Math Of Institutional Racism

An anarchist’s derivation and solution for the phenomenon of inherent prejudice in the hands of a powerful monopolistic state.

Say P=100. Group A,B, each 50. X=30, where X is a negative quantity like crime. a , b are the proportions of respective populations to crime such that aA + bB = X = 30.

  1. Suppose a>>b. For instance, a=50%, b=10%. This unfortunately translates into general human psychology, but more damagingly law enforcement, treating each individual of A as having a 50% chance of being X, and every B as having a 10% chance. When instead each individual is either 100% X or 100% not, the way a sock is either green or red, and not 20% green when in a drawer containing 2 green socks and 8 red ones. The mind is not evolved to understand probability, it is evolved to understand frequency, which we’ve manipulated mathematically to signify a notional quantity known as probability, the danger of which Hume so presciently realized. Does the mind know that an individual from A is 50% a criminal, does it think he is a criminal 50% of the time, does it think he is 5 times the criminal that someone from B is? The idiocy of internalizing statistics into psychology that is then responsible for generating predictions is acknowledged, but its true magnitude is left as a hopeful question mark.
  2. Suppose we attach a cost Y to law enforcement equal to the cost suffered by an individual who is 100% not X when treated like X. As a young male, I belong to the cohort most responsible for almost every crime, and therefore singled out. But if I am pulled over, and they find nothing, then I am owed this price Y.
  3. Given a department budget, Price Y makes law enforcement more cautious when pulling me over, in order to reduce the payout Z = Y*(iA + iB), to innocent members of A, B. iA, iB depend on the standards of evidence used to accost somebody. Since a, b (probabilities of X) are inadmissible as a standard of evidence, any lowering of an actual standard of evidence regresses the probability of i to non-objective prejudices, and therefore the ratio iA:iB approaches a:b.
  4. Hanlon’s Razor: In its more benign form, this is a lesser injustice, since it is less a sin of commission (active institutional prejudice) than it is a reflection of a sin of omission (lazy lack of standard of evidence). Suppose the standard of evidence is different from A and B, where A has much lower standard of evidence, then this is the active sin of commission and more egregious institutional prejudice, the failure of Hanlon’s Razor.
  5. If a>>b, it is bad both for A and B, because it provides a lazy alternative to an actual standard of evidence, and if this laziness is applied uniformly, then it results in higher iB as well, not just iA. It is a loss to society, and not just to A, although they lose disproportionately more. Obviously none of this applies in a state that fails in the far more basic task of properly integrating its population such that any one section does not bear more of the cost of crime (and more of the benefit of law enforcement).
  6. Suppose standard of evidence is such that the hit rate of an arrest is HR%. That means for catching X=30, I’ve accosted X/HR people. Of this, total i= (1-HR)*X, how these i’s are distributed across A and B depends on my hit-rate. When HR=X, the floor value, that means I have zero standard of evidence and also zero prejudice, I am simply rounding up every single person in the population, of which whoever is X gets caught along with all the healthy coral and poor dolphins. I arrest 100, I catch 30, there are 70 innocents to be paid. It’s unjust because as an innocent I have a 100% chance of being accosted anyway, and incompetent because I bear the tax-burden of paying out the innocents needlessly accosted, but fair because every single person has a 100% chance of being accosted, and it’s as if the X has funded my compensation payout.
  7. As hit rate increases from X, but still at low levels, I regress to the mean of a, b, so at HR=50, i=30, iA is likely 25 and iB 5, ie i’s are distributed in the same ratio as X. So imposing a price Y to a rational actor seeking to reduce overall payouts Z would appear to force an increase in the standards of evidence.
  8. Unfortunately, Z can equally be reduced by keeping standards of evidence the same, and simply catching fewer X. I reduce i=20, keep the same HR of 50% and only catch 20 X. This is a loss to society also. Where is the equilibrium between Z and dX? If society were asked what it would be willing to pay in order to maximize X caught, it would likely generate a number, that is the cost it suffers when being accosted while innocent, and the humiliation that brings. It will be willing to suffer a certain value of this, given the benefit of X being caught. But what if the cost of this suffering is borne 5x times by A over B? Then society can never agree on the principles of law enforcement that maintain both justice as well as fairness, because while A can agree on the 5x cost as being objectively worth the benefit of catching dX, it is colored by the knowledge that B is only paying a fifth of that cost for the same benefit.
  9. The gap in this cost is what is being covered by the suggestion of a compensation price Y, something that A would agree to much more than B, (funded by tax money by A+B equally, but enjoyed disproportionately by A) arriving at the tail-eating snake. These are co-factors and that makes it an unanswerable solution if seeking a compromise. We end up with statements like this, which would solve the problem but sound inherently unfair or implausible: ‘pay compensation to iA and iB, without reducing X caught’. Unfortunately this is indeed the correct solution, regardless of how have-and-eat-cake it sounds.
  10. Here’s an even more unfair and implausible demand. Pay compensation Y to all iA and iB. Minimize overall payouts Z (this is, after all, taxpayer money) by increasing hit-rate (reducing i) without any drop in X. In a brand new society, this would be the entire point of law enforcement, to catch X, and leave iA, iB alone, ie Xc (caught) = X, and HR=100, ie i=0. Law enforcement isn’t free, it costs LE. At low levels of X, society can simply compensate the victim of all X and still have it be lower than LE. When X rises such that HR*X (Xc) > LE, we accept their monopoly in the use of force, and the cost that comes from HR being less than 100%.
  11. Since HR<1, if iA, iB inadvertently get caught in the crossfire, then they are owed compensation which directly reflects either the incompetence of law enforcement or the irreducible floor of uncertainty in complex human affairs. So the equilibrium is where Xc > LE + Z, that law enforcement needs to measure up to. If LE + Z > Xc, they either need to reduce their own costs (defund the police) or decrease Z (reduce i) or increase Xc (increase HR, which also reduces i)
  12. But this neglects an inconvenient fact. As X increases and the cost of X activities rises, society becomes more and more accepting of the cost suffered by iA and iB that would come from law enforcement overreach. iA, iB then suffer the dual costs of X as well as from law enforcement overreach. We can bear this dual cost in 2 ways, 1) if we are thoughtful creatures, we seek to use the money we’ve just spent so stupidly in order to reduce X, or 2) we accept a lower HR as the necessary evil of fighting increased X, because as a creature that does not understand probability I estimate my cost as a victim of crime to be higher than the cost of being accosted and humiliated.
  13. Both of these groups can indeed converge on the same solution though. For group 1, you take the difference in a, b as the extent to which environment impacts X and therefore the most indicative opportunity to reduce it. The VC’s and financial modelers in us realize that integrating society and eliminating reducible crime (environmentally caused) pays off far more over the long term, and costs far less right now. Instead of using tax money to address iA, iB and LE, you use it to address a, b and X.
  14. For group 2, you make the prospective values clearer by compensating all victims of crime equal to the loss suffered, equating law enforcement to a guarantor or an insurer against crime rather than a protector. That means Xc must be > LE + Z + I (insurance payout). This isn’t an arbitrary hamstringing of law enforcement, but society’s more exact relationship with the costs of crime and the benefits of fighting it. If crime costs less than fighting it, then the fighting is being done wrong, and HR needs to increase. If the argument is that there is no viable solution for Xc such that it is > LE + Z + I, the classical deterrent model, that’s fine too, we just take the first derivative of this entire equation, dXc > dLE + dZ + dI, eliminating the need for a deterrent variable and guiding only change rather than viable solutions. Any solution of this derivative leads back to Point 13.
  15. Irreducible Crime: The anarchist society now faces an uncomfortable floor. Suppose society is perfectly integrated, and A=B, a=b, all reducible crime from environmental factors has been reduced. Xc=X, HR=optimal, i=minimal. How do you deal with these members of X now? They are different from everyone else, on account of them being X and therefore judged unworthy of civil society. The fact that their differences cannot be easily understood using filters of income, race, age, gender and choice of fizzy beverage is irrelevant to the fact that society has judged them different ‘in some way’, and having integrated A+B into a united C, it is now prejudiced against X. The strongest utopian will not conceive of a society with 0 criminals, asserting that crime is 100% environmental. So what of these X? Since civil society has been tacitly agreed by all of us to the fundament of humanity, we’d rather cast X outside of humanity than suffer the effort of examining the rules of a society that cannot fully integrate every single human being. For now, since resources are still our biggest constraint and even the goal of integrating A,B (removing reducible crime) seems impossible, we don’t even bother about thinking about a society’s problem of irreducible crime. It is as if they exist simply to validate the existence of free will and therefore define the respectability of the rest of us actively choosing to be non-criminals. Here lies the unfortunate difference between flaming idiot racists and their more benign cousins the institutional racists. The former think we’ve already reduced irreducible crime, that the environment is just an excuse. The latter accept the dominant environmental factor of crime, but are just unaware of how that same environmental factor has manifested in their opinion on crime.

A novel insightful exercise to determine the pragmatic difference in intellectual payoff between a novel insight and an obvious fact mistaken for novel insight.