When BEA revises the gross product by industry estimates to make them consistent with the comprehensive revision of the NIPA's, estimates of real gross product by industry will be calculated using fixed 1987 price weights. Updating the fixed weights from 1982 to 1987 will have the largest impact on manufacturing—specifically, on the nonelectrical machinery industry, in which computers and peripheral equipment are produced. BEA will also prepare alternative estimates of real gross product by industry using benchmark-years weights. (BEA does not plan to calculate chain-type annual-weighted indexes—the other alternative measure presented in the article—for real gross product by industry, because less product detail is available annually than for benchmark years.)
Because of the substantial change in the relative price structure—traceable largely to the declining prices of computers and peripheral equipment—the use of fixed 1987 price weights for the gross product by industry series will not adequately portray the course of manufacturing in the late 1970's or early 1980's. Likewise, use of fixed 1977 or fixed 1982 price weights will not adequately portray manufacturing in the late 1980's. The benchmark-years-weighted index, which allows for change in the relative price structure, will present a more accurate picture.
Exhibit 1 shows the differences in growth rates for manufacturing that result from the use of prices of either 1977, 1982, or 1987 as fixed weights in calculating an index of real manufacturing gross product. The exhibit also shows the growth rates that result from the use of benchmark-year prices as weights. The index based on 1982 price weights essentially corresponds to the 1982 dollar series released in April 1991, before the comprehensive revision of the NIPA's (it incorporates small revisions that have little effect on the growth rates). The other three indexes are calculated from the same price and quantity information as used for the 1982 dollar series. When the gross product by industry indexes are revised, the growth rates may differ from those in the exhibit; however, the pattern of differences in growth rates will remain about the same.
The lack of additivity of the benchmark-years-weighted quantity index may present a problem to some users of the gross product by industry series. Within the framework of the benchmark-years-weighted quantity indexes, it may not be possible to calculate, in an exact sense, a time series for manufacturing as a share of total output. However, by comparing the growth rates of manufacturing with those of total GDP, it will be possible to determine whether manufacturing gross product increased more or less rapidly than total GDP.
