Wednesday 9 March 2022

Robert Sugden & why behavioral econ turned to shit

As a species, it makes sense for us to have diverse preferences and endowments because this enables faster & more exhaustive exploration of the fitness landscape. Individuals have preferences as well as mimetic drives to ignore those preferences and mimic those of more successful individuals because that is a regret-minimizing strategy for the species. 

It is useful to define different metrics for different purposes on any actual landscape. Thus we may want to know which is the quickest way to walk from a to b, or else which is the most scenic route between a to b, or which route is best for a pub crawl. This is because the landscape itself exists objectively in the same way for all of us. However, 'connectivity' is problematic. The next shortest path or next most scenic route may be far away from what was selected and thus one can't easily switch between them. However, precisely because the landscape is fixed, concrete steps can be taken to increase connectivity. 

Sadly fitness landscapes for our species don't have this feature. They are defined on coevolved processes. Thus 'quickest' routes can become 'longest' routes because of traffic jams with no way of increasing connectivity. Scenic routes can become fucking eyesores. Pub crawls don't last very long if drunken girls get tired of glassing each other and turn on us. 

One reason having consistent or alethic preferences is foolish is because you make yourself vulnerable to a predator or a parasite. Thus 'mixed strategies' dominate. 

Why did Amartya Sen get a Nobel for worthless shite? The answer is that if there is a hegemonic power- a rule setter- then there can be a useful metric for Social Choice under Globalization. However, instrumentalizing that metric predictably caused it to become mischievous. Co-evolution, not talking virtue signaling bollocks, is what determines what types of Society will exist and why some will shrink as others expand. This has become obvious to us as Putin's troops pillage Ukraine. 

Is Welfare Econ utterly useless? No. It is the sort of thing we all do all the time. We just don't make any great claims for it. The same is true of behavioral econ.  However, claiming that behavioral welfare econ has some magical property is silly. The thing is wholly arbitrary and was soon shown to be utterly mischievous. 

A better way forward was available. Rather than indulging in 'mathsy' bollocks, people could have just focused on existing mechanisms and how to make them better. Where behavior is diverse, structural causal models can be investigated such that we get an idea of how a particular outcome can be more parsimoniously achieved. But making normative judgments about behavior serves no worthwhile purpose or else can itself be castigated in the same manner. Welfare Econ, as taught by pedants, tended to get sillier and sillier just as O.R and Fintech got smarter because the wooly minded and diversity hires got corralled into it. For a brief while- when Russia was sliding into chaos and China was still struggling to rise, Sen-tentious shite was in joint-supply with other types of complacent, virtue signaling, displays of the soft power and exorbitant privilege of the 'rule-setter' or 'nudger' of some supposed 'Rules based World Order'. Then came a collapse in first financial, then military, hegemony and suddenly everybody realized that the Ivy League had been producing cretins and charging an arm and a leg to do so. 

Robert Sugden, whose textbook on CBA I recall using, has an excellent paper here titled Normative Econ without Preferences.

From the 1930s to at least the 1980s, there was a clear consensus among economists about the theoretical framework within which normative analysis was conducted. This framework rested on a very parsimonious conceptual scheme, essentially as follows.

A society consists of a set of n individuals. A social state is a possible description of society. This may be a single description (for example, ‘the socialist party forms the government’) or an n-tuple of individual-specific descriptions (for example, specifying each individual’s bundle of consumption goods).

What about each person's description of society? Should that not be included? What about each person's description of other people's descriptions of society? Surely that is relevant because if affects expectations? What about the outcomes of social processes which change such descriptions? Surely they too would be relevant? The problem is that none of these things are unique. My description of society is contingent on my not having been mugged on the way to the gym. Given that today is a sunny day and a pretty girl smiled at me, I'm in quite a good mood and have a correspondingly warm feeling towards society. 

A social state is surely something more than an economic state of society.  Even if it isn't, it is a fact that some goods and services are only consumed because they change the way we'd describe society. The trouble is that whereas we can say with certainty 'you ate that apple' or 'you didn't eat that apple. It has been saved for future consumption', we can't really tell whether we consumed a 'non-excludable' service or not. Moreover, it is perfectly possible to hold two or more inconsistent 'descriptions' or attitudes to society. This may affect a future decision and we could speak of this as a 'damming up of capacitance diversity' which may be released later. However for the moment there is no unique 'n-tuple' specifying our consumption bundle. Thus there can't be a non-arbitrary 'binary relation'. In other words, whatever consensus in this field obtained, it was empty. 

 For each individual, there is a binary relation of preference over the set of all relevant social states.

But if one feasible 'relevant social state' is optimal for that individual, why can't that information be available to him directly? If there really is a 'binary relation' of the type described above, why hasn't someone specialized in spotting the supremum and selling that information so that nobody need bother to rank social states? 

Consider a society composed of mathematicians. They are all beaten to death unless the can prove the Reimann hypothesis within a specified time frame. Clearly, because mathematicians genuinely want to advance Maths, they would all prefer the social state where this is exactly what happens. The optimal Social Welfare Function would assign different tasks to different mathematicians and would coordinates their activities by setting Conferences on specific topics with specific invitees. But knowing it would give us a lot of clues as to how the problem can be solved. This in turn would change the SWF with the result that even more clues become available. 

This is not particularly far fetched. A bunch of guys tasked with beating the Nazis to the secret of the Atom bomb might spend a lot of time discussing how they should go about it. These discussions themselves spur research so new vistas open. In the process, entirely new and very fecund horizons might be opened up. Progress might not just be faster towards a specific end, it may be faster to every possible end- including undesirable ones like blowing up the Earth. 

My point is that people want to get together to work out ways to promote the Social outcomes they desire. Social Welfare is an activity of a spontaneous and non-deterministic type. Its purpose is to change what its participants think and do and prefer. The thing is not an algorithm or a mapping. It is foolish to assume the opposite. One might as well say 'all humans end up dead. Life is a function which turns babies into corpses'. 

 Preferences are complete (that is, the individual can rank each pair of social states in terms of preference or indifference) and integrated (that is, non-stochastic, context-dependent and internally consistent in the sense defined by axioms such as transitivity and the sure-thing principle).

Thus preferences can't be 'adapted processes'- i.e. ones that 'can't see into the future'- because they are non-stochastic. However many goods and services are stochastic. The Ito integral was being used in mathematical finance by the Seventies. It seems the 'consensus' upon which normative economics had settled on was decidedly out of date. One 'sure thing' everybody knew about was that using traded options could be beneficial for everybody. Why define preferences in such a manner an actual 'sure thing' was ruled out? 

 An individual’s preferences are revealed in the choices she makes from opportunity sets of feasible objects.
But this was not observable for free goods, some types of self-produced goods, non-excludable services etc. Revelation only occurred when a transaction occurred. But most choices have to do with deciding not to transact. Thus what is revealed isn't even the tip of the iceberg- it is a piece of that tip which was hacked off and which floated away.

 Normative economics is concerned with the social ranking of social states. The Pareto principle prescribes that if all individuals weakly prefer some social state x to another social state y, then x is socially ranked at least as highly as y; if in addition at least one individual strictly prefers x to y, then x is ranked strictly above y. Most applications of normative economics accepted this principle.

It was silly to do so. I suppose the 'sure-thing principle' or the 'independence of irrelevant alternatives' militated for the notion that the thing was harmless or inconsequential. This involved a more or less implicit 'axiom of truth' whereby what people knew, or thought they knew, was assumed to be true. Yet, knowledge is strategic. There are lots of things we could, and maybe should, know but which it is better for us to remain ignorant of. 

However, this conceptual scheme was open to a range of different interpretations, each of which was compatible with the standard neoclassical theory of individual behaviour. It was possible, and indeed quite common, for economists to use the scheme without saying how they interpreted it.

I think the context made it obvious which ideological camp or affiliative group was being appealed to. 

Crucially, the concept of preference could be interpreted in different ways. On some readings of revealed preference theory, preference means choice (subject to some qualifications about indifference): to say that a person prefers x to y is to say that if she faces the opportunity set {x, y}, she in fact chooses x. On another interpretation, a preference is a mental state that tends to cause choice: to say that a person prefers x to y is to say that she is psychologically disposed to choose x from {x, y}. On a third interpretation, a preference is an individual’s subjective judgement about her welfare: to say that a person prefers x to y is to say that, in her judgement, x would be better for her than y. If individuals are self-interested and rational in the sense of neoclassical theory, these interpretations are observationally equivalent. The person’s judgement that x is better for her than y provides her with good reason both to be disposed to choose x rather than y and to act on that disposition.

I think the motivation of Revealed Preference was to get rid of subjective talk and arbitrary stipulations of things inside the heads of other people. However there is no 'observational equivalence' under different interpretations. Inconsistency means Preferences changed in the first case. In the second case, psychological factors gain salience. There may be 'signaling' or 'strategic behavior' or something more complex going on without any change in Tastes. The third case makes Social Welfare impredicative upon individual notions of what it should be. 

It is perfectly possible to be rational and self-interested as well as to have a rich psychological life in which strategic behavior plays a part while also holding decided views on what welfare ought to mean and how it ought to be implemented. 

But it would be foolish to try to fit psychology or moral philosophy into a formalism invented for a very different purpose- viz. to permit econometricians to help enterprises get better at predicting demand and supply. 

One might say that this formalism represents 'mathematical economics'. But, do the great mathematicians of the world come to mathematical economists and ask them to figure out how they should best coordinate their activities so as to solve the Reimann hypothesis more quickly? No. Of course not. The fact is this type of 'mathsy' econ is only done by guys who couldn't go the distance in pure fields. There may have been exceptions behind the Iron curtain- guys like Kolmogorov who had to do some O.R type work for plywood factories, etc, in order to evade being sent to a Gulag for the crime of 'bourgeois idealism'. 

The concept of a social ranking was equally open to different interpretations. 

Only one mattered. If we've gotten the 'Law & Econ' mechanism design right and preference and endowment diversity is not too great or too small then whatever outcome the market arrives at with the judiciary's imprimatur is optimal if legislators are getting Public Finance right. No doubt, we could always incorporate any substantive, properly budgeted measure, into the Law or Fiscal Policy such that some obvious benefit is received or danger avoided. 

On one interpretation, it is a normative judgement made from a neutral viewpoint by a ‘social planner’ or ‘ethical observer’: to say that x has a higher social ranking than y is to say that, as judged by a neutral observer, x is better for society than y. 

Fuck 'neutral observers'. They have no skin in the game. Who cares what the UN Special Rapporteur says? We know she is a virtue signaling ninny who will suggest that Scottish women don't have enough access to arable land to feed their wee bairns. 

On another interpretation, it is an aggregation of individuals’ preferences: to say that x is ranked above y is to say, roughly speaking, that there is a greater weight of preference on the x side than on the y side. (Needless to say, the meaning of such an aggregation depends on how the concept of preference is interpreted.) 

98 % of all people polled agreed that cats should be allowed to marry dogs provided Putin shoves his nukes up his ass. 

On a third interpretation, a social ranking is the actual result of some (or perhaps of a recommended) collective choice mechanism: to say that x is ranked above y by some specific mechanism is to say that that mechanism would choose x from {x, y}.

Yup. That's how Westminster works right enough- thinks nobody at all. 

The open-endedness of the concepts of preference and social ranking left room for many different interpretations of the Pareto principle. If a social ranking is a ‘better for society’ judgement made by a neutral observer, the Pareto principle might be a normative principle prescribing respect for individuals’ choices, dispositions or subjective judgements.

There is no way to know that a butterfly beating its wings in China won't, by some extraordinary chain of events, result in nuclear Armageddon. A person could claim that every action by any and everybody worsened her welfare. Those who appealed to the Pareto principle in the manner quoted above could be told that they ate dog turds even if they didn't eat dog turds because nobody but the author of this ignorant screed cares one way or another about such a ridiculous allegation. But this one crazy cretin cares about nothing else. Thus ascribing dog turd devouring to welfare economists represents a Pareto improvement for Society. 

 Or it might be a normative principle requiring the observer to take account of each individual’s welfare, defined in some objective sense, combined with the assumption that individuals in fact tend to prefer what is objectively good for them. If a social ranking is an aggregation of preference, the Pareto principle might be a minimal, non-normative aggregation principle. (If some component of an aggregate increases in quantity and if no component decreases, the aggregate increases.) If a social ranking is the result of a recommended collective choice mechanism, the Pareto principle might be a minimal normative condition of democracy. (Voting rules should be such that, if a social decision has to be made between x and y, and if no one votes for y and at least one person votes for x, then x should be declared the winner.)

All this is silly. Aggregation is good because of the Law of Large numbers and the Condorcet Jury theorem and the fact that we have evolved using a regret minimizing strategy on an uncertain fitness landscape. It would be foolish to vote where few are involved. On the other hand Hohfeldian incidents, in Law, take care of 'immunities' and 'obligations' we care about. 

The variety of interpretations of the Pareto principle allowed a corresponding variety of interpretations of the First Fundamental Theorem of welfare economics—the theorem that every competitive equilibrium is Pareto-efficient.

Easily disproved. I am made better off and nobody is made worse off by a rider to the effect that dog turd devouring is rampant among Welfare Economists. I may mention, I have carried out an exhaustive correspondence with every single such pedant and they have all refused to enter this debate. This shows they are indifferent to the outcome. 

In particular, it could be interpreted as a conclusion about how competitive markets contribute to social welfare, or as a conclusion about the tendency of such markets to respect consumer sovereignty by responding to individuals’ actual demands. Thus, economists could agree that the First Fundamental Theorem picked out a desirable feature of markets without actually agreeing about what that feature was.

Incentive compatibility. The folk theorem of repeated games. Coase's theorem. Law & Econ. Mechanism Design. Stuff that motivated useful work. Binmore made 10 billion for the exchequer with some nifty auction design. That was super cool. 

Some of the tensions in economists’ apparent consensus about normative analysis began to emerge in the 1970s and early 1980s, when social choice theory was one of the most fashionable areas of economic theory and a topic of interest to many philosophers. But developments in behavioural economics revealed more fundamental problems. The long-standing consensus was based on a shared sense that choice, preference and welfare were closely related. 

They are but cognition is costly and mimesis is cheaper than mathematics and, anyway, Knightian Uncertainty obtains and why was I not taught about Volatility surfaces when I was at skool? 

But a recurring finding of behavioural economics is that individuals’ choices are sensitive to features of the ‘context’ or ‘framing’ of decision problems that seem to have no relevance to individuals’ interests or well-being. The implication is that individuals often do not have the kind of preferences that economists have traditionally assumed when treating preference as the central concept in their normative analyses. Something in the consensus position has to be given up. Choosing what that should be, and so finding a way forward for normative economics is the reconciliation problem.

If so, this is a 'reverse mathematics' project. Find a more parsimonious system within which the difference can be quantified and then do 'reconciliation'. I suppose this will be category theoretical. 

Non-meddlesome preferences and the problem of the Paretian liberal

My approach to this problem builds on what I learned through my involvement in debates about the interpretation of the concepts of ‘individual preference’ and ‘social ranking’ that took place among social choice theorists in the 1970s and 1980s. A theorem due to Sen (1970), ‘The impossibility of a Paretian liberal’, crystallised some of the central issues in these debates and was the subject of my first work in the philosophy of normative economics. Let me explain.

For Sen, this theorem was part of a research programme grounded on the belief that normative economics, as practised at the time, had too narrow an ‘informational base’.

Sen came from India where there was a 'Gandhian economics' tradition as well as various Marxist economic traditions. However, the mainstream econ he had himself taught had a wide 'informational base' because the Indians were quite good at sampling theory and had a lot of data on nutrition and educational outcomes so forth. All Indian econ was normative. Sen's colleagues at the D.School joined the Planning Commission or presided over the Reserve Bank or became Finance Minister or, in one case, a two term Prime Minister. Every action taken was justified on the basis of CBA and, moreover, there were innumerable Commissions and Committees and Official Inquiries and so forth. 

 Sen shared the consensus view that normative economics was about social rankings of social states,

But this was not the consensus Indian, or even British, view. Some Social States were better than others because little kiddies got more milk and elderly people didn't have to freeze to death. The Government was expected to ensure these outcomes and if people were willing and able to pay enough in taxes, that's exactly what they did. 

 but challenged the ‘welfarist’ assumption that, in arriving at judgements about social rankings, the only relevant information was information about individuals’ preferences. 

But all information may be encoded in some individual's preferences and, even if this person does not exist, as an ideal there may be Tardean mimetics of his preferences. This is the 'field theory of Buddhas'. But it also explains why Mahatma Gandhi- though an obvious crack-pot financed by greedy banias- is considered the Indian George Washington- or, in the context of Brexit, perhaps I should say the Indian BoJo. 

His theorem is intended to show that the welfarist approach cannot take account of widely shared moral intuitions about the value of individual liberty.

No account given by a cretin can account for shit. 

Since Sen is presenting an impossibility result, he defines what he sees as a minimal condition of social respect for liberty. His condition of minimal liberty is that there should be some protected personal sphere for each of at least two individuals in society. Which individuals should have this privilege, and what should belong to their personal spheres, is left open, so on the face of it this is an extremely weak requirement.

No. It is far too strong. If even one individual has a 'personal sphere', they could assemble a Doomsday device there. There can't be a Hohfeldian immunity of this sort. Either you can have Liberty under the Rule of Law- in which case every Hohfeldian incident is justiciable and the underlying right is defeasible, or else Liberty is for those whom none dare fuck with and also those they take under their protection. 

A personal sphere is to be understood as a nonempty set of pairs of social states. If some pair of social states {x, y} is deemed to belong to person i’s sphere and if i prefers x to y, then x must be socially ranked above y. 

Either this social ranking is 'justiciable' in which case it is defeasible, or else it is merely descriptive in which case there is no reason to rank it. We don't say society ranks Beyonce as more beautiful than me. She just is more beautiful. Society has got nothing to do with it. 

Sen (1982, p. 286) says that the minimal liberty condition is intended ‘to permit each individual the freedom to determine at least one social choice, for example, having his own walls pink rather than white, other things remaining the same for him and the rest of the society’.

In that case the thing should not show up in the social ranking. It is a fact that some people may do disgusting things to themselves in their own bedrooms with the curtains closed. But, as a society, we are not obliged to give a higher rank to their doing so rather than their not using their vacuum cleaner in a manner which the manufacturer never intended. 

 The theorem shows that, for any society and for any specification of personal spheres that satisfies the minimal liberty condition, there is some profile of individual preferences such that either the Pareto principle is violated or the social ranking contains a cycle.

It also shows that Amartya Sen eats dog shit- unless it is a fact there is even one other person in the world who is not utterly indifferent to this assertion of mine. 

In my contributions to the debate about the theorem, I gave prominence to the following example, based on one originally presented by Gibbard (1974). In a strict sense, it is not a proof of Sen’s theorem, but it illustrates the theorem’s underlying normative logic. The example should be read in the context of the marriage norms of the 1970s. There are three unmarried individuals, Annie, Bill and Charlie. Three alternative social states are feasible: that no one marries anyone, x; that Annie marries Bill, Charlie remaining single, y; and that Annie marries Charlie, Bill remaining single, z. On any normal conception of liberty with respect to marriage, every individual has the right to stay single if he or she so chooses, and any adult man and adult woman (if not close relatives and not already married to anyone else) have the right to marry if they both choose to do so. The first of these propositions fits naturally into Sen’s conception of the personal sphere. Consider any two social states that differ only in respect of whether a particular man is married to a particular woman—for example, x and y, which differ only in respect to the marriage or non-marriage of Annie and Bill. It would be in the spirit of Sen’s characterisation of liberty to say that either Annie’s preferring x to y or Bill’s preferring x to y is a sufficient condition for x to be socially ranked above y. Indeed, Sen (1982, pp. 299–302) seems to endorse this formulation in a discussion of liberty in relation to marriage. An analogous formulation of the second proposition would be to say that if the only difference between two states (such as x and y in the example) is whether a particular man and woman are married to one another, and if both of them prefer being married to one another to remaining single, then this is also the social ranking.

Society may want a particular couple to get married because it is affected by the genetic quality of the offspring. It may want another couple to stop getting jiggy coz their spawn will be murderous sociopaths. 

As I like to tell the story, Annie and Bill are long-standing friends. Annie would like the two of them to marry. Bill would prefer the relationship to continue without any commitment, but would rather marry Annie than lose her altogether. Charlie admires Annie and would like to marry her. Annie is fond of Charlie and would settle for him if it was clear that Bill would not marry her. Representing this familiar triangle in terms of transitive individual preferences: Annie prefers y to z and z to x; Bill prefers x to y and y to z; and Charlie prefers z to x. 

This is not the case. The choice menu has been mis-specified. Annie wants to marry any X better than Bill or Charlie. This is like the solution to the Secretary problem. Interview three candidates and then hire the next who is at least as good. Society should encourage better solutions to the stable marriage problem by creating mechanisms whereby candidates can 'do discovery' without compromising themselves before pairing off in a time delimited fashion so the 'market clears'. Marriages matter because babies matter. The 'affluent society' and the 'welfare state' depended on creating economic conditions favorable to pairing off for the purposes of mortgage, matrimony and being made miserable coz your kids think you are totes uncool. 

Bill wants milk but not the cow. He represents a 'hold-out' problem. Kaushik Basu has a little play 'Crossings at Benares Junction' where an aggressive Iyengar girl tries to trap the Bengali hero into marriage by pointing out that by marrying her, her other suitors too benefit because they find other partners. For some mysterious reason, Iyengar girls were depicted as sexually predatory back then. Anyway, the stupid Bengali game theorist wrongly applies 'backward induction' to scare off the Iyengar. I need hardly say that an Iyer maiden would have raped the fellow with a strap-on so incessantly that he would marry his violator out of a desperate desire to 'restore his honor'. 

In one possible version of the story, Charlie (perhaps modelling himself on the hero of Dickens’s Tale of Two Cities) self-sacrificingly prefers y to z. This leads to a variant of Sen’s impossibility result. Socially, z is ranked above x (because this is the shared preference of Annie and Charlie) and x is ranked above y (because this is Bill’s preference). But if the social ranking is also required to satisfy the Pareto principle, y is ranked above z, producing a cycle.

So what? Why should we care? What matters is whether Annie gets preggers. How it happens affects Society. Charlie may prefer to withdraw his offer of marriage if he thinks he will be raising another man's child. 

However, I would rather have Charlie believing that he is the right man for Annie, and so I tell the story with Charlie preferring z to x and being indifferent between x and y. Forget about social rankings for the moment and ask what would happen if these preferences were common knowledge and if Bill and Charlie were both free to make proposals of marriage to Annie. Charlie has nothing to lose by proposing. If Bill expects Charlie to propose, the best he can do it to propose too. Having received both proposals, Annie will accept Bill. (If the story were set on 29 February, when by tradition women propose to men, Annie’s best strategy would be to propose to Bill, making it clear that if he rejected the proposal she would propose to Charlie. The result would be the same.) So, by permitting marriages to be arranged by a propose-and-accept rule, society allows y (Annie marries Bill) to come about when x (no one marries anyone) is feasible. But Bill prefers x to y. If we accept Sen’s characterisation of liberty, respect for Bill’s liberty requires that x is socially ranked above y: we must conclude that the workings of the propose-and-accept rule have violated Bill’s protected personal sphere. Now the paradox is not an inconsistency between liberty and the Pareto principle; it is that Sen’s characterisation of liberty is telling us that, if society is to protect individual liberty, it cannot allow marriages to be formed by free choice and mutual consent. Another way of expressing the sense of paradox is to say that the liberty that (supposedly) has not been respected is Bill’s freedom not to marry Annie, but Bill has in fact chosen to marry her. Has he invaded his own personal sphere?

If he feels so, Annie may change her mind about marrying him. After all, her preference is to be married to a willing partner not a guy who thinks you ruined his fucking life and OMG you just don't understand me. I could of been a rock star! 

The stable marriage problem is one which should concern Society. But, clearly, welfare economists can say nothing useful in this context. Nor can I. On the other hand, one look at me will convince most men my age why they need to clean up their act if they don't want wifey to divorce them once the kids move out. 

Sen’s impossibility result initiated a huge literature in which social choice theorists tried to find coherent ways of representing individual liberty. Most contributions to this literature framed this problem in the same way that Sen had done—as a problem of arriving at a social ranking of social states, using information about individuals’ preferences over those states. For many writers, the starting point was the recognition that that Sen’s result (along with a whole range of related puzzles) is possible only if some individuals have strict preferences over pairs of states that belong to other individuals’ personal spheres. (In the first version of my example, {x, z} belongs to Annie’s and Charlie’s spheres, but Bill prefers x to z; {x, y} belongs to Annie’s and Bill’s spheres, but Charlie prefers y to x.) Such preferences were often characterised as ‘meddlesome’. One response was to look for ways of editing (or ‘laundering’) individuals’ preferences so as to remove elements of meddlesomeness before using them as inputs to the formation of social rankings.

On the other hand, it was possible that Sen was simply stupid. He mis-specified the set of social states. He himself cites no 'meddlesome' preferences. Lewd doesn't really want Prude to read porn. He just wants Prude to be more open-minded. Similarly Prude would like Lewd to consider the harmful consequences of a hedonistic lifestyle. 

Human beings have great plasticity in behavior and have plenty of cognitive capacity to enrich understanding of  'intensional' states of the world


 Sen’s liberty principle might then be retained by applying the Pareto principle to laundered rather than actual preferences.

The bigger problem is of 'strategic' preferences. If you object to something I do, I find something you do which I can claim causes me great offense and psychic pain.

 Another response was to argue that if, in some domain of life, an individual is sufficiently meddlesome, his or her right to a protected personal sphere should be deemed to have lapsed.

e.g. a vexatious litigant being barred from commencing civil proceedings without prior permission.

The Pareto principle might then be retained by restricting the scope of the liberty principle. (Compare the popular view that the rights of free speech and assembly should not be given to anti-democratic organisations.) A third response, in the spirit of Sen’s critique of the overly narrow informational base of welfare economics, was to argue that the social ranking should take account of the intentions and motivations that lie behind meddlesome preferences. (Does Bill’s preference for x over z reflect spite, jealousy, or just a desire to maintain a long-standing friendship?) Increasingly complex—not to say baroque—formulations of liberty were proposed, discussed, and found to generate yet more paradoxes.

It is sufficient that people can easily enter or exit a 'Tiebout' model to say that for those who reside there Liberty exists to the extent they desire it or are willing to invest in it. The problem is that Liberty can become very very expensive to defend. Doing so means creating threat points. In other words, the set of feasible economic states must be vulnerable in specific ways for Liberty to itself be robustly feasible. 

I suppose, for John Stuart Mill, who was born a year after Trafalgar and who died 20 years before Germany began to challenge British naval supremacy, it seemed obvious that utilitarianism- such as that which supposedly characterized the 'good despotism of John Company or, indeed, that Benthamite nonsense over which Raja Ram Mohan Roy had gone ga ga- could generate sufficient power to protect itself without any 'excess burden'- i.e. the Royal Navy could more than pay for itself while the colonized heathen could pay for the standing Army and provide the Red Coat with plenty of target practice. 

During the First World War,  it was realized that Liberty meant conscription even for England. But, this was equally true of any political philosophy. The lives of young men have to be sacrificed on the battlefield so that Plato's Academy can survive to corrupt youth. 

An initially small minority of theorists, of whom I was one, argued that the source of the problem was the presupposition that respect for liberty should be represented as a relationship between individual preferences and social rankings (Nozick 1974, pp. 164–166; Sugden 1978, 1985; Gärdenfors 1981). Although we never managed to persuade Sen,

because he thought endless public discussion would magically make everybody nice as pie. 

 we did convince three of the leading social choice theorists of the time (Gaertner et al. 1992).

Our argument was that liberty is a property of the procedure that determines which social states come about, and so is a relationship between individuals’ choices (not their preferences) and social outcomes (not rankings). 

I suppose, the meaning is 'determines that portion of which social states come about' which are under the control of those individuals who are choosing to fight for liberty'. 

As Peter Gärdenfors pointed out, procedural properties of social choice can be represented using the theoretical framework of game forms—that is, games with the same properties as those of classical game theory, except that outcomes are described in physical terms, rather than as profiles of players’ utilities. In this conceptual scheme—sometimes called the game form approach—individuals’ preferences play no role; ‘social’ judgements are not about the outcomes of a society’s decision-making procedures, but about the procedures themselves.

Mechanism design is 'reverse game theory'. Choice of mechanism is part of Social Choice. This seems perfectly reasonable. 

 In the case of marriage, respect for individual liberty makes certain requirements of the procedure by which marriages come about. 

Why? What's so special about marriage? Why not treat it simply as a contract? 

Specifically, the marriage of any single and adult man and woman should occur if and only if they both choose to marry.

It may be more equitable to treat them as if they were married if such were the case for all practical purposes and great harm would result if an exception were not made

 Either version of the propose-and-accept rule satisfies this requirement. In the example, the social state that comes about—Annie marries Bill, Charlie remaining single—does so through the workings of a procedure that respects individual liberty. 

Does it though? Greater liberty exists if the form of a contract is not arbitrarily limited. Bill may prefer some other type of alliance which is just as good for her purposes but falls short of what Charlie offers. Indeed, Bill's preferred offer may be superior in some particular- e.g. he may stipulate that he is prepared to provide financially for her elderly mother. However, because Charlie has made an offer, Bill has to match it. Instead of a 'customized' solution or 'incomplete contract' which Annie may find superior, she is stuck with choosing between two 'contracts of adhesion'. 

That is all that needs to be said. Viewed in the perspective of liberty, Annie, Bill and Charlie’s preferences, intentions and motivations have no bearing on the case.

But Liberty has been restricted because of a narrow definition of marriage which may only subsist for historical reasons of a religious type. 

In some respects, the game form approach is aligned with a way of thinking that became important in moral and political philosophy around the same time as the debates I have described. This was the idea that each individual’s well-being or ‘advantage’ should be assessed in terms of the opportunities from which she can choose, rather than in terms of the outcomes she experiences.

This is foolish. The fact that Obama became President doesn't mean his generation of African American men had equal opportunities to White males (ceteris paribus). Indeed, if 'opportunity' appears better than outcome, chances are there is some type of discrimination going on. 

 This idea is central to John Rawls’s theory of justice, in which holdings of ‘primary goods’—things that ‘normally have a use whatever a person’s rational plan of life’—are the metric of advantage (Rawls 1971, p. 62).

But a person's rational plan of life is likely to embrace reproductive opportunities. In that case positional goods are more important than primary goods because of the way discounted utility works. Perhaps, if instead of spending money on College fees, I'd bought a sports car when I was 17, I'd have many more progeny. By the Price equation, I'd be much better off because I'd have a high chance of having distant descendants. 

 It is also central to Sen’s (1985) theory of capability, in which a person’s advantage is assessed in terms of the set of combinations of ‘beings and doings’ from which she can choose.

This has the advantage of being wholly vacuous. 'Primary goods' still ties back to stuff which can be measured- e.g. how much electricity is used by the household- while 'beings and doings' can't be measured at all. 

 This new interest in opportunity prompted a line of research whose aim was to find ways of measuring the extent or value of opportunity that is provided by a person’s opportunity set. As a contributor to this research programme, I argued that the extent of an individual’s opportunity should be assessed without reference to her actual preferences, but only in terms of what is ‘normally’ or ‘reasonably’ preferred by people in general (Jones and Sugden 1982; Sugden 1998).

This is what a Judge or Tribunal might do to assess damages. Sadly, it doesn't allow much scope for endless pi-jaw. In particular, testimony from bisexual Patagonian penguin handlers regarding the horrors of the Holodomor might be ruled irrelevant. This flies in the face of Sen's insistence on bringing in outside perspectives. 

However, the example of Annie, Bill and Charlie illustrates the limitations of an approach to normative analysis that assesses each individual’s opportunity set independently of those of other people and then looks for a fair or equal distribution of opportunity.

A smart person- perhaps a marriage counsellor- might be better able to tell Annie what her actual opportunity set is. What made both Game Theory and Social Choice stupid was that the guys doing it were too stupid to properly specify available options. 

The fact is, people thinking of getting married would do well to think carefully about stable marriage dynamics. Try picking a couch together. That's the point where men suddenly realize that marriage means soul murder. 

The problem is that the contents of each individual’s opportunity set depend on the choices that other individuals make from theirs. 
But these choices can feature added cognitive complexity which change the opportunity set. It is one thing to marry someone who because they are the best option available and another to marry them because they have the intensional property of being your soul-mate.

Given the propose-and-accept rules for forming marriages, an individual’s opportunities to marry are not simple properties of his or her opportunity set.

This is foolish. What people look for is a soul-mate. Sure, things might just work out that way but you need to be more invested in this. You're not buying a couch. 

For example, Charlie’s opportunity to marry Annie is an opportunity for him to marry her if she also chooses to marry him. 

Which he could leverage to get sex from her hot, but seriously messed up, sister. 

Given Annie’s preference for Bill, the opportunity set that Charlie faces as an individual is less rich than Bill’s.

Unless he makes the most out of being known to be in the running for Annie's hand. Fuck, hot messed up sister is a minor. Also her ex just got acquitted of multiple homicide after all the witnesses fell off tall buildings. 

 Nevertheless (and contrary to the position that seems to be taken by adherents of the ‘involuntary celibacy’ movement), I think we should say that the propose-and-accept rules treat Bill and Charlie equally. 

If they were equal in other relevant respects- sure. But this is a function of 'maximum uncertainty'- i.e. we become less confident of this assertion the more we know about those to whom it applies. 

The interrelatedness of opportunity sets is crucial for the analysis of market opportunities in The Community of Advantage.

It is also crucial to the existence of language. Indeed, it crucial for a theory of evolution. 

But let me take a step back and consider some of the general methodological features of the controversy about how to formulate principles of individual liberty. The original problem was to give a formal representation of an intuitive normative principle—the principle that society ought to respect individuals’ personal spheres.

Unless it has a good reason not to do so. This should be a justiciable matter, not something given a 'formal representation'. Thus freedom of religion is all very well till we find guys are recruiting for ISIS. Similarly, we may feel private property is cool till we discover Russian oligarchs own a lot of the coolest property in our city while the hike in the base rate means we're in negative equity. 

 Social choice theorists tackled this problem within a pre-existing conceptual framework in which normative judgements are expressed in social rankings and social rankings depend on individuals’ preferences. 

Notional preferences- sure. I suppose one could say there is a 'revealed' aspect to this as well. Foreigners seem keen to come to liberal countries. But they are keen to get the fuck away from liberal countries which are going down the shithole economically or in terms of being nuked to buggery. Meanwhile, various tyrannies which can reward talent have always attracted it. 

Accordingly, they tried to represent the intuitive principle as a particular kind of relationship between preferences and social rankings. It turned out that this representation captured the intuition only if individuals’ preferences had certain properties of non-meddlesomeness.

and were as stupid as shit. Why stipulate that there will always be some sphere beyond the reach of the law such that a nutter might use it to assemble a doomsday device? It's better to let people believe what they want and then fuck them up just as they are about to turn totes Bond villain. 

Take the supposed inviolability of embassies. The CIA did an air-strike on the Chinese embassy in Belgrade in 1999 just to clarify matters. That's one reason Biden closed his embassy in Kyiv a full ten days before Putin's tanks started rolling. 

 Many theorists seem to have concluded from this that the principle itself was applicable only if preferences had those properties. 

It was a stupid principle. 

They responded by proposing analyses which either ‘corrected’ individuals’ preferences to fit the representation, or disqualified non-fitting preferences as normatively unacceptable, or required investigations into individuals’ intentions or motivations as a means of deciding whether particular preferences should be respected. Intuitively, none of these proposals seems compatible with the simplicity of the original principle—the principle that in certain personal matters, individuals should be left free to make their own choices. 

unless there is a compelling public interest not to do so. But that is true of everything. 
The game form approach cuts through these problems by representing that principle in a different conceptual framework—a framework that does not refer to individuals’ preferences and does not look for social rankings of social states.

The problem is one of arbitrariness. If the game form is what it is only because that's what you say it is though others disagree, then no 'principle' of a canonical or 'natural' type has been represented.

Latent preferences and the reconciliation problem


I now return to the reconciliation problem. In 2003, two remarkably similar proposals for tackling this problem were published in American legal journals (Camerer et al. 2003; Sunstein and Thaler 2003). Each of these papers had a prominent legal scholar (Cass Sunstein in one case, Samuel Issacharoff in the other) as one of its authors. The other authors were leading American behavioural economists: Richard Thaler (writing with Sunstein) and Colin Camerer, George Loewenstein, Ted O’Donoghue and Matthew Rabin (all writing with Issacharoff). The titles of these papers were similar in intent: ‘Libertarian paternalism’ and ‘Regulation for conservatives’. The common idea was that behavioural economics could justify interventions in the economy (paternalism, regulation) that neoclassical economists had traditionally opposed, and that those justifications were immune to the kinds of objections that might be raised by thinkers on the political right (libertarians, conservatives). 

This was never the case. The interventions were arbitrary. The other side could always add unacceptable riders. Suppose a cognitive bias of a racist type exists. Should critical race theory be taught? Or should Christian teaching take the place of Darwinian Science on the syllabus? In other words, even if everybody accepts that an intervention is justified, the shape and thrust of it may be anything at all. Mathematics looks at canonical or non-arbitrary solutions which everybody would accept. There is little point having a mathsy type of analysis where no such thing is available.


Both papers were written as manifestos for an approach that I will call behavioural welfare economics. This approach has subsequently become mainstream within behavioural economics and, especially after its popularisation in Nudge (Thaler and Sunstein 2008), widely accepted in policy-making circles.

Then we discovered there could be a Trumpian, or Putinian, or 'Proud Boy' behavioural welfare econ. Why not simply say 'name-calling and making Q-anon type conspiracy theory is a type of Welfare Econ?' The UN may go in for Sen-tentious shite. But it may be wholly ignored and disintermediated as we go back to a world where the truth is a function of which power bloc you are the subject of 

This approach is premised on the assumption that (in Thaler and Sunstein’s words) ‘individuals make pretty bad decisions—decisions that they would not have made if they had paid full attention and possessed complete information, unlimited cognitive abilities, and complete self-control’.

But Governments make worse decisions. 

Public policy, it is argued, should be designed to counter individuals’ tendency to make bad decisions. Crucially, its aim should be to ‘make choosers better off, as judged by themselves’ (Thaler and Sunstein 2008, p. 5; italics in original). Or, as Thaler (2015, p. 326) puts it, he and Sunstein ‘just want to reduce what people would themselves call errors’.

Which is cool, till wokeness turns into your worst nightmare and all sorts of rando nutjobs think they have a right to cancel you coz of stuff you yourself really ought to consider your own cognitive biases and systematic errors of judgment. Were you born with a penis? Boy are you fucked! Cut it off now! Do you have a vagina? Girl, you iz a TERF and don't know it! Chop your boobs off immediately. 

 The implication is that an individual’s judgements about what makes her better off are expressed in the choices she herself would make in the absence of errors induced by inadequate attention, information, cognitive ability and self-control.

Or having been born with a penis or being cursed with light pigmentation. 

By abstracting from the effects of these errors on observable choice behaviour, analysts can reconstruct individuals’ latent (or ‘underlying’ or ‘true’) preferences and use those preferences as data for normative economics. 

Very true. We can see that people ought really to be worshipping Satan in between clamoring for the defunding of the Police. Also, thieves must be rewarded not punished and how come cats don't got no votes? That's fucking Fascism right there!

Similarly, Camerer et al. (2003, pp. 1217–1218) say: ‘[A] large part of behavioural economics describes ways people sometimes fail to behave in their own best interests… It is such errors—apparent violations of rationality—that can justify the need for paternalistic policies to help people make better decisions and come closer to behaving in their own best interest’.

By chopping their goolies off or lopping off their tits. Also how come we aint nuking ourselves in solidarity with the good peeps of Ukraine? How myopic is that? 

Camerer et al. present behavioural economics as the latest step in a series of advances in economics, each of which has relaxed some assumption that had previously been a standard feature of economic theories. They characterise the ‘simplest models in economics’ as assuming perfect competition, perfect information and perfect rationality. From the 1930s, economists began to develop models of imperfect competition. From the 1970s, economists began to develop models of imperfect information. Now (as viewed from 2003), behavioural economics is making a further advance: ‘Relaxing the assumptions of perfect rationality represents a logical next step in this productive progression. The scientific consolidation of psychological findings into a new brand of behavioural economic theory breathes new life into the rationales for paternalistic regulation’ (p. 1218, italics in original).

And for chopping off goolies and lopping off tits. Also we must nuke ourselves immediately after defunding the police. 

Although Camerer et al. do not quite say this explicitly, the suggestion seems to be that the assumptions of the pre-1930 models supported the view that the market system, if left alone, would tend to produce socially desirable results, and that governments should therefore take a broadly laissez faire approach to the regulation of markets. Models of imperfect competition and imperfect information in turn provided rationales for new and more interventionist forms of regulation. Camerer et al. are presenting behavioural economics not only as a further stage in the progress of scientific understanding, but also as a further stage in the recognition of sources of market failure and of the need for interventionist economic policies. As I sense it, the sub-text—one that is common to a lot of work in behavioural welfare economics—is that the findings of behavioural economics undermine traditional arguments in favour of markets.

And in favor of enforcing the laws upon which markets rely. 

How do Sunstein and Thaler’s and Camerer et al.’s approaches relate to neoclassical welfare economics? It seems clear that both sets of authors are trying to arrive at normative judgements about what is better or worse for society, made from a neutral viewpoint. 

Cats are neutral. All judgments should only be made by miaow-people. 

In Sunstein and Thaler’s (2003) paper, this is the viewpoint of a ‘planner’—someone who ‘must design plans for others’ (Sunstein and Thaler 2003, p. 1190); in Nudge, it is the viewpoint of a ‘choice architect’. Camerer et al. do not say explicitly who (apart from themselves) ought to make these judgements but, from the first sentence onwards, it is clear that their topic is ‘regulation by the state’ (p. 1211). By implication, their viewpoint is that of a publicly authorised regulator. 

Authorized by Trump- sure. Things are working out swell in Putin's Russia coz of regulators- right?

However, much of the evidence that Sunstein and Thaler present in support of the claim that individuals make bad decisions is not evidence of identifiable errors of knowledge or reasoning. Some of it is merely evidence that people make decisions that, as judged by those authors (and probably by many of their readers) are imprudent or foolish. For example, in arguing for nudges against obesity-inducing diets, drinking and smoking, Thaler and Sunstein (2008, pp. 7, 44) simply report familiar statistics about the associated health risks and then conclude: ‘With respect to diet, smoking, and drinking, people’s current choices cannot reasonably be claimed to be the best means of promoting their well-being’ 

The same thing could be said of what these two nitwits were up to. Welfare Econ is just sloppy thinking by credentialized shitheads. Whatever discoveries it claims to make are two edged swords. There's nothing wrong with CBA and mechanism design. People can agree that policies based on number crunching can be quite useful. But this only true if those paid to do this boring work focus entirely on what William Blake called 'minute particulars' rather than 'the General Good.' 

No comments: