The Role Of The Ask Gap in Gender Pay Inequality

Motivation

Although the raw gender pay gap has narrowed in the US from 40% to 20% in the last 60 years, the residual pay gap, defined as the portion of the pay gap that cannot be accounted for by gender differences in measured qualifications, experience, industry, and occupation, has stagnated at around 10% for the past 30 years (Blau and Kahn, 2017). This dynamic is heightened due to women’s lower salary expectations than comparable men, especially at the top of the income distribution (Reuben, Wiswall, and Zafar 2017; Bergerhoff et al. 2021). Understanding the extent to which salary demands influence the total wage gap may provide advantageous insights to tease out if wage gaps result from differences in negotiation skills across genders between men and women rather than actual gender discrimination in the labor market.

This research delves into unexplored components of the salary negotiation process by leveraging the role of the ask gap as a determining factor of the gender wage gap, using data from Hired.com, a leading online U.S. recruitment platform for full-time engineering jobs in the tech sector. In particular, we investigate the extent to which the ask gap affects the bids and final offers received by candidates’ using data from over 113,000 unique candidate profiles on almost 40,000 jobs and nearly 6,550 firms.

Setting

In Hired.com , every candidate seeking a job must provide the salary they are looking for in their next job. This ask salary is visible to recruiting firms, along with the candidate’s resume information. Secondly, companies select potential candidates’ profiles by sending them interview requests along with a bid salary, indicating how much they would be willing to pay the candidate before interviewing them. Lastly, the platform records whether the candidate accepts or rejects the interview request. While interviews are conducted outside the platform, Hired.com gathers information on whether the company makes a final offer of employment to the candidate and at what salary, known as the final salary.

Documenting the existence of the Ask Gap

Figure 1 plots the distribution of women’s and men’s ask salaries. Although the shape is quite similar across genders, female distribution is shifted to the left, which aligns with the fact that women, on average, ask for $6,826 USD less than men. When estimating the difference, we find a significant 6.6% of raw ask gap. However, this might reflect other aspects driving the gap, such as occupation, job title and experience. When controlling for resume characteristics from candidates profiles, the adjusted ask gap narrows to 2.9%, which encompasses a difference of over $3,830 USD from the annual salary. The size of the gap increases considerably with work experience: for candidates with 15-20 years of experience, the gap reaches 5.4%.

Figure 2, on the other hand, shows the distribution of bid salaries received by candidates and reflects the existence of a gender bid gap, as the distribution for females is skewed to the left; showing that women receive bids that are $5,430 lower than men. This raw bid gap is estimated at 3.3% and narrowed to 2.2% when adjusting for candidates’ resume characteristics. More interestingly, the figure shows that the shape of the bid and ask distributions is strikingly similar, which means firms appear to closely follow individuals asks when forming their bids.

Given this noticeable similarity in their distribution, we explore the role of the ask gap in explaining the bid gap. A linear model conditioning solely on candidates’ resume characteristics explains 82% of the variation of bid salaries. When adding the asking salary to the set of controls, the percentage of variation explained increases to 95%, leaving little room for omitted variable bias. The ask gap also explains nearly all the gender gap in final salaries, and there is no evidence that women are discriminated against on other margins, such as receiving fewer bids or offers. In fact, conditional on their resume characteristics, women get slightly more bids than men; and conditional on interviewing, women are just as likely as men to get a final offer. All in all, women on the platform received bid salaries that are lower than those received by men, which is almost entirely explained by the fact that women asked for lower salaries.

Closing the gap: An experimental design to anchor beliefs

The previous evidence shows that women’s negotiation skills at the bargaining table are a crucial determinant of salary differences between men and women. Therefore, policies aimed at reducing these gendered differences in negotiation could serve as a vital step toward more equitable labor markets.

We leverage this by estimating the market-level effects of an unanticipated policy that increased women’s ask salaries by changing the way in which salaries were asked to candidates. Before the reform, the asking salary was an empty text field that candidates were asked to submit. After the reform, the field was pre-filled with the median bid salary over the last twelve months for a comparable candidate on the platform. This shift gave candidates information on the typical offers received by similar candidates on the platform, giving them an anchor to benchmark their own ask salary. Using an Interrupted Time Series design, we compare individuals who created a profile before and after the change to isolate the reform’s effect on ask salaries and bid salaries.

Figures 3 and 4 plot the time series of average ask and bid salary after netting out for resume characteristics. Before the feature change, the evolution suggests that salaries would have continued to evolve in parallel but at different levels in the absence of the feature change. Then, we observe a striking jump in female ask, bid, and final salaries to the level of men’s salaries. The narrowing of the gap between the two lines persists several months after the change.

In the pre-reform period, the ask gap was 2.9%, which essentially went to zero after the reform. This pattern is led by women asking for more rather than by men asking for less. In particular, the reform led women to ask for 3.2% more while men continued asking for roughly the same as they would have otherwise. On the same line, the bid gap narrows from 2.5% to -0.3% after the reform. This result is mainly driven by the fact that women are offered 2.6% more, and men are offered about the same as they would have been offered absent the reform. Women were not penalized for increasing their ask salary as there was no decrease in the number of bids they received.

This evidence indicates that a simple policy change that benchmarks women’s expectations can eliminate the gap. As observed, women did not experience a reduction in the number of bids received after the reform the results suggest that women face a minor penalty for demanding wages comparable to men. Understanding the contexts and conditions under which asking for higher pay benefits women rather than harm them is essential for future research and to level the playfield across genders.