Business Employment Dynamics

Business Employment Dynamics Size Class Data:
Questions and Answers

GENERAL QUESTIONS

  1. What is the source of these data?
  2. What are the uses of these data?

  3. ISSUES IN METHODOLOGY: FIRM VERSUS ESTABLISHMENT

  4. Why was “firm” level chosen?
  5. What is the difference between establishment data and firm data?
  6. How are firms identified?
  7. How are employment changes at the firm calculated?
  8. Are net and gross job flow statistics by size class different if the establishment or the firm is used as the unit of analysis?
  9. Do the totals for size class data equal the reported totals from other Business Employment Dynamics data?
  10. How are large conglomerates with more than one Employer Identification Number (EIN) handled in these size class data?
  11. What percentage of U.S. firms have only one establishment? What percentage of U.S. firms operate more than one establishment?
  12. How many firms are in each size class? What percent of total US employment do they constitute?
  13. Is there information on the industry breakdown of these size classes?

    ISSUES IN METHODOLOGY: DYNAMIC SIZING

  14. What sizing methodology is used?
  15. What is dynamic sizing?
  16. If a gain of one employee pushes a firm into a new size class, does that job gain go under the size class before or after the employment change happened?
  17. How was the dynamic sizing method chosen?
  18. What other sizing methodologies were considered?
  19. What are the advantages of dynamic sizing?
  20. Why are the numbers of firms by direction of employment change not published for this series (as in other Business Employment Dynamics series)?

    ISSUES IN METHODOLOGY: SIZE CLASSIFICATIONS

  21. Why are there 9 size classes?
  22. Can data for the different size classes be combined?
  23. Can you add gross gains or losses within a size class across four quarters to get annual gain or loss?
  24. Can you add net change within a size class across four quarters to get annual net change?

    ISSUES IN METHODOLOGY: OTHER QUESTIONS

  25. How does BLS handle Professional Employer Organizations (PEOs) in these data?
  26. How are the rates for gross job gains and losses calculated?
  27. Why is it that shares of gross job gains and gross job losses do not sum to the share of net?
  28. What employers are excluded in this Business Employment Dynamics series?
  29. What is seasonal adjustment? How are these figures adjusted?

    QUESTIONS ON PUBLICATION

  30. Will BLS publish size class data using other methodologies (such as base sizing or mean sizing)?
  31. Are size class data available for individual states? Metropolitan Statistical Areas? Counties?

    OTHER DATA ON SIZE CLASS

  32. How do these size class data differ from other QCEW size class data?
  33. What are the unique features of the new series?
  34. Does BLS have any other data on business size?

GENERAL QUESTIONS

  1. What is the source of these data?
    The new Business Employment Dynamics size class data are derived from the Quarterly Census of Employment and Wages (QCEW), also known as the ES-202 program. This program is a quarterly census of all establishments covered under state and federal unemployment insurance programs, about 8.5 million business establishments, representing about 98 percent of employment on nonfarm payrolls.
  2. What are the uses of these data?
    These data can be used to better understand how businesses of different sizes contribute to changes in employment, how employment growth or loss patterns differ over shorter and longer time periods, and how employment patterns by size class change with the business cycle. The unique quarterly frequency allows a clearer view of economic change, and the timely release (only eight months after the reference quarter) make these data useful to examine relatively current trends.


  3. ISSUES IN METHODOLOGY: FIRM VERSUS ESTABLISHMENT

  4. Why was “firm” level chosen?
    While all of the other Business Employment Dynamics series use the establishment as the unit of analysis, these size class data are based on the firm level. The firm level is more consistent with the role of corporations as the economic decision makers than each individual establishment.

    For more information about the difference between firm level and establishment level data, see the article in the July 2004 issue of Monthly Labor Review, "Why size class methodology matters in analyses of net and gross job flows," (PDF) by Cordelia Okolie.
  5. What is the difference between establishment data and firm data?
    An establishment is defined as an economic unit that produces goods or services, usually at a single physical location, and engaged in one or predominantly one activity. A firm is a legal business, either corporate or otherwise, and may consist of one establishment, a few establishments, or even a very large number of establishments.
  6. How are firms identified?
    For these data, firms are identified by the unique Employer Identification Number (EIN) issued by the IRS. Multiple establishments that share the same EIN are counted as a single firm, and employment from each establishment is aggregated to determine employment for the firm as a whole.
  7. How are employment changes at the firm calculated?
    Linked establishments are aggregated to the firm level using the Employer Identification Number (EIN) issued by the IRS. Employment at a firm in the third month of the current quarter is compared to employment at that firm in the third month of the prior quarter to measure openings, expansions, closings, and contractions.
  8. Are net and gross job flow statistics by size class different if the establishment or the firm is used as the unit of analysis?
    Yes, gross job flow statistics and the contribution to the net employment growth for large multi-establishment employers depends upon whether the unit of analysis is the establishment or the firm. Similar data calculated using establishments as the unit of analysis would shift employment into the smaller size classes.
  9. Do the totals for size class data equal the reported totals from other Business Employment Dynamics data?
    Net changes in total private employment are the same for not seasonally adjusted data and are similar for seasonally adjusted data. However, these size class data report changes at firm level, whereas data from other Business Employment Dynamics releases report changes at the establishment level.
  10. How are large conglomerates with more than one Employer Identification Number (EIN) handled in these size class data?
    Firms are identified by IRS-issued Employer Identification Numbers (EINs); therefore a large conglomerate with multiple EINs is counted as a number of individual firms. Gross job flows and size class designation are estimated independently for each subsidiary firm.
  11. What percentage of U.S. firms have only one establishment? What percentage of U.S. firms operate more than one establishment?
    In March 2005, 95.2 percent of U.S. firms (measured using IRS-issued Employer Identification Numbers) had a single establishment. These businesses covered 44.6 percent of employment.

    Distribution of firms by number of establishments, first quarter 2005, not seasonally adjusted, in thousands
    Number of firms Firms as a
    percent of total
    Employment Employment as a
    percent of total

    1 establishment

    4,706 95.2 48,164 44.6

    2 establishments

    112 2.3 5,823 5.4

    3 establishments

    36 0.7 3,324 3.1

    4 establishments

    20 0.4 2,504 2.3

    5 - 9 establishments

    38 0.8 7,650 7.1

    10 or more establishments

    30 0.6 40,438 37.5

    TOTAL

    4,942 100.0 107,902 100.0


  12. How many firms are in each size class? What percent of total US employment do they constitute?
    In March of 2005, 54.4 percent of all firms were in size class 1 - 4 employees, which constituted 5.2 percent of employment. Firms with 1,000 or more employees represented only 0.2 percent of total firms; however they constituted 37.4 percent of employment.

    Number of firms and employment by size class, first quarter 2005, not seasonally adjusted
    Firm size class (number of employees) Number of firms Employment
    Number of firms Cumulative total Share oftotal Cumulative share Number of employees Cumulative total Share oftotal Cumulative share
    In thousands Percent In thousands Percent

    1 - 4

    2,687 2,687 54.4 54.4 5,606 5,606 5.2 5.2

    5 - 9

    1,006 3,693 20.3 74.7 6,613 12,219 6.1 11.3

    10 - 19

    610 4,303 12.4 87.1 8,204 20,423 7.6 18.9

    20 - 49

    392 4,695 7.9 95.0 11,801 32,223 10.9 29.9

    50 - 99

    129 4,824 2.6 97.6 8,873 41,096 8.2 38.1

    100 - 249

    75 4,899 1.5 99.1 11,310 52,406 10.5 48.6

    250 - 499

    23 4,922 0.5 99.6 7,813 60,219 7.2 55.8

    500 - 999

    11 4,932 0.2 99.8 7,334 67,553 6.8 62.6

    1,000 or more

    10 4,942 0.2 100.0 40,349 107,902 37.4 100.0


    The QCEW also publishes establishment (not firm) counts for each size class each year. These counts are available in the annual Employment and Wages publication or they can be found online at the following link (Table 3: Private industry by supersector and size of establishment (PDF): Establishments and employment, first quarter 2003)
  13. Is there information on the industry breakdown of these size classes?
    Dynamic size class data with respect to industry sectors are not currently available.


  14. ISSUES IN METHODOLOGY: DYNAMIC SIZING

  15. What sizing methodology is used?
    The method of dynamic sizing is used for these calculations.
  16. What is dynamic sizing?
    Dynamic sizing allocates each firm’s employment growth or loss during a quarter to each respective size class in which the change occurred. For example, if a firm grew from 2 employees to 38 employees, then of the 36-employee increase, 2 would be allocated to the first size class 1 - 4, 5 to the size class 5 - 9, 10 to size class 10 - 19, and 19 to size-class 20 - 49. Dynamic sizing provides symmetrical firm size estimates and eliminates any systemic effects, which may be caused by the transitory and reverting changes in firms’ sizes over time. Additionally, it has a clear conceptual foundation, allocating each job gain or loss to the actual size class where it occurred.
  17. If a gain of one employee pushes a firm into a new size class, does that job gain go under the size class before or after the employment change happened?
    When a firm moves from one size class to another either by expanding or contracting, dynamic sizing credits the single job that pushed the firm across the threshold to the higher of the two size classes. For example, if a firm has 4 employees and gains one employee, that gain is credited to the size class for firms with 5 - 9 employees. Conversely, if a firm has 5 employees and loses one employee, that loss is credited to the 5 - 9 size class.
  18. How was the dynamic sizing method chosen?
    BLS spent considerable time analyzing the existing data and methods, including research and other published papers, and reviewing internationally used methods. BLS developed size class data based on other methods such as quarterly base sizing, mean sizing, end sizing and annual base sizing. A review of the strengths and weaknesses of each method was conducted, and dynamic sizing emerged as the preferred methodology for tabulation of Business Employment Dynamics data by firm size. A BLS working paper, “Business Employment Dynamics: Tabulations by Employer Size,” by Shail J. Butani and others, will be published in an upcoming issue of the Monthly Labor Review that examines different methods and details how dynamic sizing was selected as the preferred methodology.
  19. What other sizing methodologies were considered?
    Base sizing uses a firm’s size in a previous period to determine its size class. There are two types of base sizing: quarterly base sizing and annual base sizing. Quarterly base sizing places each expansion or contraction into a size class based on the firm’s size in the previous quarter. Annual base sizing uses the firm’s size in the first quarter of the previous year to place each expansion or contraction into a size category.

    End sizing uses the firm’s employment size at the end of the period.

    Mean sizing uses an average of the beginning and ending employment.
  20. What are the advantages of dynamic sizing?
    Dynamic sizing has consistency and evenly distributes large changes across size classes. Moreover, dynamic sizing makes sense from an empirical point of view. While a firm may show an increase of 60 employees, in reality this increase occurred over the course of a quarter. Indeed, it is likely that such a firm progressed through each size class during the three months of its expansion. Again, using the above example, it is probable the firm hired a few employees each week or couple of weeks; thus it is conceivable the firm had, at one point 4, then 15, then 30, then 40, and finally 64 employees.
  21. Why are the numbers of firms by direction of employment change not published for this series (as in other Business Employment Dynamics series)?
    In dynamic sizing, an expanding or contracting firm could be counted in more than one size class. Measuring the number of firms by direction of employment would require developing a different methodology for counting firms (because only a fraction of the employment went to each size class).


  22. ISSUES IN METHODOLOGY: SIZE CLASSIFICATIONS

  23. Why are there 9 size classes?
    The Office of Management and Budget (OMB) established official size class standards for use by Federal agencies, as they do for industrial and occupational classifications.

    Data users can combine different size classes to form broader categories; since there is no single definition of what constitutes a small or large business, data users can create the categories they are interested in studying.
  24. Can data for the different size classes be combined?
    Yes, within a single quarter, different size classes together can be added to create a new, broader classification. For example, the four smallest size classes can be summed to get data for businesses with 1 - 49 employees. However, gross gains and gross losses from consecutive quarters can not be added, since this would result in double-counting.
  25. Can you add gross gains or losses within a size class across four quarters to get annual gain or loss?
    No. Annual Business Employment Dynamics statistics by size are not available at this time. Adding quarters to get annual results will lead to inaccuracies since some firms will be counted more than once. For example, a firm that expands in quarters one and two, and contracts in quarters three and four will be counted in both expansions and contractions for those respective quarters; however if that firm ended the year with more employees than it started, the difference should only be counted as an expansion. A BLS working paper, “Business Employment Dynamics: Tabulations by Employer Size,” by Shail J. Butani and others, examines these and other issues in sizing methodology. It will be published in an upcoming issue of the Monthly Labor Review.
  26. Can you add net change within a size class across four quarters to get annual net change?
    Yes. Using not seasonally adjusted data, the net change within a size class across four quarters sums to annual net change (not seasonally adjusted) for that size class.


  27. ISSUES IN METHODOLOGY: OTHER QUESTIONS

  28. How does BLS handle Professional Employer Organizations (PEOs) in these data?
    Professional Employer Organizations (PEOs) are treated like other employers under administrative and support services.
  29. How are the rates for gross job gains and losses calculated?
    Gross job gains and gross job losses are expressed as rates by dividing their levels by the average of total private employment in the current and previous quarters. This provides a symmetric growth rate. The rates are calculated for the components of gross job gains and gross job losses and then summed to form their respective totals. These rates can be added and subtracted just as their levels can. For instance, the difference between the gross job gains rate and the gross job loss rate is the net growth rate.
  30. Why is it that shares of gross job gains and gross job losses do not sum to the share of net?
    Shares are calculated for each size class as a percent of the total of gross job gains, openings, expansions, gross job losses, closings, and contractions, respectively. The shares of different size classes within the same category can be added. For example, firms with 500 or more employees account for 23.3 percent of gross job gains on average from third quarter 1992 through first quarter of 2005 (sum of 4.9 percent and 18.4 percent). Because of this, shares of gross job gains and gross job losses within a single size class can not be added, nor do they sum to the share for net change.
  31. What employers are excluded in this Business Employment Dynamics series?
    The dynamic size class data are created from the Quarterly Census of Employment and Wages. Gross job gains and gross job losses are derived from longitudinal histories of private sector employer reports. In the fourth quarter 2004, this included over 6.6 million private sector employers.

    Major exclusions from UI coverage are self-employed workers, religious organizations, most agricultural workers on small farms, all members of the Armed Forces, elected officials in most States, most employees of railroads, some domestic workers, most student workers at schools, and employees of certain nonprofit organizations.

    Gross job gains and gross job losses data do not include government employees, private households, and establishments with zero employment. Data from Puerto Rico and Virgin Islands also are excluded from the national data.
  32. What is seasonal adjustment? How are these figures adjusted?
    Over the course of a year, the levels of employment and the associated job flows undergo sharp fluctuations due to such seasonal events as changes in the weather, reduced or expanded production, harvests, major holidays, and the opening and closing of schools. The effect of such seasonal variation can be very large.

    Because these seasonal events follow a more or less regular pattern each year, adjusting these statistics from quarter to quarter can eliminate their influence. These adjustments make non seasonal developments, such as declines in economic activity, easier to spot. The adjusted figure provides a more useful tool with which to analyze changes in economic activity.

    The employment data series for opening, expanding, closing, and contracting establishments are independently adjusted using X-12 ARIMA software and the net changes are calculated based on the difference between gross job gains and gross job losses.


  33. QUESTIONS ON PUBLICATION

  34. Will BLS publish size class data using other methodologies (such as base sizing or mean sizing)?
    The BLS has plans to release data using quarterly base sizing and mean sizing methods. These alternative sets of measures will be released periodically.
  35. Are size class data available for individual states? Metropolitan Statistical Areas? Counties?
    Dynamic size class data with respect to state, metropolitan statistical areas (MSAs), and counties are not currently available.


  36. OTHER DATA ON SIZE CLASS

  37. How do these size class data differ from other QCEW size class data?
    Both data sets are derived from the same source, the QCEW program. The existing profiles are annual snapshots of the status of establishments. These profiles do not change much from year to year.

    These new data show the quarterly employment dynamics at the firm level, whereas other Business Employment Dynamics data show employment dynamics at the establishment level.
  38. What are the unique features of the new series?
    These new Business Employment Dynamics size class data are designed to exploit programmatic advances that allow these data to be issued frequently and on a current basis. These data will be released quarterly, and only 8 months after the reference quarter, which will make them more useful for assessing current economic trends.


  39. Does BLS have any other data on business size?
    Yes, the BLS QCEW program has published profiles of the number of establishments and employment and wages by size of the establishments for many years. These data are available on the QCEW website.




Last Modified Date: December 15, 2005

Recommend this page using: