F.— ECONOMIC SCIENCE AND STATISTICS 143 



denominators. Certainly, it cannot be assumed that for all industries 

 there should be the same normal profit per capital on the books. State 

 taxation policy abandoned this assumption under the Excess Profits Duty 

 and, after a temporary lapse last spring, abandoned it again in the new 

 National Defence Contribution. Thus, to be scientific, enquirers such 

 as Epstein and Paton must isolate the comparison of profits per capital 

 for different sizes of firms within single industries. Epstein gives data 

 permitting others to make this isolated comparison. His conclusion that 

 the smaller corporations are more profitable than the larger is based on 

 an analysis of corporations whose average capital was $13,500,000. These 

 corporations are not, unfortunately, broken down into size classes for 

 specific industries. But a series of separate chapters give profit rates 

 for particularly small corporations with an average capital of $171,000, 

 and for twenty-three specific industries (excluding major and all mis- 

 cellaneous groupings) the rate of profit of a sufficient number of these 

 corporations is given to form a reliable average for the industry. Com- 

 paring this average for the small corporations with the average rate of 

 profit for the large corporations as a whole ^° industry by industry, the 

 small corporations are found on the average for the five years 1924 to 

 1928 to have the higher profit in seven industries,^! the larger corporations 

 the higher profit in sixteen industries. ^^ 



Clearly, until the profit rates of different sizes of corporations within 

 the whole larger-sized group is analysed industry by industry, it is rash 

 to accept Epstein's conclusion from his unanalysed group, that the smaller 

 manufacturing corporations tend to be the more profitable. The very 

 small corporation is probably not. 



Rates of profit on capital even when compared within isolated industries 

 are a difficult measure wherewith to interpret efficiency. Comparisons 

 of costs as well as profits on output or sales are of greater direct use as a 

 guide to policy, and economists must welcome the detailed cost analysis 

 of distributive firms that are being published in America by the Federal 

 Government and various Research Bureaus. One of the most remarkable 

 results was found by the first (1930) U.S. Census of Distribution in 

 averaging the operating costs of wholesale merchants of different size. As 

 the size increased from firms selling under $25,000 a year right up to 

 firms selling over $1,000,000, operating costs per $100 of sales fell con- 

 tinuously, for the nine successive classes of size, from $26-89 to 

 $9-66. 



In retail trade the outstanding cost and profit analysis are those con- 

 ducted by the Harvard Bureau of Business Research. They certainly do 



^o The comparison is slightly biassed in favour of the profitability of the small 

 corporations by the fact that these did not include corporations suffering a deficit 

 whereas the large corporations did. During the years 1922 to 1929 Epstein 

 (op. cit. p. 350) does not consider this inclusion to be serious because of the rela- 

 tively slight number and extent of such deficits. 



*i Dairying, meat-packing, lumber manufacturing, blank paper, stationery 

 castings and forgings, sheet metal. 



'^ Bakery, flour, confectionery, men's clothing, knit goods (hosiery), furniture, 

 cardboard boxes, newspapers, job-printing, paints, petroleum refining, ceramics, 

 heating and ventilating machinery, electrical machinery, motor vehicles, tools. 



