Linking firms with establishments in BLS microdata
The percentages of establishments and employment that can be identified with the single EIN used in firms’ Form 10-K filings are least for the firms listed in the Dow Jones Industrial Average (the largest and most complex firms) and greatest for the firms listed in the Russell 2000 Index (the smallest examined in this article). This situation suggests that the larger a publicly held firm, the smaller are the percentages of its establishments and employment that can be identified with the single EIN listed in the firm’s Form 10-K filings. Still, even for the smallest of the publicly held firms on the list (those listed in the Russell 2000 Index), less than half of all establishments and a very small percentage of employment can be linked directly to the QCEW by using only the EINs listed in the firms’ Form 10-K filings. Those EINs frequently can be linked with establishments of firms classified into NAICS code 551, “Management of Companies.” The following tabulation gives the actual number of EINs used by these firms, as well as the number of states in which the firms operate:
|Category||Number of firms examined||Mean number of EINs per firm (standard deviation)||Mean number of states with nonzero employment per firm (standard deviation)||Mean number of states per EIN (standard deviation)|
|Dow Jones Industrial Average||14||29.5 (39.8)||48.9 (4.5)||13.4 (15.8|
|S&P 500 (excluding Dow Jones)||14||321.5 (869.7)||36.5 (18.8)||2.4 (6.7)|
|Russell 2000||15||5.0 (4.4)||34.2 (18.5)||16.2 (18.3)|
Finding all the establishments (in practical terms, finding all the EINs associated with the establishments) for a firm appearing in BLS data is important to researchers who want to link firm-level data with BLS establishment-level microdata. This section presents four case studies of the efforts involved in such linking. Examined are one firm listed in the Dow Jones Industrial Average, one firm listed in the S&P 500 Index, one firm listed in the Russell 2000 Index, and one firm that is privately held (and thus would not need to file Form 10-K). These firms were not chosen completely at random; rather, they were selected because information on their total employment was available (in some cases, from the firm’s website or Form 10-K filing). For each firm, a certain percentage of establishments and the percentage of employment can be found by searching for the firm name in the QCEW for the fourth quarter of 2009. Greater percentages of establishments and employment can be found through more rigorous matching efforts that use the names of all subsidiaries and all addresses of establishments of the firms listed in Form 10-K reports for 2009 and on firms’ websites. However, more time is required to find these additional names and addresses.
Searching for establishments by firm name has advantages and disadvantages for researchers. The QCEW contains both legal and trade names for each establishment, and these names can be used in computer searches. However, many of the names listed in the QCEW are older names of company plants or subsidiaries. Moreover, few names are unique, so, in addition to matching the establishments found with the firm in question, computer searches for establishments by firm name may incorrectly match the name with the establishments of hundreds or thousands of other firms. Another drawback is that the QCEW includes only the most recent version of names and addresses of establishments, complicating name searches for establishments that operated during earlier periods.