The argument that labor should receive enough to buy back the product is merely a special form of the general “purchasing-power” argument. The workers’ wages, it is correctly enough contended, are the workers’ purchasing power. But it is just as true that everyone’s income—the grocer’s, the landlord’s, the employer’s— is his purchasing power for buying what others have to sell. And one of the most important things for which others have to find purchasers is their labor services.
All this, moreover, has its reverse side. In an exchange economy everybody’s money income is somebody else’s cost. Every increase in hourly wages, unless or until compensated by an equal increase in hourly productivity, is an increase in costs of production. An increase in costs of production, where the government controls prices and forbids any price increase, takes the profit from marginal producers, forces them out of business, and means a shrinkage in production and a growth in unemployment. Even where a price increase is possible, the higher price discourages buyers, shrinks the market, and also leads to unemployment. If a 30 percent increase in hourly wages all around the circle forces a 30 percent increase in prices, labor can buy no more of the product than it could at the beginning; and the merry-go-round must start all over again.
No doubt many will be inclined to dispute the contention that a 30 percent increase in wages can force as great a percentage increase in prices. It is true that this result can follow only in the long run and only if monetary and credit policy permit it. If money and credit are so inelastic that they do not increase when wages are forced up (and if we assume that the higher wages are not justified by existing labor productivity in dollar terms), then the chief effect of forcing up wage rates will be to force unemployment.
And it is probable, in that case, that total payrolls, both in dollar amount and in real purchasing power, will be lower than before. For a drop in employment (brought about by union policy and not as a transitional result of technological advance) necessarily means that fewer goods are being produced for everyone. And it is unlikely that labor will compensate for the absolute drop in production by getting a larger relative share of the production that is left. For Paul H. Douglas in America and A. C. Pigou in England, the first from analyzing a great mass of statistics, the second by almost purely deductive methods, arrived independently at the conclusion that the elasticity of the demand for labor is somewhere between 3 and 4. This means, in less technical language, that “a 1 percent reduction in the real rate of wage is likely to expand the aggregate demand for labor by not less than 3 percent.”* Or, to put the matter the other way, “If wages are pushed up above the point of marginal productivity, the decrease in employment would normally be from three to four times as great as the increase in hourly rates”† so that the total incomes of the workers would be reduced correspondingly.
Even if these figures are taken to represent only the elasticity of the demand for labor revealed in a given period of the past and not necessarily to forecast that of the future, they deserve the most serious consideration.