The Economics of Welfare

Part III, Chapter XIV

INTERFERENCE TO RAISE WAGES IN PLACES AND OCCUPATIONS WHERE THEY ARE UNFAIR

§ 1. IT is convenient to use the term money wages in a wide sense, so as to include a money estimate of any payments in kind that there may be. Thus something should be added to the literal money wages recorded for food and drink provided for agricultural labourers, for coal provided for coal-miners, for housing and food provided for domestic servants, and so on. For the purpose of this chapter wages are taken to mean money wages corrected in this way. Provided that the wages paid to workpeople in all places and occupations were equal to the values of the marginal net product of their work—possible divergences between private and social net product are, for the present purpose, ignored—and provided that the distribution of all grades of workpeople among different places and occupations were such as to maximise the national dividend in the wide sense of Chapter IX. § 2. there would be established between different people’s wages a certain relation.
This relation I define as fair.*1 As between similar persons it is equivalent to the relation of
equality, subject to adjustment for differences in incidental advantages and disadvantages as described in Chapter IX. § 2. My
definition thus conforms to that given by Marshall when he writes that, in any given industry, wages are fair relatively to wages in industries in general, if, allowance being made for differences in the steadiness of the demand for labour, “they are about on a level with the payment made for tasks in other trades which are of equal difficulty and disagreeableness, which require equal natural abilities and an equally expensive training.”
*2 As between persons who are not exactly similar, fairness implies wages which, after adjustment for incidental advantages and disadvantages, are proportioned to “efficiency”; the efficiency of a worker being measured by his net product conceived as marginal,
*3 multiplied by the price of that product. While, however, it implies this, it, of course, does not imply only this.
*4 On the above basis, and utilising the results of Chapter IX., I shall ask, first, whether, and, if so, in what circumstances, the national dividend will be benefited by interference—in the form, for example, of the legal enforcement throughout a district or occupation of piece-rates equal to those paid by reputable firms there—designed to raise wages in a particular centre that are unfairly low relatively to those ruling in industries in general; and secondly, whether, and, if so, in what circumstances, it will be benefited by interference designed to raise wages that are already fair. In the present and following chapters I shall endeavour to answer these questions upon the assumption that the reactions produced by the earnings of workpeople upon their capacity can
be ignored, and I shall reserve for Chapter XVIII. the inquiry how far our results must be modified when that assumption is removed. Let us turn then to the theme of the present chapter, namely, the effect, apart from the reactions on workpeople’s capacity, of interference with
unfair rates of wages.

§ 4. One preliminary observation of a general character should be made. Given that the number of workpeople (of
given quality) assembled in a particular place or occupation is such that, in the existing conditions of demand there, the value of the marginal net product of labour is not sufficient to carry with it a fair wage rate, the effect which interference would have on the national dividend is entirely independent of
the reason why the conditions of demand there are what they are. On reflection this is obvious. But in popular argument it is frequently ignored. Thus manufacturers in out-of-the-way districts often urge that the inferiority of their machinery or the magnitude of their freight charges—factors which depress the level of their demand for labour—”justify” the payment of a wage lower than that paid by their competitors. In an agreement made in the coal industry of Illinois, Indians, Ohio, and Pennsylvania, the validity of this plea was formally recognised, and “the scale nicely adjusted so that the districts with the better quality of coal and the lower railway charges are required to pay enough higher wages than other districts to counterbalance their superior natural advantages.”
*5 Much argument has been expended upon the question whether pleas of this type are sound or unsound. The truth of the matter is that they are neither the one nor the other, because they are irrelevant. The effect which interference with unfairly low wage rates at any point produces on the national dividend will be good or bad according to the way in which it reacts on the distribution of labour power between different places and occupations (including the occupation of idleness). The character of this reaction is not made different by any difference in the causes through which the existing conditions of the demand for labour at the point affected have come about. It depends exclusively on the causes which have prevented the number of workpeople attached to that point from so adjusting itself to the existing conditions of demand there as to make the value of the marginal net product of labour equal to what it is elsewhere, or, in districts where the cost of living for workpeople diverges from the normal, different from what it is elsewhere by the appropriate amount.
*6 We have, therefore, to distinguish the principal
causes to which failure of adjustment may be due, and to examine in turn the effects of interference with each of the different types of unfair wages for which these several causes are responsible.

§ 5. First, the wage rate in some place or occupation may be unfairly low because the costs of movement prevent workpeople there from moving to other places or occupations where the wage rate is higher. Among unfairly low wages of this kind are included the abnormally low wages that may rule (1) in some country distant from others and differing from them in race, language, or religion; (2) in some occupation, movement from which to other occupations would involve much loss of skill; (3) in some occupation mainly filled by low-grade unskilled workpeople, the failures of higher classes, who have spent their energies in attempting a skilled trade or in confidential service to a particular firm, and are, consequently, incapable of becoming either high-grade unskilled labourers or skilled workers of another kind; (4) in some district, movement away from which would involve, for the workmen who move, the loss of special opportunities for earnings which that district offers to their women folk, or, for women workers, to their men folk; and (5) in some form of work, to wit, home work, to which more people are tied by the non-economic compulsion of family cares—a very large proportion of home workers are married women or widows
*7—than, in the existing highly developed state of factory manufacture, economic considerations alone would warrant. It can be shown in a summary manner that interference designed to raise any of these forms of unfairly low wages will, apart from reactions on capacity, be injurious to the national dividend. The only part of the effect which even
appears to be advantageous is the movement—if any such is set up—of certain workpeople away from low-waged places or occupations to others; for the diminution of work available to those workpeople who do not move away obviously involves nothing but loss. But the argument of Chapter IX. has already shown that that movement of workpeople, though it appears to be advantageous, is not
really so. For no doctoring of the wage rate can alter the costs of movement; it can only cause the obstacles set up by them to be leapt over or forced. So long, however, as the costs of movement are what they are, that distribution of labour which differs from the absolute ideal only on account of the costs of movement was shown in Chapter IX. to be the ideal distribution
relatively to the fact of those costs.*8 Any change of distribution, therefore, so long as those costs remain, must actually make the dividend smaller than it would otherwise have been. Plainly, therefore, no loop-hole is left for any gain.
*9

§ 6. Secondly, the wage rate in some place or occupation may be unfairly low because ignorance retains there workpeople, who, if costs of movement alone were in question, would find it advantageous to move. To determine the effect of interference in these conditions is more difficult than it was in the conditions discussed in the preceding section. For, as was shown in Chapter IX., while the forcing of obstacles set up by costs of movement involves a loss to the dividend, the forcing of those set up by ignorance involves a gain. Hence, so far as the effect of pushing up wages in a low-wage place or occupation is to transfer workpeople from employment in that place or occupation to employment elsewhere, the national dividend will be increased. The apparent advantage involved in movement, where movement occurs, is also a real advantage. Before, however, any conclusion can be drawn as to the
net effects of interference, we need to inquire how far, and in what circumstances, the pushing up of wages will transfer superfluous workpeople elsewhere.

When the method of engagement that prevails is the preference method, it is plain that, whether the demand for labour is elastic or inelastic, no superfluous new employees will be tempted to attach themselves to the place or occupation
where the wage rate has been raised, and that all persons dislodged from employment there will know themselves to be definitely and finally dislodged. For, under this method, certain men are formally preferred to others, and it is known that whatever work is available will be wholly concentrated upon them. In these circumstances, since, if the others do not move, they must expect to earn nothing at all, they will be under a very strong inducement to move. Hence the whole effect of the enhancement of wages will consist in the transference of some workpeople from points of less effective employment to points of more effective employment. Except as a temporary incident in the process of transition, no unemployment or partial employment will be created anywhere as a set-off against this gain. Consequently, interference designed to force up wages to the fair level
must benefit the national dividend.

When the method of engagement of labour is either the privileged class or the casual method, the effect on the national dividend is different according as the demand for labour, in the place or occupation where the wage rate is raised, has an elasticity less than, or greater than, unity. If the elasticity is less than unity, the attractiveness of that place or occupation to workpeople already assembled there will be increased, because the mathematical expectation of earnings will be increased. Consequently, under either of these two methods of engagement, there is no reason to anticipate any movement of workpeople away from that place or occupation to others; and under the casual method there will probably be some movement in the opposite direction. It is certain, however, that the amount of work available in that place or occupation will be diminished. Hence the national dividend will necessarily suffer. If the elasticity of the demand for labour is greater than unity, the attractiveness of the place or occupation to the workpeople assembled there, as well as to outsiders, will, under both these methods of engagement, be diminished when wages are raised. Therefore, it is
prima facie probable that some workpeople will move away; though it should be remembered that, in so far as workpeople are better acquainted with, and so attach
more importance to, the nominal wage than to the prospect of continuous employment, the tendency to move may meet with considerable obstruction. This obstruction is the greater in that those workpeople who could move most easily, namely, the young men, are not likely to be hit by unemployment so much as the “average” worker. To the extent that movement does occur, the national dividend will be made larger. But, on the other hand, some workpeople, who formerly were fully employed, are likely, at all events for some time, to remain in the place or occupation in partial employment. The injury which the national dividend suffers from this cause may either exceed or fall short of the benefit which it receives from the movement of the other workpeople. It is impossible to say generally whether the net result will be advantageous or disadvantageous. It will be different in different circumstances. There is, however, one case in which a definite solution is obtainable. Where the elasticity of demand is so great that the raising of the wage rate to the “fair” level reduces the demand for labour to zero, the attractiveness of the place or occupation to all workpeople is also reduced to zero, and those assembled at it
must, therefore, move away. This condition is satisfied when individual employers are so incompetent, or individual factories or mines are so badly situated, that the enforcement of fair wages causes them to collapse altogether before the competition of others. When it is fulfilled—it will be remembered that we are here discussing cases in which movement on the part of workpeople is obstructed, not at all by costs (
e.g. the cost of learning a new trade) but only by ignorance—the national dividend is bound to be increased by interference designed to enforce fair wages.

§ 7. We now turn to the second main class of unfair wage rates that was distinguished in § 3, namely, wage rates that are unfair, not because the value of the marginal net product of labour at the points where they occur is insufficient to yield a real wage equal to wages elsewhere, but because exploitation on the part of employers forces workpeople to accept in payment for their services less than the value which the marginal net product of their services has to these employers.
There is, indeed, something artificial in this statement of the problem, because, if any employer, or body of employers, exploits the workpeople in his service, he will, in general, not be able to hire as much labour as would have been available for him otherwise, and, consequently, the value of the marginal net product of such labour as he does hire will be indirectly raised. Hence, in general, if an employer exploits his men by paying them 5s. a week less than the value of their marginal net product to him, this will not mean that they are getting 5s. a week less than the fair wage obtainable for similar work elsewhere, but perhaps only 4s. or 3s. less than this.
*10 This being understood, we may proceed to investigate the may in which unfairly low wages due to exploitation may be brought about.

It was explained in Chapter VI. that, if perfectly free competition prevailed everywhere, the wage rate paid by any employer in any occupation would be determinate at a definite point. The value of the marginal net product of labour of given quality would be the same to all employers—to simplify the exposition we ignore for the moment local differences in the cost of living—and, if one employer offered a man less than others, that man would know that he could at once get as much as this value of his marginal net product from others. In so far, however, as movements of workpeople are hampered by ignorance and costs, a monopolistic element is introduced into the wage bargain. Consequently, there is created
a range of indeterminateness, within which the wages actually paid to any workman can be affected by individual “higgling and bargaining.” The upper limit of this range is a wage equal to the value of the marginal net product of the workman to the employer engaging him, it being understood that this value is not fixed from outside, but depends in part upon how, many men the employer concerned chooses to engage. The lower limit is a wage equal to what the workman believes he could obtain by moving elsewhere,
minus an allowance to balance the costs of the movement. The width of the gap
between the workers’ minimum and the employers’ maximum varies in different circumstances. It is made larger when the employers in a district tacitly or openly enter into an agreement not to bid against one another for labour, since, in that event, the alternative to accepting terms from them is to seek work, not near by, but perhaps in an unknown district. For example, in some districts the rate of pay to agricultural labourers had, before the war, become a matter of tradition and custom. Though conditions had become quite different from what they were when this tradition crystallised, nobody ventured to take the initiative in breaking away from it. “The farmer,” says the
Report of the Land Inquiry Committee, “has been accustomed to pay a certain wage and to feel that the conditions of farming would not allow him to go beyond that limit, and we have found instances of his going without labour for a time rather than grant a rise in wages…. His line of defence is greatly strengthened by the solidarity of interests among farmers. If an employer in the town wishes to make a substantial advance in wages, he can afford to be indifferent to the resentment, if any, among other employers. But the personal bonds between farmers are extremely close, and the best employer of labour is sensitive to social ostracism. From many parts of the country we have heard of cases where farmers would willingly raise wages but for fear of local opinion. Thus, a farmer told us that, to avoid the appearance of paying higher wages than the farmers round him, he had actually resorted to subterfuge and adopted a bonus method of payment.”
*11 The width of the gap is also greater the more free employers are to make use of devices likely to aggravate the ignorance of their workpeople as to the real amount of the earnings they are receiving—such devices as are combated by the “particulars clause,” the Truck Acts and the other forms of protective legislation that were discussed in Chapter IX. § 8. Whenever any gap exists, exploitation of the workpeople up to the measure of this gap is
possible.*12

Whether and how far, when the extent of the gap is given, exploitation will actually take place, depends partly on the relative bargaining power of the employers and workpeople concerned and partly on the willingness of the stronger party to exercise its power. Even when the gap is large, the occurrence of exploitation is not certain, and, in occupations where the workpeople have been able to organise themselves into strong Trade Unions, supported by a reserve fund and bargaining for their wage rates as single collective wholes, it is not even probable. But, in occupations in which the workpeople—whether because they are widely scattered in space, or because they are poor and ignorant, or because they are women who do not expect to continue in industry after marriage, or for any other reason—are unorganised, there are grounds for fearing that exploitation will often occur. The chief of these is that, when workpeople are unable to combine, an employer generally possesses considerably greater strategic strength than his opponents. First, the actual process of bargaining is one to which he is accustomed and to which he is, in a sense, trained, while these things are not true of the generality of workpeople. Women workers and children are especially weak from this point of view. Secondly, partly because he is richer and partly because he employs a considerable number of workmen, an employer usually stands to suffer a smaller loss of well-being when a bargain with an individual workman fails to be consummated than that individual workman stands to suffer. He is, therefore, in a better position to push things to extremes. The significance of the number of people employed is brought out by the comparative weakness of employers in bargains about domestic service: “The alternative to the well-to-do woman of doing without a servant for a single day is perhaps as disagreeable to her as the alternative to the servant of being out of a place; and the worry and inconvenience to
the mistress of finding another servant is at least as great as the discomfort to the servant of getting another situation.”
*13 Thirdly, in some circumstances for a workman to refuse an employer’s terms involves for him further evil in addition to loss of wages. This will happen if, besides being a workman, he is also a tenant to his employer, and so liable to eviction from his house. In view of these considerations, if an employer of unorganised workpeople chooses to exercise his bargaining power, he
can pay wages much nearer to the workman’s minimum than to his own maximum. Where employers reckon to keep the same workpeople for a long time, fear of injuring these workpeople’s future efficiency may induce them in their own interest to concede terms more liberal than they need have done. Furthermore, it is to be expected that feelings of generosity and kindliness will often prevent employers from fully exercising their power. But, when they are themselves very poor, there is little scope for generosity; and, even when they are not very poor, if they work through foremen or sub-contractors employed upon piece-profits, there is little prospect of it.
*14 It follows that, among unorganised workpeople unable to bargain collectively, a number of men and women are likely to be paid wages approaching much more nearly to the lower than to the upper end of the range of possible rates. Such wages are, in general, lower than the wages for similar work that are paid elsewhere; that is to say, they are “unfair.”
*15

First, the forcing down of wages in particular places or occupations, though it does not reduce the labour supply sufficiently to compel employers to refrain from it, may easily reduce it to some extent by driving some workpeople away.
*16 When this happens, the quantity of labour employed there will be so far contracted that the value of the marginal net product of labour there becomes greater than it is elsewhere. This involves injury to the national dividend. Consequently, a forcing-up of the wage rate, by bringing in men from other occupations which yield a smaller marginal return, would benefit it. In so far, for example, as the wages of agricultural labourers before the war were kept down by tacit understandings among farmers, the legal enforcement of a higher wage would have increased the number of these labourers in a way unambiguously advantageous to the national dividend.
*17

Secondly, it was pointed out in Part II. Chap. IX. § 16, that, “when their clients, be they customers or workpeople, can be squeezed, employers tend to expend their energy in accomplishing this, rather than in improving the organisation of their factories.” To prevent them from seeking profit along the line of bargaining power indirectly impels them to seek it along that of technical improvement. Thus Mr. Mallon writes: “An employer, compelled by the Trade Board [through the institution of a minimum wage] to scrutinise his factory, found that, through lax organisation, its workers were often kept waiting for work to their, and his own, considerable loss. Applying himself to the removal of this cause of waste, he was soon able to provide for the steady and continuous employment of the workers, the outcome being substantial gain to them and, in at least an equal degree, to himself. Such cases could be multiplied indefinitely. In many factories and workshops for the first time methods and equipment are being overhauled, with results at which many of the employers, not at the outset in favour of the Act, are pleased and astonished.”
*18 With this passage may be compared a kindred observation of Miss Black’s: “It has been shown over and over again that, when employers are prevented from developing their business along the lines of cheap labour or bad conditions, they proceed to develop it along the lines of improved methods, and that the improved methods tend both to increased output and to greater cheapness.”
*19 This view, the correctness of which nobody seriously doubts, is not, it should be understood, equivalent to the view that, if in any country labour in general is abundant and, therefore, cheap, employers are discouraged from making use of machinery. Since machinery is itself the product of labour,
that view is not correct.
*20 If, however, in a particular district or occupation employers are able to exploit a particular class of labour, they are discouraged from making use of machinery, because machinery embodies the services of other classes of labour, which, not being exploited, are more expensive relatively to their efficiency.

Thirdly, if particular employers outbargain their workmen, in such wise as to compel some or all of them to accept a wage below the value of their marginal net product, it necessarily happens that these employers are receiving more than the normal earnings of persons of their degree of competence;—a state of affairs which, in view of the imperfect mobility of employing power between occupations, may continue for some time. If the exploiting employers were persons of the ordinary competence of their grade, interference, which forced up the wages paid by them to the fair level, would simply compel them to hand over to workpeople profits formerly exacted from them by
force majeure, and would have no other effect. As a matter of fact, however, exploitation of this kind is much more often practised by incompetent or badly situated employers, who, without it, could not maintain themselves in business, than by competent and well-situated men. The small masters have, throughout history, been always the worst exploiters. Hence exploitation provides, in the main, a bounty at the worker’s expense for relatively incompetent and badly situated employers; and the prevention of exploitation would tend to hasten their defeat at the hands of more efficient rivals. This consideration, in conjunction with the others that have preceded it, makes it plain that external interference to prevent that type of unfair wages which I have described as exploitation is desirable in the interest of the national dividend as well as upon other grounds.

§ 9. All that has been said hitherto, both of unfair wages that are equal to the value of marginal net product, and of unfair wages that are exploited below the value of marginal net product, is of quite general application. It holds good equally of men’s wages and of women’s wages;
and, of course, in view of their inferior organisation, the danger of exploitation is especially great with women. Any specific plea that the wage of either sex in any place or occupation is unfair would need to be reviewed in the light of it. There still remains, however, a somewhat special problem arising out of
the relation between men’s wages and women’s wages, which our analysis does not cover. It may happen that women’s wages in some place or occupation are fair relatively to women’s wages in other place or occupations, but unfair relatively to men’s wages in that place or occupation. This statement has not, of course, anything to do with the well-known fact that women’s day wages on the average are considerably lower than men’s day wages. Women, looking forward, as they do, to matrimony and a life in the home, are not trained to industry as men are, and do not devote to it that period of their lives when they are strongest and most capable. Thus between the age-periods of 18-20 and 25-35 there is a great decrease in the percentage of women who are engaged in wage-earning occupations, and this is due, no doubt, to the withdrawal of many of them at marriage. Prof. Sargant Florence writes: “The common age at marriage being between 21 and 25 for spinsters, the typical length of industrial life for a woman would be eight years.
*21” In these circumstances, even though women’s natural endowments of mind and muscle were equal to those of men, which, on the average, they are not, it would be surprising if their day wages were not lower. Certainly, the fact that they are lower involves no unfairness in the sense in which that term is here used. In some places or occupations, however, it may happen that not only the day wages, but also the piece wages, or, more accurately, the efficiency wages, of women are lower than those of men. This state of things may come about because women’s wages in those places or occupations are
unfair relatively to women’s wages elsewhere. In that event there is nothing special about it, and the analysis relevant to it has already been given in preceding sections. But it may also come about because women’s wages in those places or occupations are unfair relatively to men’s wages there, although they are fair relatively to women’s wages elsewhere. It is this state of things that constitutes our present problem.

In order to understand the matter rightly, analysis is necessary. The common idea is that women are normally paid less than men because men’s wages have, in general, to support a family, while women’s wages have only to support the women themselves. This is very superficial. The correct line of approach would seem to be as follows. The productive efficiency of a representative woman relatively to that of a representative man is different in different occupations: in some, such as nursing and the tending of infants, it is much greater; in others, such as coal-mining and navy work, it is much less. If we knew enough of the facts, we could draw up a list of all occupations, giving for each of them the amount of normal man’s labour to which a day or week of normal woman’s labour is equivalent. The relation between the demand schedules for women’s work and for men’s work is determined by the facts embodied in this list, in conjunction with the general conditions of demand for the products of the several occupations. The relation between the supplies of women’s work and of men’s work is determined partly by the physiological fact that male and female children survive in nearly equal numbers, whatever the comparative wages ruling for men’s work and women’s work may be; and partly by the economic fact that the proportions, respectively, of the men and women in existence who offer their work in industry depend, not only on the wages offered to members of either sex separately, but also, since women are the less likely to work at industry the more money their husbands are earning, on the aggregate amount of the joint family income. These two sets of influences together govern and determine the relation between the general level of the wages per day paid to representative
members of the two sexes.
*22 In equilibrium there is one general rate of representative men’s day wages and one general rate of representative women’s day wages, the one or the other being higher according to the circumstances of supply, and according as the commodities demanded by the public are chiefly commodities for the manufacture of which the one or the other sex is specially well fitted.
*23

Men alone are employed in all occupations where the ratio of their efficiency to women’s efficiency exceeds the ratio of their day wages to women’s day wages; women alone in all occupations in opposite case; and men and women indifferently in the marginal occupations in which their respective efficiencies bear to one another the same ratio as their respective day wages. In these marginal occupations, that is to say, the efficiency wages of the two sexes are equal. This equality of efficiency wages means, with certain allowances, equality of piece wages. The principal allowances are, first, a small extra for men, because, since, at need, they can be put on night-work and can be sworn at more comfortably, it is rather more convenient to employ them; secondly, a small extra to the more skilful workers, whether men or women, because they occupy machinery for a shorter time than less skilful workers in accomplishing a given job. In equilibrium the piece wages paid to the members of the two sexes in the marginal occupations are, with these limitations, equal.
*24 This is the state of things which the play of economic
forces tends to bring about; and, so far as it in fact brings it about, it is not possible for women’s wages in any place or occupation to be fair relatively to women’s wages elsewhere and yet unfair relatively to men’s wages there.
*25

In real life, however, it happens from time to time that economic equilibrium in this matter is not attained. In particular occupations employers may pay to women workers an efficiency wage, which, though fair relatively to women’s wages elsewhere, is less than the efficiency wage they are paying to men workers, and yet may still employ some men. They may
do this either for a short time, while they are in process of substituting the one sex for the other; or for a long time, because trade union pressure or custom either compels the retention of some men, or vetoes the entry of more than a limited number of women. In these conditions is the claim “equal pay for workers of equal efficiency” justified? In what way would interference to raise the women’s efficiency wage to the level of the men’s affect the national dividend?

If the power of tradition, custom or trade union pressure is such that neither the number of women attached to the occupation under review nor the number of women employed there will be different if employers are allowed to pay them the lower rate from what it will be if they are forced to pay the higher rate, the national dividend will not be affected at all. This, however, is a very improbable state of affairs. For, even though the number of women who may be
employed in the occupation is rigidly limited, it is likely that the number trained for and attached to it will be made greater by the higher wage rate; and, if this happens, the dividend obviously suffers from the enforced idleness of those who are attached but not employed. Furthermore, in real life permission to pay the lower wage rate will seldom be without effect on the number of women who are employed there and take the place of men. For example, “in Stoke-upon-Trent it appears that women and girls are very largely employed in the pottery industry. In some branches of this trade they are being employed to an increasing extent upon work which, a few years ago, was performed almost exclusively by men; they are now acting in competition with male labour; and, as they are able to do similar work for lower wages, they are gradually driving men from certain sections of the trade.”
*26 This consideration confirms Professor Cannan’s contention, that “the most powerful lever for increasing the opportunities of women is taken away if they are not allowed to do the work cheaper.”
*27 It follows that, generally speaking, to compel the payment to them of an efficiency wage equal to that paid to men in occupations which they are seeking to enter, and in which
such a rate would give them higher earnings than similar women can obtain elsewhere, must obstruct their entry either directly, or indirectly by relaxing employers’ efforts to break down the customs and rules that hinder it. But, since,
ex hypothesi, they are more efficient, relatively to men, in these occupations than they are in marginal occupations common to both sexes, their entry would necessarily be beneficial to the national dividend. Hence, generally speaking, interference designed to enforce the payment to them of a “fair” wage, as compared with the wages paid to men, in circumstances when this means an unfairly high wage as compared with women’s wages elsewhere, would injure the national dividend.
*28 Interference may still conceivably be advocated by those who wish to exclude women from industry, as far as that can be done, on general social grounds. Such a defence for it is, however, insecure, because interference of the type here discussed not only lessens the aggregate number of women in industry, but distributes them among different occupations in a wasteful manner. The social argument for excluding women from industry generally cannot sustain a policy which has this effect.

§ 10. In view of the distinctions that have been found, in the course of this chapter, to exist between different forms of unfair wages, it is plain that interference directed indiscriminately against all forms must do harm as well as good. The procedure, which, if practicable, would most advantage the national dividend, would be to examine and deal separately with every place or occupation where there was
prima facie reason to believe that wages were unfair in any of the senses studied in the text. It may, however, be argued that this plan is an impracticable one, and that it is necessary either to interfere against unfair wages by
broad general rules or else not to interfere at all. Thus, it may be said, it is practicable to pass and administer a law like the French law of 1915, which provides that female outworkers shall be paid a piece-wage that will yield to the average outworker earnings equal to those of an average factory worker;
*29 but it is not practicable to enact that a wage lower than this shall be permitted when it is due to family ties which restrain home workers from going into factories, and forbidden when it is due to exploitation. When the issue raised is of this type, policy must be based on a balancing of conflicting considerations.