The fact that on average women earn less than men is not necessarily the result of discrimination: when given a choice between a fixed salary and performance-related pay, women choose the former far more often than men, even if they could earn more by opting for the latter. This is the result of a study carried out by the Institute for the Study of Labor and the University of Bonn.

The researchers had worked out a laboratory experiment involving a total of 119 men and 121 women. They were to multiply pairs of numbers together over a ten-minute period. They were able to choose beforehand how they wanted to be paid. Either they could opt for a fixed sum of seven euros, or they could choose to be paid just under 20 cents for each correct multiplication. Alternatively they could also take part in a kind of tournament, where the opponent was chosen at random. Whoever solved the most tasks won 20 euros, with the opponent getting nothing.

'In our experiment only 44% of all the women taking part chose the performance-related options, although many of them could have earned more if they had,' is how the Bonn economist Professor Armin Falk summarises the results. 'By contrast 68% of the men chose this option.' The results correspond to the statistical data of the socio-economic panel, a survey which the German Institute of Economic Research carries out each year. According to this, 33% of all women work in the public sector, a field in which fixed (though relatively low) pay is the norm. In contrast, only 21% of men are employed in this field.

I can add to that the anecdotal evidence that Mrs. Businesspundit had two offers when we moved to Florida in 2000, and she took the one that paid significantly less, because she liked the people more. Money isn't that important to her compared to other things. Of course, this is all just sidestepping the issue of how compensation should work. Is it a push or a pull model? Shouldn't people be rewarded for not worrying so much about pay, instead of the other way around?

  • I also recently read a study that proved a reason women earn less is that they tend to negotiate their salary less then men do.

  • I’m suspicious of this conclusion. If women are less willing to accept risk-based compensation plans, then why are there so many women in sales? Virtually all sales compensation systems are performance-based.

    I’m also not convinced that women in general negotiate for themselves less aggressively than men…certainly, there have been few shrinking-violet types among the women working for me…

  • Personal experience is helpful challenging the assumptions and results of this type of study, but is unhelpful for validation — unless you have come to personally know a statistically valid sample of working women.

    David’s reply is a good example of this; his objection is coherent and strong, but doesn’t cut across all working women.

    So my first question for David would be, “do the women working for you earn less than the men working for you?” Of course, I cannot and do not expect David to answer, but I think my point is clear.

  • You make a good point Kevin but…

    1) What’s the statistical validity of the original study?

    2) How does the original study “cut across all women, even “all” German women?

    3) What can you learn from the question “do the women working for you earn less than the men working for you?” There are so many possible reasons why the men and women could differ.

    David is (for me and you I think) presenting some intuitions suggesting that we dig further and question this study’s conclusions. That said you are right (as the maxim says) “for example is not proof.”

  • Kevin…notice that the original study was conducted (a)in an artificial environment and (b)in Germany. I’m not sure that decisions made about pursuing an opportunity for 20 Euros can be generalized to real-life career decisions. Also, the problem involved a specific skill (multiplication) so it’s possible that what was really being measured was the person’s perception of his own relative multiplication skills.

    And I suspect that attitudes toward work (and gender roles in work) are somewhat different in Germany than they are in the USA.

  • Laurence and David,

    I have to agree with you both.

    The lack of credibility is precisely why I don’t like “economic experiments”; I believe that the decisions and tasks undertaken in the lab do not OFTEN take into account the way people think about real world decisions. It’s like they’ve isolated people in a bubble for a few hours, and then asked them how much they enjoy the fresh spring air outside…

    I don’t want to denigrate the experimental work too much — it has some very good uses, and some of my best friends are economic experimentalists at GMU. I have been talking about the problems with such experiments for so long, I just wanted to start on a different direction.