Most workers say employers should not ask about salary history
July 12, 2017
• Survey finds “significantly” more working women than men say pay history should not be asked
• “The time of looking backward to go forward to determine pay is over”
More than half of U.S. workers -- employed or unemployed but looking -- believe employers should not ask candidates about their current or past salary history when negotiating a job offer, according to a new survey from jobs site Glassdoor of Mill Valley.
The survey, conducted online by Harris Poll among more than 1,300 U.S. adults ages 18 and older, comes at a time when new laws are being adopted to address this inherent gender bias in long-standing hiring practices. Several states and cities are currently considering laws that would ban employers from asking about salary history, following similar laws recently passed in New York City, Philadelphia, San Francisco, Massachusetts, Delaware and Oregon, among others, Glassdoor says.
Significantly more working women (60 percent) than working men (48 percent) believe salary history questions should not be asked, the survey finds.
On average, women in the U.S. earn about 76 cents for every dollar men earn on an unadjusted basis, according to Glassdoor Economic Research. This documented pay gap, compounded by the fact that more than two-thirds (68 percent) of women do not negotiate pay compared to half (52 percent) of men, can quickly put women at pay disadvantages, especially when prior salary history is used to determine starting pay in job offers, the company says.
"The time of looking backward to go forward to determine pay is over. Asking prior salary history questions can trigger unintended consequences and introduce bias into the hiring process that disadvantages women from day one," says Dawn Lyon, Glassdoor chief equal pay advocate and senior vice president of global corporate affairs. "We need to reframe the conversation to pay expectations around the value of the job and the skills and relevant experience required to do it.”
She says many companies are already doing this without legislation or regulation. “And, candidates can help change the conversation by offering answers that address their pay expectations based on the role and their current market value, while also taking into account how the company structures its overall pay and benefits package," she says.
While most Americans do not think employers should ask about current or past pay, most do want more pay information up front from employers. Nearly all U.S. workers (98 percent) say it would be helpful to see pay ranges included in open job listings, and 95 percent say it is important to be thoughtful and informed about a company's pay philosophy (such as how pay and pay increases are determined) prior to accepting a job offer.
Glassdoor says this is valuable for employers to consider given that nearly three in four U.S. workers (72 percent) report that a salary and compensation package is among their top considerations when determining whether to accept a job offer.
"Pay is a key area where implicit bias can creep into people processes," says Lori Nishiura Mackenzie, executive director of the Clayman Institute for Gender Research at Stanford University. "Women are often implicitly assumed to be less qualified and thus, have to work harder to demonstrate their worth, especially in roles that are male-dominated.”
She says that due to stereotypes and bias, past salary is not an accurate measure of an employee's value and putting all the onus on the candidate to negotiate their salary is not the answer either.
“It is critical to base offers on what the job is worth, starting with clear criteria and qualifications for the role when making decisions about a total compensation package," she says.
The 2017 survey was conducted online within the United States by Harris Poll on behalf of Glassdoor from March 30 - April 3, 2017 among 1,329 U.S. workers (who are employed or unemployed but looking) ages 18 and older. The online survey is not based on a probability sample and therefore no estimate of theoretical sampling error can be calculated.