Thursday, April 14, 2016

Understanding the Gender Wage Gap

This past Tuesday was Equal Pay Day, which is focused on trying to resolve the gender wage gap. The reason Equal Pay Day is on April 12, is because that's when John F. Kennedy signed the Equal Pay Act of 1963 into law. Now some 50 years later, the wage gap still exists in some form; Equal Pay Day is aimed at reminding us of this fact.

But while it is true that wage gap still exists, it is important to understand the precise details of what exactly that means. The oft-cited statistic that women make 77 cents for every dollar a man makes is as shocking as it is misleading. Only by understanding the definition and causes of a wage gap can we move on to determine what, if anything, could be an actual solution.

In this piece, we'll begin by explaining what the 77 cents figure means as well as what other measures of the wage gap can tell us. Then we'll briefly discuss the likelihood of discrimination, and we'll close by contemplating what stronger equal pay legislation might look like in practice, and whether it would conceivably work.

Lies, Damn Lies, Statistics, and Gender Wage Gap Discussions
A common claim that one hears on this subject is that women make 77 cents in the workplace for every dollar a man makes. This is true, but it doesn't mean what most people (myself included, until quite recently) imagine it to mean.

What many people think this means--and indeed, it is sometimes explicitly stated in this format, incorrectly--is that if a man and a woman have the exact same job in places with the same cost of living, the woman will make, on average, 77 percent as much. If true, that would be distressing. But this is manifestly not what this statistic means.

In reality, what it means is that if you look at the median full-time weekly earnings of all women in the US, across all fields, and divide it by the median full-time weekly earnings of all men in the US, you get a ratio of about 77 percent.

That's an interesting result perhaps, but it actually tells us very little. To see why, let's consider a quick and simple hypothetical. Suppose that the US economy contained only two professions for a moment, economists and janitors, and in this economy, all economists are women, and all janitors are men. And because this hypothetical society takes the maxim that "cleanliness is next to Godliness" very much to heart, sanitation services command a premium. In particular, all janitors earn $100k a year while economists earn just $70k a year. If we were to run the same median earnings calculation that produces the 77 percent number above, we would find that women make 70 percent as much as men in the hypothetical economy. That said, we could also say, in our hypothetical, that economists make 70 percent as much as janitors. The former could imply some kind of untoward discrimination between men and women; the latter is merely evidence that janitorial services are more highly demanded / valued than economics services.

Back in the real world, we find something similar as a cause for part of the discrepancy in median earnings. According to a recent analysis on this subject, approximately 25 percent of the gender wage gap can be explained simply by the fact that women tend to be concentrated in professions that, on average, pay lower salaries than the professions that men dominate.

What of the other 75%? Well, a few other key factors likely to be involved are the following:
  • Number of hours worked (even in full time professions, men tend to work more hours than women)
  • Work experience
  • Educational background (today there are slightly more women in college than men, but there's probably still some catching up to do for average education levels to be roughly equal in the workforce)
  • Career interruptions (for example, to have and raise young children), which tend to limit the rise in one's earnings
A 2007 study controlled for these factors and then compared hourly wages among men and women (thereby, theoretically comparing men and women that had the same characteristics, and worked the same hours), found that this could explain all but 5 percent of the overall wage gap. That is, if this study were taken as gospel, then the general talking point would be updated to say that women make 95 cents for every dollar a man makes.

And that's about as close to equality as the more honest statistical calculations will get us. We can't say for sure what explains the remaining 5%. Perhaps it is completely or partially due to discrimination. And perhaps it's due to women preferring other job benefits, like flexible hours, as opposed to exclusively focusing on pay. Things like that can't be easily captured in the data, and it's not clear how one would quantify their value even if we could identify it.

Another thing to remember here is that all of the statistics discussed above, even the studies that attempt to control for occupations, are still painting with a very broad brush. People working in business and finance might tend to make more than people working in education. But there is still a massive range of jobs that fall under the umbrella of "business and finance", probably anywhere from bank teller to hedge fund manager. To the extent that women or men are concentrated in jobs within industries that are more or less highly compensated, this won't likely be captured in the data.

Note that none of this should be taken to mean either that a) the gender wage gap is a complete myth or b) wage discrimination against women never occurs. It simply means that the usual narrative on this subject is very misleading.

Would Greed Prevent Wage Discrimination?
This is a bit of an aside, but it is useful to clarify precisely what wage discrimination would entail in practice.

As I understand it, wage discrimination implies that an employer will decide to pay a woman less than a man for equivalent work. There might be legitimate reasons for this (i.e. factors that have nothing to do with gender), but let's assume a case where the reasons are not legitimate at all. Our employer is just an avowed misogynist. He just got out of a bad relationship, and he has now decided that all women are to blame for the state of his personal life; paying his female employees less is his effort at attaining justice through collective punishment.

Now let's imagine similar views, or views that have the same practical effect, cause other employers to engage in the same practice of paying female employees less for no good reason. This is, after all, what is necessarily implied in discussions of a gender wage gap, that the pay disparity is entirely without just cause.

In this case, women are being dramatically underpaid relative to what they are worth. It then follows that a smart entrepreneur could come in and offer to pay women more than what they are currently making (90 percent of a man's wage instead of 77 percent), and get two significant benefits from this approach:
  • Since they are paying more than the competition, they could probably choose the very best female candidates for their company.
  • They would still get a discount relative to what they would need to pay a merely average male candidate.
It follows therefore that many entrepreneurs, solely out of concern for running a profitable efficient company rather than any social justice ethos, would jump at the chance to compete for women that are currently being underpaid. Indeed, the only reason that (at least some) entrepreneurs wouldn't pursue this strategy is if everyone thought that even the best female candidates wouldn't perform as well as an average male candidate. Call me optimistic, but I don't view this as a prevalent worldview in the US. And even if it were, you would only need a few business people to break this norm and take a chance on the strategy above. Undoubtedly, they would outperform their competition upon doing so. And over time, this natural competitive pressure would tend to eliminate any wage discrimination entirely.

Here again, this should not be taken to mean that discrimination never occurs. But for better or worse, self-interest tends to be a very reliable predictor of behavior. To assume that wage discrimination is happening and is widespread throughout the economy, we must also assume that the general impulse for misogyny (even among women, since they can be entrepreneurs too) is more powerful than the desire to pursue one's own material self-interest. At least in the US, I find that very difficult to believe.

What Should The Government Do About It?
It may be a frustrating answer to this question, but the best answer is probably nothing.

This is the necessary conclusion once we understand the factors that explain most of the pay disparity, as discussed above:
  • Educational attainment
  • Occupational choice
  • Hours worked
  • Career interruptions
  • Work experience
See the problem here is that the above items, except for work experience, come down to a matter of preferences for men and women. For example, many women go into teaching, which is not a high-paying profession. If the government were to try to correct for this, they only have three options, in order roughly from least to most extreme. They can either a) engage in an educational campaign to try to convince more women to pursue high-paying fields (or convince men to pursue low-paying fields), b) mandate an equal number of men and women get high-paying degrees/credentials in a given year by applying rules on public universities, for example, or c) require all occupations be compensated equally. Option A is possible, but there's obviously no guarantee of success. Meanwhile, options B and C are (hopefully) sufficiently absurd on their face that they do not require a thorough refutation here.

So if they can't really fix the preferences problem, what about work experience? Here they could assign a requirement that all businesses compensate for work experience equally (e.g. $1,000 more per year of experience) or, compensate people equally, regardless of their experience. But of course, this is really just a slightly different form of Option C from above and is in no way feasible. Moreover, who hasn't had a boss they couldn't stand at some point in their life? Would you honestly want them to always get paid more than you, just on account of them having existed longer? I submit that no reasonable person could answer that in the affirmative.

What about the prospect of preventing actual, arbirtrary discrimination? Here the prospects for government intervention would seem to be decent, but in fact, the options aren't much better. In theory, the way the government could help prevent discrimination would be to make it easier for women to sue for discrimination. And for this to happen, there would need to be a more objective and straightforward way for people to prove they had been discriminated against.

President Obama moved in this direction when he signed Lilly Ledbetter Act in 2009, which made it easier to claim discrimination. He also took another step down this path recently, when he announced plans to require federal contractors to report additional pay data. In particular, companies that are federal contractors will be required to report pay data based on ethnicity, race, and sex, and they would also report where it falls within 12 pay bands (0 - $19k, $19k - $24k, etc.). Under the new reporting scheme, each employee would also be categorized among 10 arbitrary job categories, as follows:
  • Executive/Senior Level Officials and Managers
  • First/Mid Level Officials and Managers
  • Professionals
  • Technicians
  • Sales Workers
  • Administrative Support Workers
  • Craft Workers
  • Operatives
  • Laborers and Helpers
  • Service Workers
It's probably pretty easy to see where this is going. If you have people that have the same job category land in different pay tiers, and then they also happen to have a different gender, it will appear to be an objective proof of discrimination. But that's only because these categories are so general. Should a CEO get paid more than a Vice President of a company? Well, probably yes, but they'd both fall in the same category above, so how would we know whether that was the reason or something more nefarious? This and other examples are bound to show up.

At least initially, this reporting proposal is aimed to "facilitate voluntary compliance" and is thankfully limited to federal contractors as mentioned. But we should not be at all surprised if it gets rolled out to the rest of the private employers and/or starts being used to litigate discrimination cases sometime soon. If and when that happens, it will not be helpful. It will create ever more litigation and bureaucracy. And to the extent that it does have an impact on employer pay practices, it will just mean ensuring all people in a given job category fall in the same arbitrary pay band. High-performing employees won't be able to advance as quickly; low-performing / less qualified employees will likely get fired (or never hired) instead of paid lower wages, so the employer could avoid charges of discrimination. It would be a change, but it's not at all clear, it's a change for the better. And that's about the best case scenario.

The essential problem with devising a government solution to this issue is that it assumes the government could acquire all the relevant information needed to determine what an objectively fair wage is for a given person--personality, follow-through, efficiency, experience, education, work ethic--literally everything. This is absurd on its face. It doesn't matter how many lines the government wants to add to the EEO-1 reporting form; they won't and can't capture everything that matters. Therefore, any decisions they attempt to make on what a fair wage is, will necessarily be wrong most of the time.

Ultimately, a fair wage has a much simpler definition that avoids all of these complications. A fair wage is one that an employee voluntarily agrees to work for and an employer voluntarily agrees to pay them. Those are the two parties that know the most about the relevant details of the situation, and they are also the only two parties affected by the decision. Thus, their decision is really the only one that matters.

Summing Up
The gender wage gap is an important issue for many people, and it is not hard to understand why. However, we have explained that the popular narrative about this subject obscures many important facts. The gender wage gap does exist, but it is caused primarily by preferences that are difficult and undesirable to "correct". Discrimination is surely a problem in some cases, but there are natural forces built into a market economy that are likely to eliminate this over time. Even if the government wanted to do something to prevent discrimination directly, they do not have and cannot acquire the knowledge necessarily to objectively determine what is fair, and will fail accordingly.

So it's still true that the gender wage gap does exist, but it doesn't mean what we think it means. And any government solutions designed to address it are all but certain to be more harmful than the problem they seek to solve. 

No comments:

Post a Comment