and in the supply of, different kinds of work.1 Conditions are conceivable, under which the value of all kinds of work would be equal, but in the actual world no such equality is found. The chief reasons for this are as follows.

§3. Taking first the demand side of the problem, we find that changes are continually taking place in the relative demand for different commodities and services. Some of these changes act directly through the tastes, or relative incomes, of consumers, others indirectly through changes in methods of production, and hence in the demand for different sorts of work, machines and materials. In other words, some changes in demand are in respect of consumers' goods, others in respect of producers' goods. Again, some of these changes in demand are more or less temporary fluctuations, others are of a more continuous and permanent character. In all cases an increase in the demand for one kind of work relatively to another will, at any rate for a time, increase the value of the former kind of work relatively to that of the latter. Further, even if no changes ever took place in relative demand, the demand would be greater for some kinds of work than for others and, unless the supply of different kinds of work could adjust itself to these differences, differences of value would result.

§4. The supply side of the problem is more complicated. In the first place, it is not merely the relative money income obtainable in different occupations, which determines their relative attraction for workers seeking employment in them. Other important elements in attractiveness are the general nature of the work, its agreeableness or disagreeableness, the social esteem in which it is held, the prospects which it is thought to offer of exceptional

1 It should not, of course, be overlooked that much unpaid work is still performed in modern communities, e.g., the very important household work of women in their own homes.

success, or security of employment, and its supposed chances of sickness or accident. If all workers were free, at the beginning of their working life, to choose any occupation, we should expect, apart from changes in demand, to find average individual earnings in different occupations over a year, or other sufficiently long period of time, roughly proportionate to current estimates of the relative attractiveness of the different occupations. Such estimates are, of course, liable to error. Individuals are apt to miscalculate chances, and in particular their own chances of securing the exceptional prizes of an occupation on the one hand and of escaping unemployment, sickness and accident on the other. In any case, however, differences in the attractiveness of different occupations are sufficient to account for considerable differences in the value of different kinds of work.

But a more important cause of such differences in value is that all workers are not free, at the beginning of their working lives, to choose any occupation. In order to enter certain occupations a long and expensive training is necessary, involving the expense of maintenance during training, which is beyond the means of the majority of workers. These occupations are partially barricaded by high costs of entry, and according as these costs of entry are relatively large, the number of workers able to enter the occupation in question is relatively small, and the value of the work performed is relatively high. This is the case in most of the " professions." It also follows that there is a great concentration of workers in occupations where the costs of entry are very small or nonexistent, and that the value of work performed in these occupations is relatively very low. This is the case with what are commonly known as the unskilled occupations. Skill, however, is a relative term. A clerk's work is skilled where the majority of workers is illiterate, but becomes unskilled when, through improvements in educa

tion, few are unable to read and write. Where, as in most modern communities, freedom in the choice of occupations is greatly restricted, owing to the lack of educational facilities necessary to secure such freedom, particular occupations, or groups of occupations of about equal attractiveness, become largely hereditary. This fact is a further cause of inequality in the value of different kinds of work, since, the poorer any section of a community is, the higher, generally speaking, is its birth-rate. Such differences in birth-rates cause a further increase in the number of workers in occupations, where the value of the work performed is already relatively low, and this increase makes its value relatively lower still.

§5. In addition to lack of educational facilities, there are two other causes which may raise the relative value of work performed in certain occupations. The first of these causes is certain, the second is hypothetical. The first is deliberate restriction of numbers, enforced by the workers already in the occupation. Restriction of this kind implies an effective degree of organisation among these workers. It is illustrated by trade union rules limiting the number of apprentices or learners. It is a form of monopolist action, which depends for its continued success upon the power of the monopolist organisation to control a sufficient number of the workers concerned, and to adapt itself to changes in productive processes, especially the introduction of machinery, which may diminish the demand for skilled and well organised workers and increase the demand for less skilled and probably less well organised workers.

The second cause, which is hypothetical, is the possibility that in some occupations there is needed for success an exceptionally high degree of natural ability, which training and education alone cannot give. It is sometimes suggested that ability to manage large businesses is a case in point. If there be such cases, complete

freedom of entry into such occupations would not be sufficient to reduce the value of work performed in them to a value proportionate to the relative attractiveness of such work. But, in view of the failure of all modern communities to secure even approximate freedom of entry into all occupations, it is impossible at present to test the truth of this hypothesis. Moreover, the exceptionally large incomes which some workers might secure, even if entry to all occupations were completely free, might be due to the exceptional amount of effective work which they performed, and not to any need for exceptional ability in all members of their occupation.

§6. Finally, in addition to differences in the value of different kinds of work, we must notice that at any given time there are often large differences in the value of the same kind of work. Such differences are greatest between different countries, but in this case they are not so great as they appear at first sight, the comparative value per unit being based upon different units, that is to say upon different amounts of work, according to the general productive power of different countries. In so far as the differences in value are real, they are due to the fact that, generally speaking, workers, capital and commodities all move less freely from one country to another than from one part of the same country to another.

But differences in the value of the same kind of work are also found within the same country. These are sometimes due to the methods of payment for work, especially when payment is based upon time spent rather than upon output. They are also due to the lack of mobility of certain classes of workers from one part of the country to another, to lack of trade union organisation, to ignorance of the economic conditions prevailing at a distance from their own homes, and to the alternative local employments which they are free to enter. These considerations largely account for the differences in the

earnings of English agricultural labourers in different counties, though part of such differences is to be accounted for by differences in efficiency, as measured by the amount of work performed. To some extent also local differences in the money value of the same kind of work depend upon local variations in the cost of living, but these differences are likely to be greatest, in contrast to the case just considered, when the mobility of the workers concerned and their knowledge of economic conditions in other parts of the country is greatest.

§7. From the point of view of obtaining an income from work, "it is an economic advantage to be born a boy rather than a girl." This fact has long been common knowledge, but the reasons for it were not clearly set out by economists until a few years ago. "The problem of women's wages," as it has been vaguely called, is however, only a particular case of the general theory set out above.2

The number of women who do no useful work, either paid or unpaid, is probably larger in most civilised countries than the corresponding number of men. Though it was diminished temporarily during the war, the "free list" is particularly large in this country among unmarried women of the middle class. In other countries, such as France and Italy, it appears to be considerably smaller.

1 Cannan, Wealth, p. 202.

It is so treated by Professors Cannan and Taussig in the passages referred to in the footnote at the beginning of this chapter. For interesting, but less satisfactory, discussions, compare Nicholson, Principles of Political Economy, III., pp. 158-166, Smart, Studies in Economics, pp. 107 ff, Webb, Problems of Modern Industry, pp. 46 f. Many economists, including Marshall, leave the question practically undiscussed. See also the Report of the War Cabinet Committee on Women in Industry (Cmd. 135), 1919, and especially the Minority Report by Mrs. Sidney Webb.

A Frenchwoman of the middle class, on whom I was billeted during the war, said to me: "I understand that in England the girls cannot cook, but spend all their time playing lawn tennis." The