You are viewing a javascript disabled version of the site. Please enable Javascript for this site to function properly.
Go to headerGo to navigationGo to searchGo to contentsGo to footer
In content section. Select this link to jump to navigation

A statistical system integrated in the national central administration: The French experience


The independence, impartiality and objectivity of the production of official statistics is the most important issue in official statistics. In order to ensure this independence, does the production of official statistics need to be a separate branch of government or should it remain within the executive branch of government? Insee has always been an executive branch of government, while the other 16 main producers of official statistics in France form part of various French ministries. This does not prevent the French statistical system from being reliable and independent from political interference and considered by the public as such. This independence is mostly de facto rather than de jure, knowing that the legal framework is progressively being strengthened. In the French context, this institutional set up presents more advantages than disadvantages.

For more than 70 years, Insee has been an executive branch of government. Indeed, Insee and the other 16 main producers of official statistics in France form part of various French ministries. Insee itself is part of the ministry of Economy and Finance.

Nevertheless, the French statistical system is reliable and independent from political interference and considered by the public as such. I will describe how our statistical system works and explain how our independence is ensured in practice.

Insee and other national statistical authorities are largely independent de facto

There is no political interference in Insee’s work. Moreover, Insee disseminates its statistics independently, making use of strict embargos and presenting its results to the media on its own. Insee follows the IMF SDDS+ standards for data dissemination, for instance for the pre-release rules.

Let me also underline that in no circumstances would the ministers or their political advisors dare to ask Insee for changes in the contents of these press releases (concerning their figures, or their comments). This is important as Insee has a long experience of “story telling”.

  • There are no political interferences either in management appointment or individual promotions. This applies to Insee as well as to the whole public statistical system.

  • The professional independence of Insee has been enhanced by extending the projected indicator publication calendar to all topics covered by official statistics.

  • To my knowledge, the last political interference occurred in 1990 and concerned the consumer price index (this is also often the main focus in other countries, see for example the case of Argentina). The Government had had a law adopted by Parliament against tobacco and alcohol consumption. This law brought about a strong increase of tobacco prices (which has been going on continuously since then). As several kinds of incomes were indexed on the consumer price index (the minimum wage for example, which is still indexed on the CPI to this day), the government did not want this tobacco price increase to have an inflationist impact through wage indexation. Therefore, the law prohibited Insee to include tobacco on the list of products included in the consumer price index.

Insee protested. The Minister of Finance, to which Insee reports, supported this protest and the law was quickly amended at the beginning of 1991. Instead of removing tobacco from the consumer price index, it was more sensibly decided that the indexation of the minimum wage (and of some other kinds of incomes and social transfers) would be based on a specific consumer price index excluding tobacco. The latter has been calculated by Insee since 1992, in parallel with the usual consumer price index.

Since then, that is over the last 26 years, there has not been any government interference in the work of Insee.

The robustness and credibility of Insee’s statistics are widely recognized in France

According to the latest representative survey conducted in 2016 for Insee by an independent private institute (IPSOS) with 1000 interviewees each in two waves (May and September):

  • Eighty percent of the French have a favorable opinion of Insee;

  • sixty-five percent of respondents believe that the information is independent from political power, up from 60% in 2015; although they do not know precisely the institutional set up of Insee, the French are well aware that it is largely independent de facto;

  • around two thirds of the French trust the figures published by Insee on the French economic and social situation (what is especially striking concerning this last percentage is that when you ask the same question about these economic and social indicators without mentioning that they are produced by Insee, this percentage falls drastically to 42% only);

  • of course, this trust varies according to the indicators; for example, only one third of respondents believe that the unemployment rate statistics produced by Insee accurately reflect the unemployment situation in France.

Overall, this confirms that the Insee “label” is credible for French society and that data produced by Insee, as well as studies, are considered to be produced independently and without interference from political power.

How does it work?

Insee is independent de facto but not independent de jure. Knowing that we started to strengthen the legal framework protecting Insee’s independence.

Although this is not the most important in my view, the creation of an independent authority is a first answer: this watchdog is the Official Statistics Authority, in French, Autorité de la Statistique Publique (ASP). This is the French equivalent of ESGAB (The European Statistical Governance Advisory Board).

The ASP is composed of 9 members who are appointed for 6 years.

The ASP checks that official statistics are prepared in all professional independence and according to the fundamental principles of the “European Statistics Code of Practice”. For example, it can protest publicly if the embargos are not respected by the government. It draws up an annual report on official statistical activity, which is submitted to the French Parliament and made public.

The ASP was created in 2009, that is before the revision in 2015 of regulation 223/2009 relating to the European statistics. Therefore, its mission has to be adapted and strengthened in order to take into account this revision.

Most likely, ASP will have a formal role concerning the appointment or revocation of Insee’s Director General, to ensure first of all that there is no dismissal for political purpose, and secondly that the appointed DG has adequate professional skills.

The main point is that I strongly believe that political interference would be too risky and too costly for the government, at least in the French context. Social control is very strong within Insee itself, through the management and the trade unions. Therefore, political interference would most likely be made public which might provoke a very strong reaction against government in the media and in the French society.

Is this situation precarious? Does it rely too much on the personality of the DG? Of course, the DG is a flagship with regard to independence. However, I wish to underline that the independence of Insee is strongly established and does not depend on the personality of the DG.

  • It would be difficult to appoint a DG with a doubtful reputation. Even if there is no legal provision, the DG has always been a statistician, or an economist with a statistical background.

  • It would also be difficult for a DG to change the decision process by himself at Insee, which is very decentralized in practice.

  • Concerning this last point, one has to mention the strong cohesion and homogeneity of the staff: all managers (called “administrateurs” and “attachés”), who are civil servants, share the same training, culture and ethics.

To sum it up, it is the community of statisticians-economists which provides the best guarantee for these values to prevail. One good example: the Prime Minister and the Minister of Home Affairs have often emphasized that the new statistical service of the ministry of Home Affairs is (according to the regulations) headed by a director who comes from Insee: implicitly, this means that having such a director, in itself, ensures the independence of this statistical service. And the media have also relayed this idea to the general public.


To conclude, should we change the organization of the French statistical system and the status of Insee in order to increase our independence? Well, if I had the opportunity, I would hesitate.

Of course, even if we have this independence de facto, it would be better to consolidate the legal framework, and from this point of view to be some kind of agency, with a Board, instead of being part of government.

But…There is a but. According to me, being an executive branch of government presents several advantages.

  • Being part of the ministry of Economy and Finance may well lead to statistical budgets being more adapted to the needs, and this is also the case for the statistical departments of ministries. This is because the main user (that is the ministry to which a statistical department reports) decides on the budget of its statistical provider.

  • Being part of the ministry of Economy and Finance allows Insee to offer better wages than other government institutions, and therefore to recruit better and more qualified staff.

  • Being a ministerial department provides close links to the most competent users and makes access to better administrative data easier.

  • Last of all, being close to other ministerial departments helps Insee and other national statistical authorities to develop better economic and social analyses of the statistics they disseminate. Indeed, the missions of Insee and statistical departments are a little bit wider than is the case in most other countries: they not only produce statistics but also conduct economic and social studies based on these statistics (for example, Insee publishes a quarterly short term analysis and forecast on the French economy, which has a wide visibility).

Altogether, when I sometimes argue in favor of another status, it is not mainly because of independence concerns, but because of the need of long term perspectives as regards the financing of Insee. When you are a government body, you cannot escape from the various ups and downs of the budget allocation process. Your budget changes every year according to the current law of finance. It can even change many times during the same year. When you are an agency, in the French context, you can have, or at least hope to have, a pluri-annual budget and some visibility and predictability regarding your financial constraints.

In any case, you should keep in mind that whatever the institutional set up, we are financed by public funds. When I receive letters from the public (I receive a few every year), I am often told that we are not independent because we are paid by government. This cannot be changed, whatever the legal framework.

The national legal framework does not provide a very strong independence to Insee’s Director General (DG)

De jure, Insee and the ministerial statistical departments (SSM) are branches of government like any other. Consequently, the Director General (DG) of Insee is appointed by the council of Ministers, like all DGs of ministerial departments in France. She or he can be replaced changed from one day to the next (or one Wednesday to the next at the weekly council of Ministers meetings). The DG of Insee reports to the Minister of Economy and Finance only and there is no real Board of Directors in the usual sense (at Insee, the board has no authority over the DG).