But now let us suppose that the increase in wage rates is accompanied or followed by a sufficient increase in money and credit to allow it to take place without creating serious unemployment. If we assume that the previous relationship between wages and prices was itself a “normal” long-run relationship, then it is altogether probable that a forced increase of, say, 30 percent in wage rates will ultimately lead to an increase in prices of approximately the same percentage.
The belief that the price increase would be substantially less than that rests on two main fallacies. The first is that of looking only at the direct labor costs of a particular firm or industry and assuming these to represent all the labor costs involved. But this is the elementary error of mistaking a part for the whole. Each “industry” represents not only just one section of the productive process considered “horizontally,” but just one section of that process considered “vertically.” Thus the direct labor cost of making automobiles in the automobile factories themselves may be less than a third, say, of the total costs; and this may lead the incautious to conclude that a 30 percent increase in wages would lead to only a 10 percent increase, or less, in automobile prices. But this would be to overlook the indirect wage costs in the raw materials and purchased parts, in transportation charges, in new factories or new machine tools, or in the dealers’ mark-up.
Government estimates show that in the fifteen-year period from 1929 to 1943, inclusive, wages and salaries in the United States averaged 69 percent of the national income. In the five-year period 1956—1960 they also averaged 69 percent of the national income! In the five-year period 1972—1976 wages and salaries averaged 66 percent of national income, and when supplements are added, total compensation of employees averaged 76 percent of national income.[9] These wages and salaries, of course, had to be paid out of the national product. While there would have to be both deductions from these figures and additions to them to provide a fair estimate of “labor’s” income, we can assume on this basis that labor costs cannot be less than about two-thirds of total production costs and may run above three-quarters (depending upon our definition of labor). If we take the lower of these two estimates, and assume also that dollar profit margins would be unchanged, it is clear that an increase of 30 percent in wage costs all around the circle would mean an increase of nearly 20 percent in prices.
But such a change would mean that the dollar profit margin representing the income of investors, managers and the self-employed, would then have, say, only 84 percent as much purchasing power as it had before. The long-run effect of this would be to cause a diminution of investment and new enterprise compared with what it would otherwise have been, and consequent transfers of men from the lower ranks of the self-employed to the higher ranks of wage-earners, until the previous relationships had been approximately restored. But this is only another way of saying that a 30 percent increase in wages under the conditions assumed would eventually mean also a 30 percent increase in prices.
It does not necessarily follow that wage-earners would make no relative gains. They would make a relative gain, and other elements in the population would suffer a relative loss, during the period of transition. But it is improbable that this relative gain would mean an absolute gain. For the kind of change in the relationship of costs to prices contemplated here could hardly take place without bringing about unemployment and unbalanced, interrupted or reduced production. So that while labor might get a wider slice of a smaller pie, during this period of transition and adjustment to a new equilibrium, it may be doubted whether this would be greater in absolute size (and it might easily be less) than the previous narrower slice of a larger pie.