CHAPTER IV
THE OBJECTIVES OF DISTRIBUTIONAL POLICIES
In the last chapter we proceeded on the assumption that the basic objective of economic policy was in some sense to maximise the sum total of individual utilities or enjoyments. But it is not at all obvious that this is a suitable criterion. One rather anomalous result of this procedure may serve to throw doubt upon it; the more the policymakers make their interpersonal comparisons of utility in such a way as to emphasise inequalities in total utilities as between different individuals, the less importance does the criterion of maximising total utility attach to the redistribution of income. This paradoxical result is illustrated in
Figure 4. Along the horizontal axis we measure an individual's consumption. Up the vertical axis we measure his total utility. Consider two individuals A and B with the same needs and tastes (i.e. the same utility functions) but with different consumption levels,
and
. Let the height
GH measure
i.e. the
total utility of individual A when his consumption is equal to
. Consider two utility functions passing through the point
H, both with the same slope (i.e. the same marginal utility of consumption at this point
H), but curve
having much less curvature than curve
. For the higher level of consumption of
at the point
D the total utility with curve
will thus be greater than the total utility of curve
(FD > ED). Moreover the slope of the curve
is also greater than the slope of the curve
.
Figure 4
We suppose individual B to have the same utility curve as A but to have
instead of only
units to consume. Consider now taking one unit of consumption from B and giving it to A (d
= d
). With both curves A will gain the same amount of utility d
; but with curve
B will lose much more utility than with curve
(d
> d
). The distributional effect, i.e. the possibilities of increasing total welfare through redistribution, will be much greater with curve
than with curve
, i.e. much greater in the case where the inequalities in total utilities are less.
The criterion of the sum of individual utilities is not concerned with inequalities between utilities. It has nothing to choose between a policy which will result in a given addition to the utility of a poor man and a policy which will result in an addition of the same amount to the utility of a rich man, although the former policy will result in a less unequal and the latter in a more unequal society. This criterion simply regards a redistribution of income as a possible efficient measure for increasing the sum total of utilities.
Let us proceed with the analysis for the time being on the assumption that all individuals have the same utility functions, but that for each of them the marginal utility of consumption declines as their consumption increases. In this case an equal distribution of a given total amount of real consumption will be needed to maximise the sum of individual utilities. The transfer of £1 worth of consumption from a rich man (to whom the marginal utility of money is low) to a poor man (to whom the marginal utility of money is high) will always increase the sum total of utility. Equality is an efficient tool for raising total utility.
Consider now two economic situations. In Situation I there is a given total of consumption equally distributed. In Situation II there is a larger total amount of real consumption unequally distributed among the same population. But the inefficiency of inequality in Situation II is just offset by the higher total level of consumption in Situation II, so that the sum total of individual utilities is the same in Situation II as in Situation I. The criterion of the sum total of utilities would have nothing to choose between the two situations.
But if one is concerned with inequalities as such, one would presumably prefer Situation I with a given total utility equally divided to Situation II with the same total utility unequally divided. Indeed, one can go further. If the real income of Situation I were still further reduced so that Situation I had a smaller total of utility than Situation II but a more equal distribution of that smaller total, one might still prefer Situation I to Situation II. One would have some pay-off in one's mi...