The Probability Of Being Accused Of The Hate Crime Of Disparate Impact Approaches 1
You’re an employer who has to hire 100 employees to get your Meme Factory running. You’ll hire the first 100 who come in the door.
Given this information (and what you learned in Class), what is the probability the government will roast you for the hate crime of “disparate impact”?
“Briggs, all hate crimes are political crimes.”
Yes, I know. But what about the chance of you been pilloried for this political crime?
“You haven’t given us enough information.”
Ah, you’re right. In order to fulfill the conditions of the hate crime of “disparate impact” you also need an Oppressor and a Victim.
“Disparate impacts” occur when, and only when, an Oppressor does not hire at least as many Victims as exist proportionally in the population.
Suppose your Victims are 13% of the population. Then, if you are an Oppressor, you have committed a hate crime if you do not hire at least 13 Victims.
“What if I hire more?”
If you achieve Diversity, that noble state where all 100 of your employees are Victims, this is not a crime. Indeed, this will likely win you an award.
“But if there are 13% Victims, that means there are 87% Oppressors. Don’t I have to hire 87 Oppressors to maintain balance?”
No. You do not need to hire any Oppressors. The hate crime only occurs if you do not hire sufficient Victims. And only if you are an Oppressor owner.
“You mean if I’m a Victim myself, I can hire as many Oppressors as I want?”
You can, as long as you are voluble in your declaration of your own Victim status. If you’re quiet about it, you run the risk of being treated as an honorary Oppressor.
Anyway, enough stalling. I’ve given you all the information you need to answer the question. What are the chances you’ll be nabbed for a hate crime if you hire the first 100 who walk in the door?
“I’m not sure. Thirteen percent?”
Not even close. Here’s how to get the right answer. Given our information, there’s a 13% chance any person walking into the door is a Victim. So you want the chance that you hire no Victims, or just 1 Victim, or just 2, or and so on, up to just 12 Victims. Hiring any number less than 13 can land you into hate crime territory. Hire 13 or more and you are safe.
“Oh, I get it. That’s easy. That’s a binomial. That gives about a 46% chance of hiring 12 or fewer out of 100.”
That’s right. There is a 46% chance you will hire insufficient Victims and run the possibility of being charged with the hate crime of “disparate impact”, even when, as here, you are wholly innocent of any crime.
About half of companies which are not discriminating, at all, will be guilty of hate crimes.
It also turns out, for boring technical reasons (look up the expected value of a binomial) the answer is about a 50% chance of a hate crime no matter how many employees you must hire.
I don’t know about you, but 50% is an awfully large chance to be charged with the hate crime of “disparate impact.” How would you solve that problem?
“Wait. In our culture there is more than one Victim group, though, isn’t there? Don’t they all have to be hired for proportional representation? Doesn’t that make it harder to escape being charged with a hate crime?”
Yes.
It turns out about half the Oppressors are actually Victims themselves. And that the original Victims themselves can be split into two groups, each about equally sized. That can be confusing, so here’s the breakdown:
Oppressors: 43.5%; Victim group 1: 43.5%, Victim Group 2: 6.5%; Victim Group 3: 6.5%.
We now want the chance that we hire 43 or fewer Victim Group 1s or 6 or fewer Victim Group 2s or 6 or fewer Victim Groups 3s. The number of Oppressors don’t matter, as before.
I know you don’t know the answer to this, so I’ll just tell you. It involves the multinomial distribution. We want 1 – Pr(Oppressors >= 0 & VG1 > 43 & VG2 > 6 & VG3 > 6 | all that information). This turns out to be about 94%. That is, with no discrimination at all, you stand a 94% chance of being accused of a hate crime.
We also have to be more careful about the number of employees. The fewer you hire, the more likely you can be charged with a hate crime. But once you hire about 1,000 or more, the chance settles down to about 93%.
“Dude, that means—”
Wait, I’m not done yet. For we have more than just these three Victim groups. There’s another, now exceedingly famous Victim group, that is about 3% of the population. We have to add that, too (and adjust the others lower). That gives about a 96% chance, or higher, that you can be charged with the hate crime of “disparate impact” even when you are not discriminating.
“And if we add more Victims the chance only grows?”
You got it.
“But it’s already almost certain you’ll be guilty even when you’re innocent!”
That’s right. So again I’ll ask: how do you handle this?
“I guess I’d have to have HR institute hard quotas, to make sure I hire at least as many Victims as I need, to ensure I don’t run afoul of the law.”
But that’s discriminating by Victim status, isn’t it? And isn’t discriminating by Victim status against the law? Not the hate crime law, but that Civil Rights law.
“You’re right. It is.”
So no matter what, you will almost certainly be guilty of crime. Even when you are entirely innocent.
Subscribe or donate to support this site and its wholly independent host using credit card click here. Or use the paid subscription at Substack. Cash App: $WilliamMBriggs. For Zelle, use my email: matt@wmbriggs.com, and please include yours so I know whom to thank.
“So no matter what, you will almost certainly be guilty of crime. Even when you are entirely innocent.”
Don’t think that isn’t by design. Were it possible to follow all regulations the same time, there would be no pretext for a lawfare-style shakedown, which most “prosecutions” of businesses ultimately are.
One glaring example is BP and the oil spill in the Gulf of Mexico. They faced dire legal consequences if they didn’t pay some $10billion to an extra-legal entity comprising FoB’s (Friends of Barry). No matter what you think of BP, that was a shakedown, pure and simple.
Or the “auto industry bailout”, which was clearly (to anyone paying attention) an irregular, illegal, coerced bankruptcy proceeding.
Or the bank “bailouts” where solvent banks that were in no distress were, in 2008 or 2009, forced to take government money, and bow to the onerous strings attached.
Or the shakedown of a famous manufacturer of hot sauce, because the smell coming from the factory was “harming residents” even though there were no complaints.
No doubt there are many more such examples perpetrated at state and local levels as well. Many, possibly a majority, of governments at various levels seem to be rackets — organized crime.
A PDF solver for race-huckstering, I love it. People might get caught up in the algebra, but what I love is how you explain the calculations, showing the absolute absurdity of using this kind of analysis to determine fairness. One additional element that I think might be fun to explain how variance and 'randomness' actually show up in datasets in the real world.
I'm a poker player (for fun, not a degenerate gambler) and the bane of Texas Hold'em poker is that there is no winning starting hand. Going deeper into the stats what you'll find is that profitable tournament poker players 'lose' more tournaments than they win, kind of like a hitter in baseball where a .335 average is considered 'good' but that means you are making an out 2/3s of the time.
Huh, so being a good poker player means you lose more than you win. One ends up 'living variance' if you will. And its clear that things are not evenly distributed. You'll lose 4 tourneys in a row to 'suckouts' where your opponent hits the 4% hand to knock you out of the tournament, and find so many ways to 'lose' when you are statistically ahead that you feel the randomness. Fyi, this drives some poker players to the brink of insanity. I had a 3 month run once where I lost almost every big tourney I played in, costing me thousands, it's called 'running bad' in the poker world. And then won it all back and them some in 2 weeks. That's variance for real and in living color. Fyi, the way I manage all that is bankroll management, I'm never playing with the 'rent money' so I have a detachment that keeps me rational no matter the results of any given tourney.
The chance that out of 100 people you will have anything like a distribution representative of the national average is very low. Given the confounding factors of geography and time already pushing 'bias' into the population one is sampling for a job opening (time in the sense that one has to be looking for a job at that time), there is no way one could ever expect to get a distribution that reflects some national population's racial proportions. As for geography, people live in clusters of 'ingroups' for the most part, self-segregating. Drawing from any local population for a job someone needs to physically show up for has little chance of mapping to the percentages of the national population.
I've recently semi-retired from corporate America and I can tell you how they do this diversity game for the most part. The idea is to stuff as many 'diverse' candidates into the top of the hiring process 'funnel'. This is seen as a way that isn't biased, lol. These HR idiots literally think that if you pre-select for race early in the process by 'encouraging' diverse people (whatever they hell that even means) to apply, that this isn't discrimination. Wrong, it's just discrimination that's harder to detect. Determining the race/sex of who gets to apply is no different than basing the final hiring decision on those aspects. They seem to have utterly forgotten that this is illegal.
It's hard to be in that hiring seat though. Without going into specifics (cuz I don't want to get sued), let's just say nobody is immune to all these word games and the underlying fear/guilt manipulations it relies upon. I remember finding myself wanting to give a black candidate more of a chance for a particular job (professional role, he had an MBA), it was like gravity. I also found I was not comfortable sharing my critical feedback and overall uneasiness with the candidate.
On paper he was good (not the best) and was decent in the interviews but something felt off with the guy. Fyi, I've bounced many white people out of interview processes based on this kind of intuition. It's often a feeling that the candidate isn't really being honest and rather is trying to get over on me somehow but I don't have 'evidence'. I've found throughout my life that this instinct/intuition is quite reliable. But now? I'm not allowed to operate on intuition. I found myself doubting my judgment, trying to analyze my own reasoning and seeing if I was biased. I felt like cuz he was black, I had to 'give him the tie' or a chance, ya know?
We ended up not hiring anyone for the role, but the process really brought me up short. I saw how this corrosive ideology just seeps into you, and how a subtle culture of fear becomes ambient when dealing with 'diverse' candidates. I'm sure other white folks who have done hiring can relate. Fyi, I felt all this while being a very active anti-woke guy politically and conservative. I am as unbigoted as I can be, and not woke, so why was this working on my brain so hard?
And then I got it. None of us can escape the totalizing institutional brainwashing that is ambient in society and has been for my entire life. When something gets as much of a push as 'diversity' and 'racism' etc, it will effect you. The social conditioning gets to us all eventually to some degree or the other.