Chapter 1.3: Following international standards

1. Generic strategic considerations in adopting and using international standards

Implementing an international standard requires a comprehensive strategic process to better inform statistical production and dissemination. As mentioned earlier in this chapter, there is often a strong link between the use of international frameworks and standards, the relevance for users of the data produced, and the efficiency of statistical systems. Although it is important to adopt international frameworks and standards, doing so can be a costly and demanding endeavour. Decisions about what to implement, when to implement, and how to implement must be well thought through, and must lead to realistic implementation plans.

The followings are a series of questions that NSOs should address prior to the implementation of an international standard:

• Does the international standard address the realities of the society, the economy and the physical environment of the country?
• Does the international standard inform the development of statistical programs?
• Are all components of the standards relevant to local realities and users' needs? What components should be considered and which ones should be ignored? What are the most important components to include? What is the rationale for not following certain aspects of a standard, and what would be the impact? How much does the NSO need to deviate from the standard to meet the needs of its own context while respecting the underlying principle of comparability?
• Has the NSO widely consulted its key stakeholders and data users to get their feedback? Were their points of view considered and their concerns addressed? Will the resulting data products be relevant in the national context?
• Does the NSO have the appropriate funds to implement and maintain this new standard and to reconcile what is available now with what is expected to be coming further to the new implementation?
• Does the NSO have the core infrastructure (statistical, information technology, human resources) to populate statistical data using this new standard?

It is important to emphasize the fact that applicability of any kind of standard or framework should be the object of deep consultations with key users and stakeholders. Consultations, engagement, and open and ongoing communications are crucial, because key users and stakeholders need to be consulted early in the process on potential changes and their impact; and once the NSO decides to implement a new standard or a change a standard, it is imperative that stakeholders and users be made aware, in advance, of the implementation plan and changes.

2. Strategy and process for the adoption and use of standardized statistical units, populations, concepts, definitions, variables and classifications

As previously mentioned, the adoption and use of standardized statistical units, populations, concepts, definitions, variables and classifications is key to ensuring statistical information coherence, integration and comparability across programs and geographical boundaries. Text box 1.3.1, below, provides definitions of the key concepts.

Good management practices in the use of international standards usually include the following:

• a governance structure supported by consultation mechanisms ensuring that the impacts of any decision on the internal producers and external users of the statistics to which the standards will apply are taken into account.
• the creation of specific and specialized organizational units, with the appropriate level of seniority, responsible for (1) taking the lead in the adoption and development of statistical standards, (2) supporting statistical programs/domains in their efforts to develop standards, where such standards do not exist or have become outdated, and (3) providing coherence or mapping between different versions of standards for time series.
• a process to monitor the extent to which statistical standards are used by the statistical programmes/domains and make them accountable for their application of the standards; and
• a communications strategy to ensure that all relevant staff are aware of statistical standards and any changes made to them, as well as the degree to which the application of each standard is compulsory.^{Endnote 3}

In Statistics Canada, the use of standardized statistical units, populations, concepts, definitions, variables and classifications benefits from a specific governance structure and a policy that aim for an effective and consistent approach.

Statistics Canada's Policy on Standards provides a framework for reviewing, documenting, authorizing, and monitoring the use of standard names and definitions for populations, statistical units, concepts, variables and classifications used in Statistics Canada's programs. The policy prescribes different levels of implementation requirements for standards approved by the appropriate senior committee (the Methods and Standards Committee):

• departmental standard – application is compulsory, unless an exemption has been explicitly obtained under the terms of the policy;
• recommended standard – a standard that has been recognized with or without a trial period of a specified duration, after which it may be declared a departmental standard;
• program-specific standard – a standard adopted by a statistical program.

Statistics Canada also has a mature and effective governance and management structure that ensures an integrated approach to strategic priority-setting, decision making and accountability (see Chapter 2.2: Integrated Strategic Planning). Responsibilities for adopting and using international standards are shared among three entities: the Standards Division, the Methods and Standards Committee, and program areas. The latter two are expected to ensure that the impact on end users of statistics is understood and factored into any standards-related decision.

The Standards Division's mandate is to develop, maintain and communicate statistical standards, to promote and monitor their implementation, to provide guidance on their interpretation, and to produce bridging mechanisms (for example, concordance tables) that facilitate the comparison or transfer of data across different classification schemes. The Standards Division is also mandated to develop, maintain and disseminate statistical metadata for the agency's surveys and statistical programs, under the terms of the Policy on Informing Users of Data Quality and Methodology.

The Standards Division is guided by international standard developments, by organizational needs or by program needs. The Methods and Standards Committee is the governance mechanism for strategic direction as well as for monitoring compliance and approving exceptions to the standards. It also receives guidance from a number of other internal committees, such as the Information Management Committee (see Chapter 2.7: Information Management), subject-matter committees, and external expert groups. Figure 1.3.1 illustrates the links within and outside the governance structure of the Standards Division.

The Methods and Standards Committee is a senior management committee whose role is to

• assist and advise on the development and application of statistical standards and metadata within the agency's programs;
• approve the adoption of statistical concepts, variables and classifications as departmental standards;
• approve exemptions to departmental standards where appropriate;
• advise on the development and use of sound statistical methods;
• provide guidance on priorities for statistical research and innovation; and,
• act as the focal point for the review and monitoring of corporate data-quality practices and issues.

This committee reports to the agency's senior management board; i.e., the Executive Management Board (EMB), to ensure global coherence of all management practices. Final decisions about departmental standards are the responsibility of the EMB based on recommendations from the Methods and Standards Committee.

Once standards are approved by this committee, they are implemented by survey programs. The program areas are responsible not only for implementing standards and communicating them to users, but also submitting to the Methods and Standards Committee applications for the declaration as standards of names and definitions of populations, statistical units, concepts, variables or classifications.

The illustration 1.3.1 highlights the governance structure for standards in Statistics Canada:

Figure 1.3.1: External and internal governance with regard to standards and classification at Statistics Canada

While the Policy on Standards and Governance provides control and balance over the application of international standards within the organization, Statistics Canada has developed tools to support the use of standards and to gain in efficiency in applying them.

First, to facilitate the implementation of standards by program areas, the agency developed harmonized content for social surveys. In addition to providing efficiencies to survey managers who re-use questions and programming for applications rather than re-developing them for every survey or survey cycle, the goal of the harmonized content initiative is to increase comparability of survey data across cycles and surveys. Several concepts have been harmonized. These include a set of standardized questions and response categories, pre-packaged metadata to accompany questions, variables and classifications, automatic retrieval in the Questionnaire Development Tool, pre-programmed computer-assisted applications, and processing rules. The use of harmonized content has been made mandatory for all surveys except where an exemption has been granted.

Coding, which is the procedure for classifying the provided data on a questionnaire to standardized statistical classification—has also been facilitated through the use of a programming tool (G-Code^{Endnote 4}) and a coding interface (Classification Coding System). G-Code is an automated system which assigns codes to descriptions. It does so by matching input text descriptions, which could be from a questionnaire, with "parsed" or "standardized" descriptions in a G-Code database. The CCS is an electronic, generalized, interactive coding tool designed to assist with the assignment of numeric codes for the specified classification. It is the key corporate tool for computer-assisted "manual" coding. The system offers codes for industry, occupation, instructional classifications and products. It provides a common set of files and rules and thus ensures consistency in coding. It is designed to be easily customizable to suit individual requirements. Other advantages include (1) the ability to code to more than one classification, (2) the ability to search through the reference file or the classification code, title, description, or notes, and (3) the ability to embed CCS into a coding application and import other data sources, such as questionnaire responses.

The ability to code to more than one classification is particularly important when different international organizations require the statistical agency to report the same information according to different standards. In Canada, for example, industries are classified according to the North American Industry Classification System (NAICS). This standard is used to classify enterprises from the bottom up, according to the value-added of the primary unit activity. As per its name, this classification is used in Canada, the United States and Mexico; it is negotiated and revised trilaterally.

However, the international standard for industry is the United Nations International Standard Industrial Classification (ISIC). The primary focus of the ISIC classification system is the type of activity in which establishments or other statistical entities are engaged. The main criteria used are the following: (1) the nature of the goods and services produced; (2) the uses to which the goods and services are put; and (3) the inputs, the process and the technology of production. The third criterion corresponds to the conceptual basis of NAICS. Canada is therefore required to use ISIC to report its data to international organizations such as the UNSD, the IMF, and the OECD. Although the two classifications are moving towards greater coherence, both classifications must be maintained. To achieve this in an efficient manner, concordance tables have been developed. Canada reports industrial data on an ISIC basis using a concordance between NAICS and ISIC and performs the manual resolution of the non-unique match using the CCS.

3. Strategy and process for the use of international statistical frameworks

To ensure that the application of scientifically-based statistical methodology meets the requirements of international standards and the principles of official statistics, NSOs are encouraged to follow international statistical frameworks.

The most commonly used ones are probably those related to the production of national accounts, balance of payments, consumer and production prices indexes, and population censuses. There is also a variety of other frameworks on specific topics, including environment, health, income and wealth, labour.

An effective and strategic approach to adopting and using international frameworks normally involves a five-step process: (1) participating in international forums or committees where new frameworks are developed or revised; (2) planning; (3) consulting users and stakeholders; (4) developing an implementation and maintenance plan; and (5) pro-actively communicating with users and stakeholders "before and after" the implementation of the new framework.

In fact, the starting point is to stay informed of new standards being developed or revised and to participate in their adoption. This can be done at two levels. First, NSOs can participate or be represented in expert groups consisting of international experts in the domain where best practices and new frameworks are defined. Second, NSOs are usually invited to participate in the final discussions prior to the adoption of a new statistical framework. Active participation in international working groups and meetings is an effective vehicle for the agency to make its voice heard, to raise potential concerns, and to influence, to the extent possible, the end result. By doing so, the agency minimizes the need for adjustment to its specific national context. The discussions and developments for new or improved frameworks are usually led by international organizations, such as the UNSD, the IMF, the various regional United Nations economic commissions,^{Endnote 5} the OECD, the Food and Agriculture Organization (FAO) of the United Nations, the World Health Organization (WHO), and the United Nations Educational, Scientific and Cultural Organization (UNESCO).

When it is time for the agency to consider adopting or implementing a new statistical framework, consultations with users and stakeholders are activated, to ensure it is appropriate to the national context or to discuss appropriate adjustments that can be made. In parallel, designing and developing an implementation and maintenance plan requires agreeing on key milestones, on funds and resources that need to be put in place in advance, and on communications strategies to maintain ongoing dialogue and engagement with key users and stakeholders.

At Statistics Canada, each subject-matter division is responsible for managing and implementing the appropriate international frameworks that are relevant to the data for which the division is responsible. Statistics Canada's System of National Accounts is provided as an example in text box 1.3.2, at the end of this chapter.

4. Strategy and process for the use of reference frameworks for modernization

More recently, there has been growing interest for a standards-based modernization agenda. The High-Level Group for the Modernization of Statistical Production and Services, which reports annually to the Conference of European Statisticians (CES), has reviewed and adopted the standards necessary to support collaboration between agencies. In this respect, the Generic Statistical Business Process Model (GSBPM), the Generic Statistical Information Model (GSIM) and the Common Statistical Production Architecture (CSPA) have been identified as key standards to drive the modernization of official statistics.

The GSBPM^{Endnote 6} is intended to apply to all activities carried out by producers of statistics, at both national and international levels, which result in data outputs. The GSBPM is designed to be independent of data sources; it can therefore be used for the description and quality assessments of processes based on surveys, censuses, administrative records, and other non-statistical or mixed sources.

The rapid adoption of the GSBPM (or closely related national versions thereof) by statistical agencies around the world shows a wide acceptance of the idea that all statistical production can be modelled in terms of different combinations of less than 50 generic sub-processes. By mapping current and planned statistical processes to the GSBPM, it becomes easier to see where synergies can be found between processes, both within and across agencies. This, in turn, helps to identify good practices and improve efficiency within each sub-process. The GSBPM is also increasingly being used as a tool to identify the cost of different parts of the production process, and to inform strategic decisions on resource allocation.

The GSIM complements the GSBPM and provides a link to data and metadata standards such as the Data Documentation Initiative (DDI) and the Statistical Data and Metadata eXchange (SDMX). The GSIM is a reference framework of internationally agreed-upon definitions, attributes and relationships that describes the pieces of information used in the production of official statistics (information objects). This framework enables generic descriptions of definitions, and management and use of data and metadata throughout the statistical production process. The GSIM provides a common language to describe information that supports the whole statistical production process, from the identification of user needs through to the dissemination of statistical products. GSIM is a strategic approach designed to bring together experts from different disciplines (e.g., information technology, statistics, subject-matter areas.) to modernize and streamline the production of official statistics.

The CSPA provides a blueprint for designing and developing statistical production components in a way that makes them much easier to share within and between organizations.

These frameworks, in particular the GSBPM and the GSIM, have been adopted by the Methods and Standards Committee of Statistics Canada as departmental standards. They support the implementation of the Corporate Business Architecture (CBA) principles that dictate Statistics Canada's modernization strategy. For details refer to Chapter 3.1: Corporate Business Architecture.