1Alfred Sauvy’s short paper published in 1946 in the first issue of Population has a rather strange title: “Évaluation des besoins de l’immigration française” (Assessment of French immigration needs). In fact, it concerns neither “immigration needs”, nor “French immigration”, but rather the needs of France in terms of immigration. Just after the Liberation, France lacked the necessary manpower to undertake its economic reconstruction, and the idea of recruiting foreign workers, as was done after the First World War, seemed a logical one. Sauvy assumes that this utilitarian attitude to immigration is shared by all; he mentions the major contribution to be made by farmers, construction workers and miners brought in from abroad. But he takes the idea one step further: beyond their contribution to rebuilding the economy, immigrants would also provide a solution, over the longer term, to the country’s demographic imbalance and, more specifically, to the challenge of population ageing. Like many of his contemporaries, Sauvy was haunted by this problem. In 1946, 16% of France’s population was aged over 60. While this proportion was much smaller than that observed today (24%), it was a world record at the time, as Sauvy does not fail to point out. For Sauvy, this “abnormally high proportion” was attributable to the combined effects of secular population decline and the collapse of births during the Great War.
2With considerable economy of means, Sauvy runs a simulation to determine France’s needs in terms of immigration. His stated target is a “stationary population” as defined by Alfred Lotka, namely an “ideal” population in which young people are sufficiently numerous and fertile to ensure generation replacement and to maintain the “structural balance”. The method is extremely simple. Sauvy takes the age structure of the French population on 1 January 1931, as indicated by the census of that year, then determines the number of old people aged 60 or above. He then calculates how many additional people would be needed in the other age groups in order to produce a stationary population pyramid. The difference between the actual situation and the model gives him the order of magnitude of the extra people needed. Why use the 1931 census as a basis? Because it is the last one held in a context of full employment. Sauvy’s goal of economic and demographic reconstruction does not imply taking France back to the conditions prevailing before the war, but aims rather to return to the level that preceded the crisis of the 1930s.
3The conclusion he draws is striking: France will not return to demographic equilibrium unless it brings in some 5,290,000 immigrants, among whom 2,450,000 adults! In a country with just 40 million inhabitants (a contemporary provisional estimate), this meant increasing the general population by 13%. In what timeframe? Sauvy doesn’t say. He compares two stocks without looking at the annual flows required for one to catch up with the other. So his model was of little use to political decision-makers. If we look at estimated net migration to France since 1946, we see that it was not until 2005 – some 60 years later – that the aggregate total reached the 5.3 million mark announced by Sauvy! Did he realize that his figure was grossly exaggerated? It is impossible to say, given the calmly composed tone of his announcement.
4Clearly, this 1946 paper does not cast Sauvy as a precursor of the anti-immigration lobby. He never championed the idea of a closed population summoned to reproduce by its own means. Quite the contrary, his solution to the French ageing problem relied on bringing in young people from abroad. In his eyes, the stationary regime was by no means a synonym of closed reproduction. Later in the text he mentions the need to maintain inflows over the long term in order to restore the demographic balance through immigration. He indicates two ways of doing this: through “a 15% increase in births or a constant inflow of young immigrants” with a combination of both being possible. Here, Sauvy focuses on the migration scenario. His paper lays the foundations for the simulations of “replacement migration” developed later by Didier Blanchet (1989) and UN experts (2001): if we relied solely on the arrival of young migrants, how many would be needed over the years lead an ageing population to replacement level?
5Is Sauvy hostile to settlement migration? Quite the contrary. Clearly, labour migration will be followed by family migration. Most migrants were still men in the 1920s because they were needed to make up for the military losses of 1914-1918. By 1946, the population shortfall concerned both men and women. And Sauvy concludes that no “demographic gains” will be made if immigrants are “condemned to singlehood”. To counteract ageing over the long term, a shift from labour migration to settlement migration was inevitable.
6When Sauvy wrote this paper, he was unaware of how the baby boom was affecting birth numbers. The spectacular upturn was not detected until some months later. It was in the spring of 1947, in the journal Population, that Jean Bourgeois-Pichat published INSEE’s extraordinary report for the year 1946, with its 840,000 births, a leap of 200,000 with respect to prewar levels. Bourgeois-Pichat talks of a 30% increase in “household productivity” (sic). But he is hesitant: is this a post-war catch-up effect linked to the return of prisoners of war and labour deportees, or the start of a lasting trend, as suggested by the concomitant increase in births in several other countries? It was not until 1948 that Sauvy acknowledged that the baby boom was here to stay (Lévy, 1990). In early 1946, when he wrote this paper, he had no knowledge of the change under way, and the term baby boom had yet to be imported into France.
7Sauvy’s arguments are still coloured by the crisis of the 1930s. After the prosperity of the 1920s accompanied by mass immigration, xenophobic pressure led to the expulsion of numerous migrants and to measures prohibiting foreigners from exercising certain trades and professions. This was the only peacetime period in French history when net migration became negative. Sauvy did not believe in the long-term future of a policy that attracted or expelled migrants solely on the basis of economic circumstance. The three decades that followed showed that in France, at least, the so-called “Trente Glorieuses” were a period that combined rising births, migrant inflows and economic growth: the pattern of immigration was not countercyclical, but rather supported economic growth. It was not until the oil shock of 1973, when France put an end to labour immigration, that inflows of immigrants arriving mainly for family reunification or as foreign students became disconnected from the economy while also becoming a long-term phenomenon. As the baby boom came to an end, the major contribution of immigrants to the French population, as predicted by Alfred Sauvy some 70 years ago, became clear for all to see.
