Page images



One of the ways in which the Office of Federal Statistical Policy and Standards coordinates Federal statistics is through standard classifications and definitions. The Federal Reports Act requires that "Information . . . shall . . . be tabulated in a manner to maximize the usefulness of the information to other Federal agencies and the public." That requirement is administered through Directives which specify that standards and standard classifications must be used wherever they are appropriate.

Standards are necessary wherever more than one organization defines or classifies objects or concepts. To be more precise, standardization is necessary when objects or concepts are classified by one or more characteristics that are also used in definitions by other organizations.

When a classification is based on different characteristics, it may still be very useful to use or modify existing standards to provide a relationship to other classifications of that object. An example of such a relationship is that between the hazardousness of materials to be shipped and their industrial origin characteristics. This use of related systems provides some comparability with data collected for other purposes.

A classification standard requires a methodical, broadly-based, scheduled updating. Because of the broader application of a standard, a maintenance agency is designated and more resources may be made available for their development and revision.

The Standard Industrial Classification (SIC), the Enterprise Standard Industrial Classification (ESIC), the Standard Metropolitan Statistical Area (SMSA), and the Standard Occupational Classification (SOC) are now included in the Directives issued by the Office of Federal Statistical Policy and Standards. These Directives provide the authority to prescribe the conditions where the classifications are to be used. Commodity classifications are not specifically stated, but will be when the current work on harmonization of commodity classifications is

complete. Classifications for commodities have been developed for several different purposes, making standardization difficult.

All of these classifications should be made Federal Information Processing Standards (FIPS). The SMSA is already a FIPS, the SIC is in process, and work is being done on the SOC. All other standards should be considered for inclusion.


One set of classifications (the ESIC, the SIC and commodity classifications) should be described as a family. First, the Enterprise Standard Industrial Classification is a classification of enterprises, i.e., the legal entities such as companies or firms. It was developed to collect and publish statistics such as profits, debts, interest, capitalization items and taxes which are usually not available separately for establishments of multi-industry companies.

Second, the Standard Industrial Classification which is a classification of establishments, that is, plants, stores, mines, farms, banks, offices, schools, etc. It was developed to collect statistics on numbers of employees, payrolls, inventories, capital expenditures, gross sales or receipts and consumption of energy and materials used which are usually available on the records of the establishments of various companies.

Third, commodity classifications were developed for the collection of data on the value and quantity of goods produced, stored, transported, imported, exported, consumed, sold or purchased. Data collected on the value of sales or production of specific commodities can be aggregated to the total value of all commodities for an establishment. Data for establishments can be aggregated to totals for enterprises or multi-unit companies.

Occupations of workers are frequently crosstabulated with the industry in which they work, but the two classifications have very little in common. Some occupations, i.e., elementary school teachers and farmers, are closely related to industries, but others such as secretaries, accountants, and payroll

clerks have no direct industry relationship and are typically employed in almost every industry.

Standard Industrial Classification

The Standard Industrial Classification has been a tool for providing comparability of industry-related statistics for 40 years. During this time, four revisions and three supplements have been issued to keep the classification current with the changing industrial composition of the economy. The latest supplement was issued for 1977 and a major revision is scheduled for 1982.

The users of the SIC and the ways in which the classification has been employed have changed significantly over the years. Although initially developed for government agencies, private sector firms and organizations have widely adopted the system as a means of comparing data they collect with other industry statistics. It is used to identify sectors of the marketplace for advertising, sales leads, and for sources of supply for materials used in industry.

While revisions distort comparability over time, the failure to update the system would soon make it meaningless. For instance, if the original classification had not been revised since first published in 1941, it would include 14 industries for inactive financial institutions and 2 industries solely involving the wholesaling of raw silk, plus two others to cover the rest of the raw silk wholesaling (though not exclusively raw silk).

The population census planners would like to have the major revisions done in years ending in seven rather than two to provide a more up-to-date industrial classification for the decennial census of population.

Review Procedures

The procedures and principles that the 1972 revision was based on should be revised by the Technical Committee on Industrial Classification and should be widely circulated for comment. Possible changes include:

a. Reducing or eliminating the differences in the industry size criteria between divisions;

b. Requiring that a single combined "score" be met for proposals for new industries, in place of the minimum for each of the three criteria (size, specialization, and coverage). The change to a single score would mean that some industry proposals would be accepted that would not have been accepted under the


existing system. For instance, an industry may be very large and highly specialized but slightly below the coverage minimum. To offset this, proposals that just barely meet the threshold on each of the criteria would probably not be accepted. This score should be set so that the number of new industries does not increase (or decrease) substantially, since large increases in the number of industries may be counterproductive. Agencies may not be able to tabulate and publish all industries and might, therefore, publish only at the three-digit level. Consideration of interim industries to provide compatability over time, that is, split an industry into two or more parts in one revision, then combine it in a different configuration in the following revision.

A wide array of interested parties should have an opportunity to comment on these changes to the principles and procedures. Some of the criticism of the 1972 SIC revision took issue with the procedures. A 1950's study by the Business Advisory Council on Federal Reports proposed a much more strict criteria. than either the current or proposed criteria. They would probably support less strict criteria now, closer to the existing procedures.

A new statement on procedures and principles should be developed by a subcommittee of the Technical Committee on Industrial Classification and be announced in the Federal Register as well as the Statistical Reporter for public comment.

Substantial pressures on agencies and this office to change or modify the Standard Industrial Classification Manual have been based on nonstatistical programmatic uses of the Manual. Secretaries or administrators of agencies are now required by Directive not to use the SIC unless they determine that the SIC is appropriate for their program purposes. However, there is seldom sufficient time available to develop a more relevant classification for specific new program uses. Consideration should be given to providing for a programmatic industrial classification. One of the features of the programmatic or nonstatistical industrial classification would be the identification of small but distinct industries which are now combined in "not elsewhere classified" or "residual" groups. Another possibility might be to use multiclassifications, thereby allowing possiblities such as manufacturing concerns with a substantial transportation element as part of their industries to be identified so as to make reference to both manufacturing and transportation. Service industries highly related to manufacturing or some sector might

be coded in two ways. At the present time, service industries serving a very narrow group of industrial clients, especially when the same activity is also typically done by the producing industries, are classified in the same major or industry group as the industries they serve. Service industries which provide services to a wide range of manufacturing or other users are classified in the Services Division. An alternative supplementary classification could put all these service industries together.

Segments of Business Classification

For many types of analysis it is necessary to have more detailed information from the firm indicating its activities in relation to specific segments of its business, especially in relation to major product lines. A number of agencies have expressed needs for such a breakdown, and recently both the Federal Trade Commission and the Energy Information Administration have developed surveys to obtain this information. The Securities and Exchange Commission, the Bureau of Economic Analysis, the Census Bureau, and the Internal Revenue Service, among others, have expressed needs for similar information, but no agreement has yet been reached concerning the level of detail required, the definitions to be used, or the procedures to be used for allocating overhead to the various business segments.

In recognition of its finding that a number of agencies have a need for similar information, the Commission on Federal Paperwork recommended that a coordinated effort be undertaken to define exact needs and to develop a classification system that will eliminate potential duplication and/or similar but noncomparable data collection.

The Commission recommended that the Office of Management and Budget (OMB) work with the agencies to develop a more useful and reportable classification system. Since this effort is primarily related to the development of statistical standards, the Office of Management and Budget assigned the responsibility for it to the Office of Federal Statistical Policy and Standards. In early 1978, several meetings were held to define a project which would meet the needs of various agencies. In fiscal 1979, the Office hopes to initiate a project to develop a classification system for segments of business.

Enterprise Standard Industrial Classification

The Enterprise Standard Industrial Classification is closely tied to the SIC. It should be revised simultaneously with the SIC since many of the concerns about the ESIC are similar to the SIC. One

of the major problems the ESIC must deal with is the classification of conglomerates. This problem should be continually reevaluated to determine if the present system is no longer the best alternative. One solution to the difficulty of analyzing enterprise data under the present system would be to tabulate all firms including statistics for companies in two parts. One part would be an aggregation of those companies which are specialized, say at least 45 or 55%. These statistics could be used for analytical ratios about the industry. A second part would include all firms including the conglomerates and very diversified firms. The total aggregation of all firms included in the tabulations on a "best fit" basis could be looked at, taking into account the lack of "pure" figures especially at the more detailed levels. Additional information about the kind and scope of secondary activities should be made available.

Legal-Form-of-Ownership Classification

In the 1972 revision of the Standard Industrial Classification Manual the ownership characteristics were removed from much of the classification and a classification system for a legal-form-of-ownership was developed for such characteristics. A skeletal version of this classification is shown on page 11 of the SIC Manual. It was recommended by the Technical Committee on Industrial Classification that the legal basis of the ownership of an establishment was not an industrial characteristic and should not be part of the classification. The SIC classification is now much more concentrated on the economic activity of the establishments.

The legal-form-of-ownership classification should be further developed and a manual promulgated for that structure with definitions and examples. This classification is used in the census of population, using the description "class of worker." In the population census, workers are being classified and therefore, these characteristics are associated with the worker. Its primary application, though, is to classify a company, firm, or business that the employee is working for, and therefore, should not be called "class of worker" but rather "class of employer."

This skeletal classification was developed by the Technical Committee on Industrial Classification. That committee is the appropriate resource to develop this classification further.

The legal-form-of-ownership classification should also be used in conjunction with the ESIC system, although it may be possible for one company to have both nonprofit and for-profit establishments. For example, one establishment of a for-profit corporation could be a nonprofit, or vice versa, such as a

nonprofit church operating a profit-making printing business. In these cases, the major or primary form of organization should be the deciding factor.

Standard Occupational Classification

Initiation of the Standard Occupational Classification Project

The project to develop a Standard Occupational Classification was started in 1966 on the recommendation of the Interagency Committee on Occupational Classification. The recommendation was based on the results of an inquiry circulated among government agencies concerning the desirability of establishing a standard for use in the classification of all occupational data. Most of the 28 agencies that were queried, favored such a system. As might be expected, operating agencies which needed data for their own internal and specific uses expressed more reservations than those agencies which needed data primarily for analytical purposes.

Recent legislation has increased the demands on agencies to supply labor force, employment, income and other data to appraise the occupational structure of the work force and the requirements of the


A standard classification, establishing a common terminology and a common set of definitions, has become an urgent need in order to maximize the analytical potential of statistics collected under a variety of programs and for a variety of purposes. For example, income information collected in one survey could be related to disability rates from another survey, if the same occupations were used in both.

A first edition of the Standard Occupational Classification was developed under the direction of the Occupational Classification Committee, an interagency body established and led by OMB. Technical work groups of government personnel advised, guided, and assisted the OMB and agency detailed staff in preparation of the classification.

In developing the system, a fundamental concern was to provide a common base in such a way that continuity of statistical data would be maintained. wherever possible. In this connection, considerable use was made of the Dictionary of Occupational Titles (DOT) and the Census occupational classification system as well as of the views of various specialized agencies.

During the development of this manual, government agencies, professional organizations,

labor unions, business associations and other interested parties were provided an opportunity to comment on the structure, group definitions, content of groups, and applicability to the classification of occupational statistics. These groups should continue to participate in the revising of the Manual. The international and the British classifications were references and particularly extensive use was made of the Canadian classification.

The SOC Manual has not been tested in an actual program and therefore, review of the Manual should be scheduled when data to evaluate its feasibility and usefulness are available. This should be done as early as possible, since some occupational data are likely to be collected in the mid-decade census of population.

Merging the SOC and the Dictionary of Occupational Titles into a single unified system should be seriously considered. Each has substantially different uses, but they are not incompatible in terms of a combined relationship.

Purpose and Scope

The Standard Occupational Classification was developed to provide a mechanism for crossreferencing and aggregating occupation-related data collected for social, economic, and statistical reporting programs. Such a standard classification system is required in order to be able to compare labor force, employment, income and other occupational data from different sources within a uniform system. The classification is intended to cover all occupations in which work is performed for pay or profit, including work performed in familyoperated enterprises where direct remuneration may not be made to family members.

The classification establishes groupings of occupations into distinct categories to facilitate longrange analyses, reflects the structure of the world of work as realistically as possible, and provides the mechanism by which data from disparate sources can be compared.

The SOC furnishes a coding system and nomenclature for identifying and classifying occupations within a broad framework suitable for both government and private analysis. However, because of the vast amount of occupational detail that must be considered in the utilization of such a system, and the wide variety of uses of occupational data that are collected, the system is not expected to meet all the specific needs of all organizations. It is expected, however, that interested agencies will use this classification system for manpower research and analysis, collection and analysis of census data,

« PreviousContinue »