Sunday 21 May 2023

Howard Chang's Illiberal Liberalism

Utility, or ophelimity, is just a word economists use to denote the reason why people buy and sell certain things. It is simply a blanket term for our economic motivation. Economists didn't want to use the word 'money' because that would have sounded vulgar. Some objected that 'utility' was 'metaphysical' (it isn't. It is 'intensional') and so Samuelson introduced the notion of 'revealed preference'- i.e. what is revealed about our utility when we buy and sell. But 'preference' is unrelated to 'usefulness' or 'profitability'. Things we do so as to survive aren't things we'd do if our lives were as we would wish them. For most of us, it is avoiding pain- not to mention extinction- which is the big motivator. We seek to minimize 'disutility' rather than maximize our pleasure or profit. 

One big problem with 'disutility' is that it depends on opportunity cost and can change discontinuously and disconcertingly even though nobody's tastes have changed. Moreover, the 'opportunity foregone' is not visible. It may not even be 'compossible' with out universe. It is something we imagine and fear. Consider the following. I would prefer to go to the kitchen and make myself a sandwich. I think I hear a noise downstairs. Has a thief broken in to the house? Suddenly, I don't want to go the kitchen. I want to hide under the bed. Have my preferences really changed? 

 In cases like the above, analysis must proceed by way of counterfactuals- i.e. far from escaping Metaphysics we are now committed to some sort of Stalnaker-Lewis 'closest possible world' ontology. I suppose we might say that we act as though we exist not in this world but in a place whose fabric is woven out of our hopes and fears. This gives rise to the dichotomy between  the 'ex ante' world of our expectations and the 'ex poste' world of actual outcomes. Is it desirable that they coincide such that we are in 'equilibrium'? The answer is that equilibrium is fine for certain things or in certain respects but disequilibrium- traveling in appetency- is better for others. There is no a priori way of deciding this matter. It is the fitness landscape which does the pruning. 

Economics had no business trying to turn itself into Philosophy for three reasons

1) Philosophy as Socrates said and Collingwood underlined, only has salience where there are 'open problems' in STEM subjects. But advances in mathematics and logic had closed, in an adverse manner, all the problems which kept Economists and Moral Philosophers in business save, as Godel pointed out, 'intensional paradoxes'. These aren't 'semantic' and are independent of language. But they depend on whether another problem is open or meaningless. That other problem is whether 'Absolute Proof' or 'concept of concept' is coherent and can be clarified. Godel thought that Intuitionism could be shown to be consistent and, more importantly, minds could be shown to be wholly unlike machines, if this were true. But, thanks to the work of Eduordo Sontag & Hava Seigelman, we know that our notion of what constitutes a machine can be expanded to 'Super-Turing' or things yet more rich and strange. It is now clear that there are huge economic benefits from this line of research. But, this means that the subject is 'applied' not 'pure'. Philosophical approaches are already way 'behind the curve' and thus irrelevant. 

What is strange is that post-War philosophy persisted for so long in a trough of semantic paradoxes of a type which, as Godel had shown, were trivial. 'Naive set theory' was known to have problems of impredicativity by the time of Russell. Thus, it was essential to only use sets or relations which were 'well defined'- i.e. the 'extension' was known or knowable. This was not the case for 'preferences'. The 'intensional fallacy' arose.  Moreover, problems of concurrency, complexity and computability meant there could be no partial orders over anything. Nobody could have the sort of preferences economists imputed to them.

2) Economists used models which neglected 'Knightian Uncertainty'. But evolution is a true theory only because some states of the world are unknowable. It is foolish to have a 'science of choice' whose first assumption is that we are omniscient- like God.

3) Economic laws are statistical regularities. The are neither 'ante res' or 'in rebus'- i.e. they are neither prior to things nor in them. What is observable is 'Granger causality' and that too only for a limited and pragmatic purpose. 


The economic analysis of law

treats it as a service industry like any other. The question is, does it add value? Is it, in some respects, a nuisance? Most importantly, is there entry or exit from the jurisdiction? If the justice system is shit, it will be disintermediated. Pass too many laws and none will be obeyed. 

Let us look at what happens when a Law Professor, Howard F Chang, thinks fools like Arrow & Sen represent actual economic analysis rather than systematic misology or a refusal to play by the rules of mathematics or logic.

In a paper titled 'A Liberal Theory of Social Welfare', Chang says the economic analysis of law

evaluates legal regimes by the criterion of social welfare, which economists usually take to be a function of the utility that individuals enjoy under the laws in question.

This is not the case. Policy is evaluated in that way. But policy is discretionary. There is a 'doctrine of political question'.  Only in rare instances would a matter become justiciable.  Even then, the law matters less than how it is interpreted and what polices are in place for its enforcement. But, Policy itself has to look at entry and exit from the jurisdiction or else risk becoming a dead letter. 

Economists generally define the utility enjoyed by each individual as the satisfaction of that individual's preferences over various states of the world.

Utility is an undefined 'Tarskian primitive' as is 'Regret'. Under Knightian Uncertainty, the latter is assumed to be minimized- at least at the macro level. If no Uncertainty arises, Expected Utility is maximized- i.e. aggregate behavior is assumed to have this quality. This is merely a hypothesis which nobody greatly cares about. All that matters is that quantitative work produces useful predictions or helps improve our Structural Causal Model. In other words, Econ is only good if it more than pays for itself. But the same is true of the Law. 

Economists normally assume that any reasonable notion of social welfare would conform to the Pareto principle, which holds that if each individual prefers one state of affairs over another, then social welfare must be higher in the first state than in the other state.

Sadly, because of Knightian Uncertainty, nobody knows if a voluntary trade which appears Pareto efficient won't turn out to be a fucking catastrophe for all parties concerned. 

It is not reasonable, it is stupid, to think that everybody preferring one state of affairs over every other necessarily means that society will be better off if that state of affairs comes about. The plain fact is that our preferences seldom coincide with what is best for us. We may, as an ad hoc assumption, pretend otherwise in specific contexts but they would be situations where there are no externalities, no income effects (i.e. real income is not materially  affected), no hedging (which also means there is no discoordination game associated with the transaction), no non-convexities, no 'topological holes' or discontinuities in the underlying configuration space, and 'Muth' Rational Expectations or stochastic equality between ex ante and ex poste. 

A Social Welfare function should look at actual outcomes and see how Society's welfare has been affected. Preferences don't matter because they are always over a 'characteristics space'- which is uncertain- not the available commodity space. 

For 'comparative statics' of a limited sort, we can pretend otherwise, but the moment we speak of the Law, we need a global approach. But 'general equilibriums' are 'anything goes' while a particular point in the configuration space is likely to be multiply realizable. In other words, though both econ and the law, or 'law & econ', may be useful enough in narrow, limited contexts, they are totally useless and misleading when treating of the totality of society. What you get is not reasoned analysis, but cascading 'fallacies of composition'.

There may be a relation between some of our preferences and some of our choices and some of the outcomes of those choices. But, this is only linear, or deterministic, if those preferences and those choices don't matter greatly. Only then do you get non-dissipative systems with 'Noetherians'. Otherwise, Life is 'far from equilibrium'. Campuses may be safe spaces to pretend otherwise. No doubt, Ukraine had such campuses. Then Putin sent in his goons.

We would all prefer to live in a peaceful world. But we understand that Military experts need to spend money on our behalf so as to frontload pain on any aggressor. Choices do matter which is why we ensure that convicts don't get to make a whole lot of them. 

Amartya Sen, in his influential article entitled The Impossibility of a Paretian Liberal, shows how liberal rights, such as the right to read a book of which others disapprove, can produce outcomes that everyone would prefer to avoid, thereby violating the Pareto principle.'

This is nonsense. Nobody knows what does or doesn't violate Pareto. Moreover, with 'nosy preferences' we don't know whether guys who say they don't want anybody to read x actually mean what they say. Worse still, they don't themselves. Everything is strategic- not just voting, but preferences themselves.  

Sen infers that one cannot uphold both liberal values and the Pareto principle.

He was just being silly. We don't know whether any given change might not be a Pareto improvement. We don't even know whether the outcome will change if nothing we do changes. 

Liberal values have traditionally been associated with the Rule of Law. This means there are Hohfeldian immunities such that 'nosy preferences' have no purchase. I can't sue you because I'm hurt that you won't buy my book. You have a Hohfeldian immunity with respect to what you want to read.  

Disturbed by the implication that "individual liberty may have to be revoked" under the Pareto principle, Sen concludes that the conflict that he exposes indicates "the unacceptability of the Pareto principle as a universal rule." 

This is foolish. The principle is 'intensional'. We don't know what it would entail. It can't conflict with anything anymore than the assertion that unicorns are nice conflicts with faith in Democracy. This is because nobody has seen a unicorn.  

One might say 'Pareto suggests that all uncoerced transactions absent market failure should be permitted'. This may conflict with a particular theology or chauvinistic ideology or a dark and pessimistic view of the Human Race. But it doesn't conflict with Liberalism. 

In a similar vein, Louis Kaplow and Steven Shavell identify potential conflicts between the Pareto principle and notions of "fairness," which give weight to considerations other than the overall utility level of each individual.'

Notions of 'fairness' can be very fucking unfair. There was a time when it was considered unfair for women doing the same job as men to be paid the same wage. After all, men have wives and children to support. Working women can't possibly have husbands because of biology. The fact is, vaginas seal up unless a woman is barefoot and pregnant. Moreover, if women have jobs which require them to use their brains, then they develop Adam's apples and grow penises. On the other hand, if a woman is given a little pin money and shows up at the factory or the office, there is no harm in extracting some labor from her. Her vagina will not seal up. Some kindly bunch of sperm donors may take turns putting a baby in her as she screams for help.  

Indeed, Kaplow and Shavell claim that "any conceivable notion of social welfare that does not depend solely on individuals' utilities" implies a "social welfare function" that violates the Pareto principle.

Sadly, as far as the mathematics goes, individual utilities may be said to depend solely on the social welfare function but not the other way around. Why? Individuals either gain utility by belonging to Society or else could be a Society unto themselves. Society is always already a collection of individuals. Its utility does not change when tasks are randomly reassigned between sufficiently similar individuals. Mr. Smith and Mr. Jones both get disutility if they have to do Jury duty. If Smith and Jones are similar enough, Society is indifferent between which of them has to perform that duty. 

There is an ancient notion of 'antidosis' or exchange of positions. It may be that Smith is chosen for Jury duty but can pay Jones to do it for him. Society may permit such exchanges. This is all that Pareto's version of what we would call the second fundamental theorem of Welfare Economics amounts to.

He wrote

Consider a collectivist society which seeks to maximise the ophelimity of its members. The problem divides into two parts. Firstly we have a problem of distribution: how should the goods within a society be shared between its members? And secondly, how should production be organised so that, when goods are so distributed, the members of society obtain the maximum ophelimity?

His answer was

Having distributed goods according to the answer to the first problem, the state should allow the members of the collectivity to operate a second distribution, or operate it itself, in either case making sure that it is performed in conformity with the workings of free competition.[

Free competition here just means that Mr. Brown and Mr. Black too can get in on the action. We don't care if Smith pays off Jones or Brown or Black so long as the Jury duty gets done.  

They infer that "as a matter of logical consistency, a person who embraces a notion of fairness must on some occasions favor adopting legal rules that would make every person worse off," a conclusion that they view as "a serious challenge to proponents of notions of fairness who also care about the well-being of members of society." 

This follows from having any sort of rule or algorithmic procedure. However, one can embrace a notion of fairness which is non-deterministic or 'intensional' in a particular sense. I might say 'it would be nice if everybody were nice'. You might dispute this. Soldiers shouldn't be nice to invading armies. I might say 'I meant it would be nice if everybody were nice only after everybody was actually nice'. This is perfectly acceptable though it is not informative or operationalizable.  One can have faith in a thing without having to subscribe to any particular belief regarding the thing. Such faith is indefeasible. But the law is concerned with defeasibility.  

We might call their claim "The Impossibility of a Fair Paretian." Unlike Sen, however, Kaplow and Shavell view their impossibility claim as a critique of all fairness notions (including liberal rights) rather than of the Pareto principle as a universal rule.

The problem here is that reliance on any principle regarding anything can leave everyone worse off with respect to that thing. But this is just another way of saying non-deterministic solutions dominate deterministic ones. Kaplow & Shavell say 'all principles that are not based exclusively on welfare will sometimes favor policies under which literally everyone would be worse off.' But even principles exclusively based on welfare have this same effect.

Obviously, once you add criteria- e.g. welfare plus fairness plus aesthetic merit- you get more and more sub-optimal outcomes. In any case, having to justify actions according to criteria itself wastes resources. It is obvious that the best rule would be to have no rule provided everybody obeys the next best rule set without it, in fact, being given normative force. 

Au fond, there is an 'intensional fallacy' at the bottom of all this rigmarole. The Pareto frontier is intensional. Nobody knows where it is. This doesn't matter because intensional terms like 'niceness' can be useful even if their extension is unknowable. 

In practice, different jurisdictions interpret contracts and other legal documents in different ways. In England, speaking generally, no importance is attached to ' prior negotiations or the parties' "declarations of subjective intent"'. In Germany, however, Courts apply a two-stage approach to interpretation: first, to seek out subjective intent and, secondly, to determine objective meaning in the light of negotiations and other extrinsic material.' One may say, that subjective intentions or 'intensions' have greater salience in German jurisprudence, whereas, in England & Wales, a more objective, textual, or 'extensional' approach is taken. However, this may make little practical difference. After all, hermeneuts of different types may still interpret the law to the same great end- which Hume declared to be Utility, as far as Justice was concerned. 

That is, Kaplow and Shavell use their claim to advocate “ welfarism,” which makes moral judgments a function only of the utility of individuals.

It is obvious that 'welfarism' can pauperize a country.  Why flog a dead horse?  

Welfarism includes a broader class of moral theories than utilitarianism, which takes social welfare in a given population to be equal to the sum of individual utilities.

The problem is not utility but disutility which is related to opportunity cost. Smart people flee jurisdictions where they are fleeced.  

A welfarist theory need not assume that social welfare for a given population is equal to the sum of individuals’ utilities.

Why not? An assignment of that sort can always be made. Anybody can do any arbitrary shit they like.  

Kaplow and Shavell, for example, assume a more general social welfare function that permits the distribution of a given amount of utility among individuals to affect social welfare.

Why not a SWF which permits the distribution of utility to fairies at the bottom of the garden?  

Any form of welfarism, however, including utilitarianism, ranks states of affairs entirely on the basis of utility, regardless of other information about those states.

But utility is a Tarskian primitive. It can incorporate any type of information or fantasy or arrant nonsense.  

Thus, the claims of Sen and of Kaplow and Shavell both have important implications for the fundamental normative question of what criteria we should ideally use to evaluate laws or public policies.

No they don't. The plain fact is, evaluation done by stupid Professors is going to be shit no matter what criteria are used.  

Both Sen’s claim and Kaplow and Shavell’s claim refer to the “ weak” version of the Pareto principle, which holds that if every individual prefers any alternative x to another alternative y, then x is socially preferable to y. 

Every individual may prefer every individual to be delivered from disease and death by Society. Indeed, every individual may prefer to go on strike till Society bribes or bumps off the Grim Reaper. But Society can't have a revealed preference for any such thing. Indeed to say x is socially preferable to y isn't any more informative than to say the Galactic Spaghetti Monster wants x and will shit on anyone who prefers y. 

The “ strong” Pareto principle holds that if at least one individual prefers x to y, and no one prefers y to x, then x is socially preferable to y. 

Cool! This means my farting just now was 'socially preferable' to my not farting. It is seldom that Society takes an interest in my activities but my farts truly are quite spectacular, if I say so myself.  

Thus, under the strong Pareto principle, x may be socially preferable to y even if all but one individual is indifferent between x and y. The strong Pareto principle is the stronger condition insofar as it implies the weak Pareto principle. That is, if the strong Pareto principle holds, then the weak Pareto principle must also hold.

Which only matters if you think my farts matter to Society.  

If the weak Pareto principle holds, however, the strong Pareto principle does not necessarily hold.

So, if only I benefit by farting, Society gains nothing. Sad. On the other hand, weak Pareto is still sub-optimal because it militates against individual experimentation and 'discovery'- i.e. trying something you don't prefer just in case you like it or learn something about it. My point is that preferences have no necessary connection with Pareto improvements. 

Prof. Chang

presents a critique of welfarism from a liberal perspective and proposes a liberal alternative to the utilitarian theory of social welfare.

The problem here is that such theories are based on arbitrary stipulations and thus can be bent any which way. 

I take issue with welfarism’s indiscriminate inclusion of all forms of satisfaction, including the satisfaction of racist or malicious preferences, in the calculation of social welfare.

Why bother? The fact is if the disutility (including such as arises out of repugnance) experienced by a larger number of people causes the rapist and serial killer to be punished, then... well, you have society pretty much as it is. 

The point about utilitarianism is that, once you shift the focus to 'transferable' utility, disutility and informational signaling and screening devices- you have a complete explanation for a lot of features of actual societies. 

I argue that a liberal theory would exclude the satisfaction of such objectionable preferences from our notion of social welfare.

In which case, you don't have a good structural causal model and can't do proper mechanism design. At one time, Professors had to pretend it was highly objectionable that some men like kissing other men. This meant some quacks could get away with charging an arm and a leg for 'conversion therapy' while nasty people took advantage of what was in effect a 'blackmailer's charter'. 

There are some 'intensions' about which we have intuitions- e.g. 'this is fair', 'this is elegant', 'this is beautiful'- but regarding which we can say nothing about 'extensions'- i.e. we have no rule by which to generate all the members of the set of the fair, or the elegant or the beautiful.  It is implausible that a 'verification'- e.g. 'this is beautiful'- can somehow generate what it evaluates. Otherwise, a person who recognizes beauty could also make herself beautiful. 

I argue that any plausible fairness theory includes multiple principles and that any such pluralistic theory must specify rules for when some principles take priority over others.

Is there a plausible 'Beauty' theory of this sort? No. A unconscionable contract is an unfair contract. Is there an algorithmic way of always and only picking out such contracts so as to render them void? No. Judges have to consider the facts of the case even if the law of contract has an algorithmic structure. There will have to be equitable exceptions where the law fails by reason of over-generality.  

There is nothing to prevent a pluralistic fairness theory from including the Pareto principle among its principles and giving the Pareto principle priority over other principles. 

The problem here is that the 'ratio' in any given case may be highly sensitive to how the facts of the case are determined. Priority is merely a legal fiction. The thing is not robust. It is context dependent.  

One may, in a rough and ready fashion, describe the 'pragmatics' of a particular legal regime- e.g. Competition policy in country X is more discretionary than in country Y. But this depends on received ideas about the public interest. It is not connected with jurisprudence and its 'i-language' of 'terms of art'. 

Liberalism is closely associated with what Category theorists call 'nauturality'- i.e. non arbitrariness. Yet, for the last 50 years, we have known that something like Kuhn's 'no neutral algorithm' stricture applies. Uncorrelated asymmetries give rise to eusocial bourgeois strategies. But they are arbitrary. The reason 'opitimization' is attractive is because it promises 'canonical' or non-arbitrary solutions. Sadly, because of Knightian Uncertainty, these are inaccessible save perhaps 'at the end of time'. Justice, like pedagogy or prostitution, must content itself with being a service industry. The alternative is to introduce more and more arbitrary and nonsensical principles and paradoxes till, quite suddenly, the Chinese eat our lunch. 

No comments: