There is a gender gap in pay. The average woman earns less money than the average man. The gap worsens at the top, when considering the highest earners. The gap has improved slightly over time, but it has nevertheless persisted over many decades. While researchers have pointed to a number of factors to help explain the gap, such as differences in labour market experience achieved by men and women and what occupations are chosen by men and women, a sizeable and persistent ‘unexplained’ gap remains (Blau and Kahn 2017).
An important and commonly asked question for policymakers and researchers is what other factors — even after accounting for differences in labour market experience, occupation, etc. — drive this gender gap in pay. Some well-studied possibilities point to women being more reluctant than men to negotiate (Babcock and Laschever 2003) or to compete (Niederle 2016).
In our own recent work (Exley and Kessler 2019), we consider an additional possibility, namely, that of a gender gap in self-promotion. We define self-promotion as the subjective way in which individuals describe their own ability and competence. Consider, for example, a job candidate. When the job candidate is asked about her aptitude in maths, an objective response may reference her (verifiable) score on a standardised test. A subjective response instead relates to how she personally describes her aptitude in math (e.g. “I’m just okay” or “I’m really strong in math”) We explore self-promotion because these types of subjective responses are elicited in myriad contexts that are relevant for educational and labour market outcomes – in school and job applications, in interviews and performance reviews, and in meetings and around the water cooler, just to name a few.
Using an online labour market platform called Amazon Mechanical Turk, we recruited 1,500 ‘workers’ and 300 ‘employers’ to participate in our study. These workers were split into one of four treatment conditions: the public version, the private version, the ambiguous version, and the private (information about others) version. Unless otherwise noted, we will focus on the findings from our public version.
At the beginning of our study, each worker completes a test with 20 analytical questions. Rather than their actual performance on this test (i.e. the number of questions they answered correctly) determining their pay, however, their self-promotion determines their pay. In particular, workers are asked to answer several self-assessment questions that require them to subjectively describe their performance. One of their self-assessments is then shared with a potential employer who will determine whether to hire them and how much to pay them if hired.
Self-promotion pays. Workers who provide more favourable self-assessments are hired more often and earn more money.
But, women engage in less self-promotion than men. For example, consider our results from the self-assessment question that asks workers to indicate the extent of their agreement on a scale from 0 to 100 with a statement that reads “I performed well on the test”. The average man rates himself a 61 out of 100 and the average women rates herself a 46 out of 100. Put differently, the average woman rates herself 25% lower than the average man.
One explanation for the gap could relate to performance differences. Perhaps men rate themselves higher because they performed better on the test. This is not the case. Both the average man and the average woman correctly answered 10 out of 20 questions on the test. (If anything, women slightly outperform men in the test on average.)
A second explanation for the gap could relate to confidence. Indeed, we find that when asked to guess their score on the test, men overestimate and women underestimate their true performance, leading to a gender gap in confidence widely observed in prior literature as well (Niederle and Vesterlund 2011). Perhaps men engage in more self-promotion because they think they performed better on the test. We find that this is not the case. Even after we inform both men and women of precisely how well they performed on the test – for example, both the average man and the average women learn that they answered 10 questions correctly and are told where in the performance distribution their score puts them – the average woman still rates herself 19% lower than the average man.
A third explanation for the gap could relate to strategic incentives. As noted above, self-promotion pays. Perhaps men inflate their self-assessments more than women because they (accurately) believe that doing so would result in higher pay. We find that this is also not the case. Even when self-assessments cannot influence how much money one makes — when we eliminate employers in our private version — the average woman still rates herself 22% lower than the average man.
A fourth explanation could relate to ambiguity in the decision environment. We investigated this by increasing and decreasing the amount of ambiguity in our experimental setting. In our ambiguous version, we increased the ambiguity around the optimal amount of self-promotion. Workers learn that there is some chance of ‘being caught’ if they engage in too much self-promotion. In particular, workers learn that employers may additionally learn their actual performance on the test. In our private (information about others) version, we decreased the ambiguity around the typical amount of self-promotion. As in our private version, there are no employers, and workers also learn the average level of self-promotion of others who performed the same as they did on the test, giving them more guidance on what an appropriate level of self-promotion might be. In both the ambiguous version and the private (information about others) version, however, the average woman still rates herself lower than the average man.
So, what else could explain the gender gap in self-promotion? At this point, we can only speculate. Prior work shows that women may be viewed more negatively – or even experience a backlash – when they engage in activities akin to self-promotion (e.g. Bowles et al. 2007). On the one hand, we note that there is no scope for such backlash in our setting. Employers do not know the gender of workers, so they cannot treat men and women differently. Employers do not know the actual performance of workers, so they cannot even identify who may be engaging in excessive self-promotion. On the other hand, if self-promotion does result in more backlash for women than men in settings outside of our study, it could be that women are more averse to self-promotion because they have – more so than men – internalised the potential risk of self-promotion.
While more work is clearly needed on the drivers of the gender gap in self-promotion, one implication is clear. Since women persistently engage in less self-promotion than men, how men and women describe their performance should be interpreted with caution – particularly when making important decisions that relate to hiring, promotion, or pay.
Babcock, L and S Laschever (2003), Women Don’t Ask: Negotiation and the Gender Divide, Princeton University Press.
Blau, F D and L M Kahn (2017), “The gender wage gap: Extent, trends, and explanations”, Journal of Economic Literature 55(3): 789-865.
Bowles, H R, L Babcock, and L Lai (2007), “Social incentives for gender differences in the propensity to initiate negotiations: Sometimes it does hurt to ask”, Organizational Behavior and Human Decision Processes 103(1): 84–103.
Exley, C L, and J B Kessler (2019), “The gender gap in self-promotion”, NBER Working Paper No. 26345.
Niederle, M (2016), “Gender”, in J Kagel and A E Roth (eds.), Handbook of Experimental Economics, second edition, Princeton University Press, 481-553.
Niederle, M and L Vesterlund (2011), “Gender and Competition”, Annual Review in Economics 3: 601–30.