A Model for Theories of Distributive Justice

advertisement
A MODEL FOR THEORIES OF DISTRIBUTIVE JUSTICE
Ted Everett
(draft 11/28/09)
1.0 Introduction
The term distributive justice means no more than justice with respect to distribution,
whether the thing distributed be money, services like health care, or any other scarce goods or
privileges. Who gets what, is the essential question. But the actual nature of distributive justice
is a matter of great controversy, and on several levels. Among professional philosophers, it is an
long-standing and important problem how to define distributive justice in an intuitive and useful
way. On this field Plato and Aristotle, Locke and Rousseau, Nozick and Rawls have all
contended, never conclusively. Among non-philosophers concerned with public policy,
questions about distributive justice underlie the most important controversies, whether over
health care, taxation, education, affirmative action, capital punishment, and even foreign policy.
One tends to find opinions roughly grouped by a general “left” or “right” orientation (say, into
publications like The Nation or The National Review), but the deeper principles defining these
positions are complex, and sometimes overlap the two sides. Even in the most uninformed
political discussions on television, one often finds competing notions of distributive justice
driving debate, though inarticulately, and usually much confused with loyalty to each
combatant’s own side and contempt for his opponents.
What I want to do here is to provide a neutral framework for discussing theories of
distributive justice. I will introduce a simple but flexible template for representing such theories
graphically, which I will call distribugrams until a better name shows up. I think that the
structured range of possibly stable distribugrams can show us something useful about relations
among competing pure theories of justice, and suggest some possibly attractive combinations.
(Theoretical purism itself, not to mention political fanaticism, is usually less appealing to people
who can clearly imagine sliding back and forth between their own perspectives and the views of
others.) I hope that I will be able to establish two things, anyway: the usefulness of these
distribugrams for clarifying and comparing theories of distributive justice, and the
reasonableness of my overall approach.
1.1 Some background values
When we ask people what basic values they think apply to questions of distributive
justice, a small set of very common answers is likely to appear, which we can for the moment
throw into an unordered list:
(a) liberty
(b) equality
(c) utility, or general welfare
(d) compassion for the least well-off (I’ll call this “minimality”)
There are some others, such as rewards and punishments according to merit, that I will leave out
of my main discussion here, not because they are unimportant, but because they do not fit my
diagrams. For the most part, then, I will concentrate on the four basic values I’ve just listed.
Note that all are matters of degree: we could have more or less equality in our society
than we do, more or less freedom, and so on. I think that almost all of us agree on assigning
some moral weight to each of these four things. Everybody counts equality as a factor in
1
distributive justice sometimes, and similarly for freedom, and so on. I don’t want to insist on
this; some of us may want to dismiss one or another of the four as altogether morally irrelevant.
But the great majority of people surely see each of these things as a good thing, at least if we can
get it without giving anything else up. That is, I think we’ll almost all agree on these as prima
facie principles of justice: other things being equal, it is better to have a society with more
equality, more freedom, more prosperity, or more help for people in the worst condition. We
need a good reason when distributing something not to divide it equally, or not to let each person
choose what they want, or not to do what works best for everybody overall, or to ignore the
needs of those who are the least well-off.
The internal relations among these four values are not immediately clear. Perhaps some
can be subsumed under others, for example concern for the disadvantaged under equality, or
equality under general welfare. But at least some of the time they seems like separable basic
values, each a good thing in its own right. In any case, there surely are trade-offs to be made
among these values, at least some of the time, in any realistic political setting. If we decide to
make people’s incomes a lot more equal, this is likely to cost something in terms of economic
freedom; if we adopt the policies that will lead to the greatest overall future prosperity, that may
reduce our present ability to care for the most disadvantaged; and so on. It appears, then, that not
all combinations of degrees to which these values are realized are equally possible, and some
may be entirely impossible, for example to maximize both economic freedom and equality at the
same time.
A pure theory of distributive justice picks one value to maximize, whatever the cost to all
the others. Pure libertarianism is essentially concerned only with freedom; whatever other values
may be realized at the same time are secondary, tradable-away in principle for any additional
amount of freedom. Pure socialist egalitarianism is essentially concerned only with equality.
Utilitarianism’s exclusive concern with total overall well-being is utilitarianism, while exclusive
concern with the condition of the least well-off is typically called Rawlsianism after the
philosopher John Rawls. These are undoubtedly the top four standard theories of distributive
justice, each with plenty of adherents. Complex theories of distributive justice, on the other
hand, pick more than one value to balance, seeking a satisfactory overall structure for all values
considered. In what follows, I will talk about the four main pure theories, then a certain range of
complex theories.
1.2 Distribugrams
Here is the main idea. Consider what people have, prior to any effort to redistribute for
the sake of justice as a single, measurable quantity, like money. Lay this quantity along the xaxis of an ordinary graph. Then let the y axis represent how much of this quantity a person is to
have after the distribution takes place. Now define the function that turns the initial, predistribution values into the ultimate, post-distribution values. Call this the distribution function
f(x). The resulting graphs will look like this, varying only with the shape of f(x):
2
y
postdistribution
f(x)
x
pre-distribution
This is a distribugram. Here is a formal definition:
(1) A distribution function is any function f(x) that takes initial values of some
measurable thing (x) into values of the same thing (y).
(2) A graph that represents a distribution function is a distribugram.
2.0 Some pure theories of distributive justice
Let us look at some theories of justice that can be defined using straight-line distribution
formulas, which are the simplest to describe. Imagine a total welfare system in which there are
only two ingredients, a national minimum income and a flat-rate income tax. Such a system can
be defined by just two numbers, the minimum income level (call it m) and the tax rate (t). Here
is the distribugraphic form of all such systems:
y
post-tax
income
f(x) = 1 – t(x)
m
x
pre-tax income
In the distribugram above, the minimum income m appears as the y-intercept of the distribution
function f(x). The slope of f(x) is by definition 1- t, where t is the rate of taxation. The value of
x will represent each person’s pre-tax income, and y will represent the amount that they end up
3
with after redistribution. Thus, if the tax rate is set at 25%, the “keep rate” will be (100-25)%, or
75%, which defines the slope of any distribution function with that tax rate. People who earn no
money at all will receive amount m (in the form of monthly checks, perhaps, as in Social
Security). Everybody else will end up with that same amount, plus whatever they have earned,
minus the tax on that earned income. For example, if the minimum income is $10,000, and the
tax rate is 25%, then a person earning $20,000 in pre-tax income will end up with $25,000
($10,000 + $20,000 - $5,000), while someone earning $80,000 will end up with $70,000
($10,000 + $80,000 - $20,000), like so:
$70,000
t = 25%
$25,000
m = $10,000
0
$20,000
$80,000
The economist Milton Friedman recommends just this sort of cash-only welfare system in
his book Capitalism and Freedom.1 But I do not mean to assume that cash is always the most
reasonable form in which to distribute benefits to people, or that income taxes are always the
most appropriate method of taxation. What I do mean to assume is that differences in the form
of taxation and delivery of benefits make no essential difference to the basic issues of
distributive justice that I am discussing here. So, think of the tax function as the total of all taxes
people are required to pay at different levels of initial wellness-to-do, all things considered, and
of the post-tax distribution as the total value of whatever cash and other benefits people have
received at the end of the day. The cash and income-tax format is just a convenient proxy.
Note that only certain combinations of tax rate and minimum are possible, given the total
value available in any system. As things are, for example, we obviously cannot guarantee a
minimum income of $1 million to all Americans, because the country doesn’t have that much
money. Even a minimum benefit level of $10,000 would be impossible for us without total pretax GDP of at least $3.1 trillion. Given the size of our present economy and its pre-tax
distribution of income, only a fairly high tax rate could generate enough revenue to guarantee
that level of benefits. For the sake of our discussion of justice, let us consider only such taxation
as is dedicated to redistribution. In this framework, each tax rate will completely determine its
associated minimum level of benefit. Thus, in any distribugram, the point m will be a function
of the tax rate r. In effect, then, the whole system is determined by the tax rate alone.
The most common pure theories of justice can all be represented by simple distribugrams.
Considering only taxation for redistribution and the benefit levels implied by each level of
4
taxation, we can reduce much of the complex issue of distributive justice to this simple question:
what tax rate makes for the most just overall distribution of benefits?
2.1 Libertarianism
The ideal of libertarianism is to have no taxation at all for the sake of redistributing
income. This entails, for our purposes, a minimum benefit level m = 0, and a tax rate of 0, which
means a “keep rate” of 1.
y
post-tax
income
t=0
m=0
x
pre-tax income
Here m = 0 and t = 0, so the post-distribution values are identical to the pre-distribution ones,
which we represent by giving f(x) a slope of 1.2
This simple distribugram could be said to represent anarchism better than normal
libertarianism. Actual libertarians from Locke on accept that there is a proper role for
government, limited to the protection of rights, or “property” broadly construed. Thus, the slope
of a total tax function would be something less than 1, since there must be some taxation in order
to pay for an army and police, plus other necessary government functions. But, as I said, I am
ignoring this factor here, in order to concentrate on taxes for redistribution only, which is what
matters to justice.
Some libertarians assimilate freedom to merit, on the grounds that merit can be defined as
something like the total market value of one’s talents, perseverance, energy, and so on. Thus,
letting people keep their market-determined pre-tax incomes is no more than giving them what
they deserve through hard work and talent, while redistributing income to the poor amounts to
rewarding them for laziness, lack of intelligence, and so on. Hence, redistribution is unfair not
only because each person has a right to keep what they have legally acquired, but because those
legal acquisitions are already fair inherently. Libertarians also sometimes argue on utilitarian
grounds, that curtailing redistribution leads to benefits for everyone, because it fosters selfreliance, economic growth, and other things that benefit the whole society. But libertarianism in
its pure form considers maximal freedom to spend one’s own money one’s own way as a basic,
or natural, right. Even if nothing good came of it in any other way, the strictest libertarians
would stand their ground on that basis alone.
5
I should point out that overall tax rates below zero are impossible, or at least
unsustainable, given that one cannot distribute more to people than can be acquired through
taxation. In principle, this cannot work:
y
post-tax
income
t<0
x
pre-tax income
In real life, though, two forms of negative taxation do sometimes appear: (a) partial negative
taxation for upwards redistribution, as with using income tax revenues for upper-class agencies
like NPR and NEA, or taxing tobacco and other products favored by low-income citizens, and
(2) negative tax rates for stimulation through deficit spending, which amounts to a tax on future
generations. Each can perhaps be justified sometimes: the former as part of a sustainable overall
distributive plan, the latter for emergencies like wars, or through expecting wealthier future
generations to pay the money back with little hardship. Thus, the latter is equivalent to ordinary
redistribution from the wealthy to the poor, but taking place over time rather than synchronically.
Considering everyone involved, in any case, the overall tax rate can never sustainably be set at
less than zero. We cannot distribute more goods than actually exist, all things considered.
2.2 Egalitarianism
If substantive equality were all that anybody meant by justice, then the minimum tax and
benefit levels would have to be set at whatever gave everyone the same. The minimum would
then be the same as both the average and the maximum. This means, in effect, a tax rate of
100%, all initial holdings to be redistributed in equal portions. Here is the distribugram for this
pure theory:
6
y
post-tax
income
t=1
m=e
x
pre-tax income
Here, the minimum value e stands for the benefit level that results from equalizing
everybody’s benefits. This level might be rather meager, judging from the experience of 20thCentury communism, but perhaps not necessarily so. In any event, e will surely be greater than
the level of 0 guaranteed by pure libertarianism.
Like libertarianism, strict egalitarianism constitutes one end-point of the range of
possibly stable distribution functions. Tax rates of greater than 100% are possible for temporary
redistributions from the rich to the poor that go beyond equality to reverse the sides, like so:
y
post-tax
income
t>1
m=p
x
pre-tax income
This sort of redistributive move might even be justified sometimes, perhaps as
punishment for wealthy people who have been seen as exploiters. There seem to have been
cases of redistribution that fit this description, as in some socialist revolutions where previously
privileged people have been deliberately reduced to miserable poverty, and some portion of the
previously poor have been raised to wealth. Robert Mugabe of Zimbabwe and similar dictators
have dispossessed successful farmers and businessmen, handing the confiscated properties out to
7
their own supporters as rewards for loyalty, while providing just this sort of punitive
justification. Pol Pot’s Khmer Rouge government and Mao Tse-Tung’s Cultural Revolution both
took this principle of retributive redistribution to its extreme expression. Such ptochracratic
distributions are necessarily short-lived, however, for the simple reason that the previously lesswell-off now constitute the more-well-off. More stable distribution systems representing the
new background situation will inevitably follow.
2.3 Utilitarianism
Strict utilitarians believe that justice can best be defined as maximizing average or total
welfare, and that equality counts only to the extent that it contributes to the overall well-being of
society. It is an unchallenged fact of economics that any decrement from a 100% tax rate will
tend to stimulate economic growth, which in turn will tend to produce more total benefits
available for redistribution. It is also plain to see that most people find satisfaction (hence,
enhanced well-being) in keeping for themselves at least some portion of whatever they have
earned. Thus, strict utilitarians tend to believe that some degree of inequality in benefits is a
good thing.
Here is the distribugram for a strictly utilitarian system:
y
post-tax
income
t: y = max(avg(x))
max(avg(x))
m=u
ex
pre-tax income
Note that u > e in this distribugram. Is this a reasonable assumption? Why must the
minimum level in a society whose redistribution system maximizes average well-being be higher
than the level of well-being in a society that taxes everything in order to equalize benefits?
There is no logical necessity here, but history seems to show that societies with the highest
average benefit levels also have higher minimum levels than those societies that are otherwise
similar, but which maximize equality instead. East and West Germany during the Cold War
provide a pretty clear example, as do North and South Korea. People have been materially better
off on welfare in either of the two capitalist halves of these nations than with average jobs in
their socialist counterparts.3 The commonsense reasons usually given for these correlations are
(1) that human beings will not usually work harder than necessary unless they receive some
share of the added value of their work, and (2) sufficient inequality allows for capital formation
by the best-off, who become private employers in a production system more efficient than pure
socialism.
8
Conceivably, a higher overall or average welfare could be achieved by making the tax
rate progressive, i.e. using higher rates for those who have a higher pre-tax income. The
utilitarian argument for this is that the less money (or other benefit) a person has to start, the
more each increment of money is worth to that person. Thus, a person with a million-dollar
income will lose less utility than a poor person gains when moderate amounts are redistributed
from one to the other. It stands to reason, then, that the higher one’s initial income, the higher
rate one ought to pay in transfer taxes, if the purpose is only to distribute money where it will do
the most good for the moment. There are also egalitarian arguments for progressivity, of course,
all the way up to a 100% rate. But on strictly utilitarian grounds we would presumably not bend
the tax curve all the way horizontal, because there are still benefits to the whole group in
individual satisfaction, work incentives, and capital formation by allowing people to keep
something of their own initial holdings.4 Here is the distribugram for a progressive tax:
post-tax
income
y
t: y = max(avg(x))p
max(avg(x))p
m = up
ex
pre-tax income
It remains an empirical question whether progressivity actually works in the way that its
utilitarian proponents claim. If the average overall benefit with a progressive rate is higher (in
the diagrams, if max(avg(x))p > max(avg(x))) then progressivity increases utility; otherwise, not.
Note also that there is already a built-in sort of progressivity implicit in a Friedman-style
minimum income system. A person making less than the minimum divided by the tax rate (x =
m/t) will have a negative effective tax rate, a person exactly m/t will pay zero effective tax, and a
person making above that amount will have a increasing positive effective tax rate
asymptotically approaching the nominal rate t. In our earlier example of a $10,000 minimum
and a 25% tax rate, the break-even point will come at a pre-tax income of $40,000, as in this
graph (not a distribugram):
9
25%
effective
tax rate
0
$40,000
pre-tax income
2.4 Rawlsian maximinimalism
John Rawls argued that justice is an essentially objective notion. He said that it can best
be understood when we place ourselves behind the “veil of ignorance”, imagining that we were
going to be thrown into a new society with no idea what position we will be assigned in that
society, or whether our talents and skills will have any value there. What economic laws, he asks
– in essence, what distribution function – would we choose in advance for such a society? Rawls
says that humans are naturally so risk-averse that any reasonable person from behind the veil of
ignorance would prefer a new society in which the worst overall position was as good as it could
possibly be. This defines what Rawls calls the maximized minimum or “maximin” conception
of distributive justice.
It is always controversial exactly what tax rates would be sufficient to generate what
levels of benefit, because of the evident dynamic effects of taxation itself. Low tax rates tend to
encourage investment and subsequent growth, and sufficiently high tax rates may retard
economic activity, leading to lower levels of distribution in the future. The economist Arthur
Laffer argued that there will always be one flat tax rate that maximizes possible revenues over
time; anything higher or lower will generate less total revenue. This tax rate, called the “Laffer
point”, has been estimated at anywhere from 8% to 65% for a society like ours, but is most
frequently held to be in the vicinity of 30% – high enough to generate a lot of revenue, but low
enough to keep most tax payers honest and productive. A Rawlsian distribugram would take that
Laffer point as the tax rate (lp), then, in order to maximize the minimality within the system, like
so:
10
y
post-tax
income
t = lp
m=r
uex
pre-tax income
I have placed the Rawlsian minimum (r) above both the utilitarian and egalitarian ones,
for the obvious reason that the Rawlsian one has been maximized by definition. There is no
reason to suppose that this maximum will not be higher than the other points, both of which
resulted from the maximization of other values. On reasonable assumptions, then, we ought to
expect that 0 < e < u < r.
Rawls’s theory has spawned a great deal of literature over the last generation or so. The
main objection to it is that intuitively, we do not only care about what happens to the least welloff in any situation, but also, if to a lesser extent, about the next-to-least well-off, and to some
extent about the third-least well-off, and so on. Even in thinking only of ourselves, we worry not
only about what is the worst thing that could happen to us, but also the second-worst, thirdworst, and so on. This suggests some kind of utilitarian or egalitarian loosening of Rawls’s strict
attention to the very bottom case.
3.0 Mixed theories of distributive justice
I have presented libertarianism, egalitarianism, utilitarianism, and maximinimalism as
pure theories of justice, each taking one intuitive moral value to define distributive justice
completely. An alternative approach sees all four values as somewhat independent moral forces,
each often at work in our deliberations over justice. If this second approach is right, then the
problem of distributive justice is to be solved not by picking one value as fundamental and
attempting to reduce the others to it, but by balancing these forces to produce a theory that “does
justice” to each of them.
Imagine reaching into one of these distribugrams, grasping the line or curve that
represents the distribution function, and moving and twisting it around continuously, changing
the slope and the y-intercept at will. This imaginary motion can be seen as representing a
continuous balancing and rebalancing of the moral forces involved in our most common
intuitions about justice:
11
y
post-tax
income
liberty
utility
equality
minimality
x
pre-tax income
Note that only a certain range of distribution functions is possible for stable systems.
Because overall tax rates above 100% are unsustainable, the final slope cannot be lower than 0.
Because negative overall tax rates are likewise unsustainable, the slope of the final distribution
function cannot be greater than 1. Therefore, the crucial line in any stable distribution function
cannot go lower to the right than horizontal (slope 0), or higher than an angle of 45 degrees
(slope 1). At the same time, the minimum benefit level cannot be lower than 0 or higher than the
Rawlsian minimum defined by the Laffer point.
y
post-tax
income
possibly
stable
distribution
functions
t = lp
t=0
r
t=1
e
x
pre-tax income
Let me discuss a few ways we might reasonably combine and balance the basic moral forces to
produce a more satisfactory theory of justice within this range of possibilities.
3.1 Communalism
The values of equality and minimality seem to spring from the same moral source,
though they yield very different theories in their pure forms. The common impulse behind both
is a concern for poor people, sick or disabled people, oppressed people, old people, children –
12
anybody we might care about whose needs outstrip their own resources. To the extent that we
can help such people, we believe that we should always try. This feeling is naturally expressed
in a special concern for those who are the very least well-off at one end, and in a general
suspicion or dislike of any kind of inequality at the other. But I think that it cannot be right – it
cannot even be what anybody really wants – that we should push this compassionate moral
impulse to either of those theoretical extremes.
Consider this family situation. You are the parent of three children: Susie is fine, with no
more problems than the average child; Jeff is a paraplegic in a wheelchair; and Billy is a
quadriplegic, essentially confined to his room for the duration of his life. It would be easy
enough for you to satisfy either extreme theory of justice in this situation. To produce complete
equality among your children, you need only paralyze both Susie and Jeff from the neck down.
Then all three children will be equally disabled, and equally miserable. No parent would do this,
of course. You would have to be insane to care so much about equality that you would ruin
everybody’s lives in every other way in order to achieve it. Strict egalitarianism, therefore,
simply can’t be right. Rawls’s position seems more moderate, and in a way it is, but in a way it
isn’t. Suppose that Billy’s condition is such that you can always do something or other to help
him: bring him some soda and a straw, change the channel on his television set, read him a story,
or just keep him company. Would you consider yourself morally bound to ignore Jeff’s needs
and desires completely for as long as he is better-off than Billy? Jeff needs a ride to school, but
you could do a little bit for Billy by keeping him company instead. Susie has cut her finger, but
the family’s last Band-aid is serving as a bookmark up in Billy’s room. What would you do?
Nobody really thinks that it would be fair to leave your other children suffering, provided only
that they remain less miserable than their most miserable sibling, just for the sake of some tiny
benefit to the child who is the worst-off. The natural thing in any family is to do the most for the
most needy child, the next most for the next most needy, and so on, in proportion to their need
and the total resources available. Billy can wait a half an hour for his soda while Jeff gets a lift
to school or has his wheelchair fixed. Even the best-off Susie ought to get a present on her
birthday, come though it may at some minor cost to Billy and Jeff. For her to get nothing at all
would plainly be unjust.
I think that these intuitive results apply to larger groups than families. It would not be
right for any village or school or other community to devote all of its surplus resources to
helping the worst-off members be a little less badly off, while other members who are not as
badly-off, but still have needs, are totally neglected. A proper teacher helps all of the students in
his class, proportionately both to their need and to their ability to benefit from the teacher’s
resources of attention. It is universally considered a bad thing for the smartest and best-prepared
students to get nothing from school, while the students with the least initial ability get
everything. Moral compassion on whatever scale, it seems, is naturally expressed in proportional
terms: the more needy or more helpable a person is, the more help they should get, other things
being equal and given sufficient resources. In real situations, we will attend exclusively to the
least well-off only if no one else needs much of anything, or if no one else can be helped at an
acceptable cost. At the same time, we will enforce total equality on a group only where no one is
thereby harmed a lot for very little gain to others.5 What really matters is that everybody gets a
reasonable share of what’s available, relative to need. Let me call this family-style notion of
justice communality, and the pure theory that takes it as the exclusive distributive value
communalism.6
13
The family-based moral impulse of communality is not a compromise between two basic
distributive values, I think, but a more basic, more deeply intuitive value in its own right.
Egalitarian and Rawlsian ideas are really just shadows of this central moral intuition, cast by
overly simple theories. Indeed, my friends on the left who claim to be strict egalitarians or
Rawlsians really are not. To judge by their actions or by their particular opinions, they are all
communalists if anything so simple.
The moral force of communality is probably impossible to define with much precision.
Our intuitions here no doubt developed over countless generations of family and community life,
hence they tend to be flexible, situational, and typically unconscious in their balancing of needs
and interests. But the communalist position is easily distribugrammed as falling somewhere inbetween the poles of strict egalitarianism and Rawls’s maximinimalism, like so:
y
post-tax
income
t = lp
lp < t < 1
r
m=c
e
t=1
x
pre-tax income
3.2 Subsistence minimalism
The family-model value of communality must also be balanced against the libertarian
value of freedom in a fully reasonable theory. Let me start from the position of strict
libertarianism, and filter in communalist intuitions until the result seems about right. Why
should a libertarian not feel the intuitive force of compassion? Most libertarians will say that
they do, that they are as compassionate as anybody else, but will also say that this feeling is
irrelevant to justice. The issue, they remind opponents, is not about compassion, but about
rights.
It is often held by libertarians that equality of opportunity is what ought to be guaranteed
to everybody, as opposed to anything like equality of results. People on the left and on the right
seem to agree that this objective state is a desirable goal, though they will differ on what
constitutes equality and opportunity. But I think that nobody who has a family or friends
believes in absolutely equal opportunity. For being a good friend or family member means in
part that one gives preference to ones friends and other family members. One tries to help them
do well, not just to see that they are able to compete on a fair basis with strangers. In trying to be
a good friend or family member, one tries to advantage ones family and friends over strangers.
If you are in a position to hire someone, and you know that a friend of yours is qualified and will
do a fine job, then you should try to hire him, even if some stranger has slightly better
qualifications. To do otherwise would be to fail as a friend, if not as an employer. And surely,
14
you do not want your children to have no better education than any other children, or your
parents to have only average care when they are sick. If you are a good parent and child, you
will want them all to have the best services available, and you will go out of your way, to a
greater than average extent, to make sure that this is so.7 This is not to say that you should cheat
or commit crimes in order to help those closest to you. What matters is to balance the objective
and subjective claims of justice in a reasonable way. As a good citizen, you must support rules
of fair play in hiring, college admissions, and so on. But you should try to help your friends and
family succeed within those rules, and try harder than average if you can. It would be wrong,
indeed unjust, for you not to.
The question, then, is what, if any, substantive distributions ought to be guaranteed to
everyone, along with fair rules of play. It is sometimes part of the moral explanation of
libertarianism, against the charge that it implies indifference to suffering, that libertarians expect
their totally voluntary system to yield an acceptable level of charitable giving in place of any
coercive redistribution.8 So, the total moral position of libertarians ought to be understood in
that context. As a matter of fact, any redistribution system can be similarly enhanced by way of
voluntary contributions to the badly-off. Let us distinguish, then, between each system of
distributive taxation and its charitably-enhanced counterpart, which will include expected total
redistributions through both mandatory and voluntary payments, and make appropriate
adjustments to the relevant distribugrams. A distribugram for such enhanced libertarianism
would look like this:
y
post-tax
income
tc > 0
t=0
m = lc
x
pre-tax income
Here, I am assuming for simplicity that charitable giving would function like a voluntary flat tax,
the amount of charity given and received being proportional to pre-distribution holdings, along
the lines of a traditional “tithe” or tenth. The slope of the charity-adjusted function fc(x) will be
less than 1, then, because the charity-enhanced tax rate tc is more than 1. Hence, the enhanced
minimum benefit level lc will be greater than 0.
Charitable giving is a potentially important factor in questions of distributive justice, so
we should bear in mind that any simple distribugram can be paired with a charitably-enhanced
variant. But I will continue to discuss questions of distributive justice primarily in terms of
mandatory redistribution and the associated guaranteed (not just expected) levels of benefit. This
accords with the libertarians’ (and most other people’s) conception of justice itself as a matter of
15
rights, not just good luck or the good will of others. We are entitled to what is just, regardless of
whether others are liable to give it to us anyway.
Is there any compassionate distribution, then, that a libertarian ought to grant strangers as
a matter of right, not charity? The strictest libertarians usually say no, on the grounds that there
is nothing that we positively owe to everybody else, over and above fair play. That is,
libertarians are liable to concede that it is in some way unjust not to share resources with their
own families and close friends. But they will balk at the idea that such family-model
responsibilities extend across the whole society. Your child, your parent, or your close friend in
need may naturally claim something from you as a right attendant on that relationship, but a
stranger may not. An exception must be made for contracts, of course. I may not know my
banker personally, but I still have to give him some of my money, because I borrowed that
money from his bank as part of a mortgage contract. If I hire a kid to mow my lawn, I have to
pay him. But this does not apply to the majority of strangers, with whom I have made no such
arrangement.
Some libertarians concede that we have some level of substantive responsibility to
strangers, but they will limit it to something like a right to life. If you are a member of my
society, I will not watch you starve or freeze to death, or lie bleeding in the street after a car
crash, without doing something about it. But beyond this minimal ability to live and get around,
perhaps to seek work or to emigrate, I need guarantee you nothing. I will call this theory of
justice, slightly modified from absolute libertarianism, subsistence minimalism, and it looks like
this:
y
post-tax
income
t: m = s
m=s
x
pre-tax income
Here, the minimum benefit level s will be small compared to those of more fully communalist
theories, and the tax rate t will be correspondingly close to 0.
3.3 Decent minimalism
Is it possible to argue that something more than subsistence can be demanded from our
fellow citizens (other than those being lawfully punished), just because they are our fellow
citizens or fellow human beings? I think it can, though I am not sure that it is really a matter of
argument as much as trying to balance competing moral forces in a way that ultimately strikes us
as fair. To whatever extent we find ourselves applying the family-model intuitions beyond our
immediate circle, we will tend to think that justice requires more than subsistence. But how
16
much more? Here, the idea of natural rights to a share others’ resources seems to fade gradually
into a plea for charitable gifts, as our association with other individuals attenuates from family
and friends to perfect strangers. So, where do we draw the line as a matter of justice, given the
competing forces of our background intuitions? Rawls’s vague criterion, applied in a particular
way by him to promote his maximinimalist theory, is still useful in principle: what kind of
society would you prefer to join under conditions of objective ignorance? This is not essentially
different from Locke’s original criterion of reasonableness: the just social contract is the one that
reasonable people favor when acting to preserve as much as possible of their natural freedom,
while seeking protection from society against human enemies and the harsh forces of nature.
Perhaps there is some kind of welfare aid beyond mere subsistence implicit in such reasonable
social contracts. If so, whatever substantive transfers would be part of such a contract ought to
be seen as representing justice of a sort, defined by what might be called contractually proper
entitlements rather than absolute rights.
Some of my friends on the right claim to be strict libertarians, but they are no stricter in
their actions or particular opinions than my egalitarian and Rawlsian friends are on the left. All
of them understand implicitly that communal justice applies at least within families and among
friends, and, perhaps to a lesser degree, among colleagues and contractual partners of various
sorts. But they think of those things as a separate category of personal ethics, outside of political
theory, while they base their views about justice in general on the untrue assumption that we are
all strangers in the state of nature. In practice, most libertarians do attend to communality as a
factor in our social contract; they just don’t accept it as a distributive principle in theory.
One reasonable conception of a contractually proper level of entitlements is that of a
decent minimum, something above subsistence but below either equality or maximization. There
are some situations in life so degrading that we would never wish them on anybody if they can
be prevented, even though they permit people to stay alive and legally free. Under the veil of
ignorance, reasonable people might well give up some small increment of freedom (e.g. to keep
more of what they earn) in order to guarantee that nothing so awful ever happens to themselves
or their families. To that extent, we might say are all in this together – hard workers and the
lazy, fools and the wise. (As we sometimes say, in reflexive pity for another person barely
subsisting in a prosperous society, “There but for the grace of God go I”.) This implies that we
ought to guarantee each other, to the extent we can, a minimum level of benefit sufficient to a
decent and dignified life, regardless of inheritance, skills, talents, and good fortune. What
constitutes a minimally decent life is vague, of course, and surely depends on a society’s cultural
expectations as well as its overall resources. In the poorest societies, no more than primitive
shelter and a bowl of rice may be available for all. In present-day Western societies, though, it
seems that we would like to guarantee each other as entitlements at least such things as treatment
for disfiguring burns, housing with working toilets and no rats, clothing that isn’t ragged,
nutritious food, and education to the high-school level, or the wherewithal to afford such basic
goods and services.
17
Schematically, the decent-minimum theory looks like this:
y
post-tax
income
t: m = d
rm=d
s-
x
pre-tax income
Here, the minimum-level distribution d is above subsistence, but still certainly below its
maximum possible level r. How d compares to the egalitarian minimum e depends on whether
the universal level of benefit available in materially strictly equal societies is actually decent,
given the strongly depressive effects of total economic redistribution. The experience of East
Germany suggests that the socialist minimum can be at least somewhat decent; the experience of
North Korea proves that it isn’t always.
3.4 More-than-decent minimalism
There may even be something to be said for a social contract that entitles all citizens in
good standing even to something more than a decent and dignified standard of living. Countries
like Denmark and Finland seem to operate quite well with positively comfortable, not merely
decent, minimum standards of living. Of course, something is being sacrificed in their accepting
the concomitant high tax rates, both in terms of economic growth potential and, perhaps, in
diminished self-reliance, ambition, and other personal virtues. But it seems to make these people
happy that they have this sort of understanding with each other, in essentially the same way as it
makes ordinary families happy that each member is entitled to a comfortable share of the whole
family’s resources, on the expectation that every member, or enough, will pitch in where needed
to keep those resources at an acceptable level. There is no “tragedy of the commons” on a
family ranch. And the warmth derived from this much family-feeling in itself can make life
better for such people, even if they are somewhat less prosperous materially than they could
otherwise have been.
This is not to say that the small countries of Northern Europe practice anything like
universal brotherhood in its pure form. Their people still care much more about their own
friends and families than they do about random fellow citizens. They do not even attempt to
equalize welfare across their whole societies, or to raise their minimum standards of living to the
highest possible level – though many of their citizens would claim one or the other as a worthy
ideal, or even (in the grips of theory) a requirement of justice. What they seem actually to do is
balance enough family-model justice against stranger-model justice to produce a minimum living
standard that is pretty nice. This more-than-decent minimalism can be seen, then, as another
potentially reasonable compromise between the purer theories of libertarianism and
18
communalism. Just how much more than decent a country’s entitlements can be will depend,
again, on the degree to which the requisite fellow-feeling really exists among the citizens, as well
as what resources the society can make available, plus other economic and cultural variables.
Schematically, it looks like this:
y
post-tax
income
t: m > d
rm>d
ds-
x
pre-tax income
Such more-than-decent minimalism works best for societies that naturally function much
like families: small, stable, ethnically and culturally homogeneous groups, in which it is rational
to treat strangers with a good deal of trust. It is unlikely to work as well for larger countries,
especially ones as dynamic and diverse as the United States. Here in America, we have stark
cultural differences, numerous racial and ethnic groups at odds with one another, scant fellowfeeling across economic, regional, and even generational lines, a good deal of violent crime in
cities, and a lawsuit for every complaint. Though we are similar in overall development and
some features of culture to our European counterparts, we could hardly be more different in the
degree to which we view, or rationally ought to view, our fellow citizens as the equivalent of
family members. The thing that binds American society together – perhaps more than any other
place on Earth – is not family feeling, but a shared willingness to tolerate people that we do not
know or like, provided only that they do us no harm. This is, of course, a sad thing in a way. It
is nice to spend time with your friends, it would be nice to have more friends, and, if such a thing
were possible, it would be nice if everyone you met was already a friend. Something like this
could even be achievable if we all had as much in common as the Finns. But the United States
has chosen instead to bring millions of poor, uneducated foreigners into our modern society
every decade, to foster novelty and creativity among competing groups, and to focus on mobility
and individual initiative as tickets to the good life of the Western middle class, rather than on
portioning things out among a small, cohesive group of people who are already prosperous and
culturally unified.9
4.0 Two metaprinciples of justice
In the end, I see distributive justice as arising from the balancing of three basic moral
forces or intuitions: a libertarian approach to strangers, a communalist approach to family and
friends, and overall utility. The best balance to strike in law depends on the extent to which
people within a society really are strangers or really are family and friends, and on what is liable
to work best given all the relevant contingent factors. This leaves plenty of room for dispute,
19
and such disputes absorb a good deal of our country’s political energies. How should these
conflicts be resolved? If citizens cannot agree among themselves what balance is appropriate
between freedom and compassion, or how much we can afford to spend on redistributive
programs, or what will actually work to satisfy the greatest number to the greatest degree, what
should we do?
There are two broad principles that reasonably apply to resolving disputes over
distributive justice – hence, two metaprinciples of justice itself. The first is expertise. Some
people know much more than others do about tax policy, welfare policy, and economics
generally, or about human needs and desires, or about justice itself. These people are better at
deciding what amounts to justice in particular cases than anybody else. This is why we have
doctors and other experts deciding who receives the few transplantable kidneys available, why
we have judges deciding which laws apply in which ways to which cases, and why we have
nutritionists and other experts at the FDA deciding what additives can go into our food. One
approach to deciding what constitutes distributive justice for our society is to gather a sufficiency
of experts into one place, for example the executive branch of government, and let them work it
out together.
One problem with this principle of expertise is that it is hard to prevent corruption among
any group of human beings given power, including groups of experts. Another problem is that
expertise in fields like overall distributive justice does not really exist. Such nominal experts as
we have are totally unable to concur on fundamental theory in the way that doctors or CPAs do.
Rather, each such “expert” subscribes to one or another competing school of thought (such as
libertarianism or egalitarianism), and they tend to gather into hostile clusters of like-minded
thinkers. None of these people can claim to have real knowledge of these matters as long as
equally qualified experts disagree. There is no working consensus on the nature of justice yet,
so there are no genuine experts in the necessary sense.
The other metaprinciple of justice is democracy. Let everybody vote, regardless of how
much they know about the topic, and at least the majority will be to some extent satisfied with
the results. This process cannot function in a pure way in a large society like ours as it once did
in Athens, but it can be approximated pretty well through systems of representation like the US
Congress. Over time, Congress can be expected to reflect the general will of the society,
including both majority and minority views, in such a way that some rough justice will probably
be done in balancing the moral sensibilities of the whole society. The results will not be
recognizable in terms of any pure theory, or even necessarily coherent, but everybody will have
had his say to some extent, so all felt moral forces will be factors in the final product.
The most obvious problem with this process is that people are often ignorant, selfish,
resentful, and otherwise unqualified to say what should be done. Moreover, they seem to be too
easily led into one or another self-interested faction with little concern for the welfare of
outsiders. Actual democratic politics tends to resemble more a free-for-all among competing
special interests than an accurate weighing of the moral views of individuals. A more subtle but
also important problem is that, partly as an artifact of short terms for our representatives, hence
the constant reelection pressure on them, the effective discount on the interests of future
generations is far steeper than anybody really wants. As a result, Congress tends to enact
programs that quickly please large numbers of voters, and to ignore the long-range costs. The
main current entitlement programs in the US are almost certain to go bankrupt in the next few
decades, and Congress seemingly cannot act with sufficient foresight to prevent this happening.
20
The American welfare system is, then, as it must be, a big mess. The supposed experts
exercise their arrogant power through the executive departments, the people exercise their
ignorant power through a short-sighted congress, and the resulting system is unavoidably
partisan, corrupt, and inefficient. Nevertheless: if we step back several yards and squint, we can
perceive a distribution structure that reflects the basic principles of liberty, utility, and
communality in some kind of working balance. Quite a lot of our resources go toward
guaranteeing something like a decent standard of living for all Americans, quite a lot of
economic freedom is preserved, and we remain the wealthiest large country in the world.
21
NOTES
Friedman describes his proposal as a negative income tax – in effect, a proportionate refund for
people who make less than the minimum pre-tax income. I think that the most efficient way to
implement such a scheme would be to send every citizen a standard monthly check, then exact a
flat rate tax on all income through each person’s employers, investment banks, and so on. That
way, no special calculations would have to be done for most individual taxpayers, and most
people wouldn’t have to fill out any tax forms. The overall transfer result is just the same.
1
Robert Nozick distinguishes his libertarian view from what he calls “patterned” conceptions of
justice, meaning conceptions that required particular patterns of final distribution like
egalitarianism, and calls his own theory of entitlement, by contrast, “unpatterned”. Nozick’s
theory of justice is nevertheless patterned in the sense required for representation in
distribugrams, namely, that there is a fixed relation between pre-tax and post-tax distributions.
2
3
Costa Rica and Cuba make another pairing of roughly similar pre-Cold War economies, one of
which has prospered in a Western economic system while the other has stagnated under
communism. Cuba is often held up as an example of high minimum standards for literacy, life
expectancy, and other valuable things. But it is not at all clear that Cuba’s average standard of
living is as high as Costa Rica’s minimum, all questions of freedom and general prosperity aside.
Cuba’s current life expectancy at birth is actually slightly less than Costa Rica’s, and its claimed
literacy rate of 99.8% (like that of Tajikistan, Turkmenistan, and other communist and formerly
communist countries) is not much higher than Costa Rica’s 95.9%. Malnutrition is not
apparently a major a problem for either country, though Cuba’s rationing system provides very
minimal amounts of basic foodstuffs.
The old Beatles song “Taxman” expresses outrage at the extremely progressive British tax rates
of the 1960s, as the song’s taxman insists “there’s one for you, nineteen for me”. This 95% top
tax rate drove many Britons, including the Beatles themselves, to emigrate – to the overall
advantage of nobody.
4
5
Even Karl Marx spoke of justice as accordance to need, not strict equality, and promoted in the
Manifesto a steeply progressive, but not total, income tax.
6
This is obviously not the same thing as communism. It is also not the same thing as
communitarianism, an approach that focuses on the welfare of communities as whole things,
with interests potentially distinct from those of their members. What I am calling communalism
makes no assumption that communities have moral status in themselves, or even exist other than
as collections of associated individuals.
7
Note that some years ago Peter Singer, the most prominent living utilitarian and an important
leftist, spend a great deal of his money that could have gone to needy strangers on private
nursing care for his own dying mother. He keeps apologizing for this instance of what must be
called misbehavior from his own theoretical point of view, but it seems obvious to others that he
did the right thing, because he was a good son. See Berkowitz, “Other People’s Mothers” [cite]
22
8
Recent studies [cite] seem to show that political conservatives (not quite the same group as
economic libertarians) actually give more to charity on average than progressives do. But
perhaps this can be explained by their attending church more often, where weekly donations are
expected, and are to be given more-or-less openly.
9
As some Northern European countries like Sweden have recently had trouble absorbing large
numbers of Middle-Eastern immigrants, and have also run into problems funding their very high
entitlements, they have been edging away somewhat from their communalist welfare-statism
toward distributive policies more like those of the United States.
23
Download