[Accessibility Information]

Related BLS programs | Related articles

December  2002, Vol. 125, No.12

Is the ECI sensitive to the method of aggregation? an update

Michael K. Lettau, Mark A. Loewenstein, and Steve P. Paben

The Employment Cost Index, or ECI, measures changes in employers’ cost of
compensating workers, controlling for changes in the industrial-occupational composition of jobs. Employers’ labor cost has two components: wages and salaries, and the cost of all nonwage benefits, including employer costs for workers’ health insurance, employer contributions to workers’ pension plans, and employer Social Security contributions.

The ECI is a quarterly index that is computed from survey information on a sample of establishments and jobs, weighted to represent the universe of establishments and occupations in the economy. In computing the national ECI, the quotes reporting compensation for individual jobs must be aggregated into a single index number. The aggregation process involves two key steps: (1) estimating the mean compensation for each of the various classes of labor defined on the basis of industry and major occupation and (2) weighting the cell means for the different types of labor to obtain a single index number. Using both arithmetic and geometric cell means, Michael K. Lettau, Mark A. Loewenstein, and Aaron Cushner constructed fixed-weight, current-weight, and superlative indexes of the increase in private employers’ compensation costs.¹ They found that the estimation of compensation growth is not very sensitive to the choice of index formula employed.

The Consumer Price Index (CPI) faces methodological issues similar to those which confront the ECI—issues discussed at length in the Boskin report.² In August 2002, the Bureau of Labor Statistics began publishing a new index called the Chained Consumer Index for All Urban Consumers (C-CPI-U). This index employs a Törnquist formula and uses expenditure data in adjacent periods to eliminate substitution bias across expenditure categories. An experimental version of the index for the first half of the 1990s suggests that it grew annually by 0.2 percentage point less, on average, than the CPI-U. This difference has increased significantly in the years since then.³

In their analysis of the ECI, Lettau, Loewenstein, and Cushner reported on indexes from September 1981 to December 1994.⁴ There now are 6½ years of additional data. In light of the continued interest in the CPI methodology, it is useful to update the original study.

This excerpt is from an article published in the December 2002 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 article in PDF (48K)

Footnotes

1 See Michael K. Lettau, Mark A. Loewenstein, and Aaron Cushner, 
“Is the eci sensitive to the method of aggregation?” Monthly Labor Review, June 1997, pp. 3–11.

2  Michael J. Boskin, Ellen Dulberger, Robert Gordon, Zvi Griliches, and Dale Jorgenson, Toward a More Accurate Measure of the Cost of Living, Final Report to the U.S. Senate Finance Committee (Washington, dc, December 1996). For a summary of these issues and of the Boskin report itself, see the winter 1998 issue of the Journal of Economic Perspectives and Roger J. Gordon, “The Boskin Commission Report and Its Aftermath,” National Bureau of Economic Research Working Paper No. 7759, June 2000.

3  See “Note on a New, Supplemental Index on Consumer Price Change,” Aug. 16, 2002, available on the Internet at http://www.bls.gov/cpi/superlink.htm.

4 Lettau, Loewenstein, and Cushner, "Is the eci sensitive?"

Related BLS programs

Compensation Cost Trends

Related Monthly Labor Review articles

Is the ECI sensitive to the method of aggregation?—June 1997.
The Employment Cost Index: what is it?—Sept. 2001.