The Programme of the French National Council of the Resistance (Conseil National de la Résistance – CNR) of 15 March 1944 defines the reforms to be introduced upon the liberation of the territory to allow it to rebuild on new foundations and establish an economic and social democracy. One of the pillars on which this reconstruction was based was the development and dissemination of information – a mission that was entrusted to the State.

The creation of the French National Institute for Statistics and Economic Studies (Institut national des statistiques et des études économiques – INSEE) for France and the overseas territories by the Finance Bill of April 1946 aimed to meet this imperative. INSEE serves a dual technical and research purpose, which makes it unique among its European counterparts as they focus purely on statistical production (Desrosières, 1989). With its own School of Applied Statistics, it is tasked with producing data to inform decision-making and providing economic studies. Its name is a topic of much discussion – there is talk of adding “and documentation” to an already very long title to confirm that the Institute is working in the public interest.

By tracing the development of INSEE and the Official Statistical Service from its creation until the mid-1980s (Figure), we aim to clarify the extent to which this statistical system fulfils this mission of producing and disseminating statistical information and how it feeds into the social and political debate. These first forty years are divided into two periods, during which the audience for INSEE’s work and its contribution to democratic debate were growing. Of course, the boundaries between the two periods are blurred, with no abrupt break in the story. As Amossé et al. wrote in their historical analysis of men and women in household statistics: “Our article is therefore presented more like the draft of a historical reading put forward for discussion than an actual work by a statistical historian”. 

Figure - Timeline: INSEE from 1946 to 1987

The first period covers the directorships of Francis-Louis Closon (1946-1961) and Claude Gruson (1962-1967). They set up an institution that gradually moved away from the CNR project, first by serving the State, something that was no doubt at odds with the aspirations of the young recruits from INSEE’s School of Applied Statistics (Ouvrir dans un nouvel ongletBrunaud et al., 2020), but in line with its human and budgetary resources and with a statistical understanding and democratic debate skills considered average among the majority of those in power (Porter, 1995).

Title I of the Decree of June 1946 defines its assignment. INSEE is required:
“to establish, collate and update statistics relating to the State and the movement of people, goods and property, by using, where applicable, information provided to it by the various administrations” (paragraph 1), “to give and update a continuous assessment of the economy” (paragraph 3), “to carry out approximate polling-based censuses on behalf of the public bodies and organisations referred to in (2) of this Article” (paragraph 9), “to coordinate the statistical methods, means and statistical work of public bodies and private organisations subsidised or controlled by the State, to centralise their statistical and economic documentation and to unify statistical classifications and codes”. As this coordinating and centralising function was not defined by any regulation, INSEE was slow to establish itself. Serving governments, all administrations and “all privately owned entities or private individuals” who so request, the scope of its actions is very broad. It covers all areas of “statistics and economic studies”, “provides training for specialist personnel” and “observes changes in the economic situation in Metropolitan France, its overseas territories and abroad” (paragraph 4).

These initial ambitions proved unrealistic with the failure of Mendès France’s Grand Ministry for the National Economy (Grand Ministère de l'Économie Nationale) and the attachment of INSEE to the Under-Secretary of State for the National Economy, then to the Secretary of State for Economic Affairs, which resulted in it suffering a lean period until 1961. The attachment to the Ministry of Finance then provided INSEE with the budgetary means to meet its initial ambitions (Ouvrir dans un nouvel ongletCloson, 1971). This attachment would also provide an opportunity to partially integrate INSEE into the Economic and Financial Studies Office (Service des études économiques et financières – SEEF), heralding the arrival of national accountants at INSEE. As Closon would go on to write a few years after leaving INSEE: “The bulk of the building work had been completed and statistics, in its broadest sense, had become generally accepted. However, it was still lacking a spark and a key addition that would allow it to provide economic calculations, and more active participation in the national accounts, closer ties with the Ministry of Finance and Planning, and the gradual establishment of new teams. The task had not yet been completed in 1961, but enough had been done to ensure that the road ahead was not only clearly signposted, but also open. INSEE had survived the critical period”.

Drawing inspiration from the North American review, Fortunes, and with a team of forecasters inherited from the National Statistical Service, Closon decided to publish an economic outlook called the Note verte, which was sent to a hand-picked selection of around one hundred decision-makers (Touchelay, 1993). Topical issues, such as the slowdown in economic activity or the mechanisms behind what was being described as a recovery in 1952, were considered jointly by specialists who published their reports in this confidential document. From 1953 onwards, all statistical commentary that appeared in the Monthly Statistical Bulletin (Bulletin mensuel de statistiques) was stripped out. It contained only general observations and became truly confidential since its circulation was limited to 150 copies. The list of recipients was drawn up by Closon in agreement with the responsible minister. The Note was Closon’s exclusive domain, and he was particularly demanding of authors.

His requirements took the form of relatively strict guidelines, such as the one from 1957 that criticised the lack of an overview of studies and stressed that “everything must be good as the Institute does not have the luxury of being able to publish mediocre papers”. Closon considered that “too broad a dissemination would have the very undesirable effect of significantly limiting our freedom of expression”, which explains his very strict control over the list of recipients. In October 1952, he turned down a subscription request from the Mulhouse Chamber of Commerce, specifying that this publication “is distributed free of charge and is specifically aimed at members of the government”. Between May and September 1955, the circulation of the Note increased by one hundred copies (Table). Closon reiterated several times that such volumes risked reducing its confidential nature and requests for it to be sent were often met with refusal.

The Note saw international success, since a correspondent from the Sunday Times in London regularly requested copies from August 1959. He had to make do with consulting Étude et conjoncture four times a year following Closon’s refusal. One month later, the Director of Finance and Treasurer of Esso Standard SA France came up against the same refusal. If recipients used the document in a manner that was at odds with its confidential nature, it would no longer be sent. Closon suspended its dissemination to Jean Monnet at the ECSC in Luxembourg after two years following its discovery in the library of the High Authority in 1954. He had no intention of reversing this decision until such time as Monnet confirmed that “this document would be treated as it should be”. The access granted to Pierre Locardel from French newspaper, Le Figaro, was permanently revoked as soon as he used information contained within the Note for an article on new car sales, published in July 1954.

With the Note Verte, the Institute was able to consolidate its role as an economic observatory serving a handful of decision-makers while also building its reputation. The success of the document revealed a real need for quantified information. Given the technical means available, it was a very onerous task involving writing, printing and postal distribution for an institution with few resources and that made its own decisions.

In August 1946, INSEE was entrusted by law with management of the electoral roll. However, this did not resolve the issues surrounding files, since an inspector general was tasked with looking into all associated elements in January 1947. In April, the Directorate-General informed the regional directors that “the principle of the demographic file is not doomed, but there is no need to take it any further at this time”. However, “the register of establishments will not only be maintained, it will also be the focus of the first push when it comes to files” (Ouvrir dans un nouvel ongletChevry, 1948). The following month, in a report sent to the Ministry for the National Economy, Closon suggested that “the files be retained, but with information relating to individuals limited to the greatest possible extent”. INSEE’s policy in this regard was therefore clearly defined, but it was not enough to allay criticism. In an article published by the Revue Défense Nationale in February 1947, Alfred Sauvy, who joined Statistique générale de la France (a general statistics office) in 1922, and was appointed director of the new INED following the Liberation of France, launched a violent verbal attack on these “instruments of oppression”, which is how he viewed INSEE’s files. Closon’s response shut down the debate (Box “Forty years of controversy”). Recognising that this was a “very delicate issue” and stating that he had “thought long and hard about the instrument of oppression that these systems could represent”, the Director-General specified that these files “are a necessity for the modern state and that it is preferable for their management to be entrusted to the Institute known for its independence rather than to a political ministry”. Yet, criticisms continued to appear in the general press until 1950. Closon then sent clarification to the director of Agence France-Presse, which was helping to counter these criticisms, and posed a fundamental question: “should a country only be organised in so far as such organisation will not present any disadvantages in the event of foreign occupation or, on the contrary, should it be equipped to avoid such occupation?” He also remarked that “documents as threatening as INSEE’s inventories, such as the electoral rolls or civil status registers, already existed in 1940 and were not used by the occupiers”. A year after this clarification, the statistical confidentiality law put an end to the criticism of the files.

The law created a joint committee tasked with drawing up the programme of statistical surveys and monitoring their execution – the Statistical Survey Coordination Committee (Comité de Coordination des Enquêtes Statistiques – COCOES) – and mandated statistical confidentiality and the obligation to respond; it established penalties associated with the obligation to respond and created a litigation committee. It introduced the principle of approval granted to the intermediary professional organisations into the surveys, which would allow surveys to be organised into sectors of activity. Its implementing decrees were not published until 15 September 1952, which highlights the lack of interest in statistics among political decision-makers.

The COCOES defines the annual programme of censuses and public-sector statistical surveys, examines their appropriateness with a view to limiting the number of questionnaires sent out and avoiding duplication, finalises the wording of the questionnaires and submits the programme to the responsible minister at INSEE, who publishes the implementing decree. Each operation approved by the Committee requires two signatures: one from the investigating ministry and one from INSEE. Given the genuine influence of this Committee on public quantification, its composition is the subject of multiple negotiations. The Decree of 15 September 1952 (No 52-1059), listing the administrations and professional organisations represented on the Committee, has been amended several times.

That Committee went on to become the permanent contact for employers in matters regarding official statistics. It examines any initiative likely to facilitate cooperation between companies and survey departments, as well as any dispute concerning a particular administrative investigation. In 1953, the general assembly of the CNPF gave a very positive assessment of the reform of official statistics, which promoted “cooperation between administrative departments and professionals” and it observed that the “contributions made by producers to the development of official statistics” could bring them significant benefits.

Following this reform, any statistical operation was the result of an official consultation, which sometimes ended with rejection.

The law of 1951 improved relations between statisticians and professional circles (Touchelay, 2000), but it fell short of completely eliminating reservations about statistical surveys. The attitude of journalist Robert Lazurick, who likened statistics to a “breakdown of public life” and who denounced “this excessive paperwork, which demands an excessively large workforce” in the L’Aurore daily newspaper in 1953 was not unusual. The demographic census conducted in the spring of 1954 even provoked a campaign of genuine hostility. The wording used by Robert Escarpit in the columns of the Le Monde daily newspaper two days prior to the issuing of the questionnaires is telling: “French, by birth or otherwise, the number of children I have, my successive homes, my means of existence – what will they be able to do with this information?”. In 1957, the French Union and Fraternity (Union et fraternité française) parliamentary group, which was closely linked to the Poujadist movement, suggested that the law of June 1951 be repealed. Although its members claimed that they did not dispute the utility of statistics, they justified their approach by clarifying that the law “allows for the use of inquisitorial methods affecting individuals’ professional lives and even their private lives”. They were also of the opinion that statistics “multiplies paperwork”, and went on to conclude that “time wasted on this kind of work is time that could have been spent on production”. The failure of their attempt on 18 April 1957 marked the retreat of the Poujadist movement and the success of the actions aimed at modernising employers.

During this period of construction, Official Statistics stepped up its production, above all to meet the planning needs of political decision-makers. This desire to expand its surveys would initially be hampered by budgetary constraints.

Knowledge of living conditions improved from the first budget des familles (Household Budget) survey, which was conducted among low-income employees in the Paris region in 1946. The operation became regular, the samples were stabilised and the questionnaires were diversified. It was supplemented with more targeted surveys, such as those conducted in 1949 on holidays taken by the French and on the living conditions of elderly people. The first enquête emploi (Labour Force Survey) was organised in 1950, followed by a survey on professional and social mobility in 1952, a household consumption survey in 1953, the Household Budget survey in collaboration with the Consumer Research and Documentation Centre (Centre de recherche et de documentation sur la consommation – CREDOC) in 1956, a tax income survey in 1958, etc. The number of surveys multiplied to meet the demands of social statistics and to document the issue of inequality. The involvement of CREDOC made it possible to improve statistical knowledge regarding housing in France. The 1946 census went a step further by asking questions about comfort. The 1954 and 1962 censuses used an expanded housing sheet and provided additional data. In 1955, INSEE and CREDOC surveyed the dwellings of non-agricultural households and then of all households in 1961. Indeed, there was a boom in surveys looking at living conditions as a result of the priority investment programme of the Fifth Republic. Therefore, during the 1960s, INSEE was interested in holidays, leisure activities, transport, health and time-budget and conducted the first vocational training survey. The surveys were designed to feed into planning exercises and, more specifically, employment forecasts, and then they went on to explore other fields, such as social and professional mobility, migration routes, etc. (Ouvrir dans un nouvel ongletMonso, Thévenot, 2010). The growth of surveys, the increase in the budget and the arrival of national accountants at INSEE (Desrosières, 1998) are all testament to the need for planning. According to Jacques Desabie, statisticians also wanted to analyse “social behaviours and [to study] living conditions in the broad sense” (Insee, 1996, p.83).

This was also the period during which national accountants arrived at INSEE. In the early 1960s, planning was at its peak and the work of national accountants was being used as a basis for forecasts. These national accounts were produced by the SEEF under the management of Gruson and were intended for the Treasury. The SEEF, made up of around fifty people, had a limited capacity when it came to producing the statistical information necessary for its work. The arrival of Gruson in the capacity of Director-General, and some of the SEEF’s teams at INSEE in 1961 allowed it to overcome these difficulties. However, their integration into INSEE was easier said than done, since the statisticians and national accountants needed to find the right modus operandi to allow them to work together. This complementarity is necessary to produce “information intended not only to describe and explain situations and developments, but also to inform actions that are never purely technical, but also social and political” (Ouvrir dans un nouvel ongletGruson, 1971). This would result in a reorganisation of INSEE and the creation of a directorate for economic reports (direction des synthèses économiques) in 1962. Looking beyond the questions concerning the behaviours of households and social groups that statisticians are most concerned with, it is a case of producing broad economic aggregates, better monitoring consumption by large groups and being able to inform economic regulatory and planning decisions (Ouvrir dans un nouvel ongletDesrosières et al., 1976).

During this period, INSEE’s experiments included initial tests of quarterly accounts and the regionalisation of accounts, and the development of complex macroeconomic models.

This was also a period of transformation for industrial statistics, with the additional implementation of industry surveys, sector-based surveys based on the company as a statistical unit, the first experimental annual company surveys, in particular with data collected by the Administration for these new systems, in spite of opposition from the CNPF. This period also saw the normalisation of the use of tax data such as industrial and commercial profits, the standardisation of which was established by the Decree of 28 October 1965, in line with the general accounting plan of 1957, and which were officially reported by the Directorate-General for Taxes (Direction générale des impôts) to INSEE from 1967 onwards.

By the end of Gruson’s term of office, INSEE, and Official Statistics in general, had expanded its range of statistical information, but was still far removed from the demographic debate and the public. 1965 saw the creation of the Liaison Committee between INSEE, the economic administrations and professional organisations (trade unions and social organisations), within the framework of the introduction of an income policy; however, the ambitious task of democratising the statistical information system was still outstanding (Bardet, 2000).

When Gruson took stock of the first twenty-five years of INSEE, he highlighted the scientific side of the Institute and the fact that it had moved beyond a “harvesting and gathering” activity: “As a general rule, INSEE only reaps what it sows – what it sowed not at the beginning of the season, but several years in advance. That is why the establishment of the economic information system is a typical example of an activity in which order is maintained purely through planning, i.e. provided each decision is modelled on a precise vision of the long-term future” (Ouvrir dans un nouvel ongletGruson, 1971). Jean Ripert made a similar observation when talking about users: “The statistician often has to remind users of the time scales, which are necessarily long and always counted in years, that are needed for the introduction of new statistical instruments” (Ouvrir dans un nouvel ongletRipert, 1971). In view of the challenges that this presents for the Institute, he insisted on forging a relationship between the producers and users of statistics in order to improve the system: “it is important that a constructive and trustful dialogue be established between the producers and users of statistical information in order to define priorities” (Ouvrir dans un nouvel ongletRipert, 1971). The following few years would therefore involve consolidation of the work completed by an institution with a recognised role, through the creation of spaces for discussions that would guide the statistical programme and respond to social demand more effectively.

The political context of the 1960s helped to open INSEE up to new audiences. First, there was strong demand for economic and social information at the local level. Created in the early 1960s, the Interministerial Delegation for Regional Planning and Regional Attractiveness (Délégation interministérielle à l’aménagement du territoire et à l’attractivité du territoire – DATAR) wanted to guide INSEE towards the creation of Regional and Economic Observatories (Observatoires économiques et régionaux – OER) in order to meet the growing demand for local information. In the mid-1960s, a reform resulted in the creation of administrative regions, and “regional missions” (missions régionales) that included representatives from the administrative, economic and academic fields, among others. In addition, at INSEE, discussions were being held regarding its missions and its role. In June 1967, these gave rise to a conference on economic information, held in Villemetrie. This was followed by the “events” of May 1968, which was a time of “permanent general assemblies” within the Institute, where the following questions were being asked: “How is the work experienced and organised? What are statistics used for? Who are we producing them for?” (Insee, 1996 p.102). Lastly, in 1969, there was a referendum on the creation of regions and the reform of the Senate. A great many elements affected the participation and the role of INSEE in the production and dissemination of local economic and social information (Bardet, 2000).

DATAR aimed to create places in each region that bring together all of the producers and users of statistical information in order to better coordinate the dissemination of this information and to “introduce pluralism into its production”. This involves meeting users, supplementing INSEE’s data with other local sources and passing on requests for statistical information.

INSEE’s position on its involvement, its participation and the scope of the information it provides began to change. As a result, the first OER were created in Lille and Marseille in 1967, followed by others until the mid-1970s. However, the initial desire for full openness based on a regional statistical programme led by a joint committee bringing together producers and users did not come to fruition. Nevertheless, the establishment of a place for discussions between the users and producers of statistical information was not ruled out at the national level.

As part of the reorganisation of INSEE in the early 1970s, the OERs were attached to INSEE’s regional directorate (directions régionales – DR) and contributed to the “consideration of the regional and local dimension when compiling national statistics”, something that Edmond Malinvaud had been striving for since the mid-1970s (Insee, 1996). In addition, the number of category A managers assigned to studies doubled between the late 1970s and the early 1980s in order to meet this need for regional and local studies.

Considerations concerning the local role of INSEE were accompanied by questions regarding economic information at national level, with the dual challenge of organising the statistical system and the subjects to be addressed or better addressed. These questions were of course not absent from the Villemetrie conference and the debates held in May 1968. Therefore, during his general policy speech to the National Assembly on 16 September 1969, Prime Minister Jacques Chaban-Delmas called for the establishment of a “new society” in which the economic information policy would be “reconsidered” (Bardet, 2000). Adopting one of the proposals made at the Villemetrie conference, the government established an economic information committee as part of preparations for the Sixth Republic. Conscious of the importance of this committee to INSEE, Ripert ensured that INSEE managers were included among its members. The Rapporteur General of this committee was therefore the Deputy Director of Economic Studies and Reports at INSEE, Philippe Berthet, and several of Ripert’s employees were involved in the committee’s working groups (Berthet, 1971). The areas for consideration selected were: taking account of the needs of different users of economic information; transparency through the establishment of dissemination cells in production centres; training; and increasing the efficiency of the economic and social information system. In fact, calls for the pluralism and democratisation of the Official Statistical Service based on internal and external demands were heard and confirmed in the work of the committee: “It is essential that each member of the social body, each group and each community can play its role in social life, and that requires in particular that we ensure that every member has sufficient economic and social information to allow them to actually perform that role” (Berthet, 1971). One of the flagship recommendations made by the committee was the creation of a National Council for Statistics (Conseil national de la statistique) made up of representatives from administrative bodies, professional organisations, trade unions, researchers and producers of statistics. Its purpose was to encourage discussions between the users and producers of statistical information prior to the introduction of statistical operations, nomenclatures, registers, etc. to ensure that social demand would be met and to avoid duplication between the operations, thereby reducing the number of surveys. This recommendation led to the creation of the CNS in 1972. However, in the early 1980s, the social partners criticised the functioning of this body, as they felt that they were not being listened to. They asserted “that the CNS is a form of manipulation that allows INSEE to implicitly assume the right to decide on its work programme on the pretext of differences between partners, whereas the very purpose of the CNS is to allow the social partners to have a say in this programme” (Spenlehauer, 1998). André Vanoli, former Secretary General of the CNS and then of the CNIS noted a “dual demand for statistical information to be made available to all social players, as opposed to just government and professional organisations, and for social partners to be involved in some of the mechanisms determining the development and content of this information” (Vanoli, 1989). These reactions, combined with the change in the political majority, led to the creation of a working group on the reform of the CNS, which resulted in the creation of the CNIS by the Decree of 1984. The changes included the creation of a Bureau to prepare the studies, composed of five representatives from the employee trade union confederations, five representatives from organisations representing companies, three government representatives [the Director-General of INSEE, the Governor of the Bank of France (Banque de France – BdF) and the High Commissioner for Planning] and two members elected by the other categories of members of the CNIS. This transformation made it possible to specify that the consultation relates to all official statistics, regardless of whether or not the producer belongs to the Official Statistical Service, from production to dissemination of the data. The topics covered were also expanded to include, for example, the emergence of data concerning the financial system, which explains the presence of a representative from the BdF at the Bureau. The Statistical Confidentiality Committee (comité du secret statistique) was established to grant researchers and public and parapublic bodies access to detailed business data. Links were also established between the CNIS and the National Commission for Information Technology and Civil Liberties (Commission nationale de l’informatique et des libertés – CNIL), which was created in 1978 under the Data Processing, Data Files and Individual Liberties law.

One of the areas in need of work that was raised by the economic information committee for the Sixth Republic was the improvement of the efficiency of the social and economic information system. Ripert decided to take the initiative and launched an audit of INSEE, conducted by McKinsey, prior to the drafting of a summary of the work undertaken by this committee. The recommendations made during the audit echoed the proposals already made in the conclusions of the Villemetrie conference and by the economic information committee. Beyond the need to be attentive to social demand, it became necessary to ensure the broad dissemination of the results produced.

Referring to the core business of INSEE – the production, analysis and dissemination of statistics – the audit suggested restructuring to better allow it to conduct its missions. The production department returned to the Directorate-General and production division managers were introduced into the regional directorates. The idea was to better coordinate production work within the regional directorates and to provide room for manoeuvre that can be used to work on studies. However, the most symbolic action was the creation of a dissemination department. This provided an opportunity to consolidate and continue the work to reform the publications and to establish the role of the press office created in the late 1960s. The journals became specialised in order to take better account of the diverse nature of the target audiences. There was also a desire for some journals that were not aimed solely at economic and social information specialists, but that would be “accessible to any cultured person with an interest in the economy” (INSEE, 1996). Soon thereafter, Économie et statistique, which first appeared in 1969, gained a new editor-in-chief tasked with rewriting articles to make them accessible to a wider audience. The first issue of Données sociales was published towards the end of 1973 in response to the strong demand for social statistics. This publication proved a real editorial success, achieving sales in excess of 10,000 copies in 1987. In view of this success, INSEE created a social studies division to support the editorial team and to use and enhance the numerous sources provided by the Household Population (Population ménages) department.

Another editorial success came in the form of the French Economy Tables (Tableaux de l’économie française – TEF): in 1976, an overhaul of this publication gave it a wide audience and it sold in excess of more than 24,000 copies in 1987. The surge in national publications was also mirrored at the regional level. The templates for regional journals were redesigned, the authors trained in editorial techniques and the TEF were broken down at the regional level and also achieved editorial success. And, in 1987, following an initiative by the Regional and Economic Observatory in Aquitaine, INSEE created a national videotex for the public: this is how 3615-Insee was born.

The development of these publications and the press office made INSEE’s work more visible. The press would include phrases such as “INSEE says that...” rather than “according to official statistics…”. This also led to the introduction of the concept of an embargo setting the time at which public dissemination takes place, compelling certain stakeholders receiving the information in advance (news agencies, authorities) to comply with the specified time for public dissemination to ensure that everybody receives the information at the same time. This is now a key criterion for independence under the European Statistics Code of Practice.

During the two periods, foundation and opening up, INSEE succeeded in providing statistics that best meet the information needs of decision-makers and a wider audience. The transition from manual data processing to digital data processing has shortened the time between the survey taking place and its results being published. Symbolic publications such as Données sociales aimed at undergraduates and high school students studying humanities and social sciences marked the opening of the Institute initially intended for decision-makers to a much wider audience and for democratic debate. A new period of democratisation of economic and social information, competition between the public and private sectors and of dispersal of demand was now emerging against a backdrop of internationalisation, particularly at European level, with the European Commission growing in importance in connection with national statistical programmes. But that’s a story for another day...