Harsanyi on Rawls

Harsanyi critiqued Rawls in an article published in 1975

The maximin (maximise minimum benefit) principle is silly. As Harsanyi points out you'd never cross a road because at least you can be sure you'll continue to live rather than go under a bus.

Actually both maximin and Bayesian expected utility maximisation are silly. Regret Minimisation is the way to go. Unlike Utilitarianism or Contractarianism, which focus on the rational self interest of the individual, Regret, by its nature, extends to people currently strangers whom you may come to care about. It includes future generations as well as the debt one may come to feel one owes ones ancestors or preceptors. It excludes devoting costly cognitive resources to 'maximising' anything. It provides a decision rule which militates against worrying too much about decision rules but just decides things already.l

Both Rawls and Harsanyi were wrong about what sort of Society impartial people (in a closed society) would choose. It is one which would feature an incentive compatible Social Insurance scheme. This scheme might have to be changed from time to prevent 'adverse selection' or some other sort of harmful 'gaming of the system'.
Nobody would want to go any further than this. It is silly to say every policy prescription should be evaluated on how it affects the weakest. This just opens the door to Gandhian or Eco-Feminist or other virtue signalling nutjobs and blathershites.

Still, Rawls and Harsanyi did well reasoned 'first order' work. This gives their oeuvre an imperative force lacking in 'second order discourse' critiquing first order research for a careerist purpose.

Harsyani gives the following example of how the difference

Both these views are fucked in the head. Common sense tells us that if the Doctor is in private practice, he will give the drug to the guy who will pay more, or for whom he feels more affection, or the one who can confer a greater reputational advantage. In consequence, the Doctor will enjoy more prestige, power or pelf and thus more people will be attracted into the profession. Letting some stupid Philosopher impose a decision rule is silly because then more people will study Philosophy rather than Medicine.

Suppose the two patients are identical in every respect but come from rural Bihar. Then you should toss a coin, hand over the medicine to the winner but assure both that they have received the drug and that it will take a couple of days to take effect. You immediately run away as fast as you can because the relatives of the guy who didn't get the drug will be gunning for you.  The other patient's relatives will also be chasing you- but only for the gentle purpose of holding you to ransom because, obviously, you are a good Doctor and can make a lot of money.

My point is that Human Societies have evolved so that decision makers are held responsible for their actions. The fact that you acted according to a decision rule is irrelevant. A machine that follows a decision rule may still be smashed to pieces by someone adversely affected. A Doctor, who by an act of omission or commission, causes harm may be held to account. The proper course is to assemble all concerned parties and explain the situation. If no 'decision rule' is accepted then perhaps a coin may be tossed or physical force will resolve the issue one way or the other.

Deciding not to act is still a decision. Those who decide not to denounce Ethical philosophers as worthless shitheads may have to bear severe consequences. If we don't chase 'virtue signallers' away, public policy will degenerate. It is not enough to vote for Trump, micturating mightily on Moral Philosophy too is required.

Economics can be a first order idiographic discipline. It can use maths and stats to lower costs and boost useful output. However the moment it tries to be 'moral' it fucks up.

Harsanyi's 'equiprobability' approach is as foolish as Rawls- indeed it cashes out as it in a steady state economy with no convexities, no uncertainty and diminishing marginal utility.
Harsanyi says that a wealthy capitalist who prefers capitalism isn't making a 'moral' choice.

This is silly. The capitalist will still choose Capitalism with compulsory Social Insurance because he knows this option is feasible. The unskilled worker may choose otherwise because he believes Capitalism is based on secretly sodomising workers as part of a Satanic ritual. There is nothing moral about such judgments. The reflect differences in knowledge,
Expected utility does not exist. The future is radically Uncertain. If expected utilities could really be accurately computed then so could future asset prices. Stock Market crashes would not be possible once Rocket Scientists ran Hedge Funds.

Laplace's indifference principle- assigning equal probabilites to outcomes we have no further knowledge about- doesn't really tame Uncertainty, as Harsanyi assumes because we don't know most outcomes and can't even imagine them. That's okay. We only feel regret for bad decisions- not for things that were always impossible for us. Using either maximin or expected utility implies that we can have a type of knowledge we can't possibly have if we evolved by natural selection. It's morality is the morality of the angels and therefore heteronomous and fundamentally superstitious. Human morality can't hinge on the assumption that we have Supernatural powers. Why not simply attribute sickness to the 'evil eye' and murrain to witchcraft?

Rawls thought the virtue of his approach was that it avoided mention of probability. But, human beings implicitly act on the basis of nothing else. We do not possess the agential inerrancy of the God of the Occassionalist. As Harsanyi writes-

We can't talk to our ancestors or those of our preceptors who have passed away. Nor can we talk to our distant descendants. Nevertheless, we can imagine ourselves feeling great regret for some action we take today which we could not justify to them in Heaven. Regret minimization captures this aspect of our actual, lived, human morality.

Rawlsian maximin and Harsanyi's equiprobability can't do so. Our interlocutors will be incredulous that we were ever so devoid of commonsense as to choose so obviously flawed a decision rule.

There can't be any justification for urging Society to adopt a decision rule based wholly on maximising something fundamentally uncertain. Indeed, the safer course is 'to first do no harm'- i.e. don't urge Society to do anything on the basis of some silly academic availability cascade whose sole purpose is to ration clerical employment or create a separating equilibrium on the basis of a costly and socially inutile signal.

Piecemeal, idiographic, first order research, on the other hand, can yield local improvements even if this involves the use nomothetically unsafe assumptions or procedures- e.g. that interpersonal comparisons of utility. We expect Judges and Juries and other officials to routinely makes such comparisons rather than take refuge in the Rule Book.  The market itself is based on 'transferable utility' in the shape of money. Judgments in Tort cases depend upon putting a cash value, and thus setting a cardinal scale, for disutility suffered or utility foregone. First order 'expert knowledge' can help guide such judgements. Nomothetic decision rules, by contrast, could vitiate the judicial function. It would create 'loopholes' and make the system more gameable. It would change the ethos of stare decisis defeasible jurisprudence. It would also negatively impact the Insurance industry and large portions of the Financial Sector. Capitalism is as much about this type of institutional culture as it is about Markets as an allocative device. A tradition of trust, public service and uberimae fidei professional integrity in Judicial and Financial Services would be placed at risk. Why? The Rules based approach is a type of akrebia which, in the final analysis, blames the victim for being at the wrong end of an information asymmetry.
A 'comparative' approach doesn't improve matters. All that happens is that Akrebia can choose a metric which shows it performs better than Oikonomia. My own claim that I am a more beautiful and talented woman than Beyonce- which I admit is not blindingly obvious to everyone- arises out of my use of metric which gives higher weighting to scrotum size. It is very sexist and racist and orientalist and hegemonic of you not to accept my metric.

Any nomothetic theory concerned with the real world faces ambiguities because there is a trade-off between precision and significance. A precise theory has no significance. A significant theory is imprecise and more or less 'anything goes'. Utilitarianism is no better or worse than any other decision rule or 'comparative' approach.

As a matter of fact, ordinal rankings are just as problematic as cardinal ones. Rawls thinks it is easy to identify the least well off. It isn't at all. Consider India's affirmative action program. Almost every dominant caste is clamouring to be classed as an oppressed tribe, or religious minority, or as 'educationally backward'. Farmers pay no Income tax in India and get free electricity in some States because they are all supposed to be very poor and oppressed. Very wealthy people have been discovered to be 'agriculturists' for tax or other administrative purposes.

Moreover, once policy makers or Judges use a particular instrument to identify a specific social or economic category, that instrument will lose its predictive power because the system will be gamed. Indeed, the policy measure will impose an increasing deadweight loss on Society. Consider rent control legislation. Notionally, it helps poor tenants and hurts wealthy landlords. In practice, the opposite may happen. Additionally, the housing stock degrades and a 'doughnut' effect arises whereby inner city areas become uninhabitable because of crime and crack-houses and a growing concentration of indigents who can't pay local taxes.

If 'rules based' akrebia is a bad idea, then 'second order rules' are even worse because they can be used to stalemate even the most needful and widely accepted of initiatives. There will always be antagonomic ideologues or moral entrepreneurs eager to push themselves forward by claiming that some salutary action offends against some sacred cow- like Liberty, or Fraternity, or Equality or some other such bullshit. No doubt, the actual blocking will be done by some political coalition or vested interest group but that changes nothing. Rules based Akrebia kills. Idiographic Oiknonomia, if Regret minimizing, gives Life.

Harsanyi's criticism of Rawls was published almost 50 years ago. He concludes thus-

One thing not apparent to Harsanyi was that there are important Evolutionary reasons why anything which is co-evolved will be so quick to change as to be 'unhackable' at the macro level so as to baffle a parasite or predator. The informational requirement for the application of 'moral principles' or 'Political ideologies' are the same as those for highly immoral principles and Mafias. Our true preferences are secret even from ourselves. Our fundamental values and allegiances can change in a heartbeat once a social tipping point is reached.

One can make a living pretending to be some sort of moral guide or political philosopher and, because of widespread Preference Falsification, some such pretence may appear to be widely accepted. But the thing is a sham. We consider only those theories as impinging on Morality and Ethics which counsel the most mischievous course of action under any given circumstances. This is helpful to us because when facing judgement we can appeal to these vaunted Moral philosophers to raise 'honest doubts' which are actually wholly dishonest and thus hope to stalemate the proceedings.