Women are traditionally paid less than men. There are many reasons for this. One of the most compelling of them is the undervaluing of women’s work. Usually, this argument is made in terms of the pay accruing to occupations. Comparable Worth Theory argues that the work done in female occupations is systematically undervalued. Occupations with lots of women (say, teaching or social work) pay less than male occupations (say tool and die makers or truck drivers), particularly after educational qualifications are taken into account. Jobs with female attributes such as “nurturing” are also underpaid relative to their educational qualifications. Paula England is particularly well-known for advocating this theory; she has produced substantial statistical evidence for her claims.

There is now a new and interesting confirmation of the undervaluing of women’s work. A research team led by Renee Adams at the University of New South Wales shows that women’s art consistently sells for lower prices than does men’s art. Adjusting for measures of the commercial viability of the paintings involved does not make the gender difference in pricing go away. The Adams team obtained data on all the sales of paintings in the major art auction houses of the world from 1970 to 2013. Their dataset was the Blouin Art Sales Index – which covers all of the art sales at the world’s most prestigious auction houses, including Christie’s and Sotheby’s. Using this index they were able to identify the gender of the most of the painters. They also have measures of other determinants of price, such as the size of the painting, whether it is signed, period, and style, and whether the painter has sold at least twenty other paintings in major auction houses.

The headline story is that women’s paintings sell for far less than those painted by men. Women’s art sells at a discount of some 40-60%.  An objection might be made that the 40-60% figure is inflated by the effects of superstar sales. There are a small number of highly famous successful painters who command enormous premiums for their work. Nearly all of these are male. However, removing the male superstars does not make the gender differentials go away. The work of female painters still sells for a 30% discount even if one removes the male superstars and adjusts for other determinants of value.

The difference in prices received by male and female artists is not particularly related to women painting in a “feminine” way. There are no particular attribute of paintings that mark them as male or female. To test this, the Adams team did an experiment where they showed subjects paintings and had them guess the gender of the artist. Neither members of the general public nor art sophisticates were able to identify the gender of painters merely from looking at their canvasses.

Significantly, in the laboratory, thinking that a painting was painted by a woman had little effect on how much the subject liked or disliked the painting. It would have been almost comic-book simple to have people think “Oh this painting was done by a woman – it must be bad.” Discriminatory processes are rarely this crude.

In the laboratory, art patrons were equally likely to like male and female art. Yet in actual art sales, paintings by women sold for less than those of men. What moved patrons from valuing men’s and women’s paintings equally on general principle to valuing them unequally on auction day?

I would argue that the discrimination process involves a process of rationalization, reasoning and discourse. It takes time to formulate why you are going to like or dislike something. This is particularly the case when you are choosing among viable options: what worker to hire, what worker to promote, what painting to buy. These decisions are not made in a vacuum. The decision maker talks to other people to get their input. If the decision maker has to explain his decision, he thinks about how he will tell the story of the decision he made. There is a process between initial exposure to a job candidate or a painting in which one has one’s initial impressions. Those impressions change both when one talks to other people and when one thinks about how one would explain one’s choice to other people should they ask. In the case of job hires, a personnel committee or a department discusses a potential new hire. In the case of art, collectors discuss potential acquisitions both with the staff of the gallery and with trusted friends whose opinions on art are valued.

My guess is that there was something about the thinking and talking process that moved potential art buyers from no particular bias against women artists on first exposure to serious bias against women artists when the time came to actually bid in an auction.

What that process actually looks like is entirely speculative. It could be that respected gatekeepers advocate male artists more enthusiastically than female artists. In another context, Alexander Pope uses the wonderful expression “Damn with faint praise”. That might easily apply here.

It might also be the case that women artists get longer harder looks by potential buyers looking for flaws. On first viewing, the woman’s painting might seem to be okay. But if the collector is going to be spending serious money on an art acquisition, and he wants to protect against a bad investment, he might spend more time looking for potential problems with a female canvas than he does with a male canvas.

I have no idea whether any of these processes actually occur – or whether the late-breaking bias against women artists comes entirely from another source. But the Adams et. al. article certainly made me want to know more about these deliberative processes. In social science, we don’t know very much about the rationalization process by which people justify excluding or undervaluing women. Regardless of whether the reasons given represent bona fide perceptions or are thin smokescreens for explicit anti-feminism – the process of the social construction of reasons for the exclusion or undervaluation of women appears to make a difference.

We need to know more about exactly how men talk themselves into thinking that particular women do work that is of comparatively lower quality.