TTG, Incorporated: Professional Territory Design Software Since 1988

BusinessCounts

Introduction

BusinessCounts is a geographic summary database of business establishments and employees for over twelve million companies and one hundred and ten million employees. The database is available at the block group level and higher, including all standard geographic aggregations.

BusinessCounts is a vital addition to residential demographic data, in that the success of many business establishments is dependent upon not only the residential population, but also the working population during the daytime. Based primarily on the InfoUSA business database and supplemented by various public data sources, BusinessCounts provides a clear look at the range and size of establishments and their employees within any geographic area.

As of the 2003 release, there were significant changes made to this data in order to eliminate variables which we do not feel are sufficiently well supported by source data, or have ceased to be sufficiently well supported, to merit inclusion in the database. These changes are outlined below:

Retail Sales Estimates
These have been removed from the database, for several reasons. First, these sales were necessarily based on the often weak relationship between employee count and sales by retail store type. These statistics are available only every five years, with the most recent being 1997 at this time. Second, these statistics are not well differentiated geographically and are subject to averaging error on the source side and estimation error on the geographic side. Finally, the problem is most noticeable when results are mapped for small areas, which is increasingly one of the intended uses of the data.

Payroll Estimates
In the past, the database has included estimates of payroll by detailed SIC category, again based on higher level geographic statistics. For many of the same reasons as above, these have been eliminated from the database.

Occupation
We have expanded the occupation table from thirteen to seventeen categories as a result of an improvement to the underlying industry-occupation tables released by the Bureau of Labor Statistics. The sales occupation (BCCCYSALES) has been split into two categories: Sales Professionals (BCCCYSALPR) and Sales Workers and Clerks (BCCCYSALCL). The clerical group (BCCCYCLER) has been split into Administrative Support Workers (BCCCYCLERO) and Technical, Sales, and Administrative Field Operations (BCCCYCLERF). The “Other” Services group has been broken into Site Based (BCCCYSERV1) and Field Based (BCCCYSERV2).Finally, the Production occupations group has been split into site and field based operations (BCCCYPRODS, BCCCYPRODF).

These changes will be very useful to retailers and food services companies who are most interested in the daytime population, as many of these field operations staff who may be assigned to a particular location are in all likelihood absent on most workdays from that location.

Business Functional Classification
This set of 38 business types, based roughly on land use style, is adapted from one which was used several companies during the early 1990’s and is, in our view, a very useful way of presenting business data to clients on the style of operation rather than the specific type of products produced. The variable ESCCYMFGSM, Manufacturing: Small Firms is a good indication of the presence of small industrial parks. A small manufacturer of plastics, for example, is likely to be different in its land use and employment profile than a large producer.

Administrative/Headquarters
We have added counts of employees and establishments for administrative or headquarters firms, as identified in the InfoUSA file.

SIC Classifications
The classification by two-digit SIC code has been slightly re-worked in order to merge certain small related categories while splitting other large categories. The result is a slightly shorter list of categories at this detail level. For example, the categories 01 (Crop production), 02 (Livestock production), 08 (Forestry), and 09 (Fishing) have been aggregated into a single category.

NAICS Classifications new with 2007 release

A few years ago, a new classification system known as NAICS (North American Industrial Classification System) was introduced to replace the aging and outdated SIC (Standard Industrial Classification) system, which has been used for several decades (with some modifications over time).

The NAICS classification is intended to more accurately reflect the growing reliance on information-based companies which were not well classified under the SIC system, reflect the changes in retailing towards large, multi-faceted retailing (e.g. home stores, grocery and drug stores, etc.), and to eliminate business types which simply no longer exist. In most cases, the orientation is more towards the product/service being offered and less towards the style or raw materials of its creation. For example, the textile manufacturing SIC group has been split by whether the product is an “end-user” product or is simply a raw material in further manufacturing. The manufacture of automobile seat fabrics is now classified under the automobile manufacturing group rather than the textile manufacturing group, reflecting the ultimate use of the product rather than its composition.

At the present time, the business coding systems of both public and private surveys and lists are in a state of transition between the two systems. Some Census Bureau products are fully converted to NAICS (such as the County Business Patterns and the 2002 Economic Censuses) while others (such as the Monthly Retail Trade reports) remain on the SIC system. InfoUSA has at this time fully coded their file with both SIC and NAICS codes.

For this release of BusinessCounts, we have presented data for the NAICS classification at the 3 and 4 digit detail levels. A 2 digit summary can be readily created using the formula features of Snap, should such be desirable. These are roughly comparable to the SIC “Major Industry” and 2 digit levels, although there are more categories under NAICS than under SIC.

Content

BusinessCounts is a geographic summary database of business establishments and employees for nearly ten million businesses and one hundred and thirty million employees. The database is available for all standard levels of geography including block group.

BusinessCounts is a geographic compilation of the InfoUSA business list, supplemented by occupational data from the Bureau of Labor Statistics and the County Business Patterns program. The primary variables available include:

Total Establishments, Employees,
Size Establishments by size
Occupation Employment by occupation
Major Industry Establishments, Employees
NAICS Establishments, Employees 3 and 4 digit

Methodology and Data Sources

The core source for the InfoUSA Business Database that is built from a careful integration of commercial databases, compiled white and yellow page directory data, city directories, corporate annual reports, and securities filings. The BusinessCounts file is current to January 2009.

In years past, a different data source was used by AGS to compile this database, and users should review the notes at the end of this document that outline the type and scope of the impacts of the change in source data. The primary changes that will be noted by users include:

• The ability to release establishment level data for use in mapping applications, with selection based on company name, SIC, geographic area, and company size
• A greatly expanded number of establishments, many of which are small and unclassified, but nevertheless reflect changes in the corporate landscape
• Improved SIC coding at establishments which include more than one major industrial group
• Reduced duplication of records – and subsequent over counting of employment – at companies which contain multiple legal entities at the same address

The database has been thoroughly cleansed for address consistency and geocoded. Virtually all records within the database are geocoded, although in some cases with less positional accuracy than others.

SIC Classification
A significant number of establishments are not SIC coded within business list files, most often including those small firms for which neither private nor public records exist. Many of these uncoded firms are simply individual holding companies, DBA (“doing business as”) names, and new firms that have not yet been well documented in multiple sources.

Employees
The file includes both a size classification (e.g. 1 to 4 employees) and in a significant number of records, verified employee counts. When an actual employee count was given, this was used directly.

In order to estimate employees for those establishments either only a size class range is available, the latest County Business Patterns (CBP) database, published annually by the Census Bureau was thoroughly analyzed. For each four digit SIC code, the average number of employees per establishment of each size class was computed in order to provide a base estimate. These were further refined by using major industry average sizes by county, since much of the county level detail is suppressed within the CBP in order to avoid the possible disclosure of individual establishment employee, payroll, and sales volumes.

Once the initial estimates were applied, the results were evaluated on a county level basis in order to ensure consistency with county totals for each major SIC group, and nationally to ensure consistency with the detailed four-digit SIC level.

It should be noted that the employee size estimates for the Public Administration (SIC 91-98) major group are not particularly accurate. Employee estimates for individual government offices are simply not easily obtained and are generally afforded less attention by the major business list providers than private sector establishments. Further, neither the CBP nor the Economic Census databases cover this important sector of the economy. The total employee estimate is therefore rather low for this sector as a whole.

Occupation
The occupation estimates were created using the 1996 Industry Staffing Patterns Estimate file, obtained from the Occupational Employment Statistics (OES) Survey, conducted by the Bureau of Labor Statistics. The occupational classification used by the OES was converted to the Census Occupational categorization using a translation file obtained from the NOICC Crosswalk and Data Center (NCDC), a technical resource center of the National Occupational Information Coordinating Committee (NOICC).

Comparability to Other Sources
Several additional sources of national and state level estimates from the BLS (Bureau of Labor Statistics) and the Census Bureau were used to verify summary counts in the final database. In general, the database agrees substantially with these estimates. The major sources of difference occur in several areas. These areas of disagreement are noted below:

• Manufacturing employment in BusinessCounts is higher than corresponding statistics from the BLS. In large part, this reflects the use of the “primary” industry within BusinessCounts. Often, many manufacturing companies also have wholesale trade and finance divisions. The employment within these divisions is attributed to the main SIC category in this file, hence manufacturing estimates are higher than in BLS sources.

• Agricultural establishments, specifically farms, tend to be underreported in the database, so total agricultural establishment and employment counts are low relative to other sources. The so-called “primary” sector is not typically well represented in either the economic censuses or the annual County Business Patterns files, and is subsequently difficult to estimate with reliability.

• Service employment in the BusinessCounts file is higher than in equivalent BLS sources, primarily as a result of classification issues. In official BLS reports, educational institutions and employment is reported within the appropriate level of government (e.g. state versus local) whereas in BusinessCounts, these are reported in the educational services category.
  
• In addition, many public and quasi-public agencies are coded to the type of service they provide rather than as public sector establishments. Public sector estimates in BusinessCounts are therefore lower than published figures by an amount roughly equivalent to the over-count in services. In addition, the tendency within business list products such as InfoUSA is to put more emphasis on private sector establishments than on public sector establishments. Subsequently, in many cases not only is there no actual employee count, but often no size class information as well. Since the Census Bureau surveys of establishments typically exclude public sector establishments, and what statistics are available are typically only at a state level, the public sector employment estimates are substantially underestimated and should not be relied upon for many analytical applications.