This page contains links to data on the relative importance of components in the Consumer Price Index for All Urban Consumers (CPI-U) and the Consumer Price Index for Urban Wage Earners and Clerical Workers (CPI-W). These data are to be used in conjunction with the CPI-U and CPI-W released in that same year.
Table 1 contains data for the U.S. city average. Tables 2 through 6 contain data for 27 metropolitan areas, 4 regions, 3 population size classes, and 10 cross-classifications of area and population size class. Table 7 presents the relative importance of the individual area all-items indexes in the U.S. city average all-items indexes.
The relative importance of a component is its expenditure or value weight expressed as a percentage of all items within an area. When the value weights are collectedmost recently during the 2003-2004 Consumer Expenditure Surveythey represent average annual expenditures, and their relative importance ratios show approximately how the index population distributes expenditures among the components. Relative importance ratios represent an estimate of how consumers would distribute their expenditures as prices change over time.
Relative importance ratios cannot be used as estimates of current spending patterns or as indicators of changing consumer expenditures in the intervals between weight revisions because consumption patterns are influenced by factors other than price change. These factors include income, variations in climate, family size, and availability of new and different kinds of goods and services.
Relative importance ratios of components in the national or local area Consumer Price Indexes can be used in the construction of indexes for special combinations of items. In such instances, relative importance ratios are used as weights to combine relative changes in prices of the selected components over specified periods.
For a description of the procedure for deriving index weights from consumer expenditure data, see chapter 16, Consumer Expenditures and Income and chapter 17, The Consumer Price Index in BLS Handbook of Methods. BLS publishes the expenditure weight, or "relative importance," of each component in the CPI once a year, using December data. In fact, relative importances change every month, reflecting the change in relative prices.
To estimate a relative importance for a component for a month (other than December), one can use its previous published relative importance and update it by published price changes. For example, suppose one wants to estimate the relative importance of energy for the CPI-U in Sept. 2005. To answer this, one needs the published relative importance for energy for December 2004. One also needs the Dec. 2004 and Sept. 2005 indexes for energy and for all items.
In this example, one would enter the weights and indexes for these two item categories (see the first 4 columns). The updated weight column is the December published weight times the relative change between Dec. 2004 and Sept. 2005.
Specifically, in this example, the updated weight for energy is 7.991 * (208.0/153.7) = 10.8141.
For all items, the updated weight is 100.000 * (198.8/190.3) = 104.4666.
To calculate the updated relative importance for energy where the weight for all items is normalized to 100, just divide the updated weight for energy by the updated weight for all items, times 100.
In this example, the estimated relative importance for energy in Sept. 2005 is (10.8141 / 104.4666) X 100 = 10.352.
Table 1. Estimating an updated relative importance for energy for Sept. 2005
|Item||Published relative importance, Dec. 2004||Index, Dec. 2004||Index, Sept. 2005||Updated weight, for Sept. 2005 [Dec. 2004 rel. imp. X (Sept. 2005 index / Dec. 2004 index)]||Updated weight, Sept. 2005, normalized so that all items = 100.000|
|100.000||190.3||198.8||104.4666||Normalized to 100.000|
Suppose that, in the above example, energy prices increase 1.0 percent in October, while the All Items index increases 0.2 percent. How does one figure out the contribution of the Energy component to the All Items change? Asked another way, what proportion of the All Items increase can be attributed to the Energy component?
The first thing one needs to do is estimate the updated relative importance for Energy for Sept. 2005 (see Table 1 above, last column). One can then multiply the updated expenditure weight of Energy times its relative price change in October (10.352 X 1.01 = 10.456). Similarly, the updated expenditure weight for All Items is 100 X 1.002 = 100.200.
The change in the expenditure weight for Energy in October is 10.456-10.352=0.104.
The change in the expenditure weight for All Items in October is 100.200-100.00=0.200.
The contribution of energy to the All Items change equals the change in the expenditure weight for Energy, divided by the change in the expenditure weight for All Items. Specifically, the contribution of Energy to All Items in this hypothetical example is 0.104 / 0.200 = 0.52 or 52 percent. Said another way, slightly more than half the increase in the October index was due to the increase in energy prices.
Table 2. Estimating the contribution of energy to the All items change in October. 2005
|Item||Normalized/updated weight for Sept. 2005 (from Table 1)||Price change from Sept. 2005 to Oct. 2005, expressed as a relative||Updated weight, Oct. 2005 (updated weight X price change)||Differences in weights|
|10.352||+ 1.0 percent, or 1.01||10.456||10.456 10.352 = 0.104|
|100.000||+ 0.2 percent, or 1.002||100.200||100.200 100.000 = 0.200|
Contribution of Energy to All Items
|-||-||-||0.104/0.200 = 0.52 =52 percent|
Text files of U.S. level Relative Importances for the years 1987-2002 are available from our FTP server.
Last Modified Date: February 14, 2013