The Economics of Welfare

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

Part IV, Chapter VII


§ 1. IN a community in which wage-rates are everywhere adjusted to the conditions of demand and supply, so that no wage-rates are uneconomically high and there is no unemployment beyond what is necessary to allow adjustments to be made to industrial fluctuations, for the State to subsidies wages in particular industries must, in general,*55 worsen the distribution of productive resources and damage the national dividend. A policy of wage subsidies applied to all industries would not necessarily damage the distribution of productive resources, but it could not improve this distribution; and, though in some circumstances it might increase the dividend, it would probably only do so at the cost of causing too much work to be done, and, therefore, in a manner injurious, and not beneficial, to economic welfare. Hence, subject to qualifications which the reader can readily provide, we may conclude that in a community, in which, apart from subsidies, rates of wages would be everywhere adjusted to the conditions of demand and supply, and policy of wage subsidies is likely to prove anti-social.


§ 2. In real life, however, it may happen that either in particular industries or, it may be, throughout industry as a whole, wage-rates are established at an uneconomically high level; that is to say, at a level too high to allow the demand for labour to absorb the supply, in such wise that more people are unemployed than are accounted for by the movements they have to make in consequence of industrial fluctuations. Thus there is some reason to believe that in England during the post-war depression, partly through direct State action and partly through the extra bargaining strength given to workpeople's organisations by the development of unemployment insurance, wage-rates over a wide area were set at a level uneconomically high in the above sense. Where conditions of this kind prevail and where public opinion insists that unemployed persons shall be somehow provided for, a policy of wage subsidies is no longer prima facie anti-social, but needs more particular consideration.


§ 3. The possibility of social gain is made clear most easily by means of a highly simplified imaginary case. Consider an agricultural community in which farmers own the land and employ labourers, all of whom are of equal skill. Let nothing else be produced except wheat, and let wages be paid in kind. Let the conditions be such that, with wages at one bushel of wheat per day, all the labourers would find employment, but that, when the rate is put at one and a quarter bushels per day, 10 per cent of them are out of work and the aggregate output of wheat, instead of being A bushels, is cut down to (A - a) bushels. Let the State insist, for humanity's sake, that a man out of work shall, nevertheless, receive, say, one-third of a bushel of wheat for maintenance, and let it take from farmers whatever amount of wheat is needed to permit of this. In such a case it is easy to see in a general way that, if a tax is imposed on the income of farmers or on the rental value of their land, and the proceeds used to give a subsidy of so much per cent on wages, the labourers are bound to gain, and the farmers—when their loss through the tax is balanced against the extra output of wheat and their savings in respect of unemployed labour—may gain. For a full understanding of the situation it is, however, helpful to make use of a few symbols.*56


§ 4. Let (x + h) workpeople be attached to a given industry, whose products are not exported. Let w2 be the wage at which all of them would find employment; w1 the wage which is actually established, and x the number of men that are actually employed. If then things are allowed to take their "natural" course, h workpeople will be unemployed in the industry. For humanity's sake these must be somehow provided for; so we suppose that a payment r is made to each of them, and—to make the case as strong as possible—that the whole sum hr is taken from non-wage-earners. This is the position in the absence of any subsidy. Now let a subsidy at a rate s = (w1 - w2) be paid in respect of each workman employed; and let the funds for it be raised by taxation imposed on non-wage-earners (e.g. income tax). The wage (including the subsidy) paid to each workman will hereafter still be w1—the workmen already in work will receive no more than before—but it will now pay employers in the industry to take on (x + h) workpeople instead of x workpeople. The output of the new h workpeople taken on will have a value equal to some amount (dependent on the slope of the demand curve for labour) intermediate between hw1 and hw2. Let it be {hw2 + hc}; which, in the special case where the demand curve for labour is a straight line, = {hw2 + ½h(w1 - w2)} = {hw2 + ½hs}. From these data it is easy to calculate loss and gain. Workpeople as a body obviously gain; for h more of them are employed at the full wage w1 for which they stipulated. Non-wage-earners neither gain nor lose in respect of the x workpeople who would be employed anyhow. In respect of the others they make a payment in wages plus subsidy equal to hw1: they obtain an extra product of a value equal to (hw2 + hc), which is less than hw1; and they save a payment to unemployed workpeople equal to hr. Their net gain is, therefore, equal to {hw2 + hc + hr - hw1} = h(r + c - s). This is necessarily positive, provided that the rate of subsidy required is less than the rate of contribution which would have been paid to unemployed workmen. When this condition is satisfied it is obvious that both the absolute receipts of labour and also the national dividend as a whole must be larger than they would have been had other things been equal but no subsidy paid.*57


§ 5. The foregoing analysis was explicitly confined to industries whose products are not exported. If the policy of subsidies were applied to export industries, the balance of gain and loss would work out less satisfactorily, because foreigners, instead of domestic users, would get the benefit of the price reduction due to the subsidy; in effect British non-wage-earners would be paying a part of the costs of work done for foreigners, which, had there been no subsidy, foreigners themselves would have paid. Here, therefore, there would only be a net gain to British non-wage-earners if the foreign demand were so extremely elastic that employment would be increased from x to (x + h) by a subsidy s, such that (x + h)s is less than hr. The case for subsidies as a means of mitigating the ill-effects of uneconomically high wage-rates is, therefore, substantially weaker for export industries than for others. Even so, however, it is clear that the subsidy device is applicable over a considerable range. It will lessen the volume of unemployment in any industry (with uneconomically high wages) to which it is applied; and, so long as it is on an appropriate scale and is confined to industries whose products are not exported, it will correspondingly increase the real income of the country.


§ 6. The foregoing analysis is in principle favourable to a policy of wage subsidies, at all events in industries other than export industries, provided that the maintenance of uneconomically high wage-rates is taken for granted. When, however, we pass from generalities to more detailed considerations, pitfalls are revealed. The most obvious difficulty has to do with the comparative treatment of workpeople in different occupations. If all occupations were rigidly separated from one another, so that, not only could no one pass directly from one to another, but also the choice among them to be made by each new generation coming to industrial age was rigidly fixed, everything would be quite simple. Each occupation could be treated as a single problem. In real life, however, different occupations are not rigidly separated, and account must, therefore, be taken of possible effects of a policy of subsidies in modifying the proportions of workpeople attached to different occupations. If exactly equal fiscal encouragement were given to all occupations, no effects of this kind would tend to come about. In practice, however, it can hardly be doubted that larger subsidies would be paid in industries with low wage-rates and large unemployment than in others. For example, at the present time the relatively distressed engineering and ship-building industries would certainly demand more favourable treatment than, say, the railway industry. As the demand for the products of any industry fell off and distress became more pronounced, higher subsidies, both absolutely and relatively to those ruling in other industries, would always be called for. Such pleas would often be acceded to. As a consequence, too many people would be set to and kept at work in some industries and too few in others. Extraordinary strength and competence on the part of the Government would be needed to prevent a policy of wage subsidies from acting in this way. If these were not forthcoming the resulting social loss might well be large. There is also a second serious danger. If the Government were in a position to control the wage demands of the workpeople as well as the amount of the subsidies, and if it were absolutely impervious to political pressure, the adoption of the above policy would not lead to any change in the rate of wages demanded. In practice, however, once the policy was adopted and, as a result of it, unemployment reduced to a low level, there would be a strong temptation to workpeople to demand higher wage-rates, while employers, hoping to recoup themselves from an increased subsidy, might not resist these demands very strenuously. In this way both wage-rates and the rates of subsidy would be subjected to a continuous upward pressure. This tendency, which would exist even in a static community, would be accentuated in the actual world; for in times of boom wages would tend, as now, to go up; and, when, subsequently, depression came, there would be a powerful demand, very likely on the part of employers and workpeople acting together, for an addition to the subsidy to prevent them from falling again. The annual revenue required to provide the subsidy would thus tend to grow larger and larger continually. The burden imposed on non-wage-earners would be raised above the benefit accorded them, and the gap would grow always wider. The supply of the services rendered by them in work and saving would be discouraged; and in the end both the national dividend and the real absolute share of it enjoyed by workpeople would be diminished.


§ 7. The broad result then is this. If wage-earners insist on maintaining a real rate of wages above the economic level in the sense defined above, and if no mitigating action is undertaken by the State, an abnormal volume of unemployment, with all the material and moral waste that this implies, is the inevitable concomitant. In principle it appears that this evil is susceptible of large alleviation, of a kind not involving injury to society at large, by a system of wage subsidies. But in practice it is probable that the application of such a system would be bungled, and that a community which relied upon it would lose more than it gained.

Notes for this chapter

Cf. Part I. Chaps. IX.-XI.
The analysis which follows was suggested to me by Mr. Ramsey of King's College, Cambridge.
In the special case where the demand curve for labour is a straight line the net gain is equal to h(r - ½s); which is necessarily positive, provided that the rate of subsidy is less than twice the rate of contribution to unemployed workmen.

Part IV, Chapter VIII

End of Notes

64 of 73

Return to top