74 ECONOMIC ESSAYS IN HONOR OF JOHN BATES CLARK 
phi Le of 
for the purposes of exchange to reduce them to common units 
of value. Due to the tendency of labor and capital to seek the 
most profitable lines of employment and the consequent tendency 
of wages of similar laborers to approach equality and of the 
interest rate to approach uniformity, the margin of production 
for both labor and capital runs through all industries and must 
therefore be measured in terms of value. It is idle to deny that 
this introduces grave complications into the theory of marginal 
productivity. 
Thus the value of a worker to an employer is measured by the 
money addition which he makes to the income of the concern 
rather than in terms of the benefit which he creates for society. 
Thus the process of reasoning by which the manufacturer of a 
quack patent medicine decides whether he shall engage another 
chemist is precisely the same as that by which a dairyman 
decides whether he shall employ another milkmaid. Each worker 
will tend to yield less profit to his employer than his predecessor 
although in one case the profits will come from conveying worth- 
less articles to adults and in the other from producing milk for 
children. Similarly those who are employed in giving an employer 
a competitive advantage over his fellows without increasing the 
national product as such all come under the principle of marginal 
productivity as do those who may actually decrease the total 
product in which they share. 
Yet even this dilemna may be mitigated if we measure the 
output of industry in the form of a composite of physical goods 
and of services as has been attempted in the various indexes of 
production compiled by Day, Stewart and others. Granted that 
there are parasitical elements within each volume of output, it 
will be enough if we assume that the proportion of parasitism 
remains the same, and consider the change in the product as a 
whole which accompanies changes in the quantity of the factors. 
And if it be objected that it is impossible to construct such an 
index of production because the values of commodities change 
from year to year, it can be pointed out that the problem has 
been virtually solved by Professor Irving Fisher's “ideal” index 
number whereby the weighted geometric mean of the ratios of 
value in the base year and in succeeding years may be secured.’ 
1 See Thorstein Veblen’s paper, “Industrial and Pecuniary Employments,” 
reprinted in The Place of Science in Modern Civilization. 
2 Fisher. The Making of Index Numbers.