There is a 23 percent wage gap between men and women in the US – women working full time still earn just 77 percent of what men earn. For years now we’ve been told that the gap narrowed considerably in the 1980’s due largely to women’s progress in education and workforce participation and that progress has since stalled, but a new study from the University of Georgia’s Jeremy Reynolds and Jeffrey Wenger has revealed some shocking results.
While writing a paper about how couples deal with health insurance arrangements when sick, Reynolds came across an interesting fact: when a spouse reports on the health of their husband, they tend to say their husband is less healthy than their husband believes himself to be and the same is true for husbands reporting on the health of their wives. This unexplainable quirk got the professors wondering about what other issues self-reporting affected.
As it turns out, it’s made much of the data about the gender wage gap seem unreliable at best.
The Trouble with Proxy Reporting
According to Reynolds’ and Wenger’s study He Said, She Said: The Gender Wage Gap According to Self and Proxy Reports in the Current Population Survey (CPS), roughly half the labor force data in the CPS are provided by proxy respondents (people in your household who answer on your behalf) and since 1979, men’s reliance on proxies has dropped dramatically while women’s reliance on proxies has increased. Obviously, this raises questions about the accuracy of obtaining information from a second party rather than the individuals themselves.
Think about it this way: if you’re away at work, how much does your spouse, teenage child, parent, or other family member really know about your earnings and hours during any particular week? Chances are, not much, but these are the people who are reporting your income information to those at the population survey. Most disconcerting is that proxy reporting doesn’t account for a small percentage of the information gathered by the CPS. Rather, it accounts for a rather large chunk. As recently as 2009, proxy respondents provided 46 percent of the data collected. If it seems so ridiculously unreliable, why is the CPS still collecting data this way? According to Wenger, it comes down to one issue: cost.
“The CPS collects data every month and the assumption is that one person in the household can provide data for every person in the household. Speaking to every person personally or conducting several interviews to compare the data received just isn’t an option because of the extra time and manpower it would require. It would simply be too costly,” Wenger said.
50 Percent Off
Back to the point: How is all of this affecting the gender wage gap? The information gathered from proxy reporting is the information used to calculate the gender wage gap and according to the researchers, statisticians have been as much as 50 percent off in tracking the progress of women’s wages in the work force.
This means that gender parity has progressed much faster than previously believed, but it also means that the gap was wider to begin with than anyone realized, especially in the 80s when it was assumed women were making major strides to close the gap. This is because women were the ones providing the information – they were usually the ones who happened to be home to answer the call. As a result, they over-reported their own earnings and under-reported their spouse’s wages. And though it’s believed that everyone over-reports their income, it’s also believed that men in particular do this more.
“We have theories as to why men do this more often than women,” Wenger said. “Men are more socialized to take on the breadwinning role. If the ‘breadwinner’ isn’t making as much as his wife, he may feel threatened and subconsciously – or perhaps consciously – underreport what she’s making. We can’t say for sure, all we do know is that men especially tend to overreport for themselves and underreport of their spouses.”
Reynolds and Wenger identified a set of biases in the existing data on the gender wage gap, revealing that people tend to underestimate the wages of others and overestimate their own. These biases were identified by examining data from the CPS. Respondents are interviewed multiple times, one year apart. When the researchers looked at how responses to these questions changed across the subsequent interviews, they found that people answered more generously for themselves than other people had for them.
After calculating these biases, the researchers went back to the data routinely used to measure the wage gap and readjusted it. It was discovered that as more women entered the labor force, men became more likely to answer surveys for themselves. The existing analysis suggests that the wage gap in America between 1979 and 2009 closed by about 16 percent (or $1.19 per hour). The researchers, however, put that number at 22 percent (or $1.76), meaning this basic calculation was off by 50 percent.
“In no way are we saying that the gender wage gap is smaller than previously believed or that it doesn’t matter. Surely it’s little comfort to women today who are still being paid less that women in the past were being paid significantly less than they were even aware of,” Wenger said. “Our research tells you nothing about the gender wage gap today; all we’re saying is that women were worse off than believed and they’ve come a lot further than we even knew. We’re academics; we just want more accurate answers to important questions.”
Wenger says that another source of inspiration for revealing this information was the fact that a lot of ink has been spilled on the gender wage gap with many of the results being dispiriting. Wenger cites the now infamous 1997 study Swimming Upstream: Trends in the Gender Wage Differential in the 1980s by Francine D. Blau and Lawrence M. Kahn, which implied that despite women making major advances in education, the gender wage gap actually worsened in the 90’s.
“The thesis was that women were ‘swimming upstream,’ swimming against a tide of discrimination and fighting against the gap. It was a disheartening story and I’m sure it left many women wondering what the point was. Why even pursue an education? These were important findings and we wanted to know if the research was true,” Wenger said. “While conducting our own research we found that there are some things that might have gone wrong; we found less evidence of ‘swimming upstream’ than Kahn and Blau did. Our findings won’t change how data is collected, chances are proxy reporting is going to continue for a long time, but maybe our study will inform policy or bring awareness to some of the issues surrounding proxy reporting.”