TRANSACTIONS OF SECTION F. 579 
This relation corresponds to a percentage movement of retail prices equal to 
three-quarters of the percentage movement of wholesale prices. Column A in the 
following table shows the movements of the wholesale prices of the goods 
selected ; Column B the movement of retail prices from this equation; Column C 
the movement on the hypothesis that cost of distribution remains unchanged ; 
Column D the weighted average of the Board of Trade’s index-numbers for the 
same commodities, the numbers in brackets showing the effect of subtracting the 
tea and sugar duties from retail prices. Column EK gives the Board of Trade’s 
number for London including all food. 
— Price Movements on Various Hypotheses. Quinquennial Averages. 
ee B C D E 
ae 132 193) |. . 121 a as 
1, 112 108° =|" 107 ae did 
1888-92 . . 104 102 | ~=102 dies = 
or o7 . 93 ga) ||) 9g 914 (93) 914 
1898-1902... 97 Spee MEE ogg 97 (974) 964 
oer |) 100 100 | 100 100 (100) 100 
mango. 111 107 107 107} (1072) 1063 
oy hia 113 109 108 fas esi 
It is suggested that retail measurements are in some cases so uncertain that it 
is necessary to explain and harmonise such differences as are shown between 
Golumns C and D before either can be used as minutely correct, and that it is 
not improbable that the figure even in Column C for 1893-7 is too low. 
(ii) The Construction of Index Numbers to Show Changes in the Cost 
of the Principal Articles of Food for the Working Classes. By 
Mrs. Frances Woop, B.Sc. 
At present the only food index numbers published annually are those given 
for London by the Board of Trade in their ‘ Abstract of Labour Statistics,’ but 
as no information is given as to the data, methods, &c., used in their construc- 
tion, it is impossible to judge whether they are trustworthy. In spite of this 
they are almost universally quoted as applying not only to London but even to 
the United Kingdom as a whole! 
Weekly or monthly records of the prices charged by retailers for the prin- 
cipal articles of food are very much to be preferred to records of the prices paid 
by consumers, i.e., family budgets, as data upon which to base food index 
numbers, and, if the results are to be taken as applying especially to the 
working classes, the records should be obtained, if possible, from working-class 
firms. Family budgets must, however, be obtained to give the articles com- 
mouly consumed by the working classes as well as the relative importance of 
the different articles. 
Series of index numbers for bread, prepared by the author from the returns 
of seven large London firms, show that the prices charged even by those firms 
which deal with the same class of customer do not show similar variations from 
year to year, and point to the conclusion that reliable index numbers for a 
eg town can only be obtained from the combined returns of a number of 
rins. ; 
The kind of the commodity studied can and must be kept constant during the 
course of an investigation, e.g., Australian mutton one year must not be com- 
pared with English mutton the next year. But the quality of Australian mutton, 
for example, may vary from year to year, so that the actual quality of the article 
studied will not necessarily be kept constant. Since, however, the investigation 
is concerned with changes in the cost of living, it is doubtful whether the index 
numbers need be modified on account of such changes in quality only. In the 
case of commodities such as butter and eggs, especially among working-class 
firms, no practical attempt is made to differentiate between the different kinds. 
It is accordingly impossible to follow changes in the price of any individual 
PPR2 
