Part III, Chapter XVI
FAIRNESS AS A VARIABLE RELATION

§ 1. As between persons exactly similar in quality fair wages mean, in accordance with what is said in § 1 of the last chapter but one, equal wages, subject only to adjustment for differences in the incidental advantages enjoyed; and they mean this in all circumstances whatever. As between dissimilar persons fair wages mean, in like manner, in any given set of circumstances, wages which are proportioned to efficiency and which, therefore, bear to one another a certain definite numerical relation. At first sight it may be supposed that this numerical relation, like the relation of equality required for fairness between similar persons, will be the same in all circumstances. This, however, is not so; for the reason that the comparative efficiencies of different persons possessing given qualities and capacities are not the same in all circumstances. It is the task of the present chapter, to make this matter clear. For that purpose it is convenient to distinguish differences between workpeople under two heads: (1) differences in the degrees of a given kind of ability; (2) differences in the kinds of ability that they severally possess. I shall begin by ignoring differences of kind, and assume that there is only one kind of ability distributed to different people in different degrees.

§ 2. The key here is given by Marshall in his discussion of the fertility of land. He shows, it will be remembered, that, even when we are considering only fertility in respect of one kind of crop—which corresponds to my one "kind" of ability—the fertilities of two dissimilar pieces of land do not bear to one another a single numerical relation which is the same in all circumstances. On the contrary, this relation will be different in different states of demand, different conditions of capital supply, and so on; and it may even happen that a piece of land, which is more fertile than another in one set of circumstances, will be the less fertile of the two in another set. In like manner we now find that the net product (conceived as marginal) and, therefore, the efficiency of workman A, is liable to bear a numerical relation to the net product (similarly conceived) of workman B, which varies as the supply of tools and so on, with which their labour is assisted, varies. Thus, though in any given set of conditions there is some definite numerical relation between the wages of the two men which is fair, this relation does not depend simply upon their personal qualities, but is liable to change as external circumstances change.

§ 3. Marshall's analysis of fertility suggests the further inference that progress tends in a general way to level up the efficiency of workers possessing less ability relatively to that of workers possessing more ability, and so to make the fair ratio between their wages approach more closely towards unity. He writes: "Independently of any change in the suitability of the prevailing crops and method of cultivation of special soils, there is a constant tendency towards equality in the value of different soils. In the absence of any special cause to the contrary, the growth of population and wealth will make the poorer soils gain on the richer."^*59 If this is so, it would seem that an analogous proposition ought to hold good of workpeople possessing different degrees of ability engaged in producing the same thing. Marshall's proposition can be rigidly proved for pieces of land which differ from one another in such wise that the yield to an r^th unit of investment in one exceeds the yield to an r^th unit in the other by the same constant amount for all values of r. Marshall illustrates this case by means of a diagram.^*60 But, of course, pieces of land need not be related to one another in this way. It may be that acre A yields a lower return than acre B for each does of investment up to the R^th dose, and thereafter yields a progressively greater return. In this case acre B is the worse when (R + k) doses are being invested, but its rentability—which is a measure of its fertility—so far from approaching towards that of acre A, is likely to fall further behind it, when a rising demand for the products of land causes the number of doses to increase beyond (R + k). Hence the question whether or not the rentabilities of acres of different capacities for producing a given crop tend towards equality as demand grows depends upon what sort of relation usually holds between what Marshall calls the produce-curves (i.e. the lists of yields to various doses of investment) of different acres picked at random. In my judgement the usual (probable) relation is such as to warrant Marshall's conclusion.^*61 If this is so, we may add, by parity of reasoning, that the wage rates, that tend to be paid and are "fair" between workpeople possessing different degrees of the same ability, will tend towards equality as population and wealth increase.

§ 4. Account has next to be taken of the fact that men are born with different kinds of ability as well as with different degrees of it. No doubt, it is possible for two men or two acres of land to differ from one another in such a way that the ratio between their capacities, when the price of capital stands at any given level, is the same for all purposes. This, however, is not likely to happen often. With capital at a given price, one man will generally stand in one relation to another in respect, say, of physical strength and in a different relation in respect, say, of mathematical ability; just as one acre will be 10 per cent better than another for growing barley but 20 per cent better for growing wheat. Indeed, it may easily happen that the order of merit between two men is different for different purposes. Mr. Tunney has the advantage of Herr Einstein in a boxing ring but not in a laboratory. It follows that, when either the public demands for the services which different sorts of ability can render change relatively to one another (whether as a result of changes in taste or of changes in the distribution of income among people with different tastes), or when developments of technique alter the relative importance of different sorts of ability in producing services and things, (1) it will be profitable to alter the relative amounts of capital invested in training different types of men, and (2) the relative efficiencies of different types, both net and gross (i.e. reckoned as including interest on capital), and so the relative earnings that are "fair" between these types, will change. Thus, when the war raised greatly the demand for the services of common soldiers, the efficiency and earnings of unskilled workers rose while those of musicians fell; and, had the war gone on for ever, children born with a special faculty for soldiering would have had a relatively high, and those born with a special faculty for music, a relatively low, expectation of earnings. On the other hand, technical developments enabling a great many industrial operations to be performed by elaborate machines associated with highly skilled attendants cause the efficiencies and the earnings expectation of intelligent children to rise relatively to those of children born with the temperament and physique of an ox. Again, if an invention causes blindness to be less of a handicap in some occupation than it has been hitherto, blind men will tend to flow into that occupation and other men out of it, until finally, their relative efficiency having risen, the earnings of blind men are everywhere slightly higher than before relatively to those of other men. With a wider sweep but a like aim Bateson wrote: "The fact that families or individuals rose into prominence or dropped into obscurity when the great industrial development of this country began, does not prove that the strains from which they came ought previously and in differing circumstances to have been in different relative positions. In various circumstances various qualities are required for success."^*62 The efficiencies of different types of people are thus determined, not merely by their natures alone, nor yet by their natures in conjunction with the supply price of capital, but by these things together with the state of demand for different sorts of service and the state of industrial technique in various occupations. It is not possible, I think, to say what kind of change in relative efficiencies, and so in relative earnings—a scattering or a tendency towards greater equality—future developments in tastes and technique are likely to bring about.

§ 5. The significance of the preceding analysis becomes apparent when the question is asked whether or not the wage ruling in some one industry is or is not fair relatively to that ruling in others. The main issue here is, of course, concerned with the relation between the wages of average workpeople in different industries. For, when the wage that is fair for an average worker in one industry relatively to that of an average worker in another is known, to discover the allowance that should be made for those above or below the average is a comparatively simple matter. If we knew, by direct judgment or otherwise, that these average workers were in all respects exactly alike, we should know also, in accordance with what was said in Chapter XIV. § 1, that fair wages would be equal wages. But, if we do not know that the workers in the industries we are comparing are exactly alike, the problem is much more complicated. It is true, as was shown in the section cited above, that, if wages are not proportioned to efficiency—as measured by marginal net product multiplied by price—they cannot stand in a fair relation to one another. But it is not true that, if they are proportioned to efficiency in this sense, they must stand in a fair relation. They only do this if a second condition is also fulfilled; namely, if labour as between the industries is distributed ideally, i.e. so as to maximise the national dividend in the wide sense of Chapter IX. § 2.^*63 How in these circumstances is it possible in practice to decide whether or not the wage ruling in one industry relatively to others is or is not fair? In certain conditions a workable method is available. It may be possible to find some model, or standard, year, in which there was a general agreement among employers and employed in an industry that the wage rate there was fair relatively to that ruling in other industries. This wage rate will be our starting-point. Having ascertained it, we try to discover by statistical inquiry in what proportion wage rates in other industries have changed since our standard year. Suppose that they have risen by 20 per cent. Then, if no obvious change has occurred in the comparative average qualities of the workpeople in our industry and in other industries, we conclude that the fair wage for our industry now is the wage that ruled there in the standard year plus a rise of 20 per cent. This method, which during the pre-war period was, in effect, pursued for a long time by Conciliation Boards in the coal industry, often enables a reasonably close determination of fair wages to be made. The analysis of the present chapter shows, however, that it only does this, and is, therefore, only applicable, provided that nothing has happened meanwhile to modify appreciably the relative aggregate demands for the kinds and degrees of ability that are chiefly utilised in our industry and in other industries respectively. If the popular taste for horse-racing had greatly expanded and that for litigation greatly diminished since the date of our model year, the abilities which go to make a good jockey would have become much more valuable relatively to those that go to make a good lawyer, and the fair wage for jockeys as against lawyers would now be much higher than it was in that year. So far, of course, as the relative supplies of the two sorts of abilities depend, not on natural endowment, but on differences in the amount of money invested in the training and nurture of persons initially similar, we should expect that the number of persons entering the occupation where the demand had risen would increase relatively to the number entering the other, and, therefore, that after a time the difference between the new and the old fair relation would be pro tanto lessened. Thus a falling off in the relative demand for skilled as compared with unskilled workers would alter the fair relation between the wages of the two classes to a less extent in the long run than it did immediately. If we may suppose that a relative rise in the demand for the services of people possessing the natural endowments of a jockey, as compared with those possessing the natural endowments of a lawyer, would cause the former class to have relatively more children—sharing, we may presume, their natural endowments—the difference between the new fair relation, as ultimately attained, and the old one, would be lessened still further.