Affiliations: [a] Center for Economic Studies, U.S. Bureau of the
Census, Washington, DC 20233, USA. Phone: + 1 301 763 2337, fax: + 1 301 763
7322
Note: [1] The views expressed are attributable to the authors and do not necessarily
reflect those of the Census Bureau. We benefitted from the comments and
assistance of Raphael Corrado, Mark Doms, Timothy Dunne, Kathy Friedman,
James Monahan, Sang Nguyen, Arnie Reznek, Mary Streitwieser, Kenneth Troske
and participants at seminars at the Center for Economics Studies and the
Industry Division.
Abstract: Establishment reclassification occurs when an establishment classified in one
industry in one year is reclassified into another industry in another year.
Because of survey design rules at the Census Bureau these reclassifications
occur systematically over time, and affect the industry-level time series of
output and employment. The evidence shows that reclassified establishments occur
most often in two distinct years over the life of a sample panel. Switches are
not only numerous in these years, they also contribute significantly to measured
industry change in industry output and employment. The problem is that
reclassifications are not necessarily processed in the year that they occur. The
survey rules restrict most change to certain years. The effect of these rules is
evidenced by looking at the variance across industry growth rates which
increases greatly in these two years. Whatever the reason for reclassifying an
establishment, the way the switches are processed raises the possibility of
measurement errors in the industry level statistics. Researchers and policymakers relying upon observations in annual changes in
industry statistics should be aware of these systematic discontinuities,
discrepancies and potential data distortions.
