Principle 14: Coherence and Comparability

Statistics are internally coherent and consistent (i.e. arithmetic and accounting identities observed).

Checks are in place to ensure the internal consistency of data published. They apply to both business and household data and are defined in the different stages of the production process.

For example, accounting equalities are verified in the Annual sectoral survey (ESA) and the Annual production survey (EAP), the National Accounts, etc. In the National Accounts, equality between the three approaches (output, demand and income) is also ensured.

In systems that simultaneously produce data at national and local level, such as the Localised employment estimates (ESTEL) on localised annual employment estimates, consistency between the results obtained at the different levels (employment areas, department, region and nation) is verified.

Statistics are comparable over a reasonable period of time.

INSEE strives to guarantee users the continuity of the series it publishes. When changes, e.g. in method or legislative context, lead to breaks in time series, INSEE documents the consequences and, to the extent possible, publishes backcasted series.

In order to ensure the comparability of the series, INSEE carries out comparisons of the data before and after modification of the series, prior to the dissemination of the data, and sometimes as a test of an envisaged change (sources, methods, etc.). Thus, the new method for the imputation of non-responses in the Enquête de fréquentation touristique tourism visitor survey was subject to in-depth validation. Similarly, the project to redesign the Labour force survey had included the systematic backcasting of long series on the labour market, with a specific method developed by a dedicated working group.

When it is not possible to recalculate a series so as to ensure complete comparability of data over time, the documentation associated with the data provides all information useful for the correct assessment of changes.

In addition, INSEE’s Revision Policy specifies the principles and approach followed in the event of changes affecting data already published. For example, the National Accounts are regularly subject to basic changes, during which multiple, sometimes cumbersome (such as changes in classifications) changes can be introduced, leading to a new publication of the entirety of the series. Where backcasting does not allow a perfect correction of the changes, warnings can inform the reader of the weakness of the data.

Statistics are compiled on the basis of common standards with respect to scope, definitions, units and classifications in the different surveys and data sources.

INSEE complies with international accounting standards and applies the European Accounting System, which is a European adaptation of the United Nations-mandated economic accounting system. For statistical operations based on European regulations, compliance with European definitions and standards is compulsory.

European surveys are carried out in compliance with the rules laid down by Eurostat (framework regulations and guidelines). Changes to these rules may lead INSEE to make substantial changes to its production systems. For example, the new framework regulation Ouvrir dans un nouvel ongletEBS (European Business Statistics) has strengthened the obligations relating to the statistical units to be taken into account in business statistics. Accordingly, a new notion of a profiled company is now used in INSEE’s structural business statistics output, which defines the statistical unit as all or part of a company belonging to a group, based on a criterion of independence in making decisions.

Compatibility with international standards is also guaranteed when national needs require specific classifications or processing operations. For example, the French Consumer Price Index (CPI) is used at national level only, but INSEE also produces a Harmonised Index of Consumer Prices (HICP) that complies with Eurostat standards.

The only notable exceptions are the classifications of diplomas/qualifications and occupations, which differ from the European classifications. Nevertheless, on these two points, INSEE carries out specific processing operations that make it possible to produce the statistics expected for international comparability.

In addition, statistical metadata, i.e. concepts, classifications or information that makes it possible to describe a statistical operation, are managed through a repository based on European standards (RMéS).

Statistics from the different data sources and of different periodicity are compared and reconciled.

Comparison between sources is an important criterion in the checking-validation stage for the disseminated data.

Where different data sources produce statistics on the same topic, a significant effort is made by the services to reconcile those data.

Where it is not relevant to reduce discrepancies between sources, for example due to differences in scope, concept or processing, the reasons for these differences are clearly explained. For example, discrepancies between the change in the number of unemployed persons as defined by the ILO (from the Labour force survey carried out by INSEE) and the change in the number of end-of-month jobseekers (DEFM), published by the Pôle emploi employment agency, are analysed and communicated. The detailed presentation of the causes of discrepancies between these two indicators was published in Insee References Employment and earned income, edition 2019.

INSEE also carries out occasional in-depth work seeking to analyse, explain and possibly correct discrepancies between sources, in order to improve the quality of its statistics.

Cross-national comparability of the data is ensured within the European Statistical System through periodical exchanges between the European Statistical System and other statistical systems. Methodological studies are carried out in close co-operation between the Member States and Eurostat.

INSEE complies with international and European recommendations on the compilation of statistics, thereby ensuring that statistics are produced in line with international expectations, allowing transnational comparisons. The Institute actively contributes to the discussions leading to the formation of these recommendations and, in particular, the guidelines disseminated by Eurostat.

As it is ahead in the design and implementation of company profiling methods, INSEE is able to share its experience with other institutes. The work is necessary not only for the compilation of quality European statistics but also for the improvement of the comparability of national structural business statistics.

Another example of the transnational comparability of statistics concerns France’s foreign trade and balance of payments, which are regularly compared with the equivalent statistics of countries trading with France. Differences in mirror statistics are identified and corrected to the extent possible.