THE NONEMPLOYER UNIVERSE
The universe of nonemployer firms is created annually in conjunction with identifying the Census Bureau`s employer business universe. If the Census Bureau receives information through administrative records that a business has no paid employees, then the business becomes part of the potential nonemployer universe. Name, address, industry classification, and receipts are available for each potential nonemployer firm. These data are obtained chiefly from the annual business income tax returns filed with the Internal Revenue Service (IRS) and maintained in the Census Bureau’s Business Register.
The potential nonemployer universe is edited and reviewed to detect and remove firms that are not true nonemployers. Among the largest group of removed firms are those which are associated with large, multiunit companies, but which were not properly linked to the company. Another group consists of income records for regulated investment companies, representing mutual fund income. The Census Bureau does not include these as nonemployer firms because their income is generally not viewed as being nonemployer income. The Census Bureau also excludes firms with receipts above a predetermined cutoff, assuming that they are really employers and that their data are included in the employer database. For corporations and partnerships this cutoff is $1 million, except for service-type industries, where the cutoff is $2 million. For sole proprietorships, the cutoffs vary widely depending upon industrial classification. For some nonemployer sole proprietorships, such as those engaged in investment and entertainment, it may be possible to have well over $1million in receipts. However, it is unlikely that a sole proprietorship restaurant, for example, would have over $1million in receipts and have no paid employees.
Potential nonemployer businesses with less than $1,000 in receipts are excluded in all sectors except construction. The small receipts of these non-construction firms indicate that they may represent hobbies as opposed to normal business activities.
A certain amount of undercoverage in the universe occurs as a result of processing deadlines. If the Census Bureau receives information on a potential nonemployer business more than one year after the end of the reference year, it is excluded in nonemployer tabulations. The estimated undercount of nonemployer corporation and partnership firms because of late reporting is around 5 percent. This accounts for about 1 to 2 percent of receipts. The estimated undercount for sole proprietorship firms because of late reporting is much less.
INDUSTRY CLASSIFICATION OF FIRMSIndustry classification of businesses obtained from the IRS are self-classified by tax filers. Codes used for classification are NAICS-based industries. The NAICS system is comprised of nearly 1,200 industries, but Nonemployer Statistics are limited to approximately 300 codes that are available through administrative-record sources, and are common to all three legal forms of organization applicable to nonemployer businesses. For the 1997 – 2001 data series, each firm was classified in accordance with the industry definitions in the 1997 North American Industry Classification System (NAICS) manual. Beginning with the 2002 data year, nonemployer firms are classified according to the 2002 NAICS manual, which contains several major revisions from 1997. For more information on the 2002 NAICS codes, as well as comparisons between the 1997 and 2002 codes, go to http://www.census.gov/epcd/www/naics.html.
There is a small percentage of nonemployers that are unclassified. The percentage is at its lowest in Economic Census years. For these cases, an industry classification is assigned, or imputed, from a classified firm within the same county with a similar value of receipts and the same legal form of organization. This imputation procedure has the effect of preserving the distribution of firms and receipts by industry, while eliminating an unclassified component which varies in size each year. To control the contribution of imputed classifications, we have adopted the rule that if firms with an imputed classification account for more than 40 percent of either receipts or number of firms in a published row, both items are suppressed and assigned a flag of `S`. At the national level, less than 12 percent of nonemployer firms have an imputed classification, accounting for 3 to 5 percent of receipts.
PRECAUTIONS IN ANALYZING AND INTERPRETING NONEMPLOYER DATA
There are several considerations that must be recognized when analyzing or interpreting nonemployer data. The first is that the Census Bureau equates a business mailing address with a business firm. Under the usual definition, a firm or business is a fixed physical location or permanent structure where some form of business activity is conducted. Stores, offices, and manufacturing plants are examples of firms, but mobile vending vehicles or temporary stands are not. However, for Nonemployer Statistics, a business that sells from a mobile vehicle, temporary stand, or even one that has a separate store for conducting business, will be considered as a firm and assigned geographically to an owner`s mailing address.
A second consideration is that many businesses use leased or contract employees. In cases where all employees are leased or contracted, the payroll for the business is zero, placing it in the potential nonemployer universe. If the firm`s receipts are large, it may be dropped from nonemployer tabulations because its receipts exceed the cutoffs. Otherwise, it appears in the tabulations. Currently, the Census Bureau does not have a method to identify the universe of firms that lease all their employees.
RELIABILITY OF DATA
The nonemployer data are tabulated from administrative records data only, and are not subject to sampling error. No survey data are collected. These data undergo complex processing, editing, and analytical review at the Census Bureau to distinguish nonemployers from employers, correct and complete data items, and form the final nonemployer universe. Despite extensive review, the data compiled for Nonemployer Statistics are subject to nonsampling errors, which can be attributed to many sources: inability to identify all cases that should be in the universe; definition and classification difficulties; errors in recording or coding the data obtained; and other errors of coverage, processing, and estimation for missing or misreported data.
The accuracy of these tabulated data is determined by the joint effects of the various nonsampling errors. No direct measurement of these effects has been obtained except for estimation for missing or misreported industry classifications; however, precautionary steps were taken in all phases of the processing to minimize the effects of nonsampling errors.
The improper inclusion of possible employer establishments in the nonemployer universe is a primary source of nonsampling error. The Census Bureau takes several steps to identify and remove these establishments from the nonemployer universe. The most significant cases were identified, but the size of the universe, limited resources, and limited data on which to make a determination do not achieve the complete identification and removal of all employer establishments.
The assignment or imputation of an industry classification to nonemployer firms
with missing classification is another source of nonsampling error. The effect
of this imputation is to preserve the overall distribution of totals by industry
at the county and higher geographic levels, but certain data cells may be adversely
effected in counties or industries with a small number of firms. Data
cells with a high level of imputation are identified in the tabulations as not
meeting publication standards.
Source: U.S. Census Bureau, EPCD, Nonemployer Statistics.