Cafe Classic: 77 Cents and Gender Discrimination - The Wrong Conclusion

Editor's Note: Nearly six years have come and gone since we first published this Classic post and the "77 cents on the dollar" myth still persists.  Here once again to explain why this is the wrong conclusion - drawing on the original 2007 research from which this statistic was taken - is Stephanie Thomas.

Can you believe it's the 21st century, and we're still talking about pay equity and the gender wage gap?!?

Last week (this post originally published July 27, 2010), President Obama urged Congress to pass the Paycheck Fairness Act. During an interview discussing the Act, White House senior adviser Valerie Jarrett said "women deserve equal pay," and cited a statistic: women earn 77 cents for every dollar men earn. 

This "77 cents" statistic is certainly making the rounds- earlier this year, Acting EEOC Chairman Stuart Ishimaru stated, "The wage gap is alive and well in America, with the typical full time year round female worker making $0.77 for every dollar earned by her male counterpart."

What's the first thing you think of when you hear that statistic? If you're like most people, you think of gender discrimination. Women earn 77 cents for every dollar earned by men because of discrimination. 

I'm here to tell you that's not true. The "77 cents" statistic can't be due to gender discrimination. Here's why.

In 2007, Francine Blau and Lawrence Kahn published an article that looked at earnings of men and women, and found a difference of 23 cents per hour; this article is the original source of the "77 cents" statistic. But they didn't stop there - they tried to understand what was causing the difference. Here's what they found:

 *Blau and Kahn also included education; they estimate that education actually closes the wage gap by 6.7% (not included in my chart since it's a negative number).

Blau and Kahn found that 59% of the gender differential could be explained by non-discriminatory things: experience, chosen occupation, chosen industry, etc. So the "77 cents" statistic can't be due to discrimination:

• Estimated wage gap based on "77 cents" statistic = $0.23 per hour
• Amount explained by nondiscriminatory factors    = $0.14 per hour
• Amount NOT explained                                           = $0.09 per hour

According to Blau and Kahn, the most that could be attributed to discrimination is $0.09 per hour. And this assumes that their model accounts for ALL legitimate nondiscriminatory factors.

Are there legitimate nondiscriminatory factors that were omitted from their model? Probably - no model is perfect. Some people have argued that men are better negotiators than women, and because of this men tend to get higher starting salaries. Are differences in negotiating skills discriminatory? Perhaps, based on the way that we raise our daughters (that's a sociological issue). But the employer can't be held responsible for differences in negotiating skills, can he?

Something else to consider: overtime hours, shift premiums, etc., may cause a difference in earnings between men and women, even though their base rates of pay are the same. If a woman chooses to work fewer overtime hours than her male counterpart, resulting in lower earnings, is that discrimination?

If we were able to build the perfect model and study every single legitimate nondiscriminatory factor, the unexplained portion of the gender wage gap would likely be less than $0.09. And even then, it doesn't automatically follow that the remaining unexplained portion of the gap is directly attributable to the employer engaging in gender discrimination.

So now you know. The next time you hear the "77 cents" statistic, stop and think about what it REALLY means...

Stephanie Thomas, Ph.D., is a Lecturer in the Department of Economics at Cornell University and the Program Director for the Institute for Compensation Studies (ICS) at Cornell’s School of Industrial and Labor Relations. Throughout her career, Stephanie has completed research on a variety of topics including wage determination, pay gaps and inequality, and performance-based compensation systems. She frequently provides expert commentary in a media outlets such as The New York Times, CBC, and NPR, and has published papers in a variety of journals.

Comments

[E. James (Jim) Brennan]

A 9% pay gap sounds right to me. The gender pay differential for ACA members (WorldAtWork used to be the American Compensation Association) back in 1981 when Howard Risher, PhD, took all those variables into account was 14.3%. That research and more was cited in my "pro" Comparable Worth article in the July 2010 Workspan. Today, college women have out-graduated men and we have experienced a feminization of the professional work force, so a lot more women have entered those highly paid "male" jobs. All else held equal, women earning 9% less than the male full dollar makes sense.

Not much has changed since I first posted that comment back in 2010, except that Howard has become an occasional guest contributor here. It is still an accurate, timely and important article!

[Tony Bergmann-Porter]

But Jim, if that were the case, a smart profit-oriented employer would hire an all female workforce (of presumptively equal capability) and enjoy the competitive advantage of 9% lower labor costs.

That this just doesn't happen has always struck me as the most convincing refutation of gender-based pay discrimination.

This is certainly not to say that there isn't an earnings gap, but that's a different discussion.

[E. James (Jim) Brennan]

One cartoon parody of the older ratios that ran from 60% to 70% joked, "Don't hire a man, hire two women!" I believe those who think that way are some of the ones who actually drive female pay lower when they do indeed follow the temptation to low-ball women with rates that would insult a man. Haven't you seen that happen, Tony?

[Ross]

I have never seen women low balled on salaries. Actually, I have seen the opposite---a conscious effort to pay women the same to avoid any claims of discrimination.

[Ross]

What is the title of the 2007 study by Blau and Kahn referenced above? Couldn't find it. Thanks.

[Ann Bares]

I believe this is it -
https://web.stanford.edu/group/scspi/_media/pdf/key_issues/gender_research.pdf

[Tony Bergmann-Porter]

More or less dittos to Ross. I have never seen a lowball offer to a female candidate merely because of her gender. I have seen budget-constrained lowball offers made to (and rejected by) candidates regardless of gender.

I also have seen women (and men, for that matter) promoted into "stretch" situations for which they are paid less than a (fully qualified and experienced) external hire would be. And, as we all know, an external hire of any gender will typically require a risk premium for changing employers.

[Ross]

Thanks for sending the reference to the article quoted, Ann. Generally speaking, I believe that it is dangerous to cite one study to substantiate a major idea in human resources, such as the one proposed here...the gender pay gap is really just $.09 per hour.

Other research studies are necessary in my opinion to confirm or refute this finding, as I have found that when different methods, samples, and time periods are used, the results of research studies can vary substantially.

I would be most convinced by a study involving an in-depth examination of personnel records and interviews with HR professionals who did the hiring in a random sample of organizations of all types.

[Stephanie Thomas]

Hi Ross - Thanks for your comments. Blau and Kahn have done a number of different studies using Census and Bureau of Labor Statistics data, and their results are consistent time and time again, and are also consistent with what other researchers looking at the same questions are finding / have found.

Looking at the detailed HRIS records for a sample of organizations would be an interesting project! I suspect that when one controlled for organization, the estimated differential would be smaller, since we would be nearly eliminating the industry differential, and and occupational differences may (or may not) be smaller. Differences in the cash-benefits tradeoff, flexible scheduling, etc., would also likely be reduced since we would expect the same employer to offer the same kinds of benefits to all employees irrespective of gender.

I agree with Ross and Tony in that observed wage differentials by gender in the workplace are not necessarily attributable to discrimination, and that the 9 cent differential is likely explained by a variety of other factors (hours worked, cash/benefits tradeoffs, skills/abilities/qualifications, "stretch" versus "non-stretch" assignments, etc.)

From my perspective, it's dangerous to perpetuate the "77 cents" statistic that aggregates everyone by gender and ignores real differences between individuals that are meaningful to the compensation choices organizations make.

[Ross]

In their latest research (2016 report), Kahn and Blau concluded this about the wage gap:

A decrease in the unexplained gap over the 1980s contributed to the robust convergence in the gender wage gap over that decade, with the unexplained gap falling sharply from 21-29% in 1980 to 8-18% by 1989. However, the unexplained gap did not fall further subsequently, remaining in this range over the succeeding 20 years.

http://ftp.iza.org/dp9656.pdf

[Ross]

A bit more from the 2016 report that continues from the above quote and appears important:

We also found that both the raw and the
unexplained gender pay gap declined much more slowly at the top of the wage distribution that at the middle or the bottom. By 2010, the raw and unexplained female shortfalls in wages, which had been fairly similar across the wage distribution in 1980, were larger for the highly skilled than for others, suggesting that developments in the labor market for executives and highly skilled workers especially favored men.