Chapter 11.
Industry Productivity Measures

Methods and Sources

Output
Industry output indexes are developed from data collected by the Bureau of the Census and other sources. Output indexes are developed as a deflated value of production or physical quantity of production of an industry. Each of these methods is discussed in turn.

Deflated value output indexes. More than two-thirds of the industry output indexes are derived from data on the value of industry output adjusted for price change. The adjustment for price change is accomplished by dividing the value of output by a price index and is known as a deflated value method.The resulting indexes are conceptually equivalent to indexes that are developed using data based on physical quantities of products.

To make maximum use of the comprehensive data from the Economic Censuses, output indexes are first derived from data for two consecutive quinquennial censuses; these indexes are referred to as benchmark indexes. Annual indexes for intercensal years are adjusted to the benchmark levels for the census years.

Benchmark indexes. For manufacturing industries, with the deflated-value methodology, current-dollar values of shipments are deflated with appropriate price indexes for each of the 5-digit product class groups, resulting in quantity indexes. The Tornqvist formula presented above is used in the computation of the real-value-of-shipments index for the industry. Additionally, to arrive at the final benchmark output index of production, adjustments are made to reflect net changes in inventories, changes in industry coverage, resales, and intraindustry transfers. Benchmark indexes are developed every 5 years, based on data from Census of Manufactures.

For trade industries, benchmark indexes are computed from sales data reported in the Census of Retail Trade. Current-dollar sales are deflated with appropriate price indexes for each category of merchandise in the industry, yielding constant-dollar sales. The Tornqvist formula is then used to calculate the real sales index for the industry. Additionally, to arrive at the final benchmark output index of production, an adjustment is made to reflect changes in industry coverage. Benchmark indexes are developed every 5 years, based on data from the Census of Retail Trade.

Benchmark indexes for the services industries are computed from data reported in the Census of Service Industries. The methodology is similar to that used in developing the indexes for the retail trade industries.

Benchmark indexes for the mining industries are computed from data reported in the Census of Mineral Industries.

Annual indexes. For manufacturing industries, the value of shipments for each 5-digit primary product (wherever made) is deflated by an appropriate price index. BLS producer price indexes (PPI's) matching these 5-digit product classes are used if available.If these price indexes are not available, deflators are developed by weighting together more detailed PPI's with base-year value of shipments weights. For recent years, industry PPI's have been available for over 90 percent of manufacturing industries.The Tornqvist formula presented above is used in the calculation of the real value-of-shipments index for the primary products of the industry.

For each year, special coverage ratios for the industry (total value of industry shipments of all products to total value of primary products, wherever made) are used to adjust the wherever-made indexes to the industry basis. The resultant industry indexes are further adjusted to reflect changes in inventories. These adjustments yield the estimated industry indexes of production.

For industries in trade and services, data on the value of sales for each year are divided by an industry price index to derive a measure of the change in the industries' real output. These industry price indexes are, for the most part, producer and consumer price indexes developed by BLS. In the case of retail trade industries, the industry price index is developed by combining current-year consumer price indexes with weights based on sales for each category of merchandise in Census years.

Physical quantity output indexes. Physical quantity output indexes are, where possible, a Tornqvist aggregation of quantities of component products. The basic data on quantities are generally primary products of an industry classified into product groups. The finest level of detail is used. For some industries, the annual indexes are adjusted to deflated-value benchmark indexes by linear interpolation. The indexes for both the annual and benchmark series are developed using the Tornqvist procedure.

Data for the physical quantity output indexes come from numerous sources, including the U.S. Departments of Commerce, Energy, and Transportation and reports from various trade associations. Physical quantity output indexes are used primarily for the mining and transportation industries, and for raw commodity manufactured products.

Sources. Industry output indexes are prepared from basic data published by various public and private agencies, using the greatest level of detail available.

Data from the Bureau of the Census, U.S. Department of Commerce, are used extensively in developing output statistics for manufacturing, trade, and services industries. The U.S. Geological Survey compiles most of the information for the mining and cement industries. Other important Government sources include the U.S. Departments of Energy, Agriculture, Transportation, and Housing and Urban Development, and the Federal Railroad Administration, the Federal Reserve Board, and the Federal Deposit Insurance Corporation. Important sources of trade association data include the Textile Economics Bureau, Inc., National Association of Hosiery Manufacturers, Inc., National Canners Association, Rubber Manufacturers Association, the American Iron and Steel Institute, Association of American Railroads, Ward's Communications, Rice Miller's Association, National Automobile Dealers Association, Pharmaceutical Manufacturers Association, American Bus Association, International Sleep Products Association, American Truck Association, American Paper Institute, Anti Friction Bearing Manufacturers' Association, Fiber Box Association, Institutional Furniture Manufacturers' Association, VISA-CARD Network and Interbank Card Network Association, National Automated Clearing House Association, Agricultural Chemicals Association, Association of Oil Pipelines, and the American Gas Association.

Labor input

The labor input indexes are developed by dividing the aggregate employee hours for each year by the base-period aggregate. Because of data limitations, employee hours are treated as homogeneous and additive with no distinction made between hours of different groups of employees. For industriesin which the self-employed are important, indexes are constructed for the hours of all persons, which includes paid employees, partners, proprietors, and unpaid family workers.

Industry employment and employee hour indexes are developed from basic data compiled by the Bureau of Labor Statistics and the Bureau of the Census. For most private nonagricultural industries, BLS publishes employment and average weekly hours data for production or nonsupervisory workers and employment data for all employees. The Bureau of the Census publishes employment and aggregate hours data for production workers and employment data for all employees.

BLS and the Bureau of the Census differ in their definition of employee hours and in their sampling and reporting methods. In general, BLS data are the preferred source for measuring industry employment and hours. Census employment is the sum of an average of production workers plus the number of other employees in mid-March. The average of production workers is the average for the payroll periods for the 12th of March, May, August, and November. In contrast, the BLS employment statistics program (790 Survey) collects employment and hours monthly and the employment levels are benchmarked each year to comprehensive data from the State unemployment insurance programs. (See chapter 2.)

Only employment data are available for nonproduction workers. The average annual hours of these workers must be estimated.The estimates of aggregate nonproduction worker employee hours for the manufacturing industries are derived from published employment data, and estimates of average annual hours per nonproduction worker.

Prior to 1968, the estimates of average annual hours worked were calculated by multiplying the number of workweeks in the year times the scheduled weekly hours. From 1968 to 1977, the estimates of average annual hours for nonproduction workers were based on data collected in the BLS biennial surveys of employee compensation in the private nonfarm economy. Since these surveys are no longer conducted, the 1977 levels of average annual hours per nonproduction worker are being carried forward until other data become available.

For the mining industries, estimates for the hours of nonproduction workers are based on data collected by the Mine Safety and Health Administration. For the trade and services industries, estimates are made for the hours of partners, proprietors, and unpaid family workers using unpublished data collected in the Current Population Survey, and for supervisory workers using data from the Census of Population.

All employee hours estimates for manufacturing industries are derived by summing the aggregate hours for production workers and the estimated aggregate hours for nonproduction workers. For trade and services industries, all-person hours estimates are derived by summing the aggregate hours for paid employees and the estimated aggregate hours for partners, proprietors, and unpaid family workers.

Next: Multifactor Productivity Measures