Tuesday, March 05, 2019

Does Google Discriminate Against Men In Pay?

Google has carried out its annual pay fairness study since 2012.  The study, which covers 91% of Google's employees, compares total employee compensations while standardizing for job type, job level, performance and location.  Standardizing, in this context, means that those four variables (and perhaps others not mentioned here) are held constant in that analysis.

These studies are then used to give additional pay to some employees who appear not to be paid enough given their job type, job level, performance and location.

The results of Google's most recent study have raised lots of eyebrows, because:

Google has given raises to thousands of men after an analysis of Google's pay structure found that the company would otherwise be underpaying those men relative to their peers, The New York Times reports. The analysis also led to raises for some women.
Google determines annual pay raises in a three-phase process. First, Google adjusts every employee's compensation based on standard factors like their location, seniority, and performance ratings. Managers can then seek additional discretionary raises for their best-performing employees.
Finally, Google performs a company-wide analysis to determine whether these raises are biased in terms of race or gender. If biases are detected, the disadvantaged workers are given additional raises to eliminate the discrepancies.
"We provided $9.7 million in adjustments to a total of 10,677 Googlers," the company said in a Monday blog post describing the results of the equity analysis.
"Men account for about 69 percent of the company's work force, but they received a disproportionately higher percentage of the money," the Times's Daisuke Wakabayashi writes. "Google said it was important to be consistent in following through on the findings of its analysis, even when the results were unexpected."

So what's going on here?  Did we suddenly wake up in the Opposite Reality, where men are the demographic group which gets paid less, even in such a male-dominated field as tech?

I can't give a truly definite answer to that because I don't have access to the original data or the decisions made inside the organization, and because several of the variables the pay study standardizes for are actually not some external god-given constants, but amenable to human manipulation. 

Who gets promoted and who gets hired for certain jobs, for instance, cannot be studied within the framework of the Google pay study as it holds job levels and job types constant.  To pick an extreme theoretical example,  a system which never promotes women but pays women more than men on the lowest of levels would cause findings similar to these.

That men got most of the pay adjustments is not the only curious finding in this year's pay study.  Another is the actual number of such adjustments, compared to previous years:

In response to the study, Google gave $9.7 million in additional compensation to 10,677 employees for this year. Men account for about 69 percent of the company’s work force, but they received a higher percentage of the money. The exact number of men who got raises is unclear.
The company has done the study every year since 2012. At the end of 2017, it adjusted 228 employees’ salaries by a combined total of about $270,000.

The bolds are mine.
And here is the explanation of the giant increase in the number of compensations:

This year, new hires were included in the analysis for the first time, which Google said probably explained the big change in numbers.

Bolds, once again, are mine.

That's enough data for me to create a tentative theory* about what has been going on: 

It's that Google has been paying more for new hires who are female than for new hires who are male, always in the standardized sense, i.e., for the same job type, performance and location.

If that is correct, why would Google do something like that? 

Perhaps because of a court case three women filed against Google in 2017:

Kelly Ellis, a former Google engineer and one of the plaintiffs in the gender-pay suit against the company, said in a legal filing that Google had hired her in 2010 as a Level 3 employee — the category for new software engineers who are recent college graduates — despite her four years of experience. Within a few weeks, a male engineer who had also graduated from college four years earlier was hired for Ms. Ellis’s team — as a Level 4 employee. That meant he received a higher salary and had more opportunities for bonuses, raises and stock compensation, according to the suit. Other men on the team whose qualifications were equal to or less than hers were also brought in at Level 4, the suit says.
The Google pay study cannot directly address the allegations in the above quote, because it would have compared Ms. Ellis' compensation to the compensation of others also assigned to Level 3, even if she properly belonged to Level 4.

But a court case focusing on new hires could have made Google bend over backwards to give incoming women good compensation packages overall, even if they turned out to be too good packages.

To establish overall patterns in how much Google pays to people of different sex, race and ethnicity requires a wider framework.  That's because promotions, hiring and firing all affect someone's earnings, and the basic Google pay study does not allow us to study those.

* It's tentative, but I find it the best explanation, given the data we are allowed to see.  That so many men suddenly got a pay rise when this did not happen in previous years when new hires were excluded strongly suggests that those rises went to new hires.  The large number also suggests that the reason is the generous compensation packages offered to the far fewer new female hires.