« PreviousContinue »
planning of occupational education and training programs and counseling and placement services, analyzing mobility of workers, and other programs. In areas where the classification is too detailed, higher level groups should be used. Where the system is not detailed enough, additional detail within the existing framework should be collected. The system also allows for rearrangements of occupations that would be useful, such as combining data for managers with data for supervisors or combining data for all helpers. Also, college teachers were subdivided by the field in which they taught, so that data for many teachers could be combined with data for practitioners in the same field.
Structure of the Classification
The SOC is structured as a four-level system with the following level designations: division, major group, minor group and unit group, with each level representing successively finer detail. This structure enables users to tabulate or analyze data on several different levels, depending on the information available or the degree of detail necessary.
Basis for Code Assignment
Occupations should be classified on the basis of the most important (or primary) activity of that occupation. The most important activity will usually be that part of the work activities that accounts for the major portion of the worker's time; but where one activity requires special skills that are crucial in carrying on the occupation (although not required for as much time as other activities), then that activity will determine the classification of the occupation. Where two activities are frequently combined in one occupation, a separate occupation may be set up for that combination.
The classification system may not list a particular new or unique occupation. An occupation that is not described or listed in the classification should be classified in the group to which it is most similar.
Presentation of Occupational Data
Publication formats of occupational data should follow the SOC structure for publication, but recombinations are encouraged for supplemental analysis.
A program to coordinate business and government personnel management systems with the SOC-DOT system should be initiated. The Civil Service system is one of the classifications which should be considered early, followed by the Defense Department's system.
Several events have had a substantial impact on commodity classification. First, the Department of Transportation's Office of Facilitation has been developing a through-bill-of-lading which requires a commodity description code detailed enough to provide commodity identification relatable to all transportation rate and hazardness classifications (e.g., rail, air, ship, and truck); and to all tariff classifications (e.g., Tariff Schedule of the United States (TSUS)); the Customs Cooperation Council Nomenclature (CCCN) (formerly the Brussels Nomenclature (BTN)); and the Standard International Trade Classification (SITC).
Second, the SITC has been revised (Revision II) with substantially more detail than the previous version. The SITC in some cases will combine several CCCN classes. The one-to-one relationship will be defined by use of a sixth character in the SITC. Any data classified by the CCCN to the subchapter will be convertible to the SITC. However, data "published" on the SITC basis will not be directly convertible to the CCCN detail.
Third, the Trade Act of 1974 requires compatibility of import, export, and production classifications.
Fourth, concerns regarding data on energy have provided an impetus to standardization of fuel and related definitions.
Classification versus Description List
One aspect of classification which has not been very well described or defined in any of the literature is the difference between a classification or structured hierarchy and a list of each and every, or almost every, item being classified.'
A classification is a structured set of classes at several levels of detail-each level being a more detailed substructure of the higher level. A classification structure should be relatively static; that is, it should not change often.
On the other hand, a single list of uniquely defined commodities should be complete and it should be dynamic, in that it should be constantly updated on a demand basis. Very detailed items, including duplicates (i.e., automobiles, cars, Buicks) could be included in this descriptor list. Each of these 30,000 to 100,000 items should be uniquely identified by a
'This applies to other classifications also, such as occupations and industries, etc.
six-digit code, and should be assigned the classification codes of several commodity structures, including the Census Bureau's Numerical List of Manufactured Products, the Tariff Schedule of the United States, Annotated, the Standard International Trade Classification, and the Customs Cooperation Council Nomenclature. The most appropriate industry of origin in the SIC could be included also.
When new products become important in trade (say $200,000 annual production or consumption) the commodity descriptor should be split, with the next available sequential numbers assigned to the new product and to the remaining part. These new codes must retain the same classification code references to the major structures. The old code number would be retired and the two (or more) new descriptors would be added in chronological sequence with a reference to the old code. This sequential numbering provides a chronological timetable showing the original items and the new ones, as well as the approximate date when the item was introduced. This aspect of coding has been ignored but is probably as important as any other characteristic of coding systems, since it provides a history of all commodity coding changes.
Old code numbers would not be reused but would still be linked to each structure and would provide some comparability and overlap for the product. They should be flagged to remind users that they are obsolete numbers and current codes should be used. A particular commodity could be reclassified in any one of the more static statistical structures when it has been described independently for several years in the description list.2
The commodity descriptors would not separately identify different colors and sizes of a commodity. There may be a difference between, for instance, unflavored toothpaste in tankcar lots and in four ounce tubes, but this is a much different market, a different unit value, a different transportation characteristic-all establishing the basis for separate identification. The difference between family size and institution size containers might or might not be a criteria for separate identity. But commodities in bulk lots versus retail packages should probably be separately described.'
'Computer technology is now at a point where most systems could use detailed codes for the items and provide internal aggregations. Changing the codes of all the detail because the code for a particular portion of the structure is changed is not necessary for today's computers. Old tabulating machines needed a structure built into the codes.
'The descriptor list should probably have different descriptors for carbon and stainless steel plate and for integral versus fractional horsepower motors, as well as differentiating between single-phase and poly-phase motors.
A Standard Commodity Classification
There are problems involved in setting up a single all-purpose commodity classification to serve all major user organizations effectively in terms of their program needs and responsibilities. Classes that are significant in a production classification structure are quite different from the classes that are significant for a foreign trade classification. For instance, we import practically all of our coffee beans, nickel ore, tin ore, and many spices but produce practically none. Therefore, a tabulation which includes such classes for production statistics would have empty cells, or they would have to be suppressed because of disclosure. On the other hand, ready mixed concrete, concrete blocks, or fresh milk are of little or no importance in export trade although they are very important in our production statistics.
There are different reasons for collecting various commodity data. The International Trade Commission (ITC) is concerned with the evaluation of the impact of duties which are imposed on imported commodities. The protection of the capacity for American production of a particular item does not necessarily make it important for economic analysis by market groups.
Classification of commodities for transportation uses have a different basis for classification. For example, the railroads have difficulty separately identifying pet foods from livestock feeds. Transportation statistics now combine these two products (of two different industries) into one class. The apparel classification for railroad transportation is much more aggregated than production data. Railroads do not have differential rates for shirts versus pants, swim suits, or stockings. These are separate items in production statistics. Further, the material, type of weave, and whether or not the apparel is embroidered are criteria in the import classification, but not for production or transportation uses. Transportation interests are not concerned with the difference between wood or metal furniture, since neither type is dense enough to be near a box car's capacity in terms of weight. Transportation statistics classifications also need categories for construction equipment, military goods, empty containers, scrap, circuses, and household goods. These are not classes in any of the production classifications, so extra categories must be added.
The Office of Management and Budget published a detailed Standard Commodity Classification System in 1943-5. The Defense Supply Agency is reported to be the only agency which tried to use it. This classification is out of date, and beyond redemption.
The Patent Office also has a classification system, much of which is on a commodity basis.
Agencies may be reluctant to give up their particular classification system for a standard. However, by defining each classification structure in terms of uniform descriptions, every agency would have the option of using the detailed codes which could then be aggregated to any structure. Agencies might welcome the development of a master reference list of descriptors which could be used to define each class in their commodity classification. This unique definition of a product in terms of a common base at the most detailed level of differentiation will also provide us with substantially increased ability to relate classification structures to each other. It may encourage the various interests in existing commodity classifications to move towards more standardization when making changes in their system, since they will be able to see the differences more clearly. For instance, if one classification differs from another one by one or two small well-defined items, that agency would be more likely to shift the commodity boundary for those items to make them more comparable to other classifications. At present, these differences are not very well defined.
The basic definitional source for commodity classifications should be an international master reference descriptor list. A descriptor list developed by the United States or any other country will not receive the universal acceptance necessary for international use. The United Nations does not have a large enough staff to do this type of work. The Customs Cooperations Council has a substantial staff and expertise, but they are now engaged almost exclusively in developing their harmonized structure; therefore, progress on descriptors has been virtually at a standstill. It will probably be awhile before they will be able to put substantial resources on the descriptors development.
The creation of a master reference source extends beyond statistical policy concerns. A tariff schedule is primarily an administrative and tax system rather than a statistical device; the U.S. Customs Service is not primarily interested in statistics, they collect tariffs. The Department of Transportation is more interested in facilitating the movement of goods, primarily through the use of a thru-bill of lading, rather than the collection of general-purpose statistics. These agencies do not perceive their role as being subservient to statistical policy considerations for such classification.
As a first step in moving toward commodity classification standardization, a committee of technicians should be set up with representation from all concerned agencies. This committee would provide coordination between agencies during revisions of all commodity classifications, and also would coordinate any problems relating to industrial classifications. A lead agency could be assigned responsibility for a particular classification area and the OFSPS-sponsored committee would serve primarily in an arbitrating role to resolve interagency differences.
There are several organizations which are concerned with commodity classification such as the American Association of Railroads, the American Trucking Association, air and ship carriers, Transportation Data Coordinating Committee (TDCC), National Committee on International Trade Documentation (NCITD), shippers, bankers, and port authorities. Organizations concerned with the transmitting of shipping documents across oceans would be interested. These private organizations should be kept informed of our intentions and may contribute some of their knowledge in using commodity classifications to the development of such descriptors and classification structures.
The organizations referred to above should be used in a direct or advisory role to the technical committee. The marketing managers are interested in and concerned with the statistics available from such a program, but the traffic managers are concerned with moving goods. These two concerns must be balanced if they are to be directly involved with technical committee.
This overall project is a continuing one; it may be 5 or 10 years before various Government and major private classifications could be oriented toward this system. However, it is imperative that the basic work be started quite soon, since the UN is working on harmonizing the International Standard Industrial Classification with the Standard International Trade Classification, the Census Bureau-International Trade Commission work is continuing. In the absence of a unified effort, agencies may develop or revise their own classifications without referring to this descriptor list as a basic or additional way to define their systems.
Agencies which need only a limited number of classes could define them in terms of another intermediate classification. As an example, the Interstate Commerce Commission's (ICC) classification for freight movements is a truncated
version of OFSPS's Commodity Classification for Transportation Statistics, which is a five-digit structure. The seven-digit Standard Transportation Commodity Classification (STCC) developed by the American Association of Railroads, is a more detailed classification with the same five-digit definition. A direct relationship to the detailed descriptor list or to the STCC would not be needed for the ICC's needs.
Much of the work to develop the interrelationship of different classification systems should be done in conjunction with the Customs Cooperative Council's work as they also have a long-range plan to develop a basic descriptor list. It appears though that they are not considering a system as detailed as would be necessary to meet all U.S. needs. The best time for each agency to insure that the basic list has enough detail for their needs is while it is being developed. The Department of Transportation (with a contract to Transportation Data Coordinating Committee) has made a good start in the development of a basic set of descriptors derived from the STCC. The STCC was chosen because it has the most detailed list of items (14,000 in all). The list now has about 22,000 items, since the CCCN and Schedule B have been integrated into this descriptor list. Incorporating the TSUSA and Numerical List of Manufactured Products into it will probably bring it up over 25,000 items.
Classification of Services
The U.S. does not have any formal classification of services similar to the classification of goods. The U.N. has a draft International Classification of All Goods and Services (ICGS), which classifies both goods and services in a consistent format. The goods area is similar to the Census Bureau's Numerical List of Manufactured Products, and the services is a similar application of that concept to the services sector. It incorporates a truncated goods classification for use in the transportation, wholesaling and retailing areas.
The ICGS should be used as a starting point and modified and expanded where necessary for use in U.S. statistical programs. It should be used in the economic censuses where appropriate.
This is a difficult area because of the diversity of services and the many ways the services can be combined. Forcing a "service" classification into the "industry of origin" concept is somewhat difficult, but when used with discretion it would provide an interrelationship with the industrial classification. The correlation will provide users with statistics of
economic activity that can be related to statistics about the establishments conducting those activities. Standard Metropolitan Statistical Areas
Standard definitions of metropolitan statistical areas were first issued by the then Bureau of the Budget (predecessor of OMB) in 1949, under the designation "Standard Metropolitan Areas"; the present designation was adopted in 1959. The general concept of a metropolitan area is one of an integrated economic and social unit with a large population nucleus.
The concept of a standard metropolitan statistical area is based upon a body of published objective criteria. The criteria for the establishment and definition of SMSA's have undergone several modifications since 1949. The current criteria used to define and designate standard metropolitan statistical areas as well as a list of the titles and definitions of the SMSA's are contained in the publication Standard Metropolitan Statistical Areas, 1975 Revised Edition. OFSPS issues amendments to update this publication when changes in SMSA definitions
Information reported in a census taken by the Bureau of the Census is the usual basis for designating a standard metropolitan statistical area. Population estimates prepared by the Bureau of the Census which have been accepted for use in the distribution of Federal benefits are also used as a basis for designating a standard metropolitan statistical area.
All standard metropolitan statistical areas include the county or counties in which the qualifying population resides. Other counties may be added to the SMSA definition provided that information from a census shows that they meet certain criteria of metropolitan character and integration.
The criteria used to designate standard metropolitan statistical areas are subject to continuing review, and changes are made as appropriate. They represent a reasoned judgment as to how metropolitan areas may be defined statistically in a uniform manner, using data items that are (1) widely recognized as indicative of metropolitan character (population, urban character, nonagricultural employment, population density, commuting ties); and (2) available from a body of Federal statistics which has been collected at the same time in all parts of the country and processed and tabulated according to consistent standards. Further changes in the criteria may be expected in the future as additional statistical data become available and as the
nature and structure of metropolitan areas themselves become better understood. Currently the Federal Committee on Standard Metropolitan Statistical Areas is reviewing the criteria to determine what changes should be made in preparation for the 1980 census.
In some parts of the country, metropolitan development has progressed to the point that adjoining metropolitan areas are themselves socially and economically closely related. The concept of Standard Consolidated Statistical Area (SCSA) recognizes this phenomenon. SCSA's defined under criteria adopted by OMB in August 1975, include two or more contiguous SMSA's which meet certain criteria of size, urban character, integration, and contiguity of urbanized areas.
The Base Year definition provides for a consistent base so that indexes can be compared more directly. The base year should be redesignated every 10 years or so. A base should preferably be a year in which economic activity is relatively normal and it should not have unusual changes in inventories, unemployment or prices from the beginning to the end of the year.
A one-year base is easier to calculate and understand, and if a single year can be found that meets the guidelines it should be designated rather than multiple years. The period used for base weights should be the same as the base year, but if it is not then the weights must be adjusted to the base year.
An important consideration in the choice of a base year should be a year in which the economic census is taken because those years are used to benchmark many statistical series. These years are therefore not as likely to be revised to the extent that other years may be.
A new base year should not be designated until the data for the base year is made final or very near final. Gas Pressure Base
The standard is necessary because of the fluctuation of the volume of gases due to temperature changes. It should also show the metric equivalent. Race/Ethnicity Standards
A standard classification for race and ethnicity has been developed and is included in OFSPS Directives. This standard is described in the chapter on Civil Rights.
Relationships to International Standards
There are several international statistical standards for which the U.S. is regularly asked to provide data to international bodies. These standards are not identical to U.S. standards. In many cases, international categories can be derived by adding together several U.S. categories, in other cases, a small amount of distortion is introduced because the categories have definitions which are not identical. In a few cases, the definitions are so different that only more summary data can be compiled.
Many of these differences in classification are due to differences between U.S. and other countries' industrial structure or process used in manufacturing. Other differences may not be based on any substantive reason and a major effort should be made to resolve these differences either in the U.S. or international classifications.
In order to improve the comparability, the OFSPS, with advice of concerned agencies, should make a determination whether to change the U.S. classification or to propose a change in the international standard. It is recognized that, at least in the industrial classification, major differences in structure exist between the United States and other countries. This effort should only be concerned with the capability to aggregate data to the international standards. It would not involve using the international standards in either detail or structure.
In commodity classifications where there appear to be strong reasons to retain the differences, it should be proposed to define subgroups which will make it possible to recombine U.S. data to international standards. In any case, we should strive to eliminate differences. It is much more important to have international comparability in a commodity classification than industrial or occupational classifications. Therefore, every difference should be resolved and we should decide which classifications should be proposed for revision.
A new Directive, formerly Circular No. A-91, "Prompt Compilation and Release of Statistical Information" (originally issued on April 26, 1972 and revised Oct. 21, 1974) was issued to insure that principal statistical series which are compiled by agencies annually or more frequently are compiled and released without unnecessary delay.
Further work on the tradeoffs between early release and accuracy need to be done, as well as on the frequency and timing of revisions.