But the existence (or the importance) of such costs is one thing, their relation to money costs is another. In Chapter VIII we saw how difficult it was to establish any clear relation between the rate of interest and the sacrifice of saving. The costs of labor present similar difficulties. The relative irksomeness of two occupations may affect the relative wages which will rule in the two cases; so, certainly, will the differences in the cost of education and training which they require. But these are matters which concern the _apportionment_ of labor between different employments. There is no good reason to suppose that the general wage-level would be reduced, merely because work as a whole became less irksome, or involved a smaller physical or mental strain. The supply of people is not determined by the same kind of influences as is the supply of a commodity. Parents do not produce children for the sake of the wages which the children will receive when they go out to work; or, if this happens, we rightly regard it as a horrible anomaly. In so far as parents are affected by economic conditions it is by their own economic conditions; the question is rather one of how many children they can afford to have, than of a balancing of the cost to them against the incomes which their children may subsequently acquire. But other considerations enter in; and, in fact, it is doubtful how the aggregate supply of labor will react to changes in prosperity. Finally, the supply of land involves neither effort nor sacrifice; and, among our money costs, we have to account for the item of the rent of land. To dispose of this difficulty by arguing that rent does not enter into marginal costs (in any sense which is not equally true of wages and profits) is to lose contact with reality. Thus the attempt to explain money costs in terms of the costs of producing the ultimate agents of production leads us into a quagmire of unreality and dubious hypothesis. For a systematic theory, which will rest on firm foundations, we must interpret money costs in very different terms. The real costs which the price of a commodity measures are not absolute, but comparative. Marginal money costs reduce themselves in the last analysis to the payments which must be made to secure the use of the requisite agents of productions. These payments _tend_ to equal the payments which the same agents could have commanded in alternative employments. The payments which they could have commanded in alternative employments, tend in their turn to equal the derived marginal utilities of their services in those employments. It is thus the loss of _Utility_ which arises from the fact that these agents of production are not available for alternative employments that is measured by the money costs of a commodity at the margin of production.