[dalke]

How would you measure the effectiveness?

I think they have, based on reading through a dozen or two discrimination court cases, the EEOC guidelines (the EEOC "is a federal agency that administers and enforces civil rights laws against workplace discrimination", quoting Wikipedia), and comparing historical accounts of the workplace environment to the modern one. (For example, the detailed EEOC policy guide on sexual harassment is at https://www.eeoc.gov/policy/docs/currentissues.html . See https://www.eeoc.gov/laws/types/age.cfm for an overview of the age related issues.) Plus many offices have EEOC training so people are aware that these laws and restrictions exists.

Your point about finding a "lawful excuse" is a bit of an oxymoron. If it's lawful then it's not a violation of the relevant civil rights laws, so it's not really an excuse.

People do use an excuse which they think is legal, and would be legal if the person affected weren't part of a protected class. (For example, Disney had a costume policy that employees were not allowed to have a beard. Sikhs wear a beard as a reminder of their commitment to their faith. This gives a protected status that does not apply to most beard-wearing people. See http://money.cnn.com/2015/07/10/news/sikh-turban-beard-disne... .)

However, most employers don't really know the relevant laws, and go more on what they think is justifiable rather than what is actually legal. That's why there are successful lawsuits.

Regarding rationality, people simply are not 100% rational. The last decade or two of behavioral psychology has convinced me of that. If people are rational, then why do you think discrimination exists at all?

Otherwise, if you don't believe that discrimination exists, then I think it's pointless to continue any conversation.

You ask for empirical data. That's of course a hard thing for many laws, with gun control laws, minimum wage laws, and education policies like charter schools being three of the more notoriously debated.

My experience when I see people ask for empirical data is that no matter what evidence I can provide, it's insufficient. That is, they use it as a way to externalize their doubts, without saying what level of data would overcome their objections.

I'm not saying you are like this, I've just been burned too many times by others doing this. So before I look for empirical evidence on this topic, can you point to other social policies you think there is sufficient empirical data to justify them? What would valid empirical data look like to you?

[tominous]

Thanks, I appreciate your points: (a) that in many cases discrimination laws have resulted in a positive outcome for the person discriminated against, (b) that the modern workplace is better than in the past and (c) that people are trained on the relevant laws, which presumably influences behaviour in a positive way.

I can't argue with point (a) because that's obviously true. But as with the Linux/SFC case I mentioned, sometimes a win is not worth the wider cost. The risk of a licensing lawsuit causes companies to shy away from free software. Does the risk of a discrimination lawsuit cause managers to shy away from protected classes?

Point (b) is also true but we don't know if it's due to the laws, or due to the general change in society that prompted the laws in the first place. Yes this is a hard question. I was hoping there might be some natural experiments out there: similar jurisdictions with different laws or different timings.

Point (c) is interesting. You say that many offices are trained on the law, but then shortly afterwards you say that most employers don't know the relevant laws. In addition, we don't know whether the training improves the situation. For example, a recent Australian Defence Force study [1] suggested that "The level of anti-Muslim sentiment among individuals who have received cultural sensitivity training is, if anything, higher than among those who have not" (with caveats about sampling bias, but in the end leaning towards "no relevant differences" in the sample groups).

It would be interesting to see any similar studies regarding attitudes before and after anti-discrimination training in the workplace. Are managers more or less likely to hire from a protected class after such training?

My point about lawful excuses and rationality is that people who are truly discriminatory will find ways to get around the law, so there will be minimal positive impact, while people who are rational will find that the law disincentives hiring from a protected class due to the increased risk.

[1] http://www.army.gov.au/~/media/Army/Our%20future/Publication...

[dalke]

> "Does the risk of a discrimination lawsuit cause managers to shy away from protected classes?"

I do not understand. A company with a pattern of shying away from a protected class is in violation of the law, and risks a lawsuit. How does a manager know which is more likely to incur a lawsuit?

> "or due to the general change in society that prompted the laws in the first place"

The laws themselves could also have changed, or accelerated the change, in society.

> we don't know if it's due to the laws

There can be, because state laws may have additional protected classes, up to and including California's Unruh Civil Rights Act, which is I believe the broadest such law in the US.

However, as the minimum wage law issue points out, even when such studies are done, the results are still contentious. This is why I asked you to describe what would be sufficiently good evidence before I start looking. If it's unrealistically high, then there's no point in me wasting my time.

This is also why I asked you to point to laws regarding social change where you think the laws actually made a difference and have a empirical support. That lets me know where you are coming from. If you can't point to any laws with sufficient empirical evidence, then I'm not likely to find something which satisfies you.

> "many offices are trained on the law, but then shortly afterwards you say that most employers don't know the relevant laws"

Yes, I can see the apparent contradiction. It's a matter of the depth of required knowledge and numbers. For an employer to work around the legal prohibitions requires a high level of understanding of the laws. Most supervisors do not have this knowledge. While for an employee to protest an illegal employer action requires a much lower understanding.

When I say "many offices", I mean numbers like 10% have good training (eg, a friend working for the state government had really good EEOC training), more (say, 40%) have mediocre training of perhaps a couple of hours by HR when starting with the job, and the rest have none other than a poster in the breakroom describing their rights. These numbers are pulled out of my ass, but roughly equal to my understanding.

This means that most supervisors don't have real knowledge of EEOC laws, other than the few hours they might have gotten years previous from HR when they started as an employee.

Suppose supervisor S fires employee E due to an illegally discriminatory reason. S might talk to a couple of people before making the decision. E, on the other hand, likely complains to friends and family about the decision. Suppose 10% of the people know the action might be illegal. If 3 people were involved in the decision, then there's a 75% chance they don't know it might be illegal. If S talks to 9 friends, then there's only a 33% chance that none of them realize it's likely illegal.

> "[ADF] cultural sensitivity training"

How is this relevant? Cultural sensitivity training is not the same as anti-discrimination laws. Australia is not the US. People can have anti-X sentiment, and even increased anti-X sentiment, and still employ people who are X in a non-discriminatory fashion.

Are you only pointing out that it's possible to carry out studies on people regarding questions of social policy?

Also, the introduction to that non-peer-reviewed paper says "Both the Evaluation Board of the Australian Army Journal, which reviews these articles, and my staff, have a number of opposing views on this article’s content and its reflection on the lived experience of Army values. ... This article is one view, of one cross section of our people, undertaking one component of our preparation for operations". Angus J Campbell and others don't seem to think the empirical evidence given is persuasive.

> "people who are truly discriminatory"

I do not understand what "truly discriminatory" means. We have plenty of lawsuits where companies have been found guilt of being discriminatory, so that's about as true as it gets.

It feels like you have another definition which is far higher than, say, an Archie Bunker level of general bigotry, and reaching more the levels of a villain in a melodrama. Are you really saying that only people who have successfully worked around the law are truly discriminatory? Because that sounds like a rather meaningless label for policy decisions.

In practice I imagine there are people who want to work around the law, but the cost (mental and financial) of figuring out how to do so is too high. That makes them practically non-discriminatory even if their true self would like to be discriminatory.

You again talk about "people who are rational". Do you have empirical evidence that people are rational? The empirical evidence from behavior psychology, which I alluded to earlier, shows that people are not rational. Do you not believe in those results?

If people are not rational, then your argument goes out the window.

[tominous]

A manager wouldn't care if their company were sued for a pattern of discrimination based solely on the numbers, because it is extremely hard to prove a pattern, even harder to prove intent, and the burden and blame would fall on the company not the individual.

By contrast a specific anti-discrimination claim would be very personal, very damaging, and (according to your scenario) completely unexpected and uncontrollable.

It's well known that (rational or not) uncertainties of this kind are weighted very highly and such situations trigger avoidance.

As to evidence, I offered an example of the type of evidence that would persuade me and you were the one that claimed it was contentious. So please don't put that label on me.

Regarding social change through laws, I can offer cases where social change came before changes in laws (e.g. blasphemy, slavery, child labour, apartheid) and cases where laws have not produced the desired social change (e.g. prohibition, drugs). I think the burden should be on those who promote certain laws to show that they actually help overall. Every law has a cost, especially one that produces so much uncertainty.

As to being rational, my other original point was that in a market those who act more rationally succeed, grow, gain influence, and crowd out their competitors. This is the kind of bottom-up change that I believe is actually responsible for social change.

[dalke]

I think you are right, that it's hard to show a pattern of discrimination. It's still a risk, and lawsuits like the age discrimination case at Google keep it in general awareness that it's something to be concerned about.

When I talk about EEOC laws with people, there is often a knee-jerk reaction toward them being "completely unexpected and uncontrollable." I can understand this, because labor rights and civil rights are so undervalued in US discourse, where the interest is more on profit and job production. EEOC training in many companies is woeful, so people end up being more scared of it than they should.

I don't understand why managers don't get real EEOC training. I've read through a dozen or so court cases, and they are all consistent and understandable. If manager are really worried, and acted rationally, then why don't they seek out the education which would greatly limit that risk?

(The answer is obvious. I don't think people are really rational in that way. But you think they do.)

In practice, what happens is the 'rational human' analyzes use what appear to be a post hoc argument that a variance from the expect optimal result is due to a personal preference for a different weighting scheme. People like religion because they place a high value on the sense of community or spirituality the practice gives. People like to discriminate because the social standing is worth the effect of a suboptimal market, etc. There's nothing which cannot be justified with "weighted very highly", making it a weak concept.

The evidence you gave looks like the sort of crap one finds with a Google search, which is easy to match. A Google Scholar search for 'anti discrimination laws effectiveness' found:

http://ilr.sagepub.com/content/56/2/244.short - Using census data from the Integrated Public Use Microdata Series, the author assesses the impact of fair employment legislation on black workers' relative income, unemployment, labor force participation, migration, and occupational and industrial distributions. In general, the fair employment laws adopted in the 1940s appear to have had larger effects than those adopted in the 1950s, and the laws had considerably smaller effects on the labor market outcomes of black men than on those of black women.

http://www-siepr.stanford.edu/conferences/gender05/racesex_s... - we find robust evidence that state equal pay laws for women–which predated federal legislation establishing both equal pay and equal employment for women–reduced relative employment of both black women and white women. With respect to race, we find some evidence of positive effects of race discrimination laws on earnings of blacks relative to whites, although no evidence of employment effects.

That second one starts "The question of the effects of sex and race discrimination laws on relative economic outcomes for women and blacks has been of great interest since the Civil Rights and Equal Pay Acts passed in the 1960s." so I presume there is a lot of literature on the topic.

If discrimination is an irrational act, then how come there has been so much of it through the centuries?

Take a look at colonist power structures. The foreign power (Britain, France, Germany, etc.) has a very small population in the country. They are outnumbered. What they do is put a minority in charge of the next rung of the power hierarchy. These people gain power and prestige by maintaining the colonial system.

At the bottom are the majority in population, even though they are the minority in power.

For a real-world example, https://en.wikipedia.org/wiki/Rwandan_genocide#Colonial_era :

> The Berlin Conference of 1884 assigned the territory to Germany[22] and began a policy of ruling through the Rwandan monarchy; this system had the added benefit of enabling colonization with small European troop numbers.[23] European colonists, convinced the Tutsi had migrated to Rwanda from Ethiopia, believed the Tutsi were more Caucasian than the Hutu and were therefore racially superior and better suited to carry out colonial administrative tasks.

This is a relatively stable system where the rational market response - the market favors those with the most market power - causes the discrimination to persist for decades. One result for Rwanda was the Rwandan genocide, which presumably was also the result of rational actions.

[tominous]

Thanks for confirming what I suspected in the first place: that anti-discrimination laws actually reduce employment in the protected classes. I.e. if it's harder to fire someone you'll be less likely to hire them.

You seem to have read an awful lot into what I said! We could have had an interesting exchange about whether the drop in employment and the added compliance burden are worth the special protections, higher average pay, etc of those who still manage to get a job. Anyway, maybe some other time!