The purpose of official statistics is to provide everybody quality data, to assist with decision-making, to feed into research and to inform debate. To ensure the credibility of those data, it is essential that the Official Statistical Service, which incorporates INSEE and the Ministerial Statistical Offices, remains independent. To that end, the French Economic Modernisation Act of 4 August 2008 created the French Official Statistics Authority with mission of ensuring “professional independence in the design, production and dissemination of official statistics”, and “compliance with the principles of objectivity, impartiality, relevance and quality applicable to the data produced”.

At the time of the establishment of the Authority, Paul Champsaur, who was in charge of it, analysed the challenges it would face and the means available to it (Champsaur, 2009). Ten years on, it is worth re-evaluating the choices made back then and examining the Authority’s contribution to the development of the Official Statistical Service. Having looked back on its origins and analysed its organisation, we go on to review the various facets of its activity from this perspective.

In 2004, when the European Union needed reliable fiscal statistics, some national fiscal notifications underwent substantial revision in connection with political cycles. Following the observation by the Economic and Financial Affairs Council (ECOFIN, 2 June 2004), on 25 May 2005, the European Commission promulgated a recommendation setting out the principles that the national statistical authorities should abide by. The first aimed to ensure independence from political and other interference, which was to be enshrined in law.

In that regard, the European statisticians who came to audit INSEE’s situation in January 2007 noted the following: “we believe that INSEE compiles and disseminates its statistics in an independent manner without political interference, despite the fact that, in contrast to the general situation of other national statistical institutes in the European Statistical System, it does not have a legal basis to this independence […]. Accordingly, it can come as no surprise that we strongly recommend that INSEE be accorded this independence as soon as practicable”.

Indeed, INSEE is a directorate of the French Ministry for the Economy. Likewise, the Ministerial Statistical Offices (MSOs) contribute to enlighten their ministry’s policy making and this activity may make impartial and independent work difficult. They may therefore be subject to pressures that would lead them, for example, to not publish the official statistics for which they are responsible in a timely manner, since the ministerial officials to whom they report may give priority to other work or may wish to postpone publications that they do not consider politically expedient.

At the same time, INSEE faced controversy in connection with unemployment figures as a result of the divergent trends that appeared between those resulting from the management of jobseekers by the dedicated agency (Pôle Emploi) and those gathered via INSEE’s Labour Force Survey, as defined by the International Labour Organization (ILO). Although being registered by Pôle Emploi does not necessarily mean that a person is unemployed under the strict criteria of the absence of employment over a certain period of time, of actively seeking employment and of being in a position to accept any possible offer, the general public and the press tended to favour the first criterion. Their influence resulted in a temporary stop to the publication of unemployment figures gathered via the Labour Force Survey.

Hence, the lack of a legal basis for the professional independence of official statistics was therefore likely to raise suspicion, even though this was established in practice. The French Economic Modernisation Act of 4 August 2008 remedied this with the creation of the French Official Statistics Authority (ASP). However, the Parliament and the Government considered that the status of INSEE and the Ministerial Statistical Offices, which are integral parts of the French Administration, was compatible with their professional independence, if subject to rigorous supervision by a trustable institution.

Professional independence is essential to the users of statistics. Without it, the statistics produced become dubious in their eyes and efforts to improve their relevance or accuracy become futile. It has therefore always been recognised as essential. As was highlighted by Edmond Malinvaud, “it is of such importance that no opportunity should be missed to make it even more obvious and, on the contrary, it should be an absolute requirement that short-term considerations that may lead to deviations from independence or ethics be discarded” (Champsaur, 2009). He went on to say that “Independence and ethics cannot be decreed; they are built over the long term through the practices of ministerial authorities, management and staff”.

Perhaps less obvious, but no less important, is the fact that professional independence is a condition for access to data. For example, this guarantee was a key factor in the decision made by major retailers to commit to the project aiming to improve the production of the price index through the use of scanner data from shops, securing their confidentiality being at the heart of the testing and subsequent industrialisation of the process (Leclair, 2019). Since the use of new data sources is bound to increase, it should be anticipated that it will require deeper interactions with economic agents for whom the guarantees regarding independence will be decisive.

Thus, the primary role of the ASP is to guarantee this independence by ensuring that any deviations or controversies such as those referred to above are dealt with at an early stage, that the Finance Ministers, for example, do not misuse good short-term figures to which they would have had privileged access under “embargo” procedures. Indeed, equal access by all to official statistics stems from the principle of independence.

The ASP’s legal framework could not be any clearer in this regard: the Authority must check that “the arrangements for the dissemination of publications by the Official Statistical Service comply with the principles of neutrality and fair treatment of users and shall, in particular, ensure that they are disseminated separately from any ministerial communications”. Professional independence therefore does not just refer to independence in the techniques and methods used to compile statistics. It also applies to the conditions under which those statistics are disseminated, since the way in which a statistic is received depends on both its content and the circumstances under which it is released: date of publication, accompanying comments, status of the person releasing the statistic, etc.

However, the ASP’s mission is much broader than this. It also covers the principles of objectivity, impartiality, relevance and quality of the data produced. Indeed, since quality is the greatest asset held by official statistics in a world where information proliferates, most often to the detriment of transparency of public debate and decisions, the Parliament had wanted that the body established to control the independence of official statistics to also be tasked with ensuring compliance with all of the principles that are key for their quality.

Decree No 2009-250 set out the operational procedures of the ASP. First of all, it establishes the independence of its nine-members college: these are appointed by the Chairs of the assemblies, the government, the major inspection and supervisory bodies; they receive no mandate from the institution that appointed them. The term of office of the president, who is appointed by the Council of Ministers by decree, is non-renewable.

The Decree bestows upon the ASP the power to issue general recommendations on the implementation of the principles that official statistics must abide by and observations with regard to any person who is not complying. To this end, it examines incidents that pose a direct threat to professional independence, such as embargo breaches or controversies surrounding certain figures. These incidents are handled such as to ensure the establishment of rules that will ensure that such incidents do not recur.

A wide range of referral channels is provided, including self-referral. The latter is the dominant channel in practice given that, when the Authority is informed of a potential problem or is asked to exercise this option, it examines whether or not to include the relevant item on its agenda. In fact, this is always the case where the request appears to be serious since it must be possible for any case that calls into question the independence, objectivity or quality of official statistics to be submitted to it and examined.

This is then always specifically reported to the public if it appears necessary to make the view of the Authority known immediately. Otherwise it will appear in its annual report. Indeed, it was not considered appropriate to equip the Authority with other arsenal, ultimately risking to undermine its credibility if generating sterile confrontations with administrative and government officials pushing them to challenge their mistakes. Conversely, the ASP makes all its views public. This is a very real power given the importance that the public places on issues associated with the independence and objectivity of statistics.

Finally, the Decree also sets out a number of specific tasks for the ASP to carry out in connection with its mission (Box 1).

The organisational framework for official statistics put in place as a result (Figure 1) is balanced:

The “executive” tasks are therefore clearly separated from the external ones of guidance and supervision, which are distinct. In this respect, this scheme was preferred to the alternative, which would have seen a supervisory body placed within the CNIS. This would have induced a risk of confusion between consultation and supervisory activities. In fact, the missions with which the CNIS and the ASP are tasked are very different: the former provides guidance and the latter ensures the integrity of the system and its independence. However, the need for consistency has not be ignored. It is reflected in particular by the obligation for the ASP to hold discussions with the President of the CNIS and the Director-General of INSEE each year.

Figure 1. Official Statistical Service and Officials Statistics: the Missions of the ASP

Furthermore, the fact that the Chair of the Statistical Confidentiality Committee is a member of the Authority’s college allows for a shared vision of the challenges associated with data protection and access, fields that are undergoing rapid change due to the development of big data and the emergence of new methods for evaluating public policies.

Another key element is that the role of the ASP covers the Official Statistical Service (Service statistique public – SSP) and official statistics in their entirety. The homogenisation of official statistics is also a guiding principle for its activity, in line with the recommendations set out in European regulations. But this has not prevented a simultaneous strengthening of INSEE’s coordination role. In fact, quite the opposite happened, as can be seen from the “MSO Charter” drawn up by its Statistical Coordination Directorate, for example (INSEE, 2019), which gives a large place to the role of the Authority.

Finally, the ASP is responsible for labelling the data produced by public operators that could be of interest to official statistics, which it does by relying on the instruction of the same Official Statistics Quality Label Committee as it does for its surveys. This organisation testifies to the desire for homogeneity of all official statistics with regard to their quality requirements.

The organisation in France appears to be roughly comparable to that of the European Union where, alongside Eurostat, there are also bodies fulfilling functions similar to those of the ASP and the CNIS.

Conversely, France is one of the few countries that has acted on the recommendation of Regulation (EC) No 223/2009 to establish a national body for ensuring the professional independence of those producing European statistics. In addition, in other countries that do make use of such bodies, they had often been created against a backdrop of heightened controversy concerning the degree of trust that can be placed in their statistical institute, while such a situation never existed in France.

Some national statistics officials claim that the creation of a body such as the ASP within their organisation rather than defending the Official Statistical Service, would constitute a gesture of distrust towards it.

Pragmatism or a head-in-the-sand policy? This should be assessed on a case-by-case basis. Nevertheless, it should be stressed that the aim is clearly not to create unfounded suspicion, but to boost trust in statistics. In this respect, any mapping of the risks for statistical systems would highlight the point at which professional independence carries major risks if it is jeopardised. From this perspective, an external supervisor such as the ASP is a valuable asset; at the very least, it constitutes a bulwark against misunderstandings or unnecessary controversy.

It is also a means of speeding up the resolution of unsatisfactory situations, because they inevitably remain so, which are difficult to overcome without external pressure to act. The reform of fishing statistics, required by the ASP following a hearing in October 2014 of the existing Ministerial Statistical Office for Fishing and Aquaculture is a good example of this.

In contrast, the exceptional nature of the ASP should be put into perspective in so far as, in different countries, the directors-general of the statistical institutes are assisted by councils that fulfil apparently similar tasks. Nevertheless, in order to assess the actual role of these councils, it is necessary to examine their missions in detail, distinguishing between whether the statistical institute is more of a compiling body or a direct producer of data; whether they are strictly focused on professional independence or have a broader scope of investigation; and whether their role is primarily supervisory or advisory.

At the same time as the creation of the ASP was being discussed, an authority was established in the United Kingdom, where the statistical institute is an “agency”. Admittedly, it was underlined the limits of the independence of such entities whose budgets remain highly dependent on political power and that independence is not an end in itself. Nevertheless, the idea that this organisation would be the “optimal” one was widely held. With hindsight, the voices that were pointing out that, within organisations, the culture and capacities are as important to performance as the apparent architecture, proved to be far-sighted. Thus, the chosen organisation in France appears pragmatic, preserving the achievements of the French Official Statistical Service, integrated into the administration and committed to excellence.

The French Statistical System is making progress and, while its situation is certainly not perfect, it appears enviable in many respects, first and foremost because its official statistics are produced by top-level statisticians. Besides, as recommended by the ASP, INSEE has now a pluriannual contract setting out objectives and resources, enabling it to manage multi-year projects. Thus, it has in fact taken on the attributes of an “agency” (Ouvrir dans un nouvel ongletBureau and Naves, 2015). However, it benefits from the synergies resulting from its administrative integration, which justifies its research capacity and allows it to offer careers to its agents. Its responsiveness in the face of the COVID-19 pandemic has allowed it to take advantage of the digitisation of its surveys in order to continue measuring economic activity and household purchasing power, and of the use of new telecoms data in order to measure population movements. This responsiveness can be firstly attributed to its internal capabilities.

Since public disclosure represents the main incentive to implement ASP’s recommendations, these must be indisputable. Otherwise, it would be easy to challenge their legitimacy or reduce their scope by creating confusion or arguing various circumstantial elements, ignorance or a lack of intention. In this respect, the Authority’s work is based first and foremost on the precise definition of professional independence set out in European legislation (Box 2).

Furthermore, at national level, the revision of Decree No 2009-250 in 2018 provided an opportunity to introduce greater clarity by stipulating that the Authority shall issue any recommendation that it deems useful “to ensure that the Official Statistical Service complies with the principles of the European Statistics Code of Practice (CoP)”. Compliance with this Code, which sets out the conditions for professional independence, together with the basic principles for statistical quality (objectivity, impartial treatment for all users, reliability, confidentiality, efficiency, sound methodology, documentation of sources and results, etc.) is therefore compulsory (Box 3).

Moreover, the supervision is thus not overly reliant on interpretation, as the “indicators”, which break down its 16 principles, are precise: for example, the stipulation that user satisfaction should be measured at regular intervals, which is directly verifiable; or that statistics originating from different sources should be compared and reconciled. Therefore, in the event of discrepancies between sources, the ASP is entitled to request that the necessary work be carried out to understand the differences.

Most of the time, there are multiple indicators of the CoP to consider together. For example, when faced with controversy over the measurement of a key-statistic (purchasing power, unemployment, poverty, etc.), the Authority will not only consider the indicator which establishes that statistical officers alone are competent to decide upon statistical methods and standards, but also those that require the methodologies to be based on statistical considerations (absence of bias, for example) and compliant with the best international standards.

By not limiting assessment to the first principle of the CoP alone, i.e. independence, the role of the Authority does not appear to be univocal, which contributes to its credibility. Indeed, it is easier, for example, to demand the strictest compliance with the embargo rules when it is clear that ASP also verifies that the dates and times of publication of the statistics have been announced in advance, and that equal access to official statistics for all is its top priority. Besides, the CoP establishes a framework that recognises that the protection of professional independence and the requirement for quality must go hand-in-hand. In this way, the monitoring of good compliance with it triggers progress beyond correcting immediate failures.

In the past, the independence of French official statistics was built around two pillars:

French Official statistics have achieved a well-deserved reputation on the basis of those two pillars. However, this did not prevent any controversy, nor did it prevent that controversy sometimes taking rather a harsh turn, as was the case in 1978 already about unemployment figures. But, in this case, a statement by the Director-General of INSEE confirming that INSEE had not made an error was all it took to put an end to any uncertainty regarding its independence. Nowadays, cracks are appearing in the general trust placed in public institutions. The ASP’s job as guarantor in this regard is therefore to ensure and convince others of their trustworthiness. In order to do so, it primarily refers to the first principle of the CoP (Box 4).

The ASP is called upon, for example, to intervene when statistical publications are delayed in relation to the dates announced. This was the case in 2011, after the President of the Federation of Parents’ Councils (FCPE) notified it that the number of publications produced during the first three quarters of 2011 by the Statistical Office for National Education was less than what could be expected in the light of its work schedule, giving rise to a suspicion of concealment. The ASP’s actions then allowed the backlog to be cleared: the schedule of works and publications for 2012 was made public; the hearing of this MSO finally provided verification that the situation had returned to normal.

More generally, the public display of the schedule of official statistics to come up represents a strong commitment to all users. From 2009, the ASP therefore aimed to expand the list of SSP statistics for which publication is announced in advance beyond the main short-term economic statistics. The objective is to neutralise any possible intervention with regard to the dates of publication, with any delays having to be exceptional, reported and justified. INSEE then followed this up in 2013 by publishing an annual schedule for official statistics online. In late 2017, the Authority noted that all MSOs had released their provisional schedules online. In addition, the ASP requested that INSEE monitor this punctuality for each MSO, which led to the observation of an average punctuality rate of just over 90%, where the delays noted did not call into question the independence of the MSOs.

All users must have equal access to statistical releases at the same time. Any privileged pre-release access granted to an external user must therefore be limited, sufficiently justified, monitored and publicised. Certain economic information is therefore communicated under embargo, in particular to journalists and to Ministerial Offices to enable them to familiarise themselves with the indicators a few hours prior to publication. Since its creation, eight embargo breaches have been detected by the ASP, six of which were governmental in origin. In addition to reminding the absolute necessity to respect the rules if wanting to keep this facility, this also led the ASP to request that INSEE restrict the rules governing the early dissemination of economic outlook indicators to limit the risk of leakage, and to draw up a framework document detailing the embargo rules for all Ministerial Statistical Offices.

Examination of these various cases has also often made it possible to identify more suitable rules, in particular to take account of the development of new media. As a result, following an embargo breach relating to the quarterly national accounts for the 4^th quarter of 2018, the Authority approved INSEE’s proposal to align the times at which the embargo is lifted for all indicators to be released on the same day.

The ASP also intervenes where disputes concerning the reliability of the figures are likely to undermine the credibility of the statistics in the eyes of the public. Recurring themes concern purchasing power, the price index, unemployment, poverty, criminality and immigration, areas in which people are quick to question INSEE’s independence. In this regard, the CoP establishes that, where appropriate, the statistical authorities “comment publicly on statistical issues, including criticisms and misuses of statistics”.

It is therefore down to INSEE (or the Ministerial Statistical Offices) to respond in the first instance. In the event that the controversies become too great, the Authority will take direct action. This was the case in June 2011, when fierce controversy, which was covered widely by the press, was triggered by the statements of the Minister for Internal Security about the academic performance of children of immigrants. INSEE therefore ensured that all journalists seeking information on the subject had access to the official statistics available on this topic and published a press release explaining what could be deduced from them. At the same time, the President of the ASP informed this Minister of the discrepancy between the estimate put forward by him and the orders of magnitude obtained from the existing statistics.

The resolution of critical controversies can sometimes result in the development of new official statistics tools. The creation of the customised inflation simulator and of enriched data on energy costs are good illustrations.

Induced transformations may also be structural, as was the case with crime statistics. Indeed, while in the case of a budget deficit, for example, the public debate questions mainly the appropriateness of the policy measures to be implemented, the controversies surrounding internal security were particularly heated at the time when the ASP was being established. Primarily, they called into question the way in which the phenomena were being measured, with the recurring suspicion that the figures were being manipulated. In the wake of this, the lack of professional independence that was associated with the production of figures provided to the public in this area appeared to be untenable. This gave rise to the creation of a new Ministerial Statistical Office, built with all the necessary guarantees of independence. The understanding that unnecessary controversy could then be avoided gradually superseded any fears that remained within the services regarding this independence.

User confidence relies not only on professional independence, but also on the objectivity, impartiality, relevance and quality of the data produced by the Official Statistical Service. In this respect, the establishment of the ASP came at a time when public mistrust was tending to extend to expertise and anything that referred to scientific standards. Moreover, digitisation has shaken up statistics, as it has the rest of the economy.

Against this new backdrop, public expectations have also become much more diverse. There is an expectation that official statistics will provide detailed data beyond the usual demographic, economic and social fields, such as on crime or sustainable development, for example. In addition, the public has access to a large amount of alternative data to those produced by official statistics in all fields, which means that the latter is compelled to convince the public of their relevance or added value, even where those alternatives are in fact very tenuous or not sufficiently precise.

The ASP is directly involved in the process of enriching official statistics, since it is responsible for labelling as such certain data coming from public agencies and private bodies exercising a public service delegation. In 2010, a procedure was put in place for that, with both objectives of ensuring the quality of these data and of increasing the scope of statistics that can contribute to public debate. The purpose is to improve the information available to users and to meet their demands in a context in which they are becoming increasingly diverse.

Such labelling concerns potential official statistics, not all the statistical output of these operators. For producers whose data are labelled, the labelling is a strategic choice that nevertheless requires professional independence during their compilation, an ongoing improvement process for statistical production and a dissemination policy that takes account of users’ needs.

Since its creation, 14 labels have been granted by the Authority for a fixed duration, generally five years. The Authority’s labelling decisions are published in the Official Journal. They are usually accompanied by recommendations to be implemented, which are followed up on by the Authority, since they constitute a condition for renewal. The areas covered concern the social field (ageing and sickness), employment and unemployment, road safety and house prices.

In 2014, the Authority thus labelled the monthly statistics concerning jobseekers registered with Pôle Emploi. This labelling was accompanied by recommendations, including in particular that the commentary should highlight the trend in recent months and that the low significance of month-to-month variation below a certain threshold be mentioned in the publications. Given the importance of this subject to the public debate, the arrangements for implementing these recommendations were re-evaluated in 2015. Finally, a consensus was reached, according to which publication was to take place quarterly, since it appeared that the commentary on the monthly figures, which had been the focus of much observer attention, was generally of no statistical relevance. As a result, these data are now published with the same frequency as the other measure of unemployment, the unemployment rate from INSEE.

To ensure that the diversity of sources is not a factor in public uncertainty, the Authority also requested that study be carried out to understand the discrepancies in the development of the two indicators. It found that the persistent or significant differences between the two data series resulted from conceptual differences in the measurement of unemployment, such as availability to accept a job, for example, which therefore reflected the diversity of the individual situations falling within the “halo” of unemployment.

In the same vein, the ASP requested an examination of the discrepancies between the different sources involved in the measurement of employment, which gave rise to an action plan to correct the bias identified with regard to youth employment in the Labour Force Survey and to strengthen communication on statistical sources that concern employment.

Although some aspects still need to be better understood, particularly with regard to the volatility of Pôle Emploi’s data, or certain biases in the evaluation of employment in the surveys, for example, the use that can be made of the different sources for labour market diagnostics has therefore progressed.

This overview highlights that the establishment of the Official Statistics Authority has resulted in balanced governance and that the expansion of its mission beyond professional independence alone has proven to be extremely fruitful given the upheaval in the data available to the public and in his expectations.

Misinformation and misinterpretations of facts and statistics are more prevalent than ever before in public discourse and often break through into the political sphere. The audience of the traditional media is shrinking as people increasingly turn to social media, where misinformation is far more likely to be shared than debunked. This places some new and demanding responsibilities on the statistical authorities. These are challenges that must be tackled by all those involved in official statistics. The ASP contributes to this by ensuring compliance with the principles of the European Statistics Code of Practice.

Ouvrir dans un nouvel ongletDécret n° 2009-250 du 3 mars 2009 relatif à l’Autorité de la statistique publique. In: Légifrance website. [online]. Modified on 20 September 2018. [Accessed 20 November 2020].

Ouvrir dans un nouvel ongletLoi n° 2008-776 du 4 août 2008 de modernisation de l’économie. In: Légifrance website. [online]. Modified on 5 July 2019. [Accessed 20 November 2020].

Ouvrir dans un nouvel ongletLoi n° 51-711 du 7 juin 1951 sur l’obligation, la coordination et le secret en matière de statistiques. In: Légifrance website. [online]. Modified on 28 June 2010. [Accessed 20 November 2020].

Ouvrir dans un nouvel ongletRèglement (CE) n° 223/2009 du Parlement européen et du Conseil du 11 mars 2009 relatif aux statistiques européennes. In: Journal officiel de l’Union européenne. [online]. Modified on 25 April 2015. [Accessed 20 November 2020].

Ouvrir dans un nouvel ongletRèglement (UE) 2015/759 du Parlement européen et du Conseil du 29 avril 2015 modifiant le règlement (CE) n° 223/2009 relatif aux statistiques européennes. In : Journal officiel de l’Union européenne. [online]. [Accessed 20 November 2020].