Printer Friendly

Statistics corner: revising classifications of economic activity.

ON NOVEMBER 6-8, 1991 the U.S. Bureau of the Census sponsored an International Conference on the Classification of Economic Activity in Williamsburg, Virginia. The conference was organized by the Bureau in response to the increasing criticism of the Standard Industrial Classification System (SIC) that is used by governmental statistical agencies to measure economic activity in the U.S.

This conference, which was organized by NABE Board Member and Associate Director of the Census Bureau Charles Waite, could prove to be a watershed event in the history of U.S. industrial statistics. The three studies conducted under the sponsorship of the Census Bureau identified innovative approaches to classifying economic activity. At the conference Charles Waite called upon all participants to offer further suggestions for "alternative classification systems that would better measure and describe the economy of the 21st century." NABE members may wish to submit suggestions to Census in the coming months.

In the opening session of the conference I participated in a panel on "Perspectives on the SIC: Problems -- Real and Perceived." This column summarizes some of the issues and suggestions that should be addressed. Copies of the full session are available from the Census Bureau.


In the late 1930s the Central Statistical Board -- a coordinating body established by President Roosevelt -- established an Interdepartmental Committee on Industrial Classification to:

"develop a plan of classification of various types

of statistical data by industries and to promote

the general adoption of such classification as

the standard classification of the Federal


Mr. V.S. Kolesnikoff was in charge of the work of the Technical Committee that was established to undertake the development of the standard classification. He led the effort for two decades beginning with the first meeting held on June 22, 1937. The first printed edition describing manufacturing industries was published in 1941 and that of non-manufacturing industries was published in 1942. As the standards were developed it took about a decade for the statistical agencies actually to adopt the proposed classifications.

The initial project was designed to classify "industry" in the broad sense of all economic activity; that is, agriculture, forestry, and fisheries; mining; construction; manufacturing; wholesale and retail trade; finance; insurance, and real estate; transportation; communication; electric, gas and sanitary services; and services. In 1952, the Advisory Council on Federal Reports undertook the first major review of the Standard Industrial Classification. For this purpose, twenty-seven committees were established to undertake a review by sector. More than 350 members of the Council participated in the review during a period of two years.

Additional revisions were published in 1958 and 1963 when changes were restricted to three-digit industry groups. A revised edition of the SIC manual was published in 1967, in 1972 and in 1987. All of these revisions have occurred within the rules and guidelines of the original classification effort.


As noted in the description of the historical evolution of the SIC, changes are narrowly permitted and the effort required to define changes takes a long time. What are the barriers to change?

The purpose of a standard classification system is to assure that statistics developed by different agencies or different sources have compatibility, so that we are assured that the same activities are being measured. Once industry groups get used to a given definition, they are reluctant to change because they believe all users of the data and, indeed, all suppliers of the data are familiar with the current system. Further, changes destroy time series and even improved classifications result in a loss of historic continuity.

A good classification system must meet a number of requirements. They are:

1. They must be clearly defined to permit the

collection of accurate statistics.

2. They need to assure comparability between data

series such as investment input and production


3. They need to reflect the current structure of the

economy in order to assist in analysis of important


Most observers of the current Standard Industrial Classification are dissatisfied on all three points. Some feel that revisions have been too modest and that the nonmanufacturing sector, in particular, is not adequately defined. The diversification of financial services with similar products being offered by banks, investment firms, pension funds, and other organizations at the international level make it particularly difficult to define "financial services" as an industry. In the area of health care a number of new forms of health service, including retirement facilities, health maintenance organizations and specialized testing operations, have totally restructured the health care delivery sector of the economy.

Another issue relates to the character of the firm. Individual establishments today often produce a wide variety of output, so that it is not clear how to classify properly the types of activities in a single location. Output is often diverted to intracompany activities while at the same time serving as specialized OEM outputs for unrelated organizations.

These problems were underscored in the 1970s when the Securities and Exchange Commission attempted to develop information on the basis of "lines of business." When similar facilities produced diverse outputs, it was difficult to assign general overheads including management costs, marketing and administrative overhead expense, physical space rental, and development costs to individual product lines. For example, how is the cost of a computer center to be allocated across functions of recordkeeping for accounting purposes, management of payroll for taxes and human resources management, monitoring of sales, inventories and production and other activities necessary to manage the business. These computer operations are often labelled the "information age," but what is their contribution to value added in a production sense?

Moving from a classification system originally designed to measure physical product to a classification system that tracks the contribution of service activities such as banking, investment, health care, and education implies a dramatic and even creative alteration in the concept of "industry." To make such a change would have major impact on several agencies, including the Bureau of Labor Statistics, which collects employment data by industry, and the Bureau of Economic Analysis, which defines value added in the economic system on the basis of industry. These organizations and many others can be expected to resist any major redefinition of the concepts that would have the effect of making their systems outmoded or, more importantly, impossible to implement with data developed on the basis of alternate concepts.


In the 1990s our accounting systems have greatly improved and business management knows much more about its activities than we characteristically assume. Furthermore, computer technology is now pervasive in businesses at all levels. It is possible to extend current recordkeeping practices to gain more insight into the business operations. In fact, one of the complaints about computer technology is that is provides so many new windows on the business operations that productivity is not really increased, but rather analysis consumes too much time.

In taking a "clean sheet" approach to industrial classification, I propose that we think in functional terms about what is happening in the economic system. By functional terms I would focus on characterizing the economic impact of specific businesses. This means, in the first instance, that considerable effort should be made to define the business structure. It is well recognized that in terms of number of enterprises the bulk of business organizations are small businesses. More than 90 percent of these businesses have fewer than twenty employees. This fact makes it possible to think about the characterization of the economic activity of these small units in a micro sense. What do these businesses contribute to the economic system functionally?

The challenge, of course, is to dissect the functional activities of the large corporations that, in fact, cut across many types of functional activities. What are these functional activities?

In a simple sense they can be characterized as follows:

1. Production of physical product. The physical

product can be measured both in terms of quantities,

as well as specific characteristics such as materials.

For example, the widget may be wood, plastic, or

metal (to take a few examples), and the functional

impact on the economic system is different because

the input resources are different.

2. Functionally, many dimensions of the economic

system represent specific services, including sales,

marketing, accounting, legal activities, health care,

education, business services, financial services,

transportation services, etc. These broad

categories, of course, have many subcategories. Within

education we recognize a wide range of activities

from elementary and secondary education in the

formal sense through job training and retraining in

the informal sense.

3. The economic system also includes a number of

basic infrastructure activities, such as the

generation of energy, the production of food, the

construction and maintenance of structures, capital

equipment required for production, and social

infrastructure, such as highways, bridges, sewers,

etc. The immediate impact of infrastructure is the

cost of building the infrastructure; yet the economic

impact extends over the lifetime of the

infrastructure -- a sharp distinction from many of the

consumption activities in the economic system.

With a "clean sheet" approach we should seek to define industry based on its functional impact. The transfer of printing from the manufacturing organization to a services organization does not change how the economic system operates. Thus, with a "clean sheet" we should try to distinguish between services inside the manufacturing firm versus purchased services. We should also try to distinguish between product use versus service required by the product; and, as noted above, we should distinguish between balance sheet items (capital goods) versus consumption of components in the production of individual products.

In approaching this new functional definition of industry, a major effort should be made to incorporate the concepts of value added that are associated with the system of national accounts, which is now being undertaken to expand upon the national income and product accounts to provide a richer perspective on both the income statement and the balance sheet of the economy. A major effort to redefine industry to assist in better accounting for the economic system should represent an important incentive to revising industry classification.


These suggestions are designed to be provocative with the hope that other members of NABE will offer further suggestions or comments. The Census Bureau is serious about stimulating a broad review. The papers in the conference offered a number of constructive suggestions and it is anticipated that the debate that has been initiated will help in the development of a creative and useful improvement in the classifications that are the framework for our basic economic statistics.

Joseph W. Duncan is Vice President, Corporate Economist and Chief Statistician, The Dun & Bradstreet Corp., New York, NY.
COPYRIGHT 1992 The National Association for Business Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Duncan, Joseph W.
Publication:Business Economics
Date:Jan 1, 1992
Previous Article:The business economist at work: the economics function at the Colorado Legislative Council.
Next Article:Research notes: financing plant and equipment expenditures.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters