Related BLS programs | Related articles
December 1994, Vol. 117, No. 12
Seasonal adjustment of quarterly consumer expenditure series
Thomas G. Moehrle
To provide more timely information on consumer expenditures, the Bureau of Labor Statistics began publishing estimates from the interview component of the Consumer Expenditure (CE) Survey on a quarterly basis in 1984.1 With quarterly data, analysts can more easily and accurately track changes in overall consumer behavior but the degree of seasonality present in a quarterly expenditure series for individual expenditure groups will render these changes difficult to interpret. For this reason, seasonally adjusted expenditure data - data with the seasonality extracted - would be the preferred series to examine. This article presents results of seasonally adjusting data from the interview component of the CE Survey.2 (For a more general analysis of seasonality, see the article by Ted Jaditz, pp. 17-22, this issue.)
The seasonal behavior of the CE interview data was first studied by Stuart Scott and James Buszuwski,3 who analyzed differences in seasonality of selected expenditure series. Their analysis highlighted the importance of seasonal variation in the interview component data, and the tendency of seasonality to vary demographically - particularly with regard to region of residence and age of the reporting household's reference person. They concluded that "consumer expenditures are among the most highly seasonal series observed by BLS." For users of the quarterly data, Scott and Buszuwski's analysis is somewhat limited because the expenditure categories they analyzed were not based on published expenditure definitions, and because the study did not cover the total U.S. sample. Rather, their data spanned the period 1980-87, and included only urban consumer units.
This study revisits the issue of seasonality, but the analysis focuses on expenditure series as they are actually published in quarterly reports. The mean expenditures in these reports are based upon the total U.S. sample of households. Because the focus is exclusively on that total sample, seasonal differences that may arise across demographic groups are not addressed. The analysis covers the period 1984-92, and employs the X-11 ARIMA methodology, developed by Statistics Canada, to seasonally adjust the data.
This excerpt is from an article published in the December 1994 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 (552K)
1 See Maureen Boyle, "BLS to publish quarterly data from Consumer Expenditures Survey," Monthly Labor Review, July 1988, pp. 27-31.
2 The CE Survey consists of an interview component and a diary component. This article reports on the seasonality of quarterly data from the interview component alone. Data from the diary component are not published or processed on a quarterly basis, and seasonal analysis thus cannot be performed on these data. For a detailed description of the CE Survey, see BLS Handbook of Methods, Bulletin 2414 (Bureau of Labor Statistics, 1992), ch. 18.
3 Stuart Scott and James Buszuwski, "Seasonal pattern Assessment of Consumer Expenditures," Statistical Note no. 32 (Bureau of Labor Statistics, 1993)
Within Monthly Labor Review Online:
Welcome | Current Issue | Index | Subscribe | Archives
Exit Monthly Labor Review Online:
BLS Home | Publications & Research Papers