CH. 5 PLURALITY OF SECONDARY PRICE-LEVELS 67

is to say, it roughly corresponds to the price-level ofwhat in subsequent chapters we shall call workingcapital, i.e. of unfinished goods.

The older wholesale index-numbers were eitherunweighted, all the principal basic commoditiesbeing treated as of equal importance, or at bestroughly weighted, wheat for example being reckonedtwo or three times as important as tin, whereas itsreal relative importance is perhaps ten times or more.More recently, however, the best official index-numbershave been elaborately and scientifically weighted,on the basis of the relative importance of differentarticles in the national economy as indicated by theCensus of Production. The culmination of these im-provements up to the present is to be found in theexcellent wholesale index of the United States Bureauof Labour, which, since its latest revision (September1927), is based on 550 individual commodities scien-tifically weighted.

Nearly all the early index-numbers—such asJevons’s, Soetbeer’s, Sauerbeck’s and the Economist' s—were of the wholesale type; mainly because thiswas, formerly, the only type for the compilation ofwhich adequate statistics were forthcoming overa series of years. Since no other index - numberswere available, these were accepted without muchqualification, both in popular and in learned discus-sions of Monetary Problems, as indicative of “ thevalue of money ”. Most of us were brought up toemploy such index-numbers as Sauerbeck’s or theEconomist' s much too light-heartedly, and withoutsufficient warning that, while there might be nothingbetter available, nevertheless the actual divergencesbetween these indexes and the Purchasing Power ofMoney might prove to be, if we could calculate them,of very great significance both theoretical and prac-tical. This light-heartedness was much encouragedby the doctrine, fallacious in my opinion, put into