(19) Life and Death in a Changing City: Mortality Patterns and Inequalities in Paris, 1890–1949 (with C. Torres, L. Kesztenbaum and F. Meslé). Forthcoming in Population and Development Review.
Abstract: Over the past two centuries, European metropolises transformed from urban graveyards to healthy cocoons. Despite its critical role in the history of longevity, this evolution remains underexplored due to limited detailed data. We leverage a newly harmonized dataset of high-quality, cause-specific mortality records to examine the rapid increase in life expectancy in Paris during the first half of the 20th century. During this period, Paris' life expectancy converged with that of the rest of France. Using a consistent cause-of-death classification from 1890 to 1949, we show that infectious diseases—particularly tuberculosis and respiratory infections—accounted for 60–70% in the 25-year gain in life expectancy at age 1, with mortality reductions among children and young adults driving most of these improvements. We also document a profound shift in the causes behind the sex gap in life expectancy, from infectious diseases to cardiovascular and cancer-related causes. Furthermore, our analysis of cause-specific and socio-economic data for Paris's 80 neighborhoods reveals pronounced social gradients in infectious mortality, with the poorest areas suffering the highest rates. These disparities widened during the early phase of the tuberculosis mortality decline, favoring wealthier areas, but narrowed in subsequent phases as public health improvements reached poorer neighborhoods.
(18) Contextualizing the Global Burden of COVID-19 Pandemic: A Historical and Geographical Exploration of Excess Mortality in France, 1901-2021 (with C.-G. Camarda), Population and Development Review, 2025. Link ; Code and data.
Media Coverage: Population Europe ; The Conversation ; N-IUSSP ; 20 minutes ; Science et Vie
Abstract: In light of the unprecedented global impact of COVID-19, comparing its demographic and spatial dimensions with past longevity crises is crucial for identifying core drivers and providing guidance for tailored health policies. This study offers the first analysis of this type, placing the recent pandemic in a historical and geographical context spanning the past 120 years and comparing its impact in 90 French regions with that of the 1911 and 2003 heatwaves and the Spanish and Hong Kong flu. Using a robust statistical model, we computed a standard excess mortality measure enabling comparison across both space and time and with the estimated effects of COVID-19 in European regions. Our results reveal that the COVID-19 burden was, on average, 90% lower than the Spanish Flu and 75% lower than the 1911 heatwave, but four times higher than the 2003 heatwave. We also reveal that the heterogeneity of the COVID-19 burden across French regions is high compared to the Spanish and the Hong Kong flu but in line with the 2003 heatwave. COVID-19's substantial spatial heterogeneity and clear spatial clusters could be the result of stringent containment measures that prevented the pandemic from spreading across the French territory, contrary to past pandemics.
(17) Unpacking occupational and sex divides to understand the moderate progress in life expectancy in recent years (France, 2010’s) (with O. Merville C.-G. Camarda and E. Cambois), International Journal for Equity in Health, 2024. Link ; Code and data.
Media coverage: médecine/sciences
Abstract: The growth in life expectancy (LE) slows down recently in several high-income countries. Among the underlying dynamics, uneven progress in LE across social groups has been pointed out. However, these dynamics has not been extensively studied, partly due to data limitations. In this paper, we utilize the recent change in French census mortality follow-up data (EDP) and apply P-spline models to explore this area. We estimate LEs across five occupational classes (OCs) and indicators of lifespan heterogeneity (edagger) within these OCs, for seven triennial periods (2011–2013 to 2017–2019). First, we found a similar ranking of OCs along the LE gradient over time and across genders, from manual workers to higher-level OCs. Noteworthy, the lowest LE in women overlaps with the highest one in men. Second, we observe varying progress of LEs. In women, LE increases in higher-level OCs meanwhile it levels off in manual workers, so that the OCs gap widens (up to 3.4 years in 2017–2019). Conversely, in men LE stalls in higher-level OCs and increases in manual workers so that the gap, which is much larger than in women (+5.7 years in 2017–2019), is tending to narrow. Finally, the lifespan homogenizes in OCs only when LE is low. Overall, the limited LE progress in France results from LE stalling in the middle of the gender-OC gradient, though LE increases at both ends. At the lower end, LE progress and lifespan homogenization suggest that laggards benefit recently improvements achieved earlier in other OCs. At the upper end, LE progress may come from a vanguard group within higher-lever OC, benefiting new sources of improvements. These findings underscore the need for further research to explore the diverse mortality dynamics coexisting in the current health landscape.
Abstract: The conventional approach to gauging mortality convergence in Europe relies on life expectancy estimates (e0) at the national level. However, mortality can differ within countries significantly. To better apprehend whether Europe's mortality patterns have been converging or diverging over recent decades, we must shift our focus to regional mortality data. Using data from statistical offices, we present annual e0 estimates for 420 regions in 16 EU countries from 1995 to 2019. In our empirical analysis, we examined whether regions with initially high mortality levels caught up with low-mortality regions, and we investigated changes in the standard deviation of Europe's regional distribution over time. Indeed, mortality variation has generally decreased from 1995 to 2019 due to larger gains in e0 for regions with initially high mortality levels. The convergence phase took place mostly during the first half of the entire time period analyzed. Over more recent periods, however, we observe more heterogeneity in the development of e0. Some advantaged regions realized further gains in e0, even as e0 improvements slowed for more disadvantaged regions. In conclusion, our analysis underscores the importance of addressing widening health inequalities. Policies should target disadvantaged regions to retard mortality divergence across Europe.
(15) An Innovative Visual Approach to Monitor Simultaneously Two Dimensions of Progress in Longevity: An Application to French and German Regions (with P. Grigoriev, S. Klüsener, F. Meslé, M. Mühlichen), Population Health Metrics, 2024. Link ; Replication kit.
Background Both enhancing life expectancy as well as diminishing inequalities in lifespan among social groups represent significant goals for public policy. However, there is a lack of methodological tools to simultaneously monitor progress in both dimensions. Additionally, there is a consensus that absolute and relative inequalities in lifespan must be scrutinized together.
Methods We introduce a novel graphical representation that combines national mortality rates with social inequalities, considering both absolute and relative measures. We use French and German data stratified by place of residence to illustrate this representation.
Results For all-age mortality we detect for France a rather continuous pace of decline in both mortality levels and variation. In Germany, substantial progress was made in the 1990s, which was mostly driven by convergence between eastern and western Germany, followed by a period with less progress. Age-specific analyses reveal for Germany some worrying regional divergence trends at ages 35-74 in recent years. This is particularly pronounced among women.
Conclusion Our novel visual approach allows evaluating easily the dynamics of societal progress in terms of longevity, and facilitates meaningful comparisons between populations, even when their current mortality rates differ. The methods we employ can be reproduced easily in any country with longitudinal mortality data stratified by relevant socio-economic information or regions. It is both useful for scientific analyses as well as policy advice.
(14) Method for redistributing misclassified causes of death (with P. Grigoriev & E. Perdrix), Population Studies, 2024. Link ; Replication kit.
Abstract: The analysis of causes of death is crucial for monitoring the epidemiologic situation and developing adequate policy responses. Yet, various methodological challenges complicate such analysis. The comparability of cause-specific mortality data depends on the proportion of misclassified deaths. To eliminate the bias due to varying proportion of such causes over time and between populations, one has to reassign ill-defined causes to other categories. In this research note, we thoroughly document and provide tools for the practical implementation of a regression-based method for redistributing misclassified causes of death proposed by a French demographer, Sully Ledermann. The method relies on the sub-national cause-specific mortality data to estimate unbiased death rates at both the national and sub-national levels. We refine the Ledermann’s method through elaborating on its mathematical properties, making additional adjustments, and evaluating the performance of the approach through simulations. To illustrate the practical application of the method, we rely on the French regional data on causes of death for the period 1979–2016. To ensure the reproducibility of the obtained results and enable potential users to adapt the program to their data, we provide the R code performing all calculations.
(13) Spatial Disparities in the Mortality Burden of the Covid-19 Pandemic across 569 European Regions (2020-2021) (with P. Grigoriev, M. Sauerberg, I. Alliger, M. Mühlichen, C.-G. Camarda). Nature Communications, 2024. Link ; Code and data ; Website app.
Media Coverage: The Conversation ; N-IUSSP ; Die Welt ; Frankfurter Allgemeine ; Podcast EU.radio ; médecine/sciences
Abstract: This article presents a detailed analysis of the global mortality burden of the COVID-19 pandemic across 569 regions in 25 European countries. We produce sex-specific excess mortality and present our results using Age-Standardised Years of Life Lost (ASYLL) in 2020 and 2021, as well as the cumulative impact over the two pandemic years. Employing a robust forecasting approach that considers regional diversity and provides confidence intervals, we find notable losses in 362 regions in 2020 (440 regions in 2021). Conversely, only seven regions experienced gains in 2020 (four regions in 2021). Most importantly, we estimate that eight regions suffered losses exceeding 20 years of life per 1,000 population in 2020, whereas this number increased to 75 regions in 2021. The contiguity of the regions investigated in our study also reveals the changing geographical patterns of the pandemic. While the highest excess mortality values were concentrated in the early COVID-19 outbreak areas during the initial pandemic year, a clear East-West gradient appeared in 2021, with regions of Slovakia, Hungary, and Latvia experiencing the highest losses. This research underscores the importance of regional analyses for a nuanced comprehension of the pandemic's impact.
(12) Spatial variation in excess mortality across Europe: a cross-sectional study of 561 regions in 21 countries (with P. Grigoriev, M. Sauerberg, I. Alliger, M. Mühlichen, C.-G. Camarda). Journal of Epidemiology and Global Health, 2024. Link ; Code and data ; Website app.
Media Coverage : The Conversation ; N-IUSSP
Objective: To measure the burden of the COVID-19 pandemic in 2020 at the subnational level by estimating excess mortality, defined as the increase in all-cause mortality relative to an expected baseline mortality level.
Design: Statistical and demographic analyses of regional all-cause mortality data.
Setting: The vital statistics systems of 21 European countries.
Participants: The entire population of 561 spatial units in 21 European countries.
Main Outcome Measures: Losses of life expectancy at ages 0 and 60 for males and females.
Results: We found evidence of a loss in life expectancy in 391 regions, while only three regions exhibit notable gains in life expectancy in 2020. For 12 regions, losses of life expectancy amounted to more than 2 years, and three regions showed losses greater than 3 years. We highlight geographic clusters of high mortality in Northern Italia, Spain and Poland, while clusters of low mortality were found in Western France, Germany/Denmark and Norway/Sweden.
Conclusions: Regional differences of loss of life expectancy are impressive, ranging from a loss of more than 4 years to a gain of 8 months. These findings provide a strong rationale for regional analysis, as national estimates hide significant regional disparities.
(11) Estimating Subnational Excess Mortality in Times of Pandemic. An application to French departements in 2020 (with C.-G. Camarda), PLoS ONE, 2024. Link ; Replication kit.
Abstract: The Covid-19 pandemic did not affect sub-national regions in a uniform way. Knowledge of the impact of the pandemic on mortality at the local level is therefore an important issue for better assessing its burden. Vital statistics are now available for an increasing number of countries for 2020 and 2021, and allow the calculation of sub-national excess mortality. However, this calculation faces two important methodological challenges: (1) it requires appropriate mortality projection models; (2) small populations implies important uncertainty in the estimates, commonly neglected. We address both issues by adopting a method to forecast mortality at sub-national level and by incorporating uncertainty in the computation of mortality measures. We illustrate our approach to French départements (NUTS 3, 95 geographical units) and produce estimates for 2020 and both sexes. Nonetheless, the proposed approach is so flexibility to allow estimation of excess mortality during Covid-19 in most demographic scenarios as well as for past pandemics.
(10) Employees in higher-level occupations outlive manual workers. How long do they live in retirement or (in) activity? (with C.-G. Camarda, E. Cambois and O. Merville), Populations & Societies, 2023. Link.
Media Coverage: The Conversation ; Slate ; Sud-Ouest ; Atlantico.
Abstract: People in higher-level occupations have a longer life expectancy than manual workers but retire later, on average. Which occupational categories spend the most years in retirement? Are any gender differences observed? How do career histories and the pension system affect time spent in retirement? To answer these questions, Florian Bonnet, Carlo-Giovanni Camarda, Emmanuelle Cambois, and Ophélie Merville measure life expectancy across occupational categories at different ages, particularly around retirement ages.
Abstract: Many recent studies show that Europe has had a lower mortality inequality for most ages than the United States over the last thirty years. However, the evolution of the income gradient in mortality all along the twentieth century remains poorly understood. This article uses a unique dataset that gives the annual lifetables and fiscal income for the 90 administrative regions of mainland France from 1922 to 2020. The income gradients in mortality are computed across regions using a traditional method with calendar ages and, alternatively, with mortality milestones to control for the increase in life expectancy over time. The study reveals a systematic reversal of the gradient that occurred around the 1970s for both sexes and all ages or mortality groups when calculated at an aggregated level. Inequality in mortality amongst the oldest age groups has however returned to a level observed at least ten years earlier because of Covid-19, even after controlling for mortality improvements over the period.
(8) Identifying age- and sex-specific COVID-19 mortality trends over time in six countries (with C. Torres, J. García et al.), International Journal of Infectious Diseases, 2023. Link.
Objective: The COVID-19 pandemic is characterized by successive waves that each developed differently over time and through space. We aim to provide an in-depth analysis of the evolution of COVID-19 mortality during 2020 and 2021 in a selection of countries.
Methods: We focus on five European countries and the United States. Using standardised and age-specific mortality rates, we address variations in COVID-19 mortality within and between countries, as well as demographic characteristics and seasonality patterns.
Results: Our results highlight periods of acceleration and deceleration in the pace of COVID-19 mortality, with substantial differences across countries. Periods of stabilization were identified during summer (especially in 2020) among the European countries analysed, but not in the United States. The latter stands out as the study population with the highest COVID-19 mortality at young ages. In general, COVID-19 mortality is highest at old ages, particularly during winter. Compared to women, men have higher COVID-19 mortality rates at most ages and in most seasons.
Conclusions: There is seasonality in COVID-19 mortality for both sexes at all ages, characterized by higher rates during winter. In 2021, the highest COVID-19 mortality rates continued to be observed at ages 75+, despite vaccinations having specifically targeted those ages.
(7) Influenza Mortality in French Regions after the Hong-Kong Flu Pandemic (with H. d'Albis and J. Thuilliez), Demographic Research, 2022. Link.
Background: Influenza mortality has dramatically decreased in France since the 1950s. Annual death rates peaked during two pandemics: the Asian flu (1956–1957) and the Hong Kong flu (1969–1970).
Objective: This study’s objective is to evaluate whether the second pandemic created a structural change in the dynamics of influenza mortality in France.
Methods: We employ a new database on influenza mortality since 1950 at the subnational level (90 geographic areas) to estimate statistical models to find out whether a structural change happened and to explain the differences in mortality rates across geographic areas. Influenza mortality increased between 1950 and 1969 and decreased from 1970 onward.
Conclusions: The Hong Kong flu is identified as the event of a structural break. After the break, geographical differences are less explained by regional characteristics such as income, density, or aging ratio. The Hong Kong flu was found to be associated with a major change in influenza mortality in France. Change in health practices and policies induced a decline in mortality that started in 1970, just after the pandemic. The health benefits are notably important for senior citizens and for the poorest regions.
Contribution: We employ a new database on influenza mortality since 1950 at the subnational level (90 geographic areas). Change in health practices and policies induced a decline in mortality that started in 1970, just after the pandemics. The health benefits are notably important for senior citizens and for the poorest regions.
(6) The demography of COVID-19 deaths database, a gateway to well-documented international data (with A. Caporali et al.), Scientific Data, 2022. Link.
Abstract: National authorities publish COVID-19 death counts, which are extensively re-circulated and compared; but data are generally poorly sourced and documented. Academics and stakeholders need tools to assess data quality and to track data-related discrepancies for comparability over time or across countries. “The Demography of COVID-19 Deaths” database aims at bridging this gap. It provides COVID-19 death counts along with associated documentation, which includes the exact data sources and points out issues of quality and coverage of the data. The database — launched in April 2020 and continuously updated — contains daily cumulative death counts attributable to COVID-19 broken down by sex and age, place and date of occurrence of the death. Data and metadata undergo quality control checks prior to online release. As of mid-December 2021, it covers 21 countries in Europe and beyond. It is open access at a bilingual (English and French) website with content intended for expert users and non-specialists. Data and metadata are available for each country separately and pooled over all countries.
(5) Income Inequality across French Departments over the Last 100 Years (with H. d'Albis & A. Sotura), Economics and Statistics, 2021. Link ; Data.
Media Coverage : Le Monde ; Banque de France
Abstract: This paper analyses the change in spatial income inequality across the departments of metropolitan France since 1922. Its most significant contribution is the reconstruction of average fiscal income per department, before and after the payment of income tax, based on an unprecedented use of archives from the Ministry of Finance. We highlight the following stylised facts: (i) a very significant reduction in interdepartmental average fiscal income inequality over the last century, with two periods of continuous decline, between 1922 and 1939 and from 1948 onwards; (ii) a significant contribution, albeit varying over time, of income tax to the reduction in inter-departmental inequality; (iii) an improvement in the situation in all departments lying along a line running from Calvados to Gard since 1948.
(4) Beyond the Exodus of May-June 1940: Internal Migrations in France during the Second World War, Demographic Research, 2021. Link.
Background: Research on annual migration between regions of the same country is scarce because statistical sources are often lacking. Consequently, little is known about migration between French regions during the Second World War.
Objective: Propose a new method to estimate the annual populations of each region and deduce the annual migration flows between regions of the same country. Use this method to present the chronology of inter-regional migration flows in France during the Second World War.
Methods: Annual regional populations are estimated using census populations and death-by-cause statistics at the local level. Death-by-cause statistics can be used to dissociate symmetrical mortality, which trends in the same way in the national territory, from asymmetrical mortality, which trends differently. This makes it possible to deduce the trend in the mortality rate and populations at the local level.
Results: Analysis of French migration reveals that (1) a large proportion of the population (up to 60%) was displaced from the north to the south in 1940 and 1941; (2) the demarcation line did not prevent population movements; (3) the demographic scars of war were still present in 1946 since not all the refugees from the northwest and northeast returned.
Contribution: This new method for analyzing annual migration flows between regions could be replicated to study other specific historical events. This paper also proposes a quantitative historical analysis of migration in France during the Second World War and shows that some territories were emptied of more than half their population in a few months.
(3) Spatial Inequality in Mortality in France over the Past Two Centuries (with H. d'Albis), Population and Development Review, 2020. Link ; Open Access Paper ; Data.
Media Coverage : N-IUSSP
Abstract : This article analyzes the evolution of spatial inequalities in mortality across 90 French territorial units since 1806. Using a new database, we identify a period from 1881 to 1980 when inequalities rapidly shrank while life expectancy rose. This century of convergence across territories was mainly due to the fall in infant mortality. Since 1980, spatial inequalities have levelled out or occasionally widened, due mainly to differences in life expectancy among the elderly. The geography of mortality also changed radically during the century of convergence. Whereas in the 19th century high mortality occurred mainly in larger cities and along a line from North-west to South-east France, it is now concentrated in the North, and Paris and Lyon currently enjoy an urban advantage.
Background: Debates concerning the French territorial divide are deep in France. However, there is little historical data on French demography at the local level for the 20th century available to study these territorial disparities.
Objective: The main objective is to present a new demographic database built at the departmental level for the 20th century, as well as the methods used to build this database.
Methods: The database was constructed by digitizing a large part of the raw data available in the French archives and by using new statistical sources for military deaths and deportees during the two World Wars. The methods used are both the methodological protocol of the Human Mortality Database and a set of methods which take into account the specificities of the French departments.
Contribution: With this database, the French departmental lifetables by sex and year including military deaths and deaths in deportation will be available, as well as the departmental populations by age, sex and year between 1901 and 2014.
(1) Inequalities in Life Expectancy and the Global Welfare Convergence (with H. d'Albis), Economics Letters, 2018. Link
Abstract : Becker et al. (2005) maintain that including life expectancy gains in a welfare indicator result in a reduction of inequality between 1960 and 2000 twice as great as when measured by per capita income. We discuss their methodology and show it determines the convergence result. We use an alternative methodology, based on Fleurbaey and Gaulier (2009), which monetizes differences in life expectancy between countries at each date rather than life expectancy gains. We show that including life expectancy has no effect on the evolution of world inequality.
Potential and challenges for sustainable progress in human longevity. Empirical evidence from subnational mortality trends in Europe (with I. Alliger, C.-G. Camarda, S. Klüsener, F. Meslé, M. Mühlichen, J. Thuilliez, P. Grigoriev). R&R in Nature Communications. Website App
Abstract: Decelerating gains in life expectancy (e0) in high-income countries have raised concerns about the future of human longevity. To enhance better understanding of these developments, we scrutinize subnational (N=450) mortality trends in Western Europe during the pre-pandemic period. Between 1992 and 2005, e0 gains were both substantial and widespread. The laggard regions experienced the fastest improvements, which led to a rapid regional convergence. However, between 2005 and 2019, e0 gains in these regions decelerated notably. Yet, they remained remarkably stable in vanguard regions, suggesting that extending longevity remains possible. The observed deceleration of e0 gains is strongly associated with mortality of the population aged 55–74, which increased by 2019 across large parts of Western Europe, particularly in Germany and France. We strongly advocate that assessing and further monitoring mortality trends at a fine geographical level are crucial for revealing potentials and challenges for sustainable progress in human longevity.
Understanding geographic disparities in mortality overtime in France, 1975-2019 (with A. Lleras-Muney, D. Yue, J. Thuilliez). R&R in Demography.
Abstract: Recent research has shown that geography is an important predictor of life expectancy at birth (e0). However, most papers have focused on place of residence at death (PoR), yet place of birth (PoB) may be a more important predictor of later mortality. To study how different measures of geography matter, we compute the first annual full lifetables by PoB for 90 French regions between 1975 and 2019 and compare them with lifetables by PoR. We show that e0 by PoR and PoB are highly correlated, but they have diverged in recent times, particularly for men, consistent with the growing importance of migration. This divergence has consequences on the trends in geographic disparities of longevity. First, geographic disparities by PoB have continuously declined, but measured by PoR they have risen since the 1990s. Second, the French geographical income gradient - the association between regional income and e0 - was negative and became positive in the 1990s when measured by PoR. When measured by PoB, the gradient has never been positive: it was negative but is now insignificant. These results are qualitatively similar if we study life expectancy at age 50.
Structural Change and Spatial Concentration of High Earners: Evidence from French Regions since 1960 (with H. d'Albis & A. Sotura). R&R in Journal of Economic Geography.
Abstract: This paper analyzes the shares of high-income earners within local populations, leveraging consistent annual income distributions for 90 French regions since 1960. The spatial dispersion of these shares has followed a U-shaped trajectory: a high concentration level in the 1960s, a decline in subsequent decades, and a marked increase since the early 2000s. In the 1960s, high-income earners were over-represented in the largest urban areas and the industrial Northeast. Today, they remain over-represented in the largest urban areas and near Switzerland. We demonstrate that shifts in regional productive structures explain this U-shaped trajectory. Deindustrialization initially reduced regional economic disparities, lowering the concentration of high-income earners. However, the shift to a service-based economy led to a resurgence of spatial concentration, as high-skilled workers increasingly clustered in large cities to benefit from agglomeration economies.
Does prosperity pay? Unraveling the relationship between economic performance and life expectancy across a large number of European regions, 2008-2019. (with M. Sauerberg, L. A. Cilek, M. Mühlichen, I. Alliger, C.-G. Camarda). Link
Abstract: Understanding the relationship between life expectancy at birth (e0) and the gross domestic product per capita (GDPpc) is relevant for cohesion policies in the European Union (EU) because it might imply that economic convergence (or divergence) is accompanied by narrowing (or widening) health gaps. Previous studies have studied the association between GDPpc and e0 almost exclusively based on national data. It is certainly more appropriate, however, to study the relationship at the subnational level because levels and trends in both variables, e0 and GDPpc, vary substantially across Europe’s regions. Accordingly, our study aims to examine whether the economic performance of regions predicts their e0 level. We build Preston curves from regression models using regional data for 21 European countries, divided into 506 regions. Mortality data comes from statistical offices and GDPpc can be obtained from the Eurostat database. The period 2008 to 2019 is particularly interesting as it follows the EU enlargement to central-eastern European countries in 2004. In our analysis, spatial units refer to NUTS-2 or NUTS-3 regions, depending on the size of the country. Our results suggest that there is indeed a positive association between GDPpc and e0. Similarly, to Preston’s original analysis, we observe an upward shift in the curve, indicating that factors exogenous to a region’s GDPpc level play an important role in explaining e0 gains as well. Yet, the relationship differs between geographical areas and we also find examples such as women in Germany, Austria, Poland, and the Netherlands where it does not seem to hold.
Distributions of fiscal income in French regions, 1960-2018 (with A. Sotura), Working paper Banque de France, 2021. Link
Abstract: This paper proposes homogeneous annual series on the income distribution of French metropolitan départements for the period 1960-69 and 1986-2018. We rely on unpublished and newly digitised archives of the French Ministry of Finance. They consist of fiscal tabulations that are a summary of households’ income tax declarations. Based on these raw sources, we interpolate the whole income distribution of French metropolitan départements after 1986. Before 1986, we need more assumptions as only households liable to French income tax filed income tax declarations at that time. We propose a methodology to estimate the number and average income of non-taxable households before 1986 that also allows us to reconstruct the income distribution of French metropolitan départements for the period 1960-69.
Exploring the Contribution of Educational Composition on Life Expectancy Changes in the United States, 2010-2018. (with J. V. Antunes Lopes, C.-G. Camarda, M. P. Bergeron Boucher, T. Riffe).
This study examines how changes in educational attainment composition contributed to trends in life expectancy (LE) at age 30 in the United States between 2010 and 2018. We ask two main questions: (1) How much of the change in LE can be attributed to shifts in the educational composition of the population? (2) How much is due to changes in mortality within educational groups? Using Horiuchi (2008)'s decomposition approach, we find that both factors played a role. The increase in the share of adults with higher education had a positive impact on LE, but rising mortality within some educational groups, especially the less educated, offset part of these gains. These results show the importance of considering both population composition and group-specific mortality trends to understand recent LE stagnation and decline in the US.
Rise and Fall of the Deaths of Despair Burden across European Regions, 2002–2019. (with I. Alliger, C.-G. Camarda, F. Meslé, M. Mühlichen, E. Perdrix, P. Grigoriev).
Concerns about stagnating life expectancy in the United States (US) have recently focused on the burden of “deaths of despair”. While extensively studied in this country, evidence for Europe remains fragmented and mostly national. We present the first systematic regional analysis of despair-related mortality across 20 European countries from 2002–2019, using age-standardized years of life lost (YLL) to capture the burden of premature deaths. We estimate trends by sex, age, and cause, identify statistically significant regional increases or declines, and compare our findings to the US as a benchmark. Our preliminary results suggest that the overall European picture is less alarming than in the US, but worrying sub-patterns emerge. Acute alcohol mortality has risen sharply in many Northern and Eastern European regions among older adults, while deaths from drugs and toxic substances have increased among younger adults in Germany, the Netherlands, the UK, Sweden, and Spain.
A Unified Innovation–Diffusion Theory of Fertility Transitions: New Insights from French Regional Fertility, 1853-2023 (with S. Florian).
We propose a theoretical framework that unifies the narratives of fertility trends into a single model with two successive innovation–diffusion processes: the spread of lower fertility (quantum) and delayed fertility (tempo), both proceeding through four phases and originating in wealthier regions before diffusing to lower-income ones. We test this framework using a new French dataset, analyzing trends in income per capita and Total Fertility Rates (TFR) by region from 1855 to 2023, decomposing TFRs into age-specific fertility rates (ASFR). During 1855–1945, TFRs were higher in poorer regions, as richer regions experienced earlier fertility decline. From 1970s, delayed fertility emerged first in high-income regions, increasing ASFRs at older ages, producing a rapid rise in mean age at birth (MAB), while poorer regions lagged behind. By 2023, results suggest that France has not yet entered the third phase of the tempo innovation, in which MAB converges across regions.
Trends in Heath Expectancies in France: to what extent the (social) cards have been reshuffled over the 2010's? (with E. Cambois).
Spatial Disparities in Disability Prevalence at the District Level in Sub-Saharan African Countries. (with A Simo-Fotso and J. Martin).
Abstract: The World Health Organization reports the highest prevalence of disability under age 60 in sub-Saharan Africa, yet evidence remains scarce due to limited and non-comparable measurements. To address this gap, we estimate subnational disability prevalence using harmonized data from 26 countries, collected between 2016 and 2022 through Demographic and Health Surveys and Multiple Indicator Cluster Surveys. Both surveys assess functional limitations in vision, hearing, mobility, cognition, communication, and self-care. We focus on adults aged 18–49 and generate estimates at the second subnational administrative level. Because direct survey estimates are often unstable at this scale, we apply small area estimation methods that borrow strength across age groups and geographic units. From these results, we derive age-standardized prevalence rates to facilitate cross-regional and cross-country comparison. Our findings reveal that 682 subnational units show prevalence above 2.5%, 233 above 5%, and 31 above 10%. The highest-burden areas cluster along the Ghana–Togo and Central African Republic–Democratic Republic of the Congo borders, in southern Democratic Republic of the Congo, and in parts of Madagascar. These patterns call for greater attention from national and international policymakers to the most affected areas.
Who lives longer in Paris? Dynamics of spatial and socioeconomic inequalities in longevity across Parisian districts, 1881--2019 (with C.-G. Camarda and C. Torres).
We combine newly digitized historical sources with modern statistical modeling to reconstruct continuous sex-specific lifetables for Paris and its 20 districts from 1881 to2023. First, we challenge the notion of “urban penalty”: by the late 19th century, life expectancy at birth (e0) was 13–14 years higher in the five best-performing districts than in the five worst-performing ones, and exceeded the national average. Second, we reveal that these gaps reflected social stratification: the five most affluent districts enjoyed e0 11-12 years higher than the five least affluent. Over subsequent decades, disparities narrowed, reaching a historical low by the mid-20th century. Yet these inequalities have widened again since the 1970s, with affluent districts regaining a clear survival advantage. Paris thus emerges from our study as a highly divided ‘longevity capital’ at the end of the 19th century, where health and life chances were deeply shaped by social geography.