Page images



Cooperative Systems' Objectives/Content

The Federal-State-Local Cooperative Statistical Systems have been defined to include "those federally initiated or sponsored statistical programs in which State and local agencies participate in the collection or compilation of nationally standardized statistics. The cooperative systems are undertaken for the mutual benefit of the participants, involve multiple States (and local jurisdictions), and contain data of a recurrent nature which are intended to have broad applicability.”

Stemming from this definition are a number of principles related to the objectives, scope, and content of the cooperative systems. 1. The concept of Federal-State-Local Co

operative Statistical Systems should be pursued and promoted whenever State and local agencies are required to participate in the collection or compilation of nationally standardized statistics. Any Federal agency considering the development of a new or revised statistical series which is capable of producing State and local area estimates should be required to consider and investigate with State/local officials the possibility of making it a co

operative program. 2. The cooperative statistical systems should be

developed for the mutual benefit of the participating (local, State and Federal) agencies. Programs which do not involve State or local agencies in the collection or compilation of statistical data should not be classified as cooperative programs. Programs which serve only Federal needs should not be classifed as cooperative programs. If, after careful examination of alternatives, it is determined that there is no mutual benefit to establishing a cooperative program, the data collection activity in question should be continued as a

Federal system. 3. The cooperative systems should be designed

with the full participation of local agencies. How extensively the cooperative systems should, can, and do meet local data needs should be examined. This does not imply that the basic data system should be designed to incorporate all local data needs, but rather that the Federal and State components should be

developed so that local data needs can be incorporated where appropriate by the relevant

agencies. 4. There should be a periodic review of common

State data needs and integration of such requirements into the cooperative systems when inclusion promotes interstate comparisons and communications or is needed by policymakers at the Federal or State levels. This review should be conducted jointly by consortia of State and Federal organizations, and should lead to recommendations for changes to the co

operative systems. 5. The cooperative systems should strive to make

better (more extensive) use of standardized local administrative data as a source for required information. Guidelines to encourage their use should be issued at the Federal level. The basic data to be obtained as a by-product of administrative record systems could be supplemented, where necessary, by special survey activities. To the extent possible, the cooperative systems should call for data which is readily available at a central source. Where necessary, methods to achieve interstate comparability of data should be developed in preference to requiring original data collection

by the data producers. 6. Where several Federal agencies require data

from the same respondent units, the Office of Federal Statistical Policy and Standards should act as an arbiter to determine the appropriate Federal sponsor. Multiple cooperative systems within a single Federal agency should be sponsored only when a single cooperative system is deemed less beneficial to the several

levels of participants. 7. To minimize costs and burden at all levels, there

should be an ongoing mechanism for the interchange of ideas, the sharing of successful strategies and unresolved problems, and, where appropriate, the coordination of Federal-Statelocal cooperative systems. An interagency work group chaired by the Office of Federal Statistical Policy and Standards, including representatives of the major agencies sponsoring cooperative statistical programs, should be established for this purpose. Courterpart organizations should be established

at the State level to examine interagency needs, priorities, and problems, and to serve as a focal point for communication on interagency issues with the Federal level.

8. Federal agencies sponsoring cooperative

statistical programs should periodically review the need for data included in their systems to determine the continuing need for the information collected. The Office of Federal Statistical Policy and Standards, in cooperation with the interagency work group (#7, above), should periodically review the several cooperative statistical programs to determine where interagency coordination of data requirements is needed. The data base containing information on the cooperative programs should be periodically updated as a

resource for this purpose. 9. Use of the data obtained through the co

operative systems should be maximized. Individual Federal agencies, as well as the interagency work group, should develop methods and work with State agencies to improve data dissemination procedures and assist users in obtaining available information.

Cooperative Systems' Support (Federal)

At the present time, the level and type of support provided by the Federal agencies to the cooperating agencies (States and localities) is highly disparate. Types of support include: personnel, funding, computer/programming support, technical advice and assistance, training programs, etc. There is no consistent policy concerning the appropriate level or type of Federal support, nor for the appropriate sharing of responsibility by system participants (Federal, State, local).

While it is not clear that a single or uniform policy with respect to Federal support is appropriate, some principles have emerged: 1. Appropriate levels of direct Federal funding

and/or field personnel should be made available to assist State and local agencies in producing data required by the cooperative systems. The degree of support and the length of time that support will be provided should be clearly specified at the outset of the program. Federal data requirements should be distinguished from data needed solely or mostly for State purposes

when defining the Federal share of funding. 2. The appropriate Federal share in supporting the

cooperative statistical programs should include

appropriate shares for systems development and

for systems maintenance. 3. Support should be provided for State and local

agencies for standardizing administrative records which are integral to producing the

statistical data required by cooperative systems. 4. The development and use of formulae for al

locating Federal resources to cooperative

statistical system participants should be studied. 5. Eligibility and State matching requirements

pursuant to Federal support of participants should be clearly established prior to the implementation of any cooperative system. Matching requirements should be broad enough to allow for alternative types of contributions from the State and/or local participants. For those cooperative systems where it is clearly established that Federal funds are available only for development costs and initial system implementation, the State/local match requirements should be increased each year. Care should be taken to ensure that the matching requirements do not preclude or work against the participation of the less well-to-do States. Eligibility requirements must include commitments to meet quality control and other standards. Continuation funding should be based on performance and periodic review

should be made of matching requirements. 6. The cooperative statistical programs should be

periodically evaluated to determine what benefits/improvements/products have been achieved as a result of Federal support. Cooperative Systems' Quality

Control/Standards It is recognized that quality control is difficult to achieve in a decentralized system. The alternative, to maintain control through direct Federal collection of the needed data, has been dismissed as an unworkable choice. Direct Federal collection of the data frequently results in duplicative collection and compilation programs at the several governmental levels, a practice which can only result in increased burden on respondents and a significantly greater expenditure of national resources to produce the required information. Thus the question becomes, what can be done to maximize quality and comparability in the data produced through the cooperative statistical systems.

Principles directed to meeting this objective have begun to develop as a result of the review of ongoing cooperative statistical programs:

1. Effective design of a Federal-State-local co

operative program must include procedures to ensure interstate comparability of data. Data standards and definitions of terms must be developed, disseminated, and monitored as an integral part of the cooperative program.

2. Whenever possible, participating States should

adopt standard data elements and data definitions as a condition of participation (and of eligibility for support) in the cooperative programs. In cases where State or local requirements cannot be made consistent with Federal standards at the point of initial recordkeeping, provisions should be made to ensure comparability in the compilation or analytical process so that national (or other) estimates and interstate comparisions can be made reliably.

data. Quality control standards should be an integral part of every cooperative system. These standards should be developed jointly by Federal and State participants, and should be applied in the collection and compilation processes at the local and State levels. The Federal sponsors of the cooperative systems should continuously monitor and periodically evaluate participants' adherence to the standards in the collection and compilation of data. Periodic evaluations of the quality of the final data produced should be carried out to surface data problems which should be addressed. Timeliness standards should be established for the cooperative systems require a time lag of no more than 12 months for monthly or quarterly data and 18 months for annual data from the reference date of the

data to their publication. 4. The respective sponsoring Federal agencies

must be given full control over funding to State and local agencies cooperating in the statistical programs if quality and timeliness standards are to be achieved.

3. The use of Federal-State-local cooperative

statistical programs as the mechanism to obtain federally required information should not detract from the quality or timeliness of these

SSEL, an annual economic data publication program provides aggregate benchmark information.

The Industrial Directory program described here is the extension of the SSEL to other agencies for similar sampling purposes. The benefits of this extension include more consistent statistics, efficiency, and some reduction in respondent burden. The Industrial Directory is simply an effort to have the Federal statistical community operate as though it were a single agency for purposes of maintaining a mailing list and sampling frame of the nation's business establishments.

Current Situation

The Bureau of the Census has been developing the processing capability and contents of an Industrial Directory for the last seven years. It now includes all establishments with employees except religious and government organizations. The extension to these areas and to businesses with no employees is now under consideration. The record format and contents of the Directory are now in place and are being used for Census programs. Efforts are being made to develop interagency access capability probably using online computer terminals.

A technical group was set up by OMB in 1972 to promote the interchange of ideas among user agencies. This group has developed a statement and proposed legislation to provide access to the Directory to other Federal statistical agencies.


Purpose One of the major problems in maintaining high quality statistical output in the United States' statistical system is ensuring that samples used for statistical purposes are as efficient and as consistent as possible. This is important for two reasons: (1) sampling from a known universe of firms or establishments permits the collection of statistics with a known and high level of accuracy imposing a minimum of reporting burden on the public, and (2) statistics collected by different agencies purporting to represent the same industry, activity, size class, or geographical region can be made compatible, thereby facilitating meaningful integrated analysis.

At present, many agencies conducting statistical surveys do not have a good list of the universe of entities which they are attempting to measure. These agencies must develop their lists of respondents from secondary sources, or from preliminary direct surveys, causing a significant respondent burden and expenditure of agency funds resulting, in many cases, in respondent lists of questionable quality and unknown relationships to similar lists developed by other agencies. In lieu of a good list from which to draw samples, most agencies now conduct inefficient surveys involving more respondents than necessary for the required accuracy of information. In some cases, the agencies are unable to conduct surveys at all, and therefore do not collect the data needed for policy decisions and for program planning and evaluation.

The Standard Statistical Establishment List (SSEL), which is the Census Bureau's name for the Industrial Directory, is an up-to-date list of all business establishments in the United States. Its purpose is to provide sufficient identifying information about a universal sampling frame to be

sed for all Federal statistical programs involving establishments as defined in the Standard Industrial Classification (SIC). It also includes similar company identification and Enterprise SIC codes of all multiestablishment companies. It is now being used internally by Census to draw samples for their programs. As an additional product of developing the

Next Steps in the Development of an

Operating Industrial Directory Legislation will be needed to authorize the Census Bureau to make this information and sampling frames available to other statistical agencies. At present, Title 13 of the United States Code does not permit the Census Bureau to make such disclosures. It is recognized, however, that it is important to maintain confidentiality of the company and establishment names, addresses and size and activity codes included in the SSEL. For this reason, the conditions in the proposed legislative initiative for releasing SSEL information will be strictly limited to

statistical purposes, and will prohibit further release of the information in identifiable form by user agencies.

The effort to provide the legal basis for the directory use should be made regardless of plans to reorganize the statistical activities of the Federal Government. It is very important that the legislation authorizing access to the Directory be done now, because attempting to reconcile various systems at the same time a reorganization is being implemented will certainly make it very difficult to keep several continuing statistical programs running smoothly.

The procedures for developing samples, the cost to agencies and clarification of which specific agencies can participate, and so forth, have not yet been developed since they may be affected by the legislation. The Census Bureau has written a set of working papers which outline the format, content, updating procedures and sources, scope, frequency of updating, and so forth, in detail.

One possibility for sample selection is to preselect several nonduplicative samples. Any particular agency which needs a sample without any special sampling characteristics would designate the particular standard error which is acceptable for their purposes. The Bureau of the Census could then provide the agency one or more segments of the preselected samples necessary to achieve those desired results. This arrangement would minimize the burden on the individual respondent by assigning establishments to one and only one of the preselected samples. (Although they might be also selected in a specialized sampling procedure.) In some cases it might make sense to utilize a particular segment more thoroughly for one or two years and then reduce the likelihood of selecting those establishments for the next two or three years. This procedure would allow for data collected in one survey to be available for summarization by characteristics collected on another survey from the identical sample. Every effort should be made to minimize the total respondent burden, even though that may mean a larger than average burden for selected establishments in one year, with a lower level in succeeding years.

annual level to make it comparable to other Census statistics.

When the Industrial Directory is operating and the reconciliation of the definitions of establishments with other agencies is complete, the collection of employment information by the Census Bureau could be reduced or minimized, and Bureau of Labor Statistics data could be used for the basic data in many cases. This would require a greater degree of cooperation and a change in the BLS confidentiality arrangement with their respondents and the cooperating States.

Additional size codes for characteristics such as number of hospital beds, number of classrooms, acres in farms, or kilowatt generating capacity could also be incorporated to maximize the directory's usefulness in specialized programs.

In these and other areas, there may be data such as receipts, production data and capital expenditures which could be transferred to Census to minimize duplication and collecting burden.

Enhancement of Industrial Directory Use

The annual Economic Data Program is based on the wealth of data backing up the Industrial Directory. The data used to determine the size codes for payroll, employment, sales receipts, and many other items can be related by Census to the establishments in the Directory. That data could be used to link establishment or company data from other agencies and summarize the individual data to disclosable levels to provide a wealth of statistics for various other agency uses.

For instance, a business survey collecting certain characteristics but not certain economic data could provide those characteristics which they collected to the Census Bureau. The data could temporarily supplement the industrial directory background data and summary economic tabulations based on business survey characteristics could be developed from Census data. This use of the Directory and its wealth of background information would require complex disclosure analysis programs, but, nevertheless, could provide very useful statistical data for other agency uses. It should be reiterated that individual company or establishment data would not be available to the using agency.

Impact of the Industrial Directory on

Other Data Collection Programs Currently, both BLS and the Census Bureau collect employment, payroll and hours worked data. This duplication is now necessary because of lack of comparability in the identity of particular establishments which cannot be reconciled under current laws. The Bureau of Labor Statistics collects monthly data which could be summarized to an

Possible Further State Use of the Industrial

Directory A possible use of the Directory beyond the uses proposed in this statement would be to widen the

« PreviousContinue »