Google Pays Female Workers Less Than Male Counterparts, Labor Department Says

What they're counting matters. It seems if Google has the goal of hiring more females than industry average, then the stats are stacked against them when you look at their worker pay distribution, even if Google isn't doing anything wrong.

Suppose the population is distributed in their experience and productivity, and that you can measure these to some extent.

Suppose your field has way more males than females. This is not your fault, more of a pipeline thing, but you want to be more equal internally than the outside world is. Most people commend this kind of thing.

If you are fairly paying both genders, wouldn't you necessarily have to take more unproductive and inexperienced female engineers to reach your goal? And if you're fairly paying based on those criteria, won't the female population be paid less as you try to take on a more-than-average number of them?

I'm not saying any of this is how it ought to be, just that it seems like a typical conclusion with napkin math, and would result in these allegations while Google is actually trying to do good.

Interestingly if equal pay were not a thing, I would expect the opposite headline: That Google would pay women much more than men to work so that they could attempt to correct a gender imbalance internally that exists in the external population (software workers).

I can't read the whole article, but does it (or DOL documents) include details like amount of experience and job function?

I've heard that much (not all) of the aggregate gap (i.e., on the national level) goes away after controlling for those variables.

Curious if that applies here, and if so, to what degree.

Edit: consider recent headlines promoting outrage at Marissa Mayer's successor earning double her salary because he's a man. Turns out he's a veteran executive with 20 years of C-level experience. Mayer didn't have that. It may be an extreme example, but it's a good reminder to be cautious about pay gap assertions.

Note that this is not originally a WSJ story; for those without a subscription, you'll learn just as much by reading from the same news source that WSJ is commenting on (and which they link to in this article):

https://www.theguardian.com/technology/2017/apr/07/google-pa...

Of course, when I say "just as much" I mean "not much", because neither party has released anything substantive to rebut the other's claims. Until then, the court of public opinion is going to be both passionate and completely uninformed. But if I'm being cynical, that's probably the point of reporting on it at this stage.

Until the government releases their methodology it is hard to know how factual this assessment is.

Google has done a lot of work internally to even up the gap and claims that they have achieved parity.

https://www.theguardian.com/technology/2017/apr/07/google-pa...

Either party could squelch the un-informed debate by releasing the data and the methods used to arrive at their conclusions.

It is hard to discern what is going on, and it feels strange to say I trust Google ahead of the government.

Google is full of engineers who prize logic and fairness above most else, and culturally most the engineers I meet in most places espouse gender parity and equality. In light of this, it is hard to believe Google would not find a way to even up payscales. That said, while non-management says one thing it is often the case that management find a way to enact the opposite.

Something in me feels like this has the air of a dirty underhanded PR-war to it, and the government is trying to pressure Google into something we don't know about...

It would be surprising and novel if the DoL statisticians had produced an unambiguous correct analysis. They're working in a complex field, and under pressure and regulated to use the wrong tools in many cases---so I'm sure plenty of the people there are excellent and could do great work---DoL statistics in general are not a good basis for action.

Example: DoL is required by law to track non-white applicants and hiring to jobs. Alice is not white, and applies to ten open jobs at Example.com, a hot new startup. She's hired for one. DoL reports this as 90% of non-white applicants being rejected. If 10 non-white applicants each apply for all those ten jobs and are all hired, one job each, that's 90% rejection.

A mix of law and regulation requires counting this way. Near one edge case (plausibly including SV today!) the error doesn't matter much. But as you get anywhere near justice, this isn't helpful.

So without seeing the details of this analysis, I'm skeptical that DoL's way of counting is helpful here.

The US Dept. of Labor is suing Google for not revealing compensation data [0].

The Office of Federal Contract Compliance Programs (OFCCP) has suspicion on pay disparities and in order to investigate further sent a request to Google to access some records (Point 9) [1]. Google denied OFCCP access to these records. Now, DoL has taken Google to court to get access to the records they need to investigate further and complete their evaluation. I wonder why Google is being so evasive about it.

[0] https://www.dol.gov/newsroom/releases/ofccp/ofccp20170104

[1] https://www.dol.gov/sites/default/files/newsroom/newsrelease...

The DoL claims "discrimination against women in Google is quite extreme, even in this industry", while Google claims "every year, we do a comprehensive and robust analysis of pay across genders and we have found no gender pay gap".

DoL is probably lacking data, and has a somewhat simplistic methodology (i.e., they see woman are paid less, but they don't know why). I highly doubt Google is intentionally discriminating against woman, I actually think they are probably trying very hard to have equal employment practices. But they still might be discriminating without knowing it.

Here is what could e happening: The DoL says "you pay women less, even when they have the same experience". Google says "we pay for performance, not experience, when you take performance into account there is no pay gap".

What Google might be missing is that there are unconscious biases causing performance reviews of woman to be more negative. I think that is highly likely. My understanding is that Google is highly reliant of performance reviews and performance pay, so under those conditions the DoL might be correct in saying discrimination is "quite extreme, even in this industry".

Assuming labor markets are liquid, if the pay gap is true it means either: 1. Women really want to work for Google and take less pay than working at other employers or 2. Women at almost all employers are receiving less money.

There were times not too long ago when there were Jewish quotas. Nobelist Physicist Richard Feynman was denied entry into Columbia and he went to MIT instead.

Nobelist Medicine & Physiology Author Kornberg was one of 3 of 200 Jewish applicants to get into medical school from his university.

Those universities/firms that didn't have Jewish quotas and discrimination got a good boost in performance over those who had quotas and discrimination.

If they women are only getting paid less at Google and not other firms then they should change firms. If other firms are paying less, then Feds should go after them as well and not single out Google. Any firm that pays women as much as men should be able to hire the competent women from Google.

In another part of the WSJ article, there is a complaint that most of the employees are male and asian or white. They should examine the portion of people attending top engineering schools: male and asian or white.

At what point is a company responsible for discrepancies in pay due to behavioral differences that likely exist between genders. Most jobs have a salary range for each title and they present candidates an initial figure at the low end of that range. To get a figure at the top end of that range, the candidate needs to initiate negotiation and have the skills/experience/practice doing so, including the knowledge that they need to do things like talk about competing offers and avoid divulging their previous salary.

There is plenty of research showing negotiating differences between the sexes (including differences in negotiating based on the gender of the counterparties relative to one another).

Shouldn't the government instead compare discrepancies between initial offers for the same position instead of actual salary negotiated?

If they don't use initial offer, then the business needs to correct for this and every way I can think of to correct destroys the firm's ability to recruit the difficult candidates while still remaining competitive with other firms. The firms that correct for these differences will pay more on average for labor than its competitors and will likely lose out on the capable candidates that know their self worth and demand compensation that reflects that.

Can someone who actually has access to the story answer the following questions for me. Is this before or after controlling for:

- field of work. Is this the whole company or just devs?

- years at the company.

- age, as a crude proxy for total experience at all companies.

Also does it say the magnitude of the gap?

There's a number of ways this headline could be true, all may be unfortunate from a societal perspective, but it only shows bias on Google's part if this is after controlling for the things companies say they base their pay on.

This often happens in cultures - not sure Google is such a culture - were you negotiate your initial salary, raises and need to push for promotions.

When advising women on interviewing I always suggest they ask for more money than they are comfortable with.

Methodology is important here. A lot of the logic behind pay gap talk assumes that the jobs men and women do in the same company are exactly the same, as if these are assembly line jobs and worker output can be compared in a direct and straightforward manner. Anyone who works in tech knows that what each person contributes varies greatly and there's no way to directly compare what two people do or even the amount of experience they bring to the table when they're hired.

What is that Google would gain from this? Not that they are strapped for money. It could only lead to such bad publicity. And if the women engineers were hired by Google, they obviously are good at it. Can't understand the rationale behind it, if true.

My incoming is already intimate knowledge to the state and my employer. Why isn't the DoL doing this type analysis across all employers? What exactly has to be divulged that the gov doesn't already know?

Has anyone ever considered that women regardless of industry get paid less because they don't negotiate for more? As a female from the tech and other industries including talent acquisition rarely do I come across another female who asks for more or alternative compensation in either the hiring, review, or promotion stage of their career. As a benefits manager I've known the top limit companies will pay and sat quietly as female executives accepted offers that were 20-30k away from what the company was willing to give. Bottom line if you don't ask for more you won't get it. Men ask, women rarely do.

I'm not certain why anyone would be surprised. Remember the spreadsheet from 2015?

https://duckduckgo.com/?q=google+spreadsheet+compensation+le...

They should replace all their men with women and reap the profits

Companies like google are obsessed with diversity. The idea that they'd systematically pay women less is laughable - you will not find more liberal environments outside of academia.

I can't help but wonder if this is politically motivated.

Just identify those workers as male. Problem solved.

Another (non-paywall) article:

https://www.theguardian.com/technology/2017/apr/07/google-pa...

Seems like it is early days on this lawsuit. The DoL are asking for more documentation about employee's pay and Google is calling it a "fishing expedition." This is definitely one to watch, but I won't be drawing any conclusions this early in this lawsuit (since even the DoL say they need more documentation to prove their case).

Prior HN comments and no paywall Guardian story: https://news.ycombinator.com/item?id=14063043

Link to avoid paywall https://www.facebook.com/flx/warn/?u=https%3A%2F%2Fwww.wsj.c...

Non-paywalled but not identical version: http://www.latimes.com/business/technology/la-fi-tn-google-w...

There is no such thing as equal pay for equal work. Two people within the same gender aren't paid equal in the Tech or any Professional services Industry forget across genders. More importantly, can someone define what equal work is? Same role/title doesn't translate to equal work. Every person has different capabilities even within in the same role and if it happens to be that a Woman outperforms than her peers then she should be paid equal or more and I wholeheartedly support that but the sad part with Women is that they never talk about such cases but instead Just hangs on to the phrase "Equal pay for Equal work" so they know if they make enough noise they can get equal pay but they never mention a word about equal work. And after working for 10 years in the Industry (which means I am young so don't think I am talking about some B.C era) and statistically speaking most of the time it had always been men who did tasks that are above and beyond from their regular duties. Sheryl Sandberg Just spits out some none sense. She just got so Lucky to be the COO at FB in the first place. she thinks that if a company can control # years of experience, education level, and blah blah then that will be the deciding factor for Equal pay. I am like really? So two people with same amount of years of experience can perform the same way? Two people with the same education level are equally bright? What if one person went to college Doofus and another went to Stanford? I am sure going to better college doesn't automatically translate into a bright and efficient person for the company but it is an accomplishment one should acknowledge for their hard work. so in short, none of these factors can be controlled and the way most companies Judge pay depends on how well one performs in the Interview which is very much reasonable. If we do the Sheryl Sandberg way we actually don't need an HR department (which Btw, is lot of Women) we can just write a program to accept or reject applicants if is really that straight forward to control the factors she is talking about. I think it would be great if someone can actually make Sheryl Sandberg to code on a daily basis and deal with all the production issues. I went to an Ivy League school and I worked as a TA for many Computer Science courses. I had worked with many students (both Men and Women) I had seen them getting Jobs at Google and FB and whenever they told me they got into Google or FB I used to ask my students so What questions did they ask you? Typically it is Algorithms most of the time as we all know and I can for sure say that the Algorithms questions asked to Men are way Significantly harder than Women. Now why is this not highlighted? Because companies like Google, FB wants to highlight that they hired Women. I bet my entire savings (I truly mean it) that if I go Google even today and pick random sample of 50 (25 Men Vs 25 Women) and randomly ask the Algorithm questions Men will significantly outperform Women because I know the level of difficulty questions asked for a Women and Men to get into Google. This is totally unfair. And I hate Men for giving up on these issues. I am sure most companies know all of this. They want to pay equal because they don't want to deal with one more hassle. If I am a CEO of a large organization I will be all over the media asking Women define what Equal work is before they start talking about Equal Pay. And I would look at it case by case If a woman outperforms her peers then As a CEO I would actually pay more period.

I can't wait to see Googles response. I have no idea why they would even open themselves up to this criticism when they could easily align the salaries.