Business employment dynamics: Tabulations by empolyer size

Monthly Labor Review

Volumn 129, Issue 2, 2006, Pages 3-22

  Butani, Shail J ^a   Clayton, Richard L ^a   Kapani, Vinod ^b   Spletzer, James R ^b   Talan, David M ^b   Werking Jr , George S ^b  

^a Division of Administrative Statistics and Labour Turnover

^b BUREAU OF LABOR STATISTICS   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 33644813252     PISSN: 00981818     EISSN: 00981818     Source Type: Journal    
DOI: None     Document Type: Article

References (30)

1. This point was clearly documented in Cordelia Okolie, "Why size class methodology matters in analyses of net and gross job flows," Monthly Labor Review, July 2004, pp. 3-12.
2. For example, see The Employment Situation: January 2006, USDL 06-160 (U.S. Department of Labor) Feb. 3, 2006. Table A-12 of the monthly Employment Situation press release is titled "Alternative measures of labor underutilization" and lists the six unemployment rates.
3. For more information, see John E. Bregger and Steven E. Haugen, "BLS introduces new range of alternative unemployment measures," Monthly Labor Review, October 1995, pp. 19-26.
4. These data refer to the first quarter of 2004, and are from the BLS Quarterly Census of Employment and Wages (QCEW) available on the Internet at www.bls.gov/cew (accessed June 2005).
5. These data refer to the first quarter of 2004, and are from the BLS Quarterly Census of Employment and Wages, on the Internet at www.bls.gov/ cew/ (accessed June 2005).
6. A comprehensive survey article is Walter Oi and Todd Idson, "Firm Size and Wages," in Orley Ashenfelter and David Card, eds., Handbook of Labor Economics (Amsterdam, North-Holland Press, 1999), pp. 2165-2214.
7. See Charles Brown and James Medoff "The Employer Size-Wage Effect," Journal of Political Economy, October 1989, pp. 1027-59.
8. See Kenneth R. Troske, "Evidence on the Employer Size-Wage Premium from Worker-Establishment Matched Data," Review of Economics and Statistics, February 1999, pp. 15-26.
9. Steven J. Davis, John C. Haltiwanger, and Scott Schuh, Job Creation and Destruction (Cambridge, MA, MIT Press, 1996)
10. and Okolie, "Why size class methodology matters," Monthly Labor Review, 2004.
11. At some point in the near future, BLS hopes to produce BED size-class statistics at the establishment level and release these for research purposes.
12. Defining a size class for the employment change that crosses a size class threshold requires some discussion. For example, should the growth of 1 employee from 4 to 5 that moves the firm from the "14" size class to the "5-9" size class be credited to the "1-4" or the "5-9" size class? The dynamic sizing methodology classifies a firm that moves from 4 to 5 employees in the "5-9" size class. This is done for two reasons. First, we want employment change to be symmetrical - the loss of 1 job from 5 to 4 should be credited to the same size class as the gain of 1 job from 4 to 5. Second, because there is no "zero" size class, the first job credited to a birth needs to be attributed to the "1-4" size class, which would be symmetrical to the last job lost credited to a death being attributed to the "1-4" size class. Thus, when a firm moves from one size class to another by either expanding or contracting, dynamic sizing credits the single job that moves the firm across the threshold to the higher of the two size classes.
13. See Per Davidsson, "Methodological Concerns in the Estimation of Job Creation in Different Firm Size Classes," 1996, Working Paper, Jonköping International Business School, on the Internet at http://www.ihh.hj.se/eng/research/publications/wp/1996-1%20Davidsson.pdf (accessed June 2005).
14. Also see Per Davidsson, Leif Lindmark, and Christer Olofsson, "The Extent of Overestimation of Small Firm Job Creation - An Empirical Examination of the Regression Bias," Small Business Economics, November 1998, pp. 87-100.
15. It is possible to construct an example in which annual base-sizing would result in nonsymmetrical statistics. If employment in March 2003 is 3, and the December 2003 to March 2004 employment growth is from 3 to 13, followed by a March 2004 to June 2004 employment decline from 13 to 3, the annual base-sizing estimator would put the employment growth of 10 into the "1-4" size class, but would put the following employment decline of 10 into the "10-19" size class.
16. References to regression to the mean bias include Davis, Haltiwanger, and Schuh, Job Creation and Destruction, 1996;
17. Davidsson, "Methodological Concerns in the Estimation of Job Creation in Different Firm Size Classes," 1996;
18. Davidsson, Lindmark, and Olofsson, "The Extent of Overestimation of Small Firm Job Creation," 1998;
19. and Milton Friedman, "Do Old Fallacies Ever Die?" Journal of Economic Literature, vol. 30, issue 4, 1992, pp. 2129-32.
20. We also have removed a very small number of businesses involved in breakouts or consolidations at any point during the year. This is done because it is difficult to follow the employment patterns of businesses with changes in reporting configurations across five consecutive quarters. Also note that this is the only section of this article where we present new results using establishments rather than firms; we do not believe the results will change if firms rather than establishments were the unit of analysis.
21. The full sample of BED microdata for the first quarter 1999 to the first quarter 2000 period, with births, deaths, and businesses involved in breakouts or consolidations, has an average quarterly employment of 107,578,247 jobs. Thus, our sample has 93.5 percent of all employment.
22. The most frequent example of this is the 8,096 establishments that start in size class "1-4" in the first quarter 1999, are in size class "5-9" in the second quarter 1999, are back in size class "1-4" in the third quarter 1999, and then expand into size class "5-9" in the fourth quarter 1999, and first quarter 2000.
23. See Okolie, "Why size class methodology matters," 2004.
24. To be precise, annual base-sizing imposes symmetry only on changes within the year, but does not impose symmetry on transitory and reverting changes that happen before and after March.
25. The BED net and gross job flow statistics are unique in that they measure quarterly employment change, whereas almost all other similar statistics in the United States (and elsewhere in the world) measure employment change on an annual March-to-March frequency. For a further discussion of the difference between quarterly and annual net and gross job flow statistics, see Joshua C. Pinkston and James R. Spletzer, "Annual measures of gross job gains and gross job losses," Monthly Labor Review, November 2004, pp. 3-13.
26. Our review of the literature, and our e-mail conversations with Per Davidsson, have not found any references beyond the two cited earlier: Davidsson, "Methodological Concerns in the Estimation of Job Creation in Different Firm Size Classes," 1996,
27. and Davidsson, Lindmark, and Olofsson, "The Extent of Overestimation of Small Firm Job Creation," 1998.
28. For a thorough description of the BED program, see James R. Spletzer, R. Jason Faberman, Akbar Sadeghi, David M. Talan, and Richard L. Clayton, "Business employment dynamics: new data on gross job gains and losses," Monthly Labor Review, April 2004, pp. 29-42.
29. The first cross sectional tabulations appear in the article by Okolie, "Why size class matters," 2004.
30. One other point warrants mention. BLS is not planning on publishing firm counts for the dynamic-sizing statistics. Firm counts are trivial to calculate for tables of employment change by industry and by geography. Firm counts are also trivial to calculate for tables of employment change by size class, when size class is defined by the mean-sizing or either of the base-sizing methodologies. However, it is difficult to calculate firm counts when using the dynamic-sizing methodology. Recall the example of a firm expanding from 3 to 13 employees, in which size class "1-4" is credited with the growth of 1 employee, size class "5-9" is credited with the growth of 5 employees, and size class "10-19" is credited with the growth of 4 employees. In this example, into which size class would this expanding firm get placed? This is a question being researched.