Amartya K.

Sen's Contributions to the Study of Social Welfare

Author(s): Kenneth J. Arrow
Source: The Scandinavian Journal of Economics, Vol. 101, No. 2 (Jun., 1999), pp. 163-172
Published by: Wiley on behalf of The Scandinavian Journal of Economics
Stable URL: https://www.jstor.org/stable/3440690
Accessed: 24-08-2018 13:00 UTC

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/3440690?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

The Scandinavian Journal of Economics, Wiley are collaborating with JSTOR to digitize,
preserve and extend access to The Scandinavian Journal of Economics

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
Scand. J of Economics 101(2), 163-172, 1999

Amartya K. Sen's Contributions to the

Study of Social Welfare

Kenneth J Arrow
Stanford University, Stanford, CA 94305-6072, USA

I. Introduction

The oeuvre of Amartya K. Sen is very considerable in magnitude, nearly

twenty books (with two forthcoming) and about 250 articles, mostly in
learned journals but also in the higher level of the popular press, and
lectures. It has nevertheless a very considerable degree of unity. It is
motivated throughout by concern for the welfare of the individuals in an
economy, with special interest in the lower part of the income spectrum;
starvation, poverty, and economic development have been the objects of
fruitful, important, and socially valuable study and have motivated his basic
research into the concepts of individual and social welfare.
All of this work is of very high quality and makes important contributions
to our understanding and to the formation of policy. But I will isolate for
special consideration that large subset which revolves about the conceptual
question: what is meant by, "better off socially"? A major portion of his
work falls directly into this category. The value of Sen's analysis of social
welfare and the conceptual questions relevant to it have now been recognized
by the award of the Nobel Memorial Prize in Economic Science.
Unlike some other work which has been awarded this Prize, especially in
economic theory, we cannot do justice to Sen's work in social welfare on the
basis of one or two seminal papers, although there have been several such, as
I will point out. Rather it is the work as a whole and the way the various parts
interplay that must be understood to see the importance of Sen's contribu-
tion. His exploration of the notions of social welfare takes place at every
level of analysis, formal-mathematical, conceptual, and empirical. It is by
far the most comprehensive study of its kind, drawing on profound under-
standing of both economics and moral philosophy. It is useful to note that he
is regarded by philosophers as one of their leading lights, and he has
published extensively in philosophical journals.
In view of the comprehensive and far-flung nature of his work, it is not
surprising that his role has been somewhat different in various specific
aspects. In some, indeed, he has opened and defined new fields with strong
results. In others, he has perfected and enriched earlier results or contributed
along with others. In still others, he has raised new inquiries whose chief
value has been the stimulation to others. What is striking is the consistency

?D The editors of the Scandinavian Journal of Economics 1999. Published by Blackwell Publishers, 108 Cowley Road,
Oxford OX4 1N, UK and 350 Main Street, Malden, MA 02148, USA.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
 164 K.J Arrow

of his thought, so that even the most formal work is motivated by philosophi-
cal and ethical considerations, while his most interpretive and broadly
ranging papers remain compatible with rigorous technical analysis.
Because of the wide range of his work on social welfare, I will discuss it
in more detail under five headings: (1) formal theories of social choice; (2)
the formal theory of individual preference and choice; (3) the conceptual
meanings of choice, welfare, and utility; (4) the measurement of social
welfare as reflected in inequality, poverty, and real social income; and (5)
empirical analyses of famine and nutrition related to Sen's work on poverty
and its moral implications. In Section VII, I briefly review a sample of his
work in fields other than the analysis of social welfare.
One general characteristic of Sen's work must be stressed. He is especially
concerned with the distribution of welfare however that concept is under-
stood. Averages for him conceal information relevant to the formulation of
sound ethical judgments. He has therefore emphasized the measurement and
meaning of inequality and in particular poverty as a morally and economic-
ally special category.

II. Formal Theory of Social Choice

A major part of Sen's work has dealt with the formal theory of social choic
The basic paradigm in this field has been that of Arrow (1951).1 It is in t
tradition of welfare economics, which has sought to provide a rational
justification for choice among alternative possible economic policies based
on the preferences of individual members of society. 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. 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. In other words, social
choice is defined by a mapping from profiles to social orderings. 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;
if social welfare functions do exist, one can attempt to characterize them in
some useful way.
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); (I) the choice between two alternatives
depends only on the individual preferences between those two alternatives

' Dates in parentheses refer to references listed at the end of this article.

? The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
 Amartya K Sen 's contributions to the study of social welfare 165

(independence of irrelevant alternatives); and (ND) the social ordering does

not always agree with the preferences of any single individual (non-dictator-
ship). Arrow then showed that there is no social welfare function satisfying
all of these conditions.
This result has given rise to a large literature, to which Sen has been the
outstanding contributor; 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.
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. 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.
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]. As will be discussed in Section IV, Sen has
objected to the view that social choice should be completely determined by
individual welfare comparisons. In particular, this broader perspective has
drawn him into studying the conception of rights. 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. 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 very surprising; 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; 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

2Dates in brackets refer to publications by Sen listed in the Bibliography. Those preceded
Roman numeral refer to the subject headings of articles written by Sen; those without a Roman
numeral refer to his books.

?) The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
 166 K.J Arrow

individual's domain of private choice enter into the Pareto judgment. 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 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. Inada (1964) had found two conditions other than Black's single-
peakedness of preferences, and Ward (1965) had generalized Black's condi-
tion. 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. Sen thus both made a major early contribution to this
particular subfield and collaborated in establishing the definitive results.
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 deep-
ened by Kalai and Schmeidler (1977). 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). 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 correspond-
ingly lightened, and the possibility of finding an acceptable social choice
procedure increased. For a survey of this point of view in subsequent
research, see Sen [I, 1977b].
The research stemming from Sen's argument has indeed been rich. This is
one of the areas where Sen's contribution has been the formulation of the
question and the inspiration of others. It is, of course, well known that
complete interpersonal comparability of unit utility differences permits -

? The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
 Amartya K Sen k contributions to the study of social welfare 167

and makes meaningful - the sum-of-utilities criterion as satisfying conditions

(P), (ND), and (I) (the last suitably reinterpreted to admit the additional
information). Following Sen's stimulus, Hammond (1976) and Strasnick
(1976) formulated a different informational basis. They admitted interperso-
nal ordinal comparisons, e.g., that individual i in state x is better off than
individual j in state y. This assumption implies that social welfare judgments
be invariant under the same monotone transformation of all individual
preferences.
Hammond and Strasnick independently showed that under this informa-
tional constraint, the "leximin" and "leximax" principles satisfy all the
social choice conditions above. They are in fact the only possible principles
satisfying these conditions and some mild additional ones. By "leximin" is
meant the following: for each of the social alternatives, x and y, rank the
individuals in increasing order of utility (this operation is meaningful under
the informational constraint). If the worst-off under x is better off than the
worst-off under y, choose x (or vice versa). If the two worst-off individuals
are equally well off, then compare the two second-worst-off individuals and
so forth.
The leximin principle is reminiscent of Rawls's "difference principle"
(1971), but it must be acknowledged that Rawls gives it a different
interpretation. To Rawls, the comparison is in terms not of well-being bu
"primary goods," those that give rise to the possibility of well-being. Sen
has developed this theme further (see Section IV below).
Sen has contributed to the technical development, as well as setting the
terms of the argument. He notes that the leximin principle is not very
persuasive by itself for a large society. It is, however, very reasonable if ther
are only two individuals. But, in conjunction with conditions (U) and (I),
leximin for two-member societies entails leximin for any society; see Sen [I,
1977b, Section 6].
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.

III. Formal Theories of Individual Choice and Rationality

Reflection on social choice provided a new stimulus for studying individual

choice and the meaning of rationality. Gossen, Jevons, Walras, Menger, and
Edgeworth had introduced the hypothesis that consumer choice of commodity
bundles was determined by maximization of a utility function subject to a

?) The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
168 K. Arrow

budget constraint. Irving Fisher and Vilfredo Pareto observed that

nature of the utility function was irrelevant; only an ordering of
bundles need be postulated. Samuelson (1938) introduced a change in
orientation in his concept of revealed preference. The choices made, not an
underlying ordering or utility function, were primary. Conditions of ration-
ality were imposed on the choices. As shown later by Ville (1946) and
Houthakker (1950), suitably strong conditions on the demand functions
implied the existence of an underlying ordering which rationalized choice.
Arrow (1959) observed that if choice from finite sets (rather than budget
sets) was assumed to be defined, 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. 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. Define, "revealed
preference," R, to mean:

(xRy) if and only if, for some S, x belongs to C(S) and y belongs to S.
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).
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).
(/3): If S is a subset of T and x and y any two elements of C(S), then eith
both belong to C(T) or neither does.

Then the following theorem holds:

Theorem. A choice function satisfies WCA if and only if it satisfies both

property (a) and property (3).

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.

?) The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
 Amartya K. Sen s contributions to the study of social welfare 169

IV. Clarification of the Aims of Social Policy

Social choice theory and utilitarianism seek to base social judgments on the
welfares of the individual members of society; in different variants, differing
assumptions are made about the properties of welfare measurements (ordinal,
cardinal, interpersonally comparable cardinal, and so forth). In some articles,
e.g., [II, 1979], 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 inter-
ests. For the former, only information about individual preferences is avail-
able, and the Impossibility Theorem is relevant. 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. 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). Even the Pareto principle can be questioned along these
lines, 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 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. 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 infringe-
ments on capability.

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 develop-
ing 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. 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

?) The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
170 K J Arrow

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). 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.
This paper generalizes the earlier work of Atkinson (1970). 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. Thus, the Gini coefficient, which is not separable in indivi-
dual incomes, nevertheless satisfies a number of interesting conditions.
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. 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.
He has pursued this analysis into the correction of national income
comparisons for distributional reasons [III, 1976a]. Again an axiomatic
characterization leads to the use of the Gini coefficient to modify the usual
measures of per capita national income, specifically to measure distribution-
ally-adjusted per capita national income as e (1 - G), valid for large
economies, where e is mean income and G is the Gini coefficient. This is
actually a special case of a much broader set of results.

VI. Empirical Studies of Distribution and Its Consequences

The normative measures of distribution and Sen's emphasis on the func

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,

?) The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
 Amartya K Sen 's contributions to the study of social welfare 171

particularly a lowering of incomes of the very poor; it was inability to afford

the purchase of food that caused hunger and famine on massive scales.
He has also studied the distribution of food within the family, particularly
under conditions of stress, and has shown that there is strong evidence of
discrimination against female family members in some parts of the world
[VIII passim]. In particular, demographic analysis indicates the higher
mortality of women at younger age levels.

VII. Other Areas of Study

Sen has written on many other topics than social welfare, broad as that
is. To judge by the references, this work has had very important effects. I
have followed only part of this work; in this section, I make very brief
references to the questions that I have followed.
One is the meaning of rationality in individual choice. Perhaps no single
paper of Sen's has been more cited than his "Rational Fools" [IV, 1977], 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 commit-
ment is completely ignored. 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.
Another continuing interest has been the achievement of efficiency in
economic development. 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. 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].
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].
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. 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.

? The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms
172 K.J Arrow

VIII. Summary

One could point to still other accomplishments, especially in several

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.

References

Arrow, K. J. (1951), Social Choice and Individual Values, Wiley, New York.
Arrow, K. J. (1959), Rational Choice Functions and Orderings, Economica, N.S. 26,
121-127.
Atkinson, A.B. (1970), On the Measurement of Inequality, Journal of Economic Theory
2, 244-263.
Black, D. (1948), On the Rationale of Group Decision-making, Journal of Political
Economy 56, 23-34.
Dworkin, R. (1978), Taking Rights Seriously, Duckworth, London.
Gibbard, A. (1974), A Pareto-consistent Libertarian Claim, Journal of Economic Theory
7, 338-410.
Hammond, P. (1976), Equity, Arrow's Conditions, and Rawls' Difference Principle,
Econometrica 44, 793-804.
Herzberger, H. (1973), Ordinal Preferences and Rational Choice, Econometrica 41,
187-238.
Houthakker, H. S. (1950), Revealed Preference and the Utility Function, Econometrica
17, 159-174.
Inada, K.-I. (1964), A Note on the Simple Majority Decision Rule, Econometrica 32,
490-506.
Kalai, E. and Schmeidler, D. (1977), Aggregation Procedure for Cardinal Preferences: A
Formulation and Proof of Samuelson's Impossibility Conjecture, Econometrica 44,
793-804.
Nozick, R. (1974), Anarchy, State, and Utopia, Blackwell, Oxford.
Rawls, J. (1971), A Theory ofJustice, Harvard University Press, Cambridge, MA.
Samuelson, P. A. (1938), A Note on the Pure Theory of Consumer's Behavior, Econom-
ica, N.S. 5, 61-71.
Strasnick, S. (1976), Social Choice Theory and the Derivation of Rawls' Difference
Principle, Journal of Philosophy 73, 85-99.
Suzumura, K. (1978), On the Consistency of Libertarian Claims, Review of Economic
Studies 45, 329-342.
Uzawa, H. (1956), A Note on Preferences and Axioms of Choice, Annals of the Institu
of Statistical Mathematics 8, 35-40.
Ville, J. (1946), Sur les conditions d'existence d'une ophelimit6 totale et d'un indice du
niveau des prix, Annales de l'Universiti de Lyon 9, 32-39.
Ward, B. (1965), Majority Voting and Alternative Forms of Public Enterprise, in
J. Margolis (ed.), Public Economy of Urban Communities, Johns Hopkins Press,
Baltimore, 112-126.

? The editors of the Scandinavian Journal of Economics 1999.

This content downloaded from 141.201.158.104 on Fri, 24 Aug 2018 13:00:57 UTC
All use subject to https://about.jstor.org/terms