Tibor Scitovsky
Welfare and Competition
Irwin, Homewood, Illinois, 1971
Chapter 15
EQUITY AND THE DISTRIBUTION OF INCOME
(pp. 285-294)
Many comments have already been made in preceding chapters on the distribution of income that would come about under universal perfect competition, on equity as a desirable aim, and on the need to weigh considerations of equity against those of efficiency. It is time now to pull all this together and supplement it with what little else the economist can say on a subject, which he can analyze but on which he cannot, in his professional capacity, make a judgment. That this is very little, indeed, will become evident in this chapter; and the reader is warned once again that the preoccupation of most of this book with problems of efficiency, with only a single chapter and a scatter of paragraphs devoted to the problem of equity, reflects the state of our knowledge and not the relative importance of the two subjects.
It became evident in Part II that efficient resource allocation calls for certain price relations; and since the prices of productive services determine the incomes of their owners, efficient economic organization clearly has implications for the distribution of income as between owners of similar and different services, and as between members of the same and of different occupations. As pointed out already (cf. especially pp. 108-9 and 184), this has a lot to do with equity; but the main problem of equity centers on the inequality of incomes between rich and poor. To discuss this, one needs some index, or measure if possible, of inequality.
1. MEASUREMENT OF INEQUALITY
Several representations and measures of the inequality of income have been developed; of these only the simplest, the Lorenz curve, can be presented here. If one imagines all households or persons arranged in order of increasing incomes, one can divide them into groups of equal size and observe the inequality of income distribution by comparing the sum of incomes earned by the members of each group. If each group containing X percent of the population earned X percent of the population's total income, incomes would be equally divided. Inequality shows up if each successive group earns a larger total income, or a larger proportion of
285
national income, than the previous group. Inequality would be at its extreme if the last group earned all the income and the other groups earned nothing.
It is customary to distinguish five or ten groups (fifths or deciles), express the number of households or persons in each group as a percentage of the total population, the incomes they earn as a percentage of the population's total income, and to cumulate both percentages. Table 15-1 shows such an income distribution for the United States. By representing these data in a diagram, one obtains a so-called Lorenz curve, whose divergence from a straight-line diagonal gives a visual image of the degree of inequality. A simple way to express this numerically is to calculate the Gini coefficient of concentration, which shows the area between the diagonal and the Lorenz
TABLE 15-1
U.S. Income Distribution by Households in 1967
Households Share of Cumulated Cumulated
Grouped by Total Income Share of Total Proportion
Income Earned Income Earned of Households
Poorest tenth 1 % 1 % 10 %
Second tenth 3 4 20
Third tenth 5 9 30
Fourth tenth 6 15 40
Fifth tenth 8 23 50
Sixth tenth 9 32 60
Seventh tenth 11 43 70
Eighth tenth 13 56 80
Ninth tenth 16 72 90
Richest tenth 28 100 100
Source: Statistical Abstract of the United States, 1969 (Washington, D.C.: U.S. Government Printing Office, 1969), Table 471, p. 321.
curve as a proportion of the total area under the diagonal. The Gini coefficient, therefore, can have values between 0 for complete equality and 1 for extreme inequality or complete concentration. Figure 15-1 shows the Lorenz curve corresponding to the income distribution for 1967 given in Table 15-1; its Gini coefficient has the value .481. Another and even simpler numerical index of inequality is the percentage of society's total income that would have to: be redistributed from rich to poor in order to equalize incomes. This has a' value of 28 percent for the same income distribution. It hardly needs adding that all these measures of inequality can be used for expressing the inequality not only of incomes but of wealth, of the distribution of any particular good, or of anything else. It also needs stressing that they are all very crude. Lorenz curves, for example, can differ not only in their distance from the diagonal but also in their shape, showing not only greater and lesser inequalities but also different degrees of inequality in different ranges -:of the income scale; and such detail, of course,
286 Top
FIGURE 15-1 (missing from this presentation, HHC)
gets lost in the calculation of something as simple as the Gini coefficient. There also exist many other and more sophisticated ways of expressing and measuring inequality; but we do not have the space to discuss them here.
2. WHAT IS CONSIDERED EQUITABLE
Equity is a matter of conscience, not of economics. But since almost all economic actions and policies affect the distribution of well being, economists who analyze these actions and policies, and often also advocate or criticize them, must be able to analyze and even to judge also their distributional effects.
The income distribution that instinctively seems ideal is an equal one. While this is an ethical judgment, it can be supported by the economic argument that the equal distribution of a given amount of total income would maximize the sum total of human satisfactions. This seemed obvious to an earlier generation of economists who believed in the law of diminishing marginal utility and in the equality of man, in the sense of people's equal ability to enjoy themselves. Today's economists no longer believe in the latter; but the same conclusion can also be derived from the assumption of a random distribution of people's ability to enjoy themselves.1
1 The first attempt to do this, contained in A. P. Lerner, The Economics of Control (Macmillan, New York, 1944), Chapter 3, contained an error which is corrected in M. Friedman, "Lerner on the Economics of Control," Journal of Political Economy, Vol. 55 (1947), pp. 405 16.
287
The only weakness of the economic argument for perfect equality lies in the assumption of a given amount of income. The amount of income generated and available for distribution depends on the strength of the incentives to produce income; and its equal distribution would eliminate the economic incentive to produce it. Complete equality of incomes therefore seems an undesirable goal, in view of its apparent incompatibility with the provision of economic incentive to produce income. As long as society prefers economic to other incentives as the main spur to economic activity, it must also accept some income inequality as the unavoidable cost of providing such incentive.
To find the best compromise between maximizing the national income produced and equalizing its distribution seems like a typical problem of economic analysis. Unfortunately, we lack the information to resolve it, knowing next to nothing about the effectiveness of individual income as a spur to the individual's economic activity, and nothing about the cost of a marginal addition to the national income in terms of the increased inequality this would require.
As long as total or virtually total ignorance persists on this subject, the toleration of income inequalities must necessarily remain a matter of individual conscience; but we shall discuss what to us seem the main considerations that guide people's consciences. Most people seem to accept, though with great individual variations in degree, inequalities in the distribution of income, provided these conform to certain principles and do not deprive those at the bottom of the scale of the essential necessities of life. The principles are the principle of nondiscrimination and the principle of equal opportunity. The former calls for equal pay for equal services and requires that differences in the prices earned by different services be no greater than necessary to assure their supply in quantities that conform to society's needs and preferences. This is the kind of income distribution implied and brought about by an efficient allocation of resources and was discussed as such in section 6 of Chapter 5. It does not involve a uniquely determined degree of inequality, partly because the distribution of skills depends on the cost and availability of education and training, partly because the distribution of income is only one of several determinants of the distribution of welfare, and most importantly because income is received not only for personal services rendered but also for the use of such resources as land and capital.
The second principle, the principle of equal opportunity, requires that if inequalities are justified by differences in occupation, skill, and training, there should at least be equality of opportunity to enter the favored occupations and acquire the highly paid skills. This principle is at the root of the argument for free and equal education, of John Stuart Mill's early advocacy of confiscatory inheritance taxes, and of the high and highly progressive actual inheritance taxes of most capitalist countries.
288 Top
The third condition of equity is the assurance of essentials even to the poorest. This can be looked at in a variety of ways and is being approached differently in different countries. In the United States, with its long tradition of expressing everything in terms of money, there is a growing belief in the need to provide everybody with a minimum of money income, without regard to his contribution to the social product. Experiments are under way with systems of income taxation that would levy a tax above, pay a subsidy (levy a negative tax) below the subsistence minimum and provide a minimum income without destroying the incentive to work.
The alternative approach is to look behind money, at the distribution of goods and services, both those that money can buy and those available free. One can then judge the degree of equity by the number and importance of the goods and services that are more-or-less equally distributed and specify the requirements of equity in terms of the goods and services one wishes to see so distributed.
As a general rule, goods and services that are free or, while scarce, are made available free, are distributed equally or according to people's needs and wants. This explains why, other things equal, a society seems the more egalitarian the larger the number of amenities with free access. The poor are not so poor in London, Paris, or San Francisco, where they have free access to many beautifully kept public parks, than they are in Chicago, Washington, or small-town America, where these amenities are almost nonexistent and one must own a garden to look at a flower. Again, the poor are not so poor in societies where much of everybody's life is lived in public places and generates external economies for all to enjoy. Walking by a sidewalk cafe is a free good almost as enjoyable as sitting in it, provided lots of people also sit in it. Some collective services, however, though available free, are quite unequally distributed; we mentioned already the many municipal services whose distribution is proportional to living space and so discriminates against the poor, who live in crowded conditions.
As to the distribution of goods and services that have to be paid for, one must note that a given inequality in the distribution of money income does not result in a similar distribution of all the goods and services on which that income is spent. On the contrary, it generally leads to a different degree of inequality in the. distribution of each item of consumption, ranging from an equal or near-equal distribution of the cheaper necessities to extreme inequality in the , distribution of the most expensive luxuries. For example, if statistics were available of the consumption of bread by income groups in the United States, they would probably show a virtually equal distribution, with the top 10 percent of households consuming no more (and possibly less) bread than the bottom 10 percent. By contrast, similar statistics on the purchase or ownership of yachts would almost certainly show extreme concentration, of virtually all yachts in the top decile. Between
289
such extremes, there is a continuous range of in-between items with varying degrees of inequality.
The interesting question to which we must now address ourselves, is why some commodities, and which ones, are distributed equally or near equally in an economy where money incomes are unequally distributed. It is obvious to begin with, that these will always be necessities, which even the poorest will buy if they can afford them, and which even the richest will not buy too much of, because their demand is biologically limited. Secondly, for the poorest to afford a given necessity in adequate quantities, its price must be low enough for their total expenditure on it not to absorb too large a part even of the poorest budget. Further factors are the average level of income and the inequality of its distribution, since these determine the size of the poorest budget. To assure, therefore, the equal or near-equal distribution of a given necessity not yet so distributed, either its price must be lowered, or average income raised, or income inequalities mitigated. The last is obvious, the first two are less so. Yet, a rise in average incomes, other things unchanged, does increase equity, at least in the limited sense of making the distribution of many necessities more nearly equal and increasing the number of necessities distributed in a near-equal way; also, a policy aimed at lowering the prices of necessities can have very similar effects to, and serve as a substitute for, a policy of mitigating inequalities in the distribution of money income.
One lesson to be learnt from the above analysis is that equity is not a matter merely of income distribution but something much more complex, of which income distribution is only one determinant. If we are right in believing that general access to essential necessities is an important element of equity, then any given inequality of money incomes becomes more acceptable in a rich than a poor economy, it also becomes more acceptable when necessities are relatively cheap than when they are relatively dear.
This brings us back to the alternative way of assuring egalitarian access to essential necessities. If a necessity is the more accessible the lower its price, then Government can assure general access to it by subsidizing its production if necessary. Some European countries subsidize the baking of bread in order to assure access to a basic food even to the poorest.2 Rent control, i.e., a ceiling on rents, has often served a similar purpose and was often retained for long periods, until its disadvantages on efficiency grounds seemed to outweigh equity considerations.
A limiting and extreme form of subsidy is the free provision of a good by the State. In Chapter 13, where we dealt with indivisible products, which by nature are available in equal quantities to all, we argued, from the point of view of economic efficiency, that when it is not feasible to let people buy at a uniform price in accordance with their preferences
2 The U.S. Food Stamp Program, discussed in Chapter 3 in another connection, differs from this in that the subsidy is not universal but confined to the poor.
290 Top
and ability to pay, the next best thing is to make them pay for the same quantity in accordance with their ability to pay.
Here, we are concerned with equity; and from this point of view, the second solution seems better than the first. Whenever incomes are unequally distributed, it is more egalitarian to give the same quantity to everybody and make them pay in proportion to their incomes, than to offer them the same price and let them buy according to their incomes. In other words, proportional taxation, discussed earlier as an efficient way of paying for indivisible products, which are equally accessible to all, now appears as an equitable arrangement as well. So much so that the demand often arises to make available through the State, and finance out of taxation, essential goods and services to which it is particularly desirable to give equal access, whether or not they are indivisible, whether or not private enterprise and the market could provide them with equal efficiency. The obvious examples are education and medical care.
The argument for free and equal access to education was mentioned earlier in this chapter. It is necessary for assuring equality of opportunity to enter any occupation and acquire any skill, which in turn appear as requirements of equity. Since education is a divisible product, equal access leads not to an equal distribution but to one according to wants and needs, which is better. The free availability of any divisible product tempts people to fill their demand to saturation and so cause an uneconomical diversion of resources from other products. In the case. of education, this tendency can be offset by rationing access to each educational stage on the basis of academic performance at the previous stage. 3
More instructive in many ways is our other example, medical care. As medicine becomes more effective, it also becomes more costly especially in the United States where almost all the advance has been of the quality improving kind, with virtually no attention paid to advance that would reduce costs. 4 As a result, the cost of medical care in the United States
3. Whether this is actually done or is politically feasible to do is an altogether different question, which is important, but with which we are not concerned here.
4 The probable explanation is highly relevant to the subject matter of this book. The market for medical services in the United States is organized in a way that has eliminated virtually all the downward pressure on prices that would normally exist in a competitive market. The ultimate buyer, the patient who foots the bill, is seldom given a choice of treatment, or hospital, or the specialist to consult, all of which are prescribed to him by his physician in a fairly authoritarian manner. In addition, the intercalation of Blue Cross and Blue Shield type agencies between buyer and seller, with these agencies largely controlled by the medical profession, removes what vestiges remain of the buyers' bargaining power, which was small to begin with, given the highly uninformed nature of this market. Hence the apparent paradox that all the innovations to reduce costs are limited to countries with some form of national health insurance and therefore some body or agency with an interest in keeping costs low. Examples are the use, in many of these countries, of less highly trained personnel for dealing with routine care (feldschers in Russia, general practitioners in Great Britain, midwives in the Scandinavian countries, lower-grade dentists side by side with dental surgeons in Germany, etc.).
291
increased much faster than the Consumer's Price Index, and somewhat faster than per capita disposable income ;5 thus pricing it out of the poor man's reach and making its distribution increasingly unequal, despite an unchanged distribution of money income. This explains the ever more urgent demand in the United States for some form of national health service or insurance, as a means of making medical care available to all. A major problem, here again, is how to prevent an excessive diversion of resources from other uses in response to people's tendency to push to saturation point their use of anything made available free. The French try to solve the problem by national health insurance refunding only 80 percent of the cost of medical care; the British charge a nominal fee and try to ration through queueing and other informal means. Nowhere is the problem completely solved; but neither is it too great, to judge by the fact that no country with a national health service spends as great a percentage of its national product on medical care as the United States, where medical care is still a market good, the demand for which is rationed by price.
We can now bring together all the various factors that together determine inequalities in the distribution of welfare. The first of these was the distribution of money income; and to this must be added the distribution of wealth, since the two together determine the distribution of the ability to pay. The progressivity or regressivity of taxation is a closely related factor, since it too influences the distribution of income and wealth after taxes. The average level of income and accumulated wealth is also relevant, being one of the determinants of the number of necessities everybody is able to afford.
Next in importance is the availability of free goods and services, whether these are free by nature, represent the external economies of someone else's activity, or are collective goods or services paid out of taxation and made available free by the State. As mentioned in Chapter 1, the dividing line between free and scarce resources can shift in time; it can also be different in different countries and civilizations, depending on climate, density of population, availability of natural resources, the nature of social institutions, the legal system, etc. etc. In some countries, wives had to be bought and poor men were condemned to celibacy; in others, private property in land is limited in the interest of giving everybody free and equal access to nature (e.g. by restricting private property over ocean beaches and by making it subject to various rights of way). All this clearly affects the welfare of the poor in relation to the welfare of the rich.
External economies were discussed in the last chapter, where we argued that by internalizing them that is, by making the beneficiaries pay for their benefits economic efficiency can always be improved. While this is so, the internalization of external economies also has another effect: it
5 I am referring to the trend of the 1950s and 1960s, as this appears from the official statistics published in the Survey of Current Business and the Monthly Labor Review.
292 Top
restricts their benefits to those who can afford to pay for them. This means that the more highly commercialized is a society and the more its every activity and manifestation is subject to economic transactions, the fewer become its free benefits with equal or random access, and the more the distribution of purchasing power becomes controlling for the distribution of goods and services. The hankering of some people, therefore, for a less commercial and less money minded world turns out to be not merely a romantic notion but also a preference for a world in which some, perhaps many of the good things of life were distributed independently of the distribution of income and wealth.
No less relevant to equity is the number and nature of the third category of free goods and services: those that are scarce by nature but provided free or at nominal cost by the State. We saw that most indivisible products are necessarily provided in this way; but so to provide some other products as well has become an important tool of social policy in many countries. The examples of free education, free medical care, and subsidized food have already been mentioned; there are many others, ranging from subsidized public transportation, public theatres, concerts and some other cultural goods in Europe to free legal assistance in Scandinavia and Britain.
Finally, the relative prices of the different goods available in the market are also highly relevant, since they determine the number and identity of the goods that all can afford. Here again, the average level of prices and discriminatory pricing as a tool of social policy are equally relevant. 6 The changing pattern of prices, brought about by economic development, can increase as well as diminish equity; in the latter case, it can also increase the pressure of public opinion for social policies mitigating inequities.
BIBLIOGRAPHICAL NOTE
Far too little has been published on this important subject. For a discussion of income distribution in the United States, .see Selma Goldsmith and others, "Size Distribution of Income since the Mid Thirties" Review of Economics and Statistics, Vol. 36 (1954 ), pp. 1-32. More recent statistical data are occasionally published in the Survey of Current Business and reprinted in the Statistical Abstract of the United States.
For the argument in favor of an equal distribution of income, see A. P. Lerner, Economics of Control (New York: Macmillan Co., 1944), Chap. 3; M. Friedman, "Lerner on the Economics of Control," Journal of Political Economy, Vol. LV (1947), pp. 405 16, reprinted in his Essays in Positive Economics (Chicago: University of Chicago Press, 1953 ) ; and W. Breit and W. P. Culbertson, Jr., "Distributional Equality and Aggregate Utility: Comment" American Economic Review., Vol. LX (1970), pp. 435 43. The rest of section 2 is based on my paper on
6 See section 1 of Chapter 25 for the effects of price discrimination on equity.
293
"Equity," printed in my Papers on Welfare and Growth (Stanford, Calif.: Stanford University Press, 1964 ) .
There is a voluminous if not very satisfactory literature on what determines the distribution of income. For an introduction to this, see the survey papers in L. Soltow (ed.), Six Papers on the Size Distribution of Wealth and Income (New York: National Bureau of Economic Research, 1969) and Income Distribution by Size: Selected Theoretical and Empirical Issues (Princeton, N.J., National Bureau of Economic Research, 1964).
294