Changes in Quantity Weighting Factors

The last major revision of the indexes of retail food prices was made in March 1943, when in addition to other adjustments the quantity-weighting factors used in combining prices to compute index numbers were revised to take into account wartime consumption patterns. In February 1946, the quantity weights were readjusted to eliminate most of the previous changes since the current average per capita consumption for most foods, as estimated by the Department of Agriculture, approximates the average for the period 1935 through 1939. Weights were reduced for cereals and bakery products, pork, chicken, milk, eggs, lard, oleomargarine, and peanut butter. Weights were increased for beef, lamb, fish, butter, canned and dried fruits and vegetables, tea, coffee, shortening, sugar, and sweets. Weights for some fresh fruits and vegetables were increased, others decreased. In general, the new weights represent the consumption pattern for moderate-income families as determined by the Bureau's surveys of consumer expenditures in 1934-36, with minor adjustments to take care of 7 foods added in March 1946.

Table 1 shows the percentage of the index represented by each component food for January 1946, as originally computed with the wartime weights and as computed with the new quantity weights. The relative importance of the individual foods in the base period, 1935-39, has been included for comparison."

For a complete description of all the changes made in March 1943, see Monthly Labor Review, July 1943-Bureau of Labor Statistics Cost of Living Index in Wartime; and also LS 43-5481 (mimeographed) Description of Changes in the BLS Cost of Living Index as of March 15, 1943.

The quantity weights for these foods had been assigned to others having similar price characteristics during the period January 1935 through February 1943, and the separation of their weights at that time made a difference in the weights of foods to which they had been assigned. These 7 foods have been retained in

the index.

'The relative importance of a food in the index must not be confused with the quantity weights, as the former is derived from the relationship of the price aggregates for individual foods to the total price aggregate for all foods. The price aggregate for a food is the product of quantity weight times price for that period, and therefore is a function of price, varying as the trend of prices for that article varies from the average trend for all foods combined. Differences in the relative importance for the various foods in January 1946, revised, and the base period are primarily due to the price variable. Because of space limitations the quantity weights cannot be printed with this article, as there is a separate quantity-weighting factor for each food in each of the cities included in the index.

TABLE 1.—Relative importance of individual foods in the retail food price index Percentage distribution for January 1946 as computed with original and revised quantity weights and for base period, 1935–39

TABLE 1.—Relative importance of individual foods in retail food price index-Continued Percentage distribution for January 1946 as computed with original and revised quantity weights and for base period, 1935-39—Continued

Revised store samples were incorporated into the index city by city as they were completed. All the changes in procedures were introduced simultaneously in February 1946, with prices for January recomputed in accordance with the new methods. The four changes which affected average prices in individual cities were (A) Revisions of food store samples; (B) Elimination of State and local taxes from average prices; (C) Changes in methods of combining quotations; and (D) Changes in editing procedures.

The results of the introduction of these new procedures are illustrated in table 2, which shows differences between average prices in Seattle, Wash., for January 1946 recomputed by introducing each one of the new procedures separately and average prices for Seattle as originally computed for January 1946.

The effect of these changes is minor in character except for a relatively few foods. With the exception of the differences resulting from elimination of sales taxes, about three-fourths of the average prices were not changed or were changed by 0.5 cent or less. Even when the obvious reductions in average prices caused by the removal of sales taxes are taken into account, prices for 40 of the 77 foods remained unchanged or differed by 0.5 cent or less from the January averages computed before the revisions. The relatively small differences are due, in part, to the small units of quantity on which prices are reported. The price of 15 pounds of potatoes, however, the

largest single price in the list of foods, had a net decrease of only 1.1 cents from the former average of 71.3 cents, and the changes resulting from each element of the revised procedures, except tax elimination, were less than 1 cent. Differences in average prices for the 40 cities in which no sales taxes are in effect were smaller than in Seattle.

TABLE 2.-Distribution of differences between average prices for Seattle as originally published, and average prices for Seattle resulting from sample revision and from each phase of revised procedures, January 1946

1 Since the same article of food did not necessarily fall within the same range of differences in all four cases, figures in the first four columns are not additive to arrive at the number in the last column. For various reasons, not all foods priced could be included in all comparisons.

The net effect of all these changes was less in the averages for the 56 cities combined than for the individual cities. The differences between the revised and original average prices were all less than 10 percent of the original. For 18 of the 78 foods priced, there was no difference between the original and revised January averages; revised prices were higher for 36 of the foods and lower for 24 of them. For more than half of the 60 foods for which differences in prices occurred, the amount of the difference was 0.3 cent or less. The revised average prices for January for each of the 56 cities and for all cities combined are published in the Bureau's monthly release Retail Food Prices by Cities, February 1946.

The changes in the levels of prices resulting from revised samples and procedures were treated so as not to affect the levels of the indexes of retail prices of foods or the consumers' price index. The new prices and weights were introduced into the index by a linking process so that the indexes constitute continuous series and may be used for the study of time trends in food prices at retail.

Index of Consumers' Prices in Large Cities, November 19461

RETAIL prices of consumer goods in large cities rose 2.2 percent between mid-October and mid-November, bringing the consumers' price index for the United States to the highest level on record. On November 15 the index was 151.7 (1935-39 equals 100), 1.5 percent higher than in June 1920, the peak after World War I.

In mid-November consumers' prices were 13.8 percent higher than on June 15, 1946, and 53.9 percent higher than in August 1939. In the year between November 1945 and November 1946 food prices rose 34 percent, clothing 13 percent, and housefurnishings 15 percent. Residential rents, however, advanced only slightly, and gas and electricity costs declined about 2 percent.

Retail food prices to moderate-income families in large cities advanced 4.3 percent between mid-October and mid-November. Higher prices were reported for all major food groups except eggs and dairy products, but sharp increases in prices of fats and oils, meats, and fruits and vegetables were chiefly responsible for this advance. Average retail prices of fats and oils rose 65 percent, as prices for lard increased 104 percent, shortening 66 to 85 percent, and oleomargarine 56 percent. Meat, poultry, and fish averaged 9.1 percent higher in 'The "consumers' price index for moderate-income families in large cities," formerly known as the "cost of living index," measures average changes in retail prices of selected goods, rents and services, weighted by quantities bought by families of wage earners and moderate-income workers in large cities in 1934-36. The items priced for the index constituted about 70 percent of the expenditures of city families whose incomes averaged $1,524 in 1934-36.

The index only partially shows the wartime effects of changes in quality, availability of consumer goods, etc. The President's Committee on the Cost of Living has estimated that such factors, together with certain others not fully measured by the index, would add a maximum of 3 to 4 points to the index for large cities between January 1911 and September 1944. If account is taken of continued deterioration of quality and disappearance of low-priced merchandise between September 1944 and September 1945, which was estimated at an additional 1⁄2 point, the total large-city adjustment would be 432 points. If small cities were included in the national average, another 1⁄2 point would be added, making the total approximately 5 points. The indexes in the accompanying tables are based on time-to-time changes in the cost of goods and services purchased by wage earners and lower-salaried workers in large cities. They do not indicate whether it costs more to live in one city than in another. The data relate to the 15th of each month, except those for January 1941, in tables 1 and 2. They were estimated for January 1, 1941, the base date for determining allowable "cost of living" wage increases under the Little Steel formula and under the wage-price policy of February 1946. January 1, 1941, indexes in tables 1 and 2 have been estimated by assuming an even rate of change from December 15, 1940, to the next pricing date.

Food prices are collected monthly in 56 cities during the first 4 days of the week which includes the Tuesday nearest the 15th of the month. Aggregate costs of foods in each city, weighted to represent food purchases of families of wage earners and lower-salaried workers, have been combined for the United States with the use of population weights. In March 1943, the number of cities included in the food index was increased from 51 to 56, and the number of foods from 54 to 61. Prices of clothing, housefurnishings, and miscellaneous goods and services are obtained in 34 large cities in March, June, September, and December. In intervening months, prices are collected in 21 of the 34 cities for a shorter list of goods and services. are surveyed semiannually in most of the 34 cities (in March and September, or in June and December). In computing the all-items indexes for individual cities and the rent index for the average of large cities, because of the general stability of average rents at present, the indexes are held constant in cities not surveyed during the current quarter. Prices for fuel, electricity, and ice are collected monthly in 34 large cities.

Rents