Printer Friendly

Integrity in official statistics.

AT A NABE Statistics Committee meeting near the end of 1992, Ed Mennis, Editor of Business Economics, suggested that we organize a special issue of Business Economics on the subject of "The Integrity of Official Statistics." This issue is the result of Ed moving forward with that creative idea.

I encouraged this special issue for two personal reasons. First, during my year as president of NABE, I've been making a special effort to call attention to the need for improving U.S. federal statistics. Second, in 1971, I was appointed Chairman of a joint subcommittee of the Federal Statistics Users' Conference and the American Statistical Association to "draw up a statement reaffirming the need for a federal statistics system of unquestioned integrity and to develop recommendations concerning procedure designed to protect the integrity of the federal statistics system."(1)

That committee was created because of concern that the federal statistical system might become politicized, a concern that was stimulated by cancellation of the press conference concerning unemployment and employment data. (This incident is discussed in Janet Norwood's article in this issue.) In 1971, the establishment of the committee was a result of several developments in addition to the cancellation of the press conferences. They included:

1. Reassignment of personnel and the reorganization of the Bureau of Labor Statistics, especially those persons previously associated with the press conferences that had been discontinued.

2. A number of specific personnel shifts and several premature retirements of top level statistical personnel in important statistical agencies, including the U.S. Bureau of the Census.

3. A reorganization of statistical agencies within the Commerce Department, which resulted in a merger of analytical and policy agencies, reducing significantly the authority and power of the major operating statistical agency.(2)

The 1973 report of the FSUC/ASA Committee on the Integrity of Federal Statistics also outlined several basic ingredients for maintaining statistical integrity. These ingredients were:

1. The statistics themselves must be accurate, consistent, and timely.

2. The public must have confidence in the statistics that are generated and in the professional ability of the people who produce them.

3. Statistical programs must be continually revised and improved to reflect new characteristics of the subjects being measure and embrace new subjects as national priorities change. These revisions must be undertaken on the basis of sound statistical principles to assure that the refinements continually result in more reliable and more sensitive statistical indicators.

4. Technical measures of reliability and sensitivity should be available to define the uncertainties and limitations associated with specific series. This requires equal attention to be given to the gathering of basic statistical data and to the compilation, adjustment, and presentation of the resulting analytical measures and statistical reports.(3)

Moving forward to 1992, we see, as discussed in detail in the articles in this issues, that integrity questions remain both an area of concern and an area of policy attention. Issues were raised about BLS revisions of employment statistics in early 1991 and about the strong report on gross domestic product growth that occurred near the November 1992 election. This issue of Business Economics was organized as an update to NABE members because of the press attention that has been given to concerns about the integrity of the federal statistical agencies. As Carol Carson notes in her introduction, "Too few business economists are aware of the systematic efforts taken by BEA to ensure the integrity of the GDP."

These articles are designed to explore current developments in maintaining the integrity of federal statistics. Each author was selected because of his or her particular knowledge of key aspects of the statistical system. All four authors were given independence in addressing the subject, although they did share ideas about the scope for individual articles, so that it would be possible to reduce (but not eliminate) overlapping discussions.

The article by Bruce Johnson and Lori Rectanus of the General Accounting Office represents a case study of governmental review of the integrity of statistical agencies. It approaches this subject by using the concerns that the first quarter 1991 GDP estimates were inappropriately developed given problems with the underlying base information from the Current Employment Statistics (CES) survey of BLS.

The next article by Carol S. Carson, Director of the Bureau of Economic Analysis, details the procedures used by the Bureau of Economic Analysis to safeguard the integrity, objectivity and professionalism of the agency that develops the GDP estimates. The final item in the series is written by Janet Norwood, former Commissioner of the Bureau of Labor Statistics. Her article reviews the broad issues of maintaining the accuracy and integrity of federal statistical programs.

Taken together, these three articles do much more than provide an update on this critical issue. They raise a number of interesting issues that merit further attention, especially by the community of professional business economists in NABE, who have a particular interest in and dependence upon the accuracy and integrity of official statistics. We encourage you to read all three articles as an excellent introduction to these issues. Let me suggest a few areas where your comments, guidance, and help would be most appropriate. They are:

1. Bruce Johnson, Assistant Director of GAO, suggests that the director of BEA should be a politically appointed director. Will this truly serve to isolate and/or protect the agency from political influence?

2. Carol Carson points out the need for a more active approach in rebutting challenges to statistical integrity, especially from representatives of the press. Will such an aggressive approach result in an appearance of "defensiveness" and thereby lead to further public concerns about system integrity?

3. Carol Carson also recommends a long-term advisory committee to oversee the development of new methodology and data revisions. How does one achieve a trade-off between the benefit from the long experience of advisors and the risk of in-bred perspectives that retard improvement?

4. Janet Norwood points out that "The statistical system has been starved for research funds for over a decade, and it is has been difficult to keep up with all of the fast changing concepts and methods." In a period of concern about federal spending, especially for overhead activities like statistics, how is public support for a topic like relatively esoteric research to be obtained? And importantly, what should be the role of business economists in building such support?

On balance, I come away from these articles with continued confidence in the professionalism of the leaders and workers in the U.S. statistical system. Based on my experience as a user for many years, and as a participant in the system for eight years, I continue to believe that the U.S. statistical system in the best official system in the world, although obviously there are opportunities for further improvement. In particular, Norwood's point about the need to adjust concepts as economic reality changes is particularly important as we prepare for the radically evolving global economic system of the twenty-first century.


For a number of years now we have included The Statistics Corner as a regular feature of Business Economics as a method for reporting upon and encouraging participation by NABE members to improve the quality of official statistics. Currently, there are two specific opportunities for input from NABE members, where you views can be very helpful and important. These are:

1. A potential revision of the concepts for economic classification policy, more popularly referred to as the Standard Industrial Classification.

2. An invitation to participate in providing comments on criteria for assessing internal designs for the Census for the year 2000.

A brief comment on each of these is presented here to interest you in obtaining more information as outlined in each section.

Economic Classification Policy. On March 31, 1993, the Office of Management and Budget issued a solicitation for comments concerning the subject of economic classification. This announcement notes that the economic classification policy committee intends "to identify the essential statistical uses of the data produced from economic classifications through research, experiences of statistical producers, information obtained through an outreach process, and information obtained from data users concerning the actual and potential uses of data that are produced with economic classification systems." The committee believes "facilitating the use of economic data should be the primary purpose of an economic classification system."

In that regard, they have issued papers for review by interested parties and they plan further publications. They are soliciting public input. While the process is already underway, and the first set of comments was requested for submission by May 28, 1993, it is expected that this process will take some time because the revision of the SIC is tentatively scheduled for implementation in January 1997.

Therefore, for those of you who missed the first round of comment, I encourage you to obtain the new issue papers as they are released and to provide additional comments. Requests for issue papers and printed information should be addressed to:

Economic Classification Policy Committee Bureau of Economic Analysis (BE-42) U.S. Department of Commerce Washington, DC 20230

You can obtain information by calling 202-523-0873. This process is being coordinated by Jack E. Triplett, Chairman of the Economic Classification Policy Committee.

Planning for Census of the Year 2000. The dicennial census is the most extensive data collection activity of the Federal Government. In each dicennial cycle there are continuing efforts to improve methodology and to refine the specific elements of the data collection process. Currently, a process is under way that will ultimately require several years of planning and testing. The first step is a request for comments on the proposed "criteria" for assessing alternate designs of the year 2000 census. Many of these proposed criteria are straightforward, including such things as the judgment that it is not feasible to seek a constitutional amendment between now and the year 2000 because the planning and testing would be required to move forward before the amendment could be achieved.

The full set of proposed criteria can be obtained from:

James L. Dinwiddie, Program Manager Year 2000 Research and Development Staff Bureau of the Census U.S. Department of Commerce Washington, DC 20233 Phone 301-763-4040

Design features of the proposed 2000 census will be evaluated in a series of design alternative recommendations (DAR's). These DARs will be widely distributed and evaluated. Mr. Dinwiddie can also put you on the mailing list for the forthcoming design alternative recommendations.

I encourage NABE members to participate in these important public debates, because the outcome will result in significant changes in official statistics that we all use. It is important that the practical needs of business users be considered along with the specific programmatic needs of government.


1 The American Statistician, April 1973. Volume 27. Number 2, pp. 58-67.

2 Ibid., p 58.

3 Ibid, p. 59.

Joseph W. Duncan is Vice President, Corporate Economist and Chief Statistician of The Dun & Bradstreet Corporation, New York, NY. He is also President of NABE and Editor of The Statistics Corner of this journal.
COPYRIGHT 1993 The National Association for Business Economists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Duncan, Joseph W.
Publication:Business Economics
Date:Jul 1, 1993
Previous Article:Chaos and Order in the Capital Markets: A New View of Cycles, Prices, and Market Volatility.
Next Article:The integrity of federal statistics: a case study from the GAO perspective.

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