Suppose you (in Ken Arrow's words)
Assume that social choice, like individual choice, is expressed by a preference ordering over (social) alternatives, so that the alternative chosen from any given feasible set of alternatives is the most preferred one.
Then it can never be the case that you would change your mind on the basis of a conversation or of an observation. In other words, you may as well be blind to what other people are doing and deaf to anything they may happen to say. You don't actually exist within a Society. You may be some sort of robot which has no 'theory of mind'. You would be incapable of strategic behavior or countering strategic behavior by others. If all robots are like you, then as a class, robot behavior would be incapable of evolution or 'self-learning'.
Hence, one formulation of social choice is that of defining a social ordering governing social choice for each profile of orderings, one for each individual.
If an individual's survival and ability to reproduce does not depend on their 'profile of ordering' then this 'formulation of social choice' would be merely a description of a steady state. However, we would not be justified in speaking of what obtains as a Society. It is merely a collection of individuals.
In other words, social choice is defined by a mapping from profiles to social orderings.
But we don't actually have any such profiles. I don't have any preference about whether Mr. Jones of Acacia Avenue, Cardiff, or Mrs. Smith, of Wordsworth Close, Newcastle, should receive the last jar of Bovril. I don't know either of them. Anyway, I don't have- and don't want- the sort of information I'd need to make such a determination.
One can then state some desirable properties of this mapping and ask if any such mapping (social welfare function in Arrow's and Sen's terminology) exists;
But because we know the thing itself doesn't exist, we also know that it can't exist in a desirable manner. The fact is, most people have two arms and one head. But there is no way of squishing people together so that you have a human being with ten heads and twenty arms. It is pointless to debate whether such a person would be better or worse at Tennis.
if social welfare functions do exist, one can attempt to characterize them in some useful way.
The don't exist. Nothing useful is achieved by 'characterizing' what doesn't exist. Why not simply debate whether Spiderman can beat up Dracula?
Arrow's conditions can be stated roughly as follows: (U) the social welfare function is defined for all profiles; (P) if the profile is such that everyone prefers alternative x to alternative y, then the social ordering sets x above y (the Pareto principle);
Yet, in life, we are constantly discovering that we can make mutually beneficial trades. Indeed, that is one reason why we socialize.
(I) the choice between two alternatives depends only on the individual preferences between those two alternatives '
But this can never be the case for any important choice- i.e. one which involves an 'income effect' or 'hedging'. In any case, if almost perfect substitutes are available, then the thing scarcely matters.
The larger problem is that we have no reason to believe that what is observed represents either 'Choice' (it may be an accident) or 'Preference' (the thing may be stochastic)
It is foolish to think an essentially biological function- i.e. stuff needed to stay alive and support inclusive fitness- is actually a mathematical function- i.e. a rule that gives the value of a dependent variable that corresponds to specified values of one or more independent variables. If a mathematical function specifies a biological function the organism would be easy pickings for a predator or parasite. Choice of a type which can be captured by a mathematical function is unlikely to be robust on an uncertain fitness landscape. It would pose an existential threat to the organism or species in question. This is not to say that, at the macro-level, we might not find that a population behaves in accordance with a Hannan consistent multiplicative weighting update algorithm or something more arcane yet. However this would not correspond to anything fully decomposable. Moreover, it would involve co-evolved processes which tame complexity in ideographic ways. No doubt, there can be better and better simulations of this but they would be 'black boxes'. You couldn't 'open them up' to find out what was happening at the individual level.
(independence of irrelevant alternatives); and (ND) the social ordering does not always agree with the preferences of any single individual (non-dictator- ship).
Yet, we find that in our families and in the places where we work and in the institutions that regulate our Society, there is generally a single individual with whom the decision making 'buck stops'. More generally, Democracies have Chief Executives whom we trust to make the big decisions- e.g. whether or not to send in the SEALs to get Osama. It is foolish to describe a person who is not despotic or tyrannical as a 'Dictator'. Why not simply say 'all cats are dogs. Dogs say woof woof. Hence my cat should say woof woof. OMG, there is something wrong with my cat!'
Arrow then showed that there is no social welfare function satisfying all of these conditions.
Because there is no Social welfare function. If we were given the choice to implement it, we would refuse. But nobody is crazy enough to say that each of us should have a preference as to whether Mr. Jones or Mrs. Smith should get the last jar of Bovril.
This result has given rise to a large literature, to which Sen has been the outstanding contributor;
this literature has been a pure waste of time.
the best survey is Sen's book Collective Choice and Social Welfare [1 970].2 There are several directions to go for changing the negative result to a positive one; they all obviously require some change in the conditions imposed. One is to enrich the informational base, so that the social ordering depends not only on individual orderings but also on cardinal values and, more significantly, on some kind of interpersonal comparisons.
Mr. Smith's wife ran off with the milk-man. Mrs. Jones has recently taken up Yoga. Knowing this will make Social Choice so much better- thinks nobody at all.
A second is to restrict the range of profiles for which the social welfare function is defined, on the basis of some a priori assumptions as to the possible kinds of orderings individuals may have.
So, this worthless availability cascade says first 'assume something impossible and then assume all sorts of other stupid shit'.
Indeed, even before Arrow's Theorem, Black (1948) had in effect shown that making pairwise choices by majority voting determines a social ordering (obviously satisfying (P), (I), and (ND)) if the social alternatives can be thought of as arrayed along one dimension and if each individual's preference ordering over this array is single-peaked.
Why not simply assume that people want more 'transferable utility'- i.e. money- not less? Oh. Rich peeps get mugged. I don't coz my Rolex is painted on. Sad.
Sen has developed both of these research directions, but, more impor- tantly, he has introduced a new statement of the problem which brilliantly combines simplicity and depth in his discussion of the impossibility of a Paretian liberal [I, 1970a].
Which simply means that people should have an extensive realm of freedom where Social Choice can't intrude. However, they could use this freedom to trade with each other or form relationships of various types such that Pareto improvements (i.e. a situation where no one is made worse off when one person becomes better off) occur all over the place.
As will be discussed in Section IV, Sen has objected to the view that social choice should be completely determined by individual welfare comparisons.
If anybody tried to interfere in our lives on the basis of 'individual welfare comparisons' we would be greatly annoyed. Thus if someone suggests to me that I should suck off homeless dudes because the repugnance I feel would be less than the pleasure I give, I will beat that person with my hockey stick.
In particular, this broader perspective has drawn him into studying the conception of rights.
But the study of rights begins and ends with Hohfeldian analysis of its incidents. A right is only a right if there is an incentive compatible remedy under a bond of law.
Each individual is taken to have the right to make certain choices among social states. To take the least controversial examples, the choice between two social states that differ only in what one individual reads or in the color of the paint in the interior of an individual's house should be made by that individual.
No. We must stop people reading child porn. We should ban lead based paints which appear more beautiful. Rights are always defeasible because everything in law is defeasible.
Sen then showed very simply that, even without requiring independence (I), the conditions (U), (P) and the "minimal liberalism" requirement, that for each individual there should be at least two social states between which that individual's choice is decisive, lead to a contradiction.
This is not the case. Sen thinks that 'lewd' really wants 'prude' to read Lady Chatterly's lover. Actually, lewd only wants prude to become more broad minded. Sen has falsely attributed a preference to 'lewd'. It is easy to show 'contradictions' arise if you tell stupid lies. Thus you might say 'the fact that you are married to a woman contradicts the fact that you spend all your time sucking off homeless dudes due to you are a big fat homo'. I reply, 'I'm not beating you with my hockey stick at all. The fact that some red liquid is leaking from your skull contradicts your assertion that you want me to stop tenderly caressing you with my hockey stick. The fact is your cup of joy runneth over with the red red wine of bliss.;
This is very surprising;
if you are a cretin.
both the Pareto judgment and the idea that each individual has some private domain of choice, even if others would make different choices over that domain, are hard to deny;
The problem here is that lewd isn't really made worse off because prude won't read porn. Lewd's actual preference has nothing to do with what prude reads or doesn't reads. Lewd would simply like prude to be more broad minded.
and independence, which on the whole is central to most variations of the Impossibility Theorem, is not assumed here. The paradox arises because "nosy" preferences of others about choices that are in an individual's domain of private choice enter into the Pareto judgment.
But only because Sen makes a very stupid assumption-viz that lewd wants prude to read a particular book whereas the guy just wants prude to be more broad minded.
The result is not only surprising analytically but also addresses profound ethical questions on the relation between even the vestigial remnant of utilitarianism contained in the Pareto principle
Bentham didn't actually go around giving beejays to homeless dudes. Nor did Pareto. Sen and Arrow may have done so.
and the existence of individual "rights," a scope (however small) over which the individual has complete control. Sen's work has sparked both a technical literature, e.g., Gibbard (1974) and Suzumura (1978), and contributed to philosophical emphasis on rights from very diverse critical viewpoints, e.g., Nozick (1974) and Dworkin (1978). A very different line of analysis pursued by Sen has concerned the possibility of resolving the social choice problem by assuming that the range of individual preferences which need to be aggregated is narrowed from the universal range postulated in condition (U). In particular, he has stated conditions on profiles under which pairwise majority voting leads to an ordering.
But those 'profiles' would still stipulate against any such voting because the thing is silly and a waste of resources.
Inada (1964) had found two conditions other than Black's single- peakedness of preferences, and Ward (1965) had generalized Black's condition.
They wasted their own and their students' time.
Sen [I, 1966] introduced a condition (extremal value restriction) which included all previous ones: in any triple of alternatives, there are one alternative and one value (best, middle, or worst) such that no one ranks that alternative at that value. Later, Sen and P. Pattanaik [I, 1969) found necessary and sufficient conditions whereby majority voting yields a well- defined first choice.
But people would prefer some mechanism other than majority voting. That's why no country- however democratic- gives votes to everybody living in the realm. There are always qualifications of various sorts.
Sen thus both made a major early contribution to this particular subfield and collaborated in establishing the definitive results.
But this subfield was useless. The fact is, once you use the axiom of choice or some other 'maximal principle', you are adding a 'non-measurable' set to the underlying configuration space. If we are speaking of improved utility, or welfare, or 'democratic' responsiveness to what voters want, the thing we are adding has no usefulness whatsoever. It is just an illusion- like the 'Banach Tarski' paradox that a pea could be reassembled such that it is bigger than the Sun.
More representative of Sen's general position is his systematization of the information constraints on social choice. As he observes in his article "Interpersonal Aggregation and Partial Comparability" [I, 1970b], Arrow's assumption of interpersonally incomparable ordinalism is an extreme case. Interpersonally incomparable cardinalism does not get us any further in avoiding impossibility results [1970, Theorem 8*2], a result greatly deepened by Kalai and Schmeidler (1977).
A wholly meaningless result. Any ordinal ordering can be turned into a cardinal ordering private to that individual. The problem is that such orderings don't exist for the set of Social States. We don't have a preference, and don't want to have a preference, whether Mr. Jones or Mrs. Smith should get the last jar of Bovril.
In effect, the assumption that individual judgments, ordinal or cardinal, cannot be compared implies that social judgments must be invariant under a wide range of independent transformations of individual preferences (all monotone transformations in the ordinal case, all affine transformations in the cardinal case).
Social judgments should not exist unless some great benefit is attained or some great danger is averted.
For any degree of interpersonal comparability (e.g., that the ratio of unit utility changes for two individuals is judged to lie between two prescribed limits), the invariance requirements on the social welfare function are correspondingly lightened, and the possibility of finding an acceptable social choice procedure increased.
But what works even better is talking and 'transferring utility'- i.e. bribing- and doing 'mechanism design' and coalition formation and so forth. Consider the very different trajectories of Sen and Prashant Kishor. The former hasn't affected Social Choice in India. The latter has used Statistical skills and knowledge of Indian conditions to emerge as a 'king-maker' in Indian politics.
Has Sen done anything whatsoever that has been useful? It had long been established in law and fiscal policy that those with disabilities need greater compensation- that's why higher damages are paid to someone who suffers a life-changing injury and also why Governments give higher tax-allowances to the seriously disabled or else provide non-means tested Disability Living Allowances etc.
On the other hand scarcely anyone has heard of Sen's 'Weak Equity axiom'
A last example of Sen's seminal role in the area of formal theories of social choice is his formulation of the "Weak Equity Axiom" in his lectures [1973]: an individual getting less utility out of any given income than another should get a higher income (incentive effects aside). That is, disabilities should be compensated for. This axiom is in general inconsistent with utilitarianism and has revived interest in equality as a criterion independent of diminishing marginal utility
This axiom is silly. Sen forgot about disutility. An able-bodied person is assumed to only get disutility while at work. Income has to be high enough, at the margin, to compensate for disutility- which is why there is Income inequality even if endowments are equal. A disabled person gets some disutility even during their leisure. Compensating for this makes sense. It is simply a matter of collective insurance. Who knows when we too may be struck down in like manner? Thus tax payers are ready to fund not just unemployment insurance- because we may lose our job- but a non-means tested Disability Living Allowance- because we may become unemployable by reason of an injury or illness and will experience disutility even in our leisure hours.
Apart from disutility, mathematical economists forgot that, because of Knightian Uncertainty, 'utility maximization' is not rational. 'Regret minimization' is. That's why able-bodied people are willing to pay for Disability benefits- they will regret not having this facility if they themselves are struck down.
Aristotle pointed out more than 2000 years ago that the fault of 'akrebia' in rhetoric consists in seeking for a greater precision than the subject matter affords. Arrow's big mistake was to treat 'Tarskian primitives' as if they could be defined with great precision. The result was an explosion of nonsense. This is the precise opposite of what the Church calls 'economia'- the mysterious management of all things such that the katechon keeps at bay the eschaton- i.e. the world wags on and the Apocalypse is delayed.
Arrow (1959) observed that if choice from finite sets (rather than budget sets) was assumed to be defined,
i.e if Tarski's undefinability theorem were assumed not to apply then true 'choice' could be defined within a first order language. But if so, just by letting the finite set be 'Peano arithmetic', there would be a way to show Godel and Turing and Tarski were silly billies. Indeed, all those smarty pants Mathematicians and Physicists and Medical Researchers were wasting their time. Just let a bunch of people chose the true theory out of the finite set constituting the first order characterization of a given discipline and, immediately, everything knowable would be known. Moreover there would be no need for Language or any type of coordination mechanism. We'd wake up in the morning knowing exactly what to do at any given moment. So would everybody else. There would be no need for words or money to change hands. Arrow had discovered a way to get rid of Language and Economics and Medicine and Science and Politics and Warfare and Diplomacy and Jurisprudence. What an amazing achievement! No wonder he got a Nobel Prize in the same year that the Swedes decided not to award anybody the Peace Prize.
the equivalence between choice functions defined by maximizing a given ordering and choices satisfying certain consistency conditions took a somewhat different form. This line of analysis invites taking apart the conditions on the choice function (the function defining the choices or choices made from any given opportunity set) and showing the equivalence of each to some form of rationality weaker than an ordering.
So, first you put in the axiom of choice or some such 'maximal principle' and then you find what you yourself put in. This has nothing to do with the actual conditions affecting the choices people make. It is merely an exercise in mental masturbation which looks a little 'mathsy' but isn't really. Given the axiom of choice, it is perfectly possible- if Choice rules can themselves be chosen- that there is some 'slingshot' choice such that all other choices are entailed. Equivalently, it could be shown that no Choice is possible save by access to an Oracle. All that obtains is a correlation between actions and outcomes.
This study has been pursued by several writers, in particular Uzawa (1956), Herzberger (1973), and Sen. Sen's article, "Choice Functions and Revealed Preferences" [IV, 1971], is the epitome of this work, summing up and extending all previous studies. There are many results; the following is a typical example. Start with a choice function, C(S), mapping each finite opportunity set S into the subset of chosen elements.
So, you are assuming some version of the axiom of choice. If 'finite opportunity sets' are interdependent you need the axiom of global choice- i.e. one can simultaneously chose an element from each non-empty set. This means either
1) there is an algorithmic solution to Concurrency problems- i.e Djikstra is wrong
2) P equals NP- i.e. the solution of any problem is as fast as verification.
But this means there is always some 'democratic' procedure which can answer all our questions. We don't need no steenkin' STEM subject researchers. Just get a big enough Condorcet Jury and they will quickly choose the Theory of Everything which exists in the Mind of God.
Define, "revealed preference," R, to mean: (xRy) if and only if, for some S, x belongs to C(S) and y belongs to S.
But this could be an intuitionistic maths- i.e. no excluded middle. How do we know xRy might not entail yRx?
State the Weak Congruence Axiom (WCA): If xRy, then for any S such that y belongs to C(S) and x belongs to S, x must also belong to C(S).
This means that Revealed Preference is an ordering. Yet, in practice we may 'mix things up' or try new stuff just for the hell of it. Indeed, if we kept abreast of mathematics, we'd know that we ought to be doing this some of the time. It would be irrational to take our habits- which is all that revealed preferences are- as optimal.
Also state properties (a) and (/f): (a): If S is a subset of T and x belongs to both S and C(T), then x belongs to C(S).
i.e. if a thing is chosen out of a bigger set of options it would also be chosen out of a smaller set of options.
(/3): If S is a subset of T and x and y any two elements of C(S), then either both belong to C(T) or neither does.
I don't know what this means.
Then the following theorem holds: Theorem. A choice function satisfies WCA if and only if it satisfies both property (a) and property (3).
But nobody would choose to have such a function.
Note: Sen also shows that WCA is equivalent to the statements that R is an ordering and that C*(S) = C(S), where C*(S) is the set of elements of S which are maximal with respect to R.
Arrow says this is a typical 'result' of Sen's. But it is meaningless. A choice function satisfies WCA if everything that can be known is known and, moreover, there is no problem of concurrency, complexity or computability. In other words, we can judge things in the manner of the immortal, omniscient, Gods. Only stupid people bother with STEM subjects and quantum computing and stuff like that. There is no need for Science or Language or Politics or Economics. What a wonderful contribution! Why not simply sit in a cave in the Himalayas saying you have achieved 'kevalya gyan'?
What are the aims of 'Social Policy'? It is to make Society better and people happier and more secure.
Sen, not surprisingly, believes otherwise-
Sen has developed a fundamental critique of this doctrine of "welfarism," as he calls it. He has argued for distinguishing aggregation of judgments of different people about social policy from aggregation of interests.
If people are irrational and make stupid judgments, then we shouldn't be 'aggregating judgments'. Nor should we be 'aggregating interests' because we have no magic ball which allows us to see what is truly in someone's interest. You may think it involves everybody sucking off homeless dudes. Others may disagree.
The fact is, no Society anywhere 'aggregates judgments' or 'interests'. Democratic countries elect representatives who engage in a lot of horse-trading. Sometimes this improves matters. Sometimes it does not. Still, it is the least bad system we know of- at least under current conditions.
For the former, only information about individual preferences is avail- able,
Nonsense! We may be able to get some information about certain preferences- but it is not reliable. This is why the results of elections often come as a shock to pundits. Few analysts thought Brexit or Trump's Presidency were at all likely.
and the Impossibility Theorem is relevant.
No it isn't. It is stupid.
For aggregation of interests, on the other hand, there is additional information beyond any measure of individual welfares, in particular, measurement of inequality in objective terms.
Rubbish! We don't know about disutility or 'global opportunity cost'. Redistributive policies have failed again and again because elasticities of supply increase very quickly and unexpectedly in some 'mission-critical' areas but not others. Indians know this very well.
Also, the forms of consumption that give rise to pleasures might be morally relevant for given utility levels (e.g., pleasures arising from sadism and masochism).
But what Arrow and Sen are doing is itself repugnant because they are wasting valuable resources on talking worthless bollocks.
Even the Pareto principle can be questioned along these lines,
No it can't. There's always a way of showing that a claimed injury to a third party is nothing of the sort.
and the paradox of the Pareto liberal (see Section II) shows that there can be principles which we regard as overriding the Pareto principle.
This is the 'nuisance' principle. People who pretend to be gravely injured by the fact that you aren't giving beejays to homeless dudes are merely virtue signaling. Using 'transferable utility'- i.e. paying people to do or stop doing stuff is the Coasian solution. Enforcing laws against repugnancy markets is the other way to go.
This line of argument has been developed further by Sen [1985, 1987, and 1992]. A given set of commodities may be utilized in different ways by a consumer, though the range of possible utilization modes may be restricted by an individual's personal limitations.
i.e. their disutility. Forgetting disutility means that you end up with an easily refutable critique of what is merely a 'straw-man'.
Utility in the usual sense may be identified with happiness or fulfillment of desires. But neither the valuation to be placed on the chosen utilization nor the choice made by the individual need be related to the utility nor indeed to well-being in any sense. Further, judgment of a given state of affairs may rationally depend, not merely on the alternative utilization chosen but on the range of alternatives available to the individual, including those not chosen. There is a preference for capability (or freedom of choice); poverty and disability are infringements on capability.
We can define anything which prevents us enjoying more utility as a type of disutility we experience. Thus, my disutility from moving my fat arse is the reason I haven't used the very expensive exercise bike I bought when the lockdown started. My great stupidity, laziness and horrible character traits means that I get much more disutility from learning to live a better life. It may be that people like me should be sent to 'fat camp' or made to undergo compulsory training in being nice. I suppose, if I break the law or become a source of public nuisance, then the Courts will mandate something of the sort.
By contrast, misfortunes which may easily befall good, hardworking, people- or which may befall those they are related to or whom they care about- then there is likely to be collective provision for the poor, the disabled, etc, etc. But this happened without any stupid, Sen-tentious, cretins getting involved.
On the other hand, it is interesting that these cretins began bleating about inequality at precisely the time that Democratic countries around the globe began giving up on this policy objective. Harold Wilson's first administration reduced inequality- but lost the elections. Labor realized that 'fiscal drag' was putting more and more upper working class people into the Income tax net. They rebelled- as they did against Scandi 'solidarity wages'. Dagenham Man turned Tory. Sen, ludicrously, would say that Britain under Thatcher might face a serious famine.
V. Measurement of Inequality and Poverty Sen's general concerns about the meaning of social policy and individual variations in the capacity for functioning have found applications in developing appropriate measures of income inequality. His paper with Partha Dasgupta and David Starrett [III, 1973b; see also III, 1978] introduced new criteria for measures of income inequality, essentially formalizing the "transfer principle" of H. Dalton, that a transfer from a rich individual to a poor one must be regarded as a reduction in inequality.
But everyone in the Seventies could see that the opposite was the case. Rich guys fuck off and take their high value adding skills with them. This does not reduce inequality. It increases non-market inequality which can't be the subject of redistributive taxation. The middle class find ways of capturing the benefits of the Welfare State- e.g use their superior knowledge to benefit from the Health and Higher Education system.
Mathematically, this leads to the implication that the function expressing inequality in terms of individual incomes must be S-concave, a weaker condition than concavity or even quasi-concavity. From this, they show that in order for one income distribution to be better than another for any S-concave inequality measure, it is necessary and sufficient that the Lorenz curve for the second lie entirely below that of the first (or, equivalently, that the first be obtainable from the second by a sequence of transfers from the rich to the poor).
This ignores incentive effects. The problem here is that the substitution effect swamps the income effect. When I was young, this was an argument for Stalinism. Shoot people who run away or who substitute leisure for work. Then I visited Russia and China and saw there was greater inequality there than in the West. The nomenklatura could live large because they had access to the foreign currency shops.
The importance of this paper lies not merely in this characterization of the strongest statements that can be made in the absence of more definite criteria, but also in its improved tests for inequality measures.
The importance of this paper lay in its showing that the Left was living in a fantasy world. Young people who studied under these nutters in the Seventies and Eighties kept their mouths shut so as to get a good degree but they voted for Thatcher and Reagan.
This paper generalizes the earlier work of Atkinson (1970).
Atkinson shat the bed by backing wage controls. I hadn't realized he was actually quite young- younger than Sen. I remember him as very old. I suppose people at the LSE considered him a relic from a past age. The notion that Wilson or Tony Benn understood technology and could make better decisions re. things like silicon chip manufacture was already laughable. Atkinson influenced Piketty who has made money out of writing shite. But David Icke has made even more money. In the end, inequality is reduced if we buy books by lunatics not highly credentialized Professors.
However, Sen has objected that Atkinson's measures, which are additively separable in individual incomes, tend not to describe inequality very well in certain circumstances.
They don't describe shit. Income is defined as what you can spend without reducing future Income. Redistributive taxes reduce the incomes of the poor if, by next year, the rich have run away. Welcome to Venezuela!
Thus, the Gini coefficient, which is not separable in individual incomes, nevertheless satisfies a number of interesting conditions.
But it could show that Chavez was a fucking genius and poor Venezuelans would soon be better off than the average Norwegian!
Closely related to measures of income inequality are measures of poverty, which have played such a large role in evaluations of public economic policies, national and international. Again, Sen [III, 1976b] has contributed importantly through an axiomatic characterization of poverty measures which take into account inequality among the poor as well as the proportion below a given level.
This was utterly useless. China and India and so forth began lifting millions out of absolute poverty when they simply copied what worked in other formerly equally poor countries.
As an approximation for a large population, he derives the poverty index, P= H[I+(I -I)G], where H is the proportion in poverty, I the ratio of total income shortfalls from the poverty level to total income, and G is the Gini coefficient of income distribution among the poor.
In India, only 15 million people pay Income tax- though the threshold is half that in the UK. For all practical purposes one could simply say H is 100 percent. On the other hand, poor people, through expenditure taxes, do have to subsidize richer, higher caste, farmers etc.
VI. Empirical Studies of Distribution and Its Consequences The normative measures of distribution and Sen's emphasis on the function- ing of individuals as the criterion for social policy have led to important studies of the interaction between income distribution and severe incapa- cities, including death. This is most strikingly brought out in his remarkable studies of famine [1981 and 1989]. His careful empirical analyses of four famines bring out clearly that neither variation in food supply due to natural causes nor physical obstacles to food distribution played significant roles in these famines. Rather, they were due to shifts in the real income distribution.
Sen assumed there was plenty of food. Why weren't some people eating it? His answer was that other people decided to eat twice as much food as before so as to have the pleasure of watching their neighbors starve. The truth is there was a 'food availability deficit' in each and every famine. But then Sen was so stupid, he thought Thatcher would preside over a big famine in the UK
Did Sen ever write anything not false, foolish or mischievous?
Arrow does not mention Sen's recent discovery that 'Public Reason' must solicit the views of Mrs. Khan from Quetta who believes all wives- not just that of Mr Jones from Cardiff- should run off with the milkman and also Bovril is yucky. However, equal respect must be shown to the views of Patti Obaweyo Golem, of Kansas City, who believes Mrs. Jones is like 'culturally appropriating' Yoga- which belongs to the Red Indians, right?- and what is Bovril anyway?
Arrow mentions Sen's stupidity re.
the meaning of rationality in individual choice. Perhaps no single paper of Sen's has been more cited than his "Rational Fools" [IV, 1977],
where the fool thinks a rational person will entrust an envelope containing money to some random stranger
in which he shows how limited the ordinary concept of rationality is in covering the range of motives in individual choice; in particular, the role of commitment is completely ignored.
Fuck commitment. Rationality requires 'regret minimization' in an uncertain world. Don't entrust money to some stranger. Make sure any 'mission critical' activity is either supervised by you, personally, or delegated to a professional who has plenty of indemnity insurance.
More recently, in "Maximization and the Act of Choice" [IV, 1997], he has shown that concepts of maximization are not necessarily connected with the implication of an ordering of alternatives.
He is merely saying that maximal elements may not be at the 'Pareto front'. I can rank job offers on the basis of salary. But I'd be a fool to decide things on the basis of the maximal on this ordering. True, the salary offered by 'Beejays for the Homeless' is much higher. But, a quick Google Search will reveal it isn't a genuine business. You are trying to trick me into giving beejays to homeless dudes for a month. Instead of paying me the promised salary, you will go around telling everybody I'm a big fat homo with an insatiable appetite for homeless dudes' jizz- just as you had always claimed.
Another continuing interest has been the achievement of efficiency in economic development.
Hilarious! Everybody knows that imitating what works in smart, similarly placed, countries is the way to go. Once you are doing smart stuff, by all means use O.R or whatever to improve efficiency.
While the general theoretical thrust has been less novel than in his work on social welfare, his contributions have been of a very high order, especially since they frequently flew in the face of then current political dogmas accepted by many economists.
But those economists were considered to be utter fools. Nobody listened to them.
This work com- menced with his early book on choosing among techniques of production [1960a] and continued with many papers on the evaluation of projects culminating in the handbook on project evaluation for the United Nations International Development Organization, written jointly with Partha Das- gupta and Stephen A. Marglin [1972].
The thing was useless. The end of the Cold War meant that the West pulled the plug on this useless bunch of tossers. Facing the loss of their well paid jobs, the UNIDO turned into a service provider. In other words, it left 'project evaluation' to the guys paying for the project. It is currently headed by a Chinese Communist.
Two papers on the rate of discount appropriate to government investment [XI, 1967 and 1982] have stressed the role of externalities in the form of public concern for the future. Closely related has been his contribution to the theory of optimal savings [IX, 1961 a, 1967, and 1975].
That theory was shit. The fact is poor countries are going to have sub-optimal savings and mobilization of savings which in turn means low r.o.i. This has to be tackled in an ideographic manner.
In a very different vein, Sen has addressed the idea that the value systems of India and other Asian countries are so different from that of the West that the polity and economy cannot be organized along Western lines, that democracy, for example, is inappropriate in South and East Asia.
The cretin doesn't get that some countries can only be run democratically because no party is strong enough to kill all its opponents.
His articles [XIY 1993a, 1993b, and 1996] have shown that a deeper knowledge of the Indian and Asiatic pasts would reveal far less uniformity and far more strands of rational analysis and democratic thinking than is asserted by self- interested participants in the debate.
Sadly, the cretin knows shit about Indian, let alone any other sort of History.
One could point to still other accomplishments, especially in several differ- ent fields of philosophy, but the area where Sen's contributions have been truly unique is his extraordinary synthesis of economic and philosophical reasoning on the bases for social policy. No one has combined different approaches, formal analysis, conceptual clarification, theory of measurement and empirical work as has Sen.
Arrow was writing this 20 years ago. Sen had just got a Nobel. He might have used his celebrity status to give good advise to India. He didn't. It was Manmohan whom History will remember- why? Singh worked hard. He may have talked bollocks but nobody could understand his mumbling- so we gave him the benefit of the doubt.
We now agree with Narendra Modi that 'hard-work is better than Harvard'. Prashant Kishor, who worked for the WHO in Chad doing (I imagine) survey or statistical work, used these skills to completely alter Indian politics. He did this through hard work and the use of simple, pragmatic, structural causal models. Meanwhile Sen-tentious Political Philosophy is considered to be as imbecilic as Spivak's 'Post Colonial Reason' or Subaltern shite.
No comments:
Post a Comment