This paper contributes not only to literature on gender wage gaps and gender gaps in salary expectations, but also to literature on gender differences in negotiation, gender discrimination and behavioral labor economics examining the role of information in the job search process and salary decisions. In particular, this paper provides a contribution in the following four issues:
- Compared to expectations measures, the ask salary plays a direct role in the salary negotiation, as it is one of the few signals voluntarily transmitted to employers. Besides, the recruitment process on the platform allows for the direct measurement of the impact of candidates’ ask gap on the firms’ offer gap, while most studies only observe either the candidate or the firm side of the market.
- Using information on both sides of the market, candidate’s and firm’s side, this paper provides evidence that women indeed ask for significantly less in high stakes environments and that this gap is consequential for resulting salary offers.
- By exploring the propensity of companies to apply to comparable candidates rather than how the probability of being hired differs across similar men and women when they apply to the same job, it finds no evidence of discrimination against women at the extensive margin. Women in fact get slightly more interview requests than men and, conditional on interviewing, women are just as likely as men to get a final offer.
- Using the quasi-experiment of the reform the platform Hired.com to estimate the market-level effects of an increase in women’s ask salaries, this paper finds that this change eliminates the ask and the bid gap, without affecting the relative number of bids received by women.
- Understanding role of information in the job search process and salary decisions in the tech-industry. In particular, it delves into the gender differences in negotiation process at the top of the income distribution, in a tech platform aimed to facilitate recruitment for well-qualified candidates.
Understanding the contexts and conditions under which asking for higher pay benefits, rather than harms, is an important avenue for future research.
Using data from Hired.com3, this paper exploits information about unexplored components of the salary negotiation process. In this platform, every candidate seeking for a job has to provide the salary they are looking for in their next job. This ask salary is visible to firms recruiting on the platform, along with the candidate’s resume information. Second, companies signal their interest to candidates with a bid salary, indicating how much they would be willing to pay the candidate before interviewing them. Last, the platform records records whether the candidate accepts or rejects the interview request. While interviews are conducted outside of 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 salary4.
3 a leading online recruitment web-page for full- time, high-wage engineering jobs.
4 It is important to note that the bid salary is non-binding, so the final salary can differ from the bid.
Measuring the gender ask and bid gaps
This paper characterizes the gender ask gap, understood as the extent to which women ask for lower salaries than comparable men. The ask gap is estimated by including candidate’s profile characteristics besides resume’s characteristics as control variables. Data is treated at the candidate level, which means there is one observation per candidate in which the first reported ask gap is kept. There’s an alternative approach to estimate the ask gap which is considering each interview request a candidate receives as a separate observation. In this approach, the ask gap is estimated considering the ask salary that the individual was reporting by the time she/he received a bid. Since individuals may change their ask salary throughout the spell, each candidate will appear as many times as he/she has received a bid from a firm. On the other hand, the gender bid gap can be understood as a measurement of the extent to which women are offered lower salaries than comparable men. In order to empirically test the relationship between the ask salary and the bid gap, the author uses a three step methodology. First, estimates the raw gender bid gap. Then, estimates how much of the bid gap can be explained with the candidates’ resume characteristics. Lastly, the effect of the ask salary on the bid gap is estimated, with and without the resume characteristics controls.
Exploiting a Reform: Understanding the gender differences on the ask gap
Finally, to empirically estimate the market-level effects of an increase in women’s ask salaries, the author studies an unanticipated feature change that affected a subset of candidates on the platform and induced women to ask for more. In mid-2018, a change on the platform affected the way some candidates were prompted to report their ask salary. Specifically, before the reform the ask salary was an empty text field, and after the reform the field was pre-filled with the median of the bid salary for a comparable candidate on the platform. Using an Interrupted Time Series design, the author compares individuals who created a profile before the change and after the change in order to isolate the effect of the reform on ask salaries and on bid salaries.
On one hand, the estimation of the ask gap indicates a 7.2% raw ask gap between men and women. When controlling for all the resume characteristics from the candidate’s profile, the adjusted ask gap decreases to 3.3%, and when adding controls for experience, location, and job title first narrows the gap down to 4.3%. On the other hand, the estimation of the bid gap indicates a 3.4% bid gap between men and women, which reduces to 2.4% when adjusting for candidates’ resume characteristics. However, when controlled by the ask salary the bid gap disappears indicating that gender differences in ask salaries can essentially explain 100% of the bid gap.
At the market-level effects, prior to the reform the ask gap was 3.3% but after the reform the ask gap goes to zero. The author provides evidence that this evolution in the ask gap is mostly led by women asking for more rather than by men asking for less, since the reform led women to ask for 2.3% more while men asked for 1.1% less than they would have otherwise.
The pre-reform gender ask gap is much larger for candidates with more experience: while the ask gap before the reform is 1% for candidates with 0-4 years of experience as a software engineer, it rises to 5.4% for candidates with 4-10 years of experience in this occupation. In terms of bid salary, the gap before the reform was of 2.4% and it goes to -0.2% after the reform. This result is driven by the fact that women are offered 1.5% more and men are offered 1.1% less than they would have been offered absent the reform. At the extensive margin, results suggest that women face little penalty for demanding wages comparable to men since they did not experience a differential effect on the number of bids received after the reform.
In particular, while resume characteristics can reduce the final offer gap to 1.8%, adding the ask salary to the controls reduces the final offer gap to an insignificant 0.5%.
The author also explores heterogeneity based on the candidates characteristics. There exists a subgroup of female candidates for whom the ask gap is close to 2.5 times as large as the APE, and there is a subgroup of women who actually ask for higher salaries than similar men. The group with the largest ask gap (7.8%) is more experienced (13 years vs. 7 years of total experience). They are more likely to be unemployed, with longer unemployment spells, less likely to have a computer science or an IvyPlus degree, and less likely to list highly-emanded coding skills.