The Economics of Welfare

Pigou, Arthur C.
Display paragraphs in this book containing:
First Pub. Date
London: Macmillan and Co.
Pub. Date
4th edition.
62 of 73

Part IV, Chapter V


§ 1. IN the latter portion of Part III. I examined at length the conditions under which an enforced increase in the wages rate paid in a particular occupation or place would injure the national dividend. We have now to consider the effect that will be produced on the real income of workpeople, and so of the poor, as a whole. For simplicity we may take for examination—no difference is made in the substance of the argument—the state of things contemplated in Part III. Chapter XVII. §§ 8-9. The wage rate was there supposed to be forced up in an occupation where it had formerly been both "fair" relatively to other industries and equal to the value of the marginal net product of the work for which it was paid. It was assumed that no difference was made either to employers' technique or to the capacity of the workpeople to whom the increased wage was given: and, on this assumption, it was proved that the national dividend must be diminished. In what, if any, conditions will the real income of labour as a whole, nevertheless, be increased?


§ 2. The first step towards answering this question is to determine in what circumstances the enforcement of an uneconomically high wage—that is a convenient term for a wage that damages the dividend—will increase the real earnings of the particular group of workpeople in whose behalf it is won. It should be noticed in passing that an uneconomic enhancement of the wage rate in any occupation may take the form either of a special enhancement of the rate per unit of labour paid to inferior workpeople—e.g. such workpeople may be given the same time-wages as competent workers—or of a general enhancement of the rate per unit of capacity paid to all workpeople. It is evident that an uneconomic enhancement of the former kind must either throw all the inferior workpeople out of work altogether, or must diminish the aggregate quantity of labour employed to exactly the same extent as an equal enhancement in the rate per unit of capacity paid to all the workpeople in the occupation. Hence it is bound to have a less favourable effect on the aggregate earnings of the whole group of workpeople concerned than the latter kind of enhancement. In what follows, therefore, it will be sufficient to consider that type of uneconomically high rate which affects equally the wages per unit of capacity paid to all workpeople in an occupation. To determine in the abstract the conditions in which the establishment of this type of uneconomically high wage will increase the real earnings of this group of workpeople is perfectly simple. It will increase them if the demand for the labour of the group has an elasticity less than unity, and it will diminish them if the elasticity is greater than unity. This result—on the assumption, of course, that the workpeople concerned are not themselves purchasers to any appreciable extent of the commodity they produce—is an obvious arithmetical truism, following at once from the definition of elasticity. To fill it out in the concrete, to investigate, that is to say, the conditions that determine whether the demand for the services of any assigned group of workpeople is likely to have a high elasticity or a low one, is the task we have now to essay.


§ 3. In the fifth section of Part II. Chapter XIV., an analysis was given of the determinants of elasticity with reference to the demands for different classes of commodities. This analysis is applicable to the demands for different classes of labour also. In that application it may be set out as follows.


First, we have the general fact that the demand for anything is likely to be more elastic, the more readily substitutes for that thing can be obtained. This fact has an important bearing on the relation between labour and machinery; for, in some industries, a very small addition to the cost of working a process by hand would induce employers to adopt mechanical appliances. Aves, for example, quotes the statement of an ex-inspectress that, in the Victorian clothing trade, where minimum wage determinations have unintentionally discriminated against home work, employment was transferred to factories using machinery, and "practically all outside work was stopped."*42 In like manner, the tanners of Victoria, commenting on the effects of the Wages Board in their industry, state: "Labour-saving machinery is forced into use, so much so that the tannery trade has been practically revolutionised since the Wages Board system was applied to the trade."*43 In circumstances of this kind the high elasticity of demand for labour is, in effect, due to the fact that there is a readily available and closely competing substitute for its services in the rival factor capital, or, more strictly, other labour accompanied by a greater amount of waiting. Since it is easier to introduce this substitute after an interval than immediately, elasticity in demand due to this cause is greater from the standpoint of a long period than of a short period.


Secondly, we have the general fact that the demand for anything is likely to be less elastic, the less important is the part played by the cost of that thing in the total cost of some other thing, in the production of which it is employed. This general fact enables us to point out certain occupations in which the demand for a particular class of labour is likely to be especially inelastic. One of these is the occupation of women in sewing on the covers of racquet and fives balls.*44 Another is that of making trouser-buttons. Lord Askwith writes: "The rich man's trousers may be cut by an expensive tailor. The buttons on those trousers may be made by sweated industry. High payment for those buttons would be but a minute part of the cost of the whole article."*45 The engineering work done by engineers engaged by building firms, since these persons are employed only incidentally and as a trivial part of the total producing force, is in a similar position. In like manner, the part played by the original labour is small in commodities for which the addition to wholesale price made by the work of the retailer is large. "For example, when we find that the maker of a lady's costume is paid 10d. or 1s., while the article is sold for 25s. to 30s., it is obvious that the wage paid is so small in relation to the retail price that, even were the wage doubled, it need necessarily affect the price but little, if at all."*46 This condition, that the part played by labour in a particular act of production shall be small, is probably fulfilled fairly often, and it is likely to be fulfilled still more often as the relative importance of installations of plant and machinery in production increases. One writer has even suggested that "the labour cost of production in most industries is usually not sufficient materially to affect the price of the finished article." It should be noticed, however, that, in the important work of coal production, hewers' labour constitutes a very large part of the total cost, and the condition stipulated for is, therefore, not fulfilled.


Thirdly, we have the general fact that the demand for anything is likely to be more elastic, the more elastic is the supply of co-operant agents of production. This fact makes it evident that the demand for labour will be specially inelastic in industries which make use of raw materials of highly inelastic supply. Apart from raw materials, the principal co-operant agents working with labour in any industry are capital instruments, managing ability and other labour. From the standpoint of a long period the supply of these to any single industry is, beyond doubt, exceedingly elastic. But, from the standpoint of a short period or a period of moderate length, it is likely to be inelastic; for specialised machinery and managing skill, and even other labour, can neither be created or brought from elsewhere, nor yet destroyed or carried off elsewhere, in the twinkling of an eye. Here again, therefore, the forces making for elasticity of demand are stronger from the standpoint of a long period than of a short period. It should be added that in some industries, notably coal-mining, Nature herself acts as a very important co-operant factor of production. In times of expanded demand new men have to be set to work on seams much more difficult and less productive than those ordinarily worked.*47 This means, from a short-period point of view, a highly inelastic demand for labour.


Fourthly, we have the general fact that the demand for anything is likely to be more elastic, the more elastic is the demand for any further thing which it contributes to produce. This fact implies that the demand will be specially inelastic for the services of workpeople engaged in the manufacture of commodities of highly inelastic demand. When the elasticity of the public demand for any commodity is given, it is obvious that, from a short-period point of view, the elasticity of the demand for new production of it will be greater or less, according as the thing can or cannot be readily made for stock. Apart from this, the circumstances upon which the elasticity of demand for various classes of commodities depends were discussed in the fourteenth chapter of Part II. The most significant of them for our present purpose is the presence or absence of foreign competition. Thus a critic comments on some of the effects of New Zealand wage regulation: "In some trades employers have not been able to cope with the extra cost of production owing to competition with the imported article. They have, therefore, had to give up the producing part of their business and increase their importations. In the tanning and fellmongering business some serious results have followed the fixing of a minimum wage. I will mention two instances. Some years ago a firm in the district of Dunedin closed down its works and removed its plant to Australia, largely owing to the conditions imposed by the Arbitration Court. A member of a Christchurch firm has informed me that, since the Court's award in the Canterbury district was made about six years ago, a much larger proportion of sheepskins have been shipped to London, without being handled by the local fellmonger, than was formerly the case. Hides, which should have been tanned here, have been shipped raw. Prior to the award, my informant's firm paid from £10,000 to £15,000 in wages; now the wages sheet amounts to only about £5000. The number of bales of wool scoured annually by the same firm since the award came into force has not been more than 2000; formerly the number was from 6000 to 8000."*48 In connection, however, with this matter of foreign competition a word of caution is necessary. Let us imagine that there are a dozen industries in this country, all of about the same size and all subject in about equal measure to foreign competition in the home market. Looking at any one of these industries singly, we conclude, perhaps, that the elasticity of demand for its product is such that an increase of 10 per cent in the cost of making it in this country would stimulate importation and reduce the demand for the home-product by 50 per cent. It is natural to infer that a 10 per cent increase in costs in all twelve industries together would reduce the demand for the home-product in all of them by 50 per cent. This, however, is not so. Foreign imports collectively constitute the foreign demand for British exports. When, therefore, for any reason it becomes advantageous to increase the sendings of one kind of import, other kinds of import will tend to fall off, the adjustment being brought about through—but not by—a change in price levels. Thus, when an extra flow of imports, stimulated by increased home-costs, helps to bring about a contraction in the demand for the home-product in one industry, the contraction will be, in part, cancelled by an expansion, made possible by lessened imports, in other industries. In other words, the demand for the whole body of British products subject to foreign competition is less elastic than the demand for a single representative item among these products. It follows that, other things being equal, the workpeople immediately affected are more likely to be benefited by interference to raise their wages if the interference is extended to several industries subject to foreign competition than they would be if it were confined to one.


§ 4. With these results in mind we may proceed to the next stage in our inquiry, and ask in what conditions the establishment at one point of an uneconomically high wage, which raises the real earnings of the workpeople there, will also raise the real earnings of the workpeople as a whole. Let us still suppose that the commodity, to the makers of which an uneconomically high wage has been assigned, is exclusively consumed by persons other than workpeople. It may be noted, in passing, that, when a factor making for inelasticity in the demand for the services of the particular group of workpeople in whom we are interested is inelasticity of supply and, therefore, "squeezability" in some co-operant group, a part of the gain to the first group will be offset by loss to the second. For the purpose of a general analysis, however, we may neglect this rather special point.


If the elasticity of demand for labour in the occupation where the wage rate has been raised is less than unity, the aggregate earnings of labour as a whole, and not merely the earnings in that occupation, will be increased, provided that either the casual method or the privileged class method of engaging labour, as described in Chapter XIII. of Part III., prevails in that occupation. Under the casual method workpeople will be attracted into the occupation from outside until the prospect of earnings per man inside and outside are brought to equality; and, since the number of workpeople left outside is diminished by this drain, the wage rate there will be raised. This proves that aggregate earnings inside and outside together must be raised. Under the privileged class method of engagement no one will be attracted from outside and no one will be driven out. Hence earnings outside will be unchanged. Since, therefore, earnings inside are, ex hypothesi, increased, it again follows that earnings as a whole are increased. If the preference method of engagement prevails, conditions are conceivable in which earnings as a whole would not be increased. For some persons must be driven out of the industry where the wage is raised, and, though those left there will be getting more than they were getting before, the influx of labour into other industries might, if the demand in these industries had an elasticity less than unity, so lower earnings outside as to outweigh the gain inside. Since, however, as was shown in Chapter III., the demand for labour in industry in general is highly elastic, the conditions necessary to this result are not fulfilled. In real life, therefore, the earnings of labour as a whole must be increased whenever an uneconomically high wage is enforced in a selected occupation, provided that the elasticity of the demand for labour there is less than unity.


If the elasticity of demand in the occupation where the wage rate is raised is greater than unity, analogous reasoning shows that earnings as a whole cannot be increased, provided that either the casual or the privileged class method of engagement prevails in the occupation. For some workpeople will be driven out of the occupation, and a new equilibrium will be established, with an expectation of earnings for every one equal to the earnings in other occupations; and these will have been made lower than before by an influx of new workers. If, however, the preference method prevails, earnings as a whole may be increased even though the elasticity of demand is greater than unity. Those who are left in the industry where the wage is raised will be getting more than they were getting before; and, though everybody else will be getting less than before, yet, if the demand for labour in other industries is sufficiently elastic, their loss need not be so great as the others' gain.


§ 5. It is now time to remove the assumption set out in § 2, that the commodity produced by the group of workpeople, whose wage is being interfered with, is exclusively consumed by persons other than workpeople. On the strength of that assumption we have been able, up to this point, to ignore the distinction between effects on money earnings and effects on real earnings. Where the assumption is unwarranted we are not justified in doing this. An increase of money earnings may be associated with a decrease of real earnings, and may, therefore, be delusive. If the commodities produced by the favoured workers are consumed by nobody except members of the working-classes, it must be delusive, for it is bound to involve a more than equivalent loss to workpeople (those inside the privileged industry and those outside it together) in their capacity as consumers. If the consumers consist partly of workpeople and partly of others, it is not possible to say absolutely whether the workpeople's gain as producers or loss as consumers will be greater. All we can lay down is that, the more important the part of the consumption for which non-wage-earners are responsible, the more likely it is that the establishment of an uneconomically high wage rate will succeed in bringing about an increase in the real income of workpeople as a whole. When, therefore, the main part of the product of any group of workers is consumed by other workers, though the establishment of an uneconomically high wage rate may enhance the aggregate real income of the favoured workers, it is not probable that it will enhance that of all workers collectively. This point is important, because, in real life, it is rich people who make, or otherwise provide, a great part of the luxuries of the rich, while poor wage-earners make things for other wage-earners. Thus, Mrs. Bosanquet writes: "Nothing strikes one more forcibly in studying the position of the lowest-paid workers than that they are almost always engaged in producing goods for the consumption of their own class.... Badly paid tailors are making cheap clothing that no rich man would look at; badly paid servants are rendering services that would not be tolerated by any one of refinement and culture; while the real requisites of refinement and culture, if by these we mean such things as art, music, and literature, are produced by professional people."*49 Again: "The great majority of wage-earners are engaged in producing for the benefit of other wage-earners, and have no direct connection with the non-wage-earning classes. The majority of builders are building houses for wage-earners; the very large majority in the clothing trades are making clothing for the wage-earners; the majority of food-preparers are preparing food for the wage-earners. More especially of the sweated trades is it literally true, almost without exception, that they are working for the wage-earners alone, and that a rise in the price of their products would be paid by the wage-earners alone. How would it be possible that the propertied classes should pay any share in the increased price of ready-made suits, or cheap blouses, or shoddy boots and shoes, or Pink's jams? The burden must fall on the consumers of these articles, and they are the wage-earners."*50 Mrs. Bosanquet would not, of course, pretend that there are no rich men's luxuries, towards which poor men's labour contributes an important part. It would seem, however, that not much of the labour of poor persons in the United Kingdom is devoted to the supply of luxuries of this sort.*51 It follows that the establishment of an uneconomically high wage rate for a particular group of workpeople is much less likely to involve a real increase in the earnings of workpeople as a whole than it appears to be when the distinction between money earnings and real earnings is ignored.*52 So far, however, the possibility that it may involve this remains.


§ 6. But against the realisation of this possibility there is at work a corrective tendency. If the real earnings of labour in the aggregate are increased by the manipulation of wages in any industry, a smaller portion of the productive power of the community is left available to provide payment for the services of capital. Had there been no alteration in the constitution of the national dividend, this would have meant a fall in the real rate of interest offered to new capital. The change in the constitution of the dividend, consisting, as it does, in a shifting of production away from its natural channels, cannot prevent this fall from coming about. It would seem, therefore, that manipulation of wage rates cannot benefit labour in the aggregate without causing the reward offered to savings to be diminished. Now, it is no doubt true that a fall in the real rate of interest will not cause everybody's savings to be diminished. The savings of those people, whose object it is to leave a definite sum to their children at death, will actually be increased; and the savings of the very rich, who merely put by what is left over after their accustomed standard of life has been satisfied, will not be affected at all. There can be little doubt, however, that, on the whole, a fall in the rate of interest will diminish savings to some extent, and, of course, a diminution of savings carries with it a diminution in the rate at which new capital equipment is provided. In this way an indirect influence is set in play tending to make the remuneration of labour in future years decrease. Here, plainly, is a tendency adverse to disharmony. Furthermore, this tendency is quantitatively important. For suppose a policy to be adopted, which, while increasing the real income of labour at the moment, causes the supply of new capital each year to be one per cent less, not necessarily than it was before, but than it would have been apart from the policy. In any one year the loss to the aggregate stock of capital equipment in the country will be small. But the annual losses are cumulative. After, say, ten years, the capital stock that is available to assist labour in its activities may be considerably smaller than it would otherwise have been.*53 Moreover, the reduction of this stock is aggravated by the fact that it must itself cause a reduction in the national dividend; that, therefore, the transference to labour of any given annual sum must throw a continually increasing burden on profits; that, therefore, the diminution (or check to the growth) of the national dividend must be greater in the second year than in the first, greater in the third year than in the second, and so on; and that, therefore, the rate of fall in the capital stock must be progressively accelerated. As this stock falls in amount below what it would otherwise have been, the annual earnings of labour also fall continuously. In the end it would seem that, as against any policy, which, in the first instance, benefits labour at the expense of injuring the national dividend, this cumulative tendency is bound to prevail, and that, therefore, from the standpoint of a sufficiently long period, any disharmony that may have been set up must disappear. Since, however, relatively to the stock of capital, the annual creation of new capital is small, and, consequently, any probable change in the annual creation very small, the harmonising tendency will work slowly. This implies that, for some time after the establishment of an uneconomically high wage in particular occupations, disharmony may prevail.


§ 7. Up to this point we have been concerned with the consequences of fixing an uneconomically high wage, as if this were a single self-subsistent act. In actual life, however, it is inevitably mixed up with State policy as regards the protection of persons in distress. If the enforcement of an uneconomically high wage in some occupation throws a certain number of people out of work for a long time, the State will have to help these people. Consequently, if we count as a part of the real income of the poor what the State provides for assisted persons, their real income in this wider sense may well be raised by a policy which lowers the earned part of their real income. Thus, by a forcing up of the wage rate in some occupations, we may suppose that the national dividend is injured, and that those workers who are left in industry gain a little less than those who are thrown out of it lose. There is then harmony between the effects on the national dividend and on the real income of the poor in the narrower sense. But, if, in consequence of increased unemployment, the expenditure of State aid to poor persons becomes £1,000,000 bigger than it would otherwise have been, there may be disharmony between the effect on the dividend and the effect on the real income of the poor in the wider sense. The way is thus opened for a somewhat special argument in favour of forcing up wage rates in low-grade occupations. It may be granted that both the national dividend and the real earnings of labour as a whole will be diminished. But, it may be claimed, at all events if the preference method of engagement prevails, that a number of people, who otherwise would have earned too little to maintain independently a decent life, will now get adequate earnings. A number of others will, indeed, earn less than before—possibly nothing at all—but, owing to the action of the State, they need not receive less than before. Thus we shall have, instead of a large body of people, all of them occasionally or partially supported by the State, one moderate-sized body fully self-supporting and another moderate-sized body scarcely self-supporting at all. From the point of view of economic welfare as a whole, particularly if the conditions are such as to make the fully self-supporting body much larger than the other, the latter state of things may be deemed the better, in spite of the fact that it involves a smaller national dividend. It may be objected, no doubt, that the persons now rendered fully self-supporting will really be sustained by the help of what is, in effect, a special tax upon the people who purchase the things they happen to make; and that the care of relatively incapable citizens is an obligation upon the whole community, and not merely upon those members of it who buy racquet balls, or whatever the article may be. It may be replied, however, that, in so far as relatively incapable citizens are responsible for products of general consumption, or in so far as they do work for municipalities or the State in connection with commodities or services not designed for sale, this objection loses the greater part of its force; and that, in any event, since every indirect tax must hit some people "unfairly," it is not of very great weight. Moreover, since the workers who benefit will not think of themselves as being in any sense "relieved by their customers," there is no danger that any injurious moral effect analogous to the "taint of pauperism" will be produced upon them. Plainly, an issue turning upon considerations of this kind is not susceptible of any general solution. Whenever it is proposed to enforce an uneconomically high wage in any occupation upon the grounds suggested in this section, a decision can only be reached by a careful balancing of conflicting tendencies, after all the relevant circumstances have been studied in detail

Notes for this chapter

Report on Wages Boards, p. 197. This "determination" fixed both an hour rate and a piece-work rate, compelling the latter to be paid to outworkers. The intention was that the two should be equivalent, but employers in practice found the hour, or wages, rate much the cheaper. The ex-inspectress added: "When the wages rate and the piece-work rate were nearly the same, as in the shirt and underclothing trade, the trouble did not occur, and, after ten years' working of the determination, these trades count many outworkers to-day." The choice between an out and an in-worker is affected by the fact that, when employing outworkers, the employer escapes charges for working space, light, firing, and so forth. "The savings upon factory rent, upkeep, and superintendence appear to be larger factors in the cheapness of home work than the lowness of wages" (Black, Makers of our Clothes, p. 44). Cf. also Marconcini, L' industria domesticu salariata (pp. 432-3). On the other hand, of course, economies of superintendence and, sometimes, of power are to be obtained in factory work.
Report on Wages Boards, p. 179. Cf. ante, Part III. Chapter XIV. § 8.
Cf. Lyttleton, Contemporary Review, February 1909.
Fortnightly Review, August 1908, p. 225.
Cadbury and Shann, Sweating, p. 124.
Cf. Hooker, Statistical Journal, 1894, p. 635 n.
Broadhead, State Regulation of Labour in New Zealand, p. 215.
Bosanquet, The Strength of the People, p. 71.
The Strength of the People, pp. 294-5.
For examples of things made by "sweated" workers and consumed by others than wage-earners, cf. Cadbury and Shann, Sweating, p. 123.
The reason why this conflict between the interests of a particular class of wage-earners and those of wage-earners as a whole, when the particular class endeavours to force up its rate of wages above the normal, is not generally recognised may well be, as Mr. H. D. Henderson suggests, that most wage movements are associated with trade cycles. In consequence of this, the wage rates of different groups generally move up and down together, and, therefore, to the casual observer, there appears to be a greater harmony of interest than there really is (Supply and Demand, p. 157).
Cf. Marshall, Economics of Industry, pp. 372-3.

Part IV, Chapter VI

End of Notes

62 of 73

Return to top