Related BLS programs | Related articles
June 1997, Vol. 120, No. 6
Michael K. Lettau, Mark A. Loewenstein, and Aaron Cushner
T he Employment Cost Index (ECI), published quarterly by the U.S. Bureau of Labor Statistics (BLS), measures changes in the price of labordefined as total compensation per employee hour worked. As a fixed-weight or Laspeyres index , the ECI controls for changes occurring over time in the industrialoccupational composition of employment. The ECI is computed from survey data on compensation by occupation, collected from a sample of establishments and occupations, weighted to represent the universe of establishments and occupations in the economy.1
The cost to employers for employee compensation has two components: wages or salaries paid to employees, and the cost of all nonwage benefits. The wage and salary component of the ECI is represented by straight-time average hourly earnings in an occupation, whether or not the employees are actually paid by the hour. Nonwage benefits, which account for about 30 percent of total compensation costs, include such things as employer contributions to employees health and other insurance, pension plans, and Social Security, as well as paid vacations and sick leave, premium pay for overtime, and nonproduction bonuses. As with the wages portion of compensation, benefit cost data are converted to an hourly basis in the ECI.
In computing the national ECI, the myriad wage quotes from the sample of individual jobs must somehow be aggregated into a single index number. This aggregation process involves two key steps. Each establishment surveyed for the ECI is placed within 1 of 73 two-digit SIC industries, and each surveyed job is placed within 1 of 10 major occupation groups.2 Because all of the occupations are not represented in all of the industries, only 720 industryoccupation cells exist (as opposed to 730). Each job quote in the survey falls into exactly one of these cells. The first step in the aggregation process involves combining all of the job quotes within a given cell to obtain a cell average. The second step involves aggregating across the cell averages to obtain the ECI.3
Another index produced by BLS using similar methodology to that used in the ECI is the Consumer Price Index (CPI). The CPI has received a great deal of attention recently concerning its alleged upward bias. Because the two indexes are constructed similarly, some of the alleged problems in the CPI also may exist in the ECI. This article examines the sensitivity of the ECI to the method of aggregation.
This excerpt is from an article published in the June 1997 issue of the Monthly Labor Review. The full text of the article is available in Adobe Acrobat's Portable Document Format (PDF). See How to view a PDF file for more information.
Read abstract Download full text in PDF (76K)
1 The ECI covers all establishments and occupations in both the private nonfarm and public sectors. As such, private sector coverage is limited to the private nonfarm economy, excluding private household workers. Public sector coverage includes employees of State and local governments, but excludes Federal workers. Finally, the self-employed, owner-managers, and unpaid family workers are excluded from coverage in the ECI.
For a more complete description of the ECI, see BLS Handbook of Methods, Bulletin 2414 (Bureau of Labor Statistics, 1992), ch. 8, "The Employment Cost Index." Also, see the following two articles from the Monthly Labor Review: Victor Sheifer, "Employment Cost Index: a measure of change in the price of labor," July 1975, pp. 312; and G. Donald Wood, "Estimation procedures for the Employment Cost Index," May 1982, pp. 4042.
2 Prior to March 1995, only nine major occupational groups were used.
3 Note that an analogous procedure is used to calculate the CPI. Specifically, the goods and services that consumers purchase are classified into 207 groups and urban areas are divided into 44 areas. The first step in constructing the CPI involves estimating the mean price of each of the 9,108 (207*44) basic CPI components. The second step involves aggregating the basic components to obtain a single index number. For a more detailed description, see the following articles in the December 1993 issue of the Monthly Labor Review: Dennis Fixler, "The Consumer Price Index: underlying concepts and caveats," pp. 312; Brent R. Moulton, "Basic components of the CPI: estimation of price changes," pp. 1324; and Ana M. Aizcorbe and Patrick C. Jackman, "The commodity substitution effect in CPI data, 198291," pp. 2533.
Within Monthly Labor Review Online:
Welcome | Current Issue | Index | Subscribe | Archives
Exit Monthly Labor Review Online:
BLS Home | Publications & Research Papers