The weak point in banks' rosy pay-gap numbers
Do “adjusted” gender pay gap metrics mean anything?
Over the past month, under pressure from the socially conscious investment firm Arjuna Capital, three banks — Bank of America, Citigroup and Wells Fargo — have become the first to publish adjusted pay gap numbers.
They all reported the same thing: Women in their organizations make 99% as much as men, after adjusting for factors like type of job and geographic location. That means two tellers in a Boston bank branch are likely to make the same salary, whether they are male or female.
This is no surprise. Legally, U.S. companies have to provide equal pay for equal work under the Equal Pay Act.
Basically, the banks have chosen a forgiving way of measuring pay gap.
There are two ways to do it. A “raw” or median pay gap calculation —the kind the U.K. government is requiring companies to report starting April 4 — is straightforward: Subtract women’s median earnings from men’s median earnings, and divide that number by men’s median earnings. It measures pay differential across the organization, not just among employees who work at the same level.
An adjusted pay gap calculation uses a statistical model that takes into account factors such as job function and tenure. This is what the banks have chosen. Critics of the method say it sidesteps one of the biggest questions of gender equity — whether women are disproportionately stuck at lower-level jobs.
At large banks, typically women make up at least half the workforce, yet only about a quarter of the much-better-paid executive ranks. So their median pay gap has to be much larger than the adjusted pay gap banks are reporting.
At least one expert questions the banks' use of the adjusted pay gap measure.
"Congratulations to the companies for being concerned about gender equity and attempting to be more transparent," said Mary Blair-Loy, founding director of Center for Research on Gender in the Professions at the University of California San Diego. "However, what they’ve released is so incomplete, there’s no way for objective outsiders to assess the usefulness of these numbers."
Depending on the controls used in the model predicting pay — like role in the organization — gender differences in pay can fluctuate widely, she said.
"It’s not clear what this announcement does other than try to make banks look good," she said. "It certainly flies in the face of decades of research. I’m not saying it’s wrong, but it looks like a piece of marketing.”
Could reporting the adjusted numbers lead these banks to a sense of complacency?
“Oh, yes,” said Natasha Lamb, managing partner and portfolio manager at Arjuna Capital.
Nonetheless, she is optimistic.
“To date, the ‘like work for like work number’ is what we’ve been able to get companies to disclose,” Lamb said.
The number is meaningful, she said, because it’s a start. The banks have been hard to persuade to do even this much. Often when Arjuna Capital begins engaging with companies on this concept, they haven’t looked at or analyzed the numbers at all.
“This is such an improvement over this ‘trust us [we have no pay gap]’ approach we’ve seen for so long,” she said.
She also points out that the 99% adjusted pay gap should be 100%. “I don’t think the work is done even there,” Lamb said.
Wells Fargo’s approach
Citi and Bank of America declined interview requests.
Wells Fargo has the best record of the three banks for promoting women into the higher ranks (and therefore paying them well). Half its midlevel managers are women, and so are 37% of senior executives.
Wells Fargo puts employee data through a model developed by Mercer to calculate its adjusted gender pay gap.
“We aggregate people by similar roles, so we’re looking at comparing people who do similar types of work,” said Mike Branca, head of compensation at Wells Fargo. “It’s not exactly the same, to make sure we’re being fair across our organization. We cluster them by role.”
Then location is figured in. “If you live in San Francisco versus Des Moines, that would be a legitimate driver of a pay differential,” Branca said.
Tenure and experience are also factors, as well as whether the person works full-time or part-time.
Mercer runs the model and identifies any gaps between the expected and actual pay for a role, accounting for the other factors.
“We take that information and we fix whatever gaps we see,” Branca said.
Wells Fargo does not use performance metrics in its calculations, as Bank of America does, according to a recent news report.
“We’ve thought about it, but in our conversations with Mercer, we think the fairest way to do it is the way we’re doing it now,” he said.
Academics have long asserted that unconscious gender bias can creep into performance metrics. Blair-Loy pointed to research conducted by Louise Roth, a professor at the University of Arizona. She studied several investment banks and found that measures of performance differ widely by the gender of the person being assessed.
“We have these biases that filter our assessment of competence,” Blair-Loy said.
Wells Fargo also measures gender equity in other ways besides pay.
“We look at how people are progressing through the organization — women, ethnic minorities or other groups of people — to make sure we’re providing fair opportunities for development,” Branca said.
Wells Fargo measures how many women and minorities it has at different levels of the organization and how promotions are being given.
“These are all things we’re super focused on,” Branca said.
Wells Fargo will disclose median pay gap numbers for its U.K. employees starting in April, as required.
Branca said the bank is committed to improving pay equity and representation of women and minorities across the organization.
“We just think it’s good business,” he said. “Our goal is to make sure we have the best people in the best roles to provide support to our customers. We want to make sure we’re attracting, developing and retaining people and it’s important to ensure that we’re doing that in a fair and equitable way. At the end of the day, it’s good for us, it’s good for our communities, and it’s good for our customers.”
Pushing for disclosure
Lamb has been on a campaign to get companies to disclose pay gap numbers for more than a year.
“It’s something that as a woman working in finance I’ve always been aware of,” she said. “Women are 20% more likely to leave a career in finance than any other career. It’s one of those fields where often more than 50% of employees are women, but they don’t hold those higher-paying positions. Something is broken.”
In finance, she noted, firms are having trouble attracting female talent.
And as an investor, Lamb views gender and racial diversity in leadership as a way to improve company performance.
“There’s a whole body of research that shows that gender diversity leads to a higher return on equity, higher profit margins and higher stock prices,” she said.
Three of the six companies Lamb has pressured to release pay gap numbers have done so. JPMorgan Chase, American Express and Mastercard are the holdouts.
Even after more banks are reporting their adjusted pay gaps, Lamb said her work in this area will continue.
“Step one is making sure women are being paid fairly for the work they’re doing now,” she said. “Step two is moving them up the ladder, attracting and retaining that talent.”
The main reason women don’t get promoted at the same rate as men in banks is still cultural, in Lamb’s view.
“If Silicon Valley is the boys club of the West, Wall Street is the boys club of the East,” she said. “They’re heavily male-dominated, male-led. Because of how entrenched that has been for so long, there hasn’t been a lot of room for women, and it’s difficult to come into a male-dominated culture as a woman.”
The bro culture also causes women to drop out.
“It’s not like they’re dropping out and going home, they’re moving to other industries,” Lamb said. “So there’s been a talent drain particularly on Wall Street because of the culture and inequity that’s present.”
How do you measure fairness?
Lamb would like to see U.S. banks start reporting the median pay gap number that the U.K. will begin mandating in April.
“That’s not an adjusted number, it reflects the structural deficiencies at these companies and it tells part of the story,” she said. “We would like to see that number and analysis expanded to their global operations and see some goals around closing that gap.”
Brian Levine, partner and leader of the workforce strategy and analytics group at Mercer, which handles pay gap analyses for several large banks, including Wells Fargo, noted that the median pay gap is also an imperfect measure, in that it doesn’t account for the fact that women and men have different roles and levels of experience, as well as for location differences.
“Women sit in different levels of the organization and that’s an issue that all organizations are working on,” he said. “But that isn’t a reflection of whether or not women and men are paid equitably for a comparable contribution or comparable role.”
Banks also use other types of gender equity analyses, like representation across levels of an organization, opportunities for advancement and promotions, Levine said.
“Are people who are sitting in the same kind of roles with the same kind of experience moving up in similar ways? That’s an analysis to address this issue of improving representation across the hierarchy,” he said.
Another type of analysis Mercer sometimes does for banks is of looks at the drivers of upward progression.
“You might find in a statistical analysis that the span of control of the supervisor is related to advancement,” Levine said. “Employees who report up to supervisors with smaller spans get more attention, better training, they’re represented better and they move up better. If that’s the case, are women and men equally likely to report up to supervisors with small spans of control?” This again, points to specific actions companies should take.
Some banks also analyze the rate of hiring by gender and ethnicity across different levels, and the rate at which women and men are promoted, Levine said.
“These organizations are spending significant dollars to eradicate any issues,” he said. “We find they do take action.”
The data behind these analyses tends to be accurate, according to Levine, because the pay data comes straight from the banks’ payroll systems and the lines of business and job titles from HR software.
“Are there holes in the data? Sure. Are there opportunities to improve the data? Absolutely,” Levine said. For instance, Mercer typically doesn’t get experience data, it creates a proxy for it using age, tenure, length of time at the company, and time in job. But women sometimes take breaks, for instance to have children, and this Mercer cannot account for with that data.
These banks may have the best intentions. It may be true they’re making every effort to create a work environment that’s fair to women.
But an adjusted pay gap isn’t a true reflection of whether women are being treated fairly in an organization.
If banks want to report that they have no gender pay gap, let’s see the raw gap numbers.
Editor at Large Penny Crosman welcomes feedback at email@example.com.