Maximin Criterion - Rawls

Download as pdf or txt
Download as pdf or txt
You are on page 1of 7

American Economic Association

Some Reasons for the Maximin Criterion


Author(s): John Rawls
Source: The American Economic Review, Vol. 64, No. 2, Papers and Proceedings of the Eightysixth Annual Meeting of the American Economic Association (May, 1974), pp. 141-146
Published by: American Economic Association
Stable URL: http://www.jstor.org/stable/1816033 .
Accessed: 08/01/2015 11:25
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
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].

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The
American Economic Review.

http://www.jstor.org

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

CONCEPTSOF DISTRIBUTIONALEQUITY
Some

Reasons

Maximin

for

the

Criterion

By JOHN RAWLS*

Recently the maximin criterion of distributive equity has received some attention from economists in connection with
the problem of optimal income taxation.1
Unhappily I am unable to examine the
merits of the criterion from the standpoint
of economic theory, although whether the
criterion is a reasonable distributive standard depends importantly on the sort of
examination that only economists can undertake.
What I shall do is to summarize briefly
some of the reasons for taking the maximin
criterion seriously. I should emphasize that
the maximin equity criterion and the socalled maximin rule for choice under uncertainty are two very different things. I
shall formulate the reasons for the equity
criterion so that they are completely independent from this rule.
In A Theory of Justice I have considered
the maximin criterion as part of a social
contract theory. Here I must assume a
certain familiarity with this conception.2
One feature of it might be put this way:
injustice exists because basic agreements
are made too late (Richard Zeckhauser).
People already know their social positions
and relative strength in bargaining, their

abilities and preferences, and these contingencies and knowledge of them cumulatively distort the social system. In an attempt to remedy this difficulty, contract
theory introduces the notion of the original
position. The most reasonable principles
of justice are defined as those that would
be unanimously agreed to in an appropriate initial situation that is fair between
individuals conceived as free and equal
moral persons.
In order to define the original position
as fair in this sense, we imagine that everyone is deprived of certain morally irrelevant information. They do not know their
place in society, their class position or
social status, their place in the distribution
of natural assets and abilities, their deeper
aims and interests, or their particular psychological makeup. Excluding this information insures that no one is advantaged
or disadvantaged in the choice of principles by natural chance or social contingencies. Since all are in this sense similarly
situated and no one knows how to frame
principles that favor his particular condition, each will reason in the same way.
Any agreement reached is unanimous and
there is no need to vote.
Thus the subject of the original agreement is a conception of social justice.
Also, this conception is understood to apply to the basic structure of society: that
is, to its major institutions-the
political
constitution and the principal economic

* Professor of philosophy, Harvard University. I am


grateful to Robert Cooter and Richard Zeckhauser for
their very instructive comments which enabled me to
improve these remarks and saved me fronm several
mistakes.
I See A. B. Atkinson, Martin Feldstein, Y'oshitaka
Itsumi, and Edmund S. Phelps.
2 In economics this sort of theory was reintroduced
by James Buchanan and Gordon Tullock.

A similar idea is found in J. C. Harsanyi (1953).

141

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

142

AMERICAN ECONOMIC ASSOCIATION

and social arrangements and how they


fit together into one system. The application of the maximin criterion to optimal
income taxation is, then, perfectly in order,
since an income tax is part of the basic
structure. But the maximin criterion is not
meant to apply to small-scale situations,
say, to how a doctor should treat his patients or a university its students.4 For
these situations different principles will
presumably be necessary. Maximin is a
macro not a micro principle. I should add
that the criterion is unsuitable for determining the just rate of savings; it is intended to hold only within generations
(Rawls, sec. 44, pp. 291-92, Kenneth J.
Arrow, 1973a, and Robert M. Solow).
But what alternative conceptions are
available in the original position? We
must resort to great simplifications in
order to get our bearings. We cannot consider the general case where the parties are
to choose among all possible conceptions
of justice; it is too difficult to specify this
class of alternatives. Therefore we imagine
that the parties are given a short list of
conceptions between which they are to decide.
Here I can discuss only two pair-wise
comparisons. These are designed to reflect
the traditional aim of contract theory,
namely, to provide an account of justice
that is both superior to utilitarianism and
a more adequate basis for a democratic society. Therefore the first choice is between
a conception defined by the principle that
average utility is to be maximized and a
conception defined by two principles that
express a democratic idea of justice. These
principles read as follows:
1. Each person has an equal right to the
most extensive scheme of equal basic
liberties compatible with a similar
scheme of liberties for all.
I This affects the force of Harsanyi's counterexamples
(Harsanyi, 1973).

MAY 1974

2. Social and economic inequalities are


to meet two conditions: they must be
(a) to the greatest expected benefit
of the least advantaged members of
society (the maximin equity criterion) and (b) attached to offices
and positions open to all under coInditions of fair equality of opportunity.
I assume that the first of these takes priority over the second, but this and other
matters must be left aside. (For more on
this, see Rawls, pp. 40-45 and 62-65.) For
simplicity I also assume that a person's
utility is affected predominantly by liberties and opportunities, income and wealth,
and their distribution. I suppose further
that everyone has normal physical needs
so that the problem of special health care
does not arise.
Now which of these conceptions would
be chosen depends on how the persons in
the original position are conceived. Contract theory stipulates that they regard
themselves as having certain fundamental
interests, the claims of which they must
protect, if this is possible. It is in the name
of these interests that they have a right to
equal respect and consideration in the design of society. The religious interest is a
familiar historical example; the interest in
the integrity of the person is another.
In the original position the parties do not
know what particular form these interests
take. But they do assume that they have
such interests and also that the basic
liberties necessary for their protection (for
example, freedom of thought and liberty
of conscience, freedom of the person, and
political liberty) are guaranteed by the
first principle of justice.
Given these stipulations, the two princiI Part of the justification for these assumptions is
given by the notion of primary goods. See Rawls, sec.
15. Of course, there are still difficulties (Arrow, 1973b,
1) 254).

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

VOL. 64 NO. 2

CONCEPTS OF DISTRIBUTIONALEQUITY

ples of justice would be chosen. For while


the principle of utility may sometimes
lead to a social order securing these liberties, there is no reason why it will do so in
general. And even if the principle often
does, it would be pointless to run the risk
of encountering circumstances when it does
not. Put formally, each must suppose that
the marginal utility of these fundamental
interests is infinite; this requires anyone in
the original position to give them priority
and to adopt the two principles of justice.
This conclusion is strengthened when
one adds that the parties regard themselves as having a higher-order interest in
how their other interests, even fundamental ones, are regulated and shaped by social institutions. They think of themselves
as beings who can choose and revise their
final ends and who must preserve their
liberty in these matters. A free person is
not only one who has final ends which he is
free to pursue or to reject, but also one
whose original allegiance and continued
devotion. to these ends are formed under
conditions that are free. Since the two
principles secure these conditions, they
must be chosen.
The second pair-wise comparison is far
more difficult. In this case the choice is between the two principles of justice and a
variant of these principles in which the
utility principle has a subordinate place.
To define this variant, replace the second
principle by the following: social and economic inequalities are to be adjusted so as
to maximize average utility consistent
with fair equality of opportunity. The
choice between this variant and the two
principles is more delicate because the
arguments from liberty can no longer be
made, at least not so directly. The first
principle belongs to both conceptions, and
so the operation of the utility principle is
hedged by basic rights as well as fair
equality of opportunity.
One reason favoring the two principles

143

of justice is this. From the standpoint of


the original position, the parties will surely
be very considerably risk-averse; if we ask
how risk-averse, we might say not less than
that of most any normal person. Of course,
this is extremely vague; but if we assume
that utility is estimated from the standpoint of individuals in society and represents, as the classical utilitarians believed,
a quantity ascertainable independent of
choices involving risk, then, given the
crucial nature of the decision in the original position, the claims of the utility principle seem quite dubious. On the other
hand, if we suppose that utility is measured from the original position and takes
account of risk, the utility criterion may
not differ much from maximin. The standard of utility approaches maximin as risk
aversion increases without limit (Arrow,
pp. 256-57). So, either way, the original
position pushes us toward maximin. However, in weighing the second pair-wise
comparison, I assume that, based on considerations of risk aversion alone, there
is a significant difference between the two
conceptions. Thus the problem is to identify other attractive features of the maximin criterion that tip the balance of
reasons in its favor.
First, much less information is needed to
apply the maximin criterion. Once the
least-favored group is identified, it may be
relatively easy to determine which policies
are to their advantage. By comparison it is
much more difficult to know what maximizes average utility. We require a fairly
precise way of comparing the utilities of
different social groups by some meaningful standard, as well as a method of estimating the overall balance of gains and
losses. In application this principle leaves
so much to judgment that some may
reasonably claim that the gains of one
group outweigh the losses of another,
while others may equally reasonably deny
it. This situation gives those favored by

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

144

AMERICAN ECONOMIC ASSOCIATION

existing inequalities an opportunity to exploit their advantage so that, as a result,


inequalities are likely to be excessive, undermining the justice of the system.
A further consideration is this: a distributive criterion is to serve as a public
principle. Citizens generally should be able
to understandit and have some confidence
that it is realized. Pattern criteria, those
that require the actual distribution to exhibit certain ascertainable features, do
well by the test of publicity. Of these,
strict equality (equal division) is the
sharpest principle. The trouble with pattern criteria is that sharpness is not the
only desideratum, and they often have
little else to commend them. On the other
hand, the utility principle is not sharp
enough: even if it were satisfied, there
could be little public confidence that this
is indeed the case. The maximin criterion
has sufficient sharpness; at the same time
it is efficient while strict equality is not.
Another ground supporting the maxi.
min criterion is based on the strains of
commitment: in the original position the
parties are to favor those principles compliance with which should prove more
tolerable, whatever their situation in society turns out to be. The notion of a contract implies that one cannot enter into an
agreement that one will be unable to keep.
By this test, also, maximin seems superior,
for the principles chosen would regulate
social and economic inequalities in the
basic structure of society that affect
people'slife-prospects.These are peculiarly
deep and pervasive inequalities and often
hard to accept.
Looking first at the situation of the less
advantaged, the utility principleasks them
to view the greater advantages of others
who have more as a sufficient reason for
having still lower prospects of life than
otherwise they could be allowed. This is an
extreme demand psychologically; by contrast, the maximin criterion assures the

MAY 1974

less favored that inequalities work to their


advantage. The problem with maximin
would appear to lie with those who are
better situated. They must accept less than
what they would receive with the utility
principle, but two things greatly lessen
their strains of commitment: they are,
after all, more fortunate and enjoy the
benefits of that fact; and insofar as they
value their situation relatively in comparison with others, they give up that much
less. In fact, our tendency to evaluate our
circumstances in relation to the circumstances of others suggests that society
should be arrangedso that if possible all its
members can with reason be happy with
their situation. The maximin criterion
achieves this better than the principle of
utility.
I have noted several reasons that support the maximin criterion: very considerable normal risk-aversion (given the
special features of the original position),
less demanding information requirements,
greater suitability as a public principle,
and weaker strains of commitment. Yet no
one of them is clearly decisive by itself.
Thus the question arises whether there is
any consideration that is compelling. I
want to suggest that the aspirations of
free and equal personality point directly to
the maximin criterion.
Since the principles of equal liberty and
fair equality of opportunity are common
to both alternatives in the second comparison, some form of democracy obtains
when either alternative is realized.Citizens
are to view themselves as free and equal
persons; social institutions should be
willingly complied with and recognized as
just. Presumably, however, certain social
and economic inequalities exist, and individuals' life-prospects are significantly affected by their family and class origins, by
their niaturalendowments, and by chance
contingenciesover the course of their lives.
We must ask: In the light of what principle

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

VOL. 64 NO. 2

CONCEPTS OF DISTRIBUTIONAL

can free and equal moral persons permit


their relations to be affected by social
fortune and the natural lottery? Since no
one deserves his place in the distribution of
talents, nor his starting place in society,
desert is not an answer. Yet free and equal
persons want the effects of chance to be
regulated by some principle, if a reasonable
principle exists.
Now when the maximin criterion is followed, the natural distribution of abilities
is viewed in some respects as a collective
asset. While an equal distribution might
seem more in keeping with the equality of
free moral persons, at least if the distribution were a matter of choice, this is not a
reason for eliminating natural variations,
much less for destroying unusual talents.
To the contrary, natural variations are
recognized as an opportunity, particularly
since they are often complementary and
form a basis for social ties. Institutions are
allowed to exploit the full range of abilities
provided the resulting inequalities are no
greater than necessary to produce corresponding advantages for the less fortunate.
The same constraint holds for the inequalities between social classes. Thus at first
sight the distribution of natural assets and
unequal life-expectations threatens the relations between free and equal moral persons. But provided the maximin criterion
is satisfied, these relations may be preserved: inequalities are to everyone's advantage and those able to gain from their
good fortune do so in ways agreeable to
those less favored. Meeting this burden of
proof reflects the value of equality.
Now the maximin criterion would conform to the precept "from each according
to his abilities, to each according to his
needs" if society were to impose a head tax
on natural assets. In this way income inequalities could be greatly reduced if not
eliminated. Of course, there are enormous
practical difficulties in such a scheme;
ability may be impossible to measure and

EQUITY

14S

individuals would have every incentive to


conceal their talents. But another difficulty
is the interference with liberty; greater
natural talents are not a collective asset in
the sense that society should compel those
who have them to put them to work for
the less favored. This would be a drastic
infringement upon freedom. But society
can say that the better endowed may improve their situations only on terms that
help others. In this way inequalities are
permitted in ways consistent with everyone's self-respect.
I have attempted a brief survey of the
grounds for the maximin criterion. I have
done this because historically it has attracted little attention, and yet it is a
natural focal point between strict equality
and the principle of average utility. It
turns out to have a number of attractive
features. But I do not wish to overemphasize this criterion: a deeper investigation
covering more pair-wise comparisons may
show that some other conception of justice
is more reasonable. In any case, the idea
that economists may find most useful in
contract theory is that of the original position. This perspective can be defined in
various ways and with different degrees of
abstraction and some of these may prove
illuminating for economic theory.6
REFERENCES
K. J. Arrow, "Rawls's Principle of Just Saving," Instit. for Math. Stud. in the Soc.
Sciences, Stanford, tech. rep. no. 106, Sept.
1973a.
"Some Ordinalist-Utilitarian Notes
on Rawls's Theory of Justice," J. Phil.,
May 1973b, 70, 254.

A. B. Atkinson, "How Progressive Should Income Tax Be?" in J. M. Parkin, ed., Essays
in Modern Economics, London 1973.
J. Buchanan and G. Tullock, The Calculus of
Consent, Ann Arbor 1962.
6 See Zeckhauser for an illustration of the use of this
sort of framework.

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

146

AMERICAN ECONOMIC ASSOCIATION

M. Feldstein, "On the Optimal Progressivity


of the Income Tax," Harvard Instit. of
Econ. Res., disc. pap. 309, July 1973.
J. C. Harsanyi, "Cardinal Utility in Welfare
Economics and in the Theory of Risk Taking," J. Polit. Econ., Oct. 1953, 61, 434-35.
,"Canthe Maximin Principle Serve as
a Basis for Morality?" Center for Res. in
Manage. Sc., Berkeley, work. pap. no. CP351, May 1973, 8-9.
Y. Itsumi, "Distributional Effects of Income
Tax Schedules," Rev. Econ. Stud., forthcoming.

MAY 1974

E. S. Phelps, "The Taxation of Wage Income


for Economic Justice," Quart. J. Econ.,
Aug. 1973, 87, 331-54.
J. Rawls, A Theory of Justice, Cambridge,
Mass. 1971.
R. M. Solow, "Intergenerational Equity and
Exhaustible Resources," dept. of econ.,
work. pap. no. 103, Cambridge, Mass. Feb.
1973.
R. Zeckhauser, "Risk Spreading and Distribution," in H. M. Hochman and G. E. Peterson, eds., Political Economics of Income
Distribution, New York 1974.

This content downloaded from 130.104.110.197 on Thu, 8 Jan 2015 11:25:56 AM


All use subject to JSTOR Terms and Conditions

You might also like