Periodic Reporting for period 1 - EPOCH (Women’s political EmPOwerment and Child Health)
Periodo di rendicontazione: 2021-10-01 al 2023-09-30
The primary objectives of the EPOCH (Women’s political EmPOwerment and Child Health) project are to: 1) systematically assess the effects of Women’s Political Representation (WPR) on child health in Lower- and Middle-Income Countries (LMIC) by using state-of-the-art statistical models combining cross-sectional and longitudinal data; and 2) to establish concrete mechanisms linking WPR with child health.
The project derives from the hypothesis that the quality of a country’s political institutions is the primary explanation for countries’ economic and health progress. Specifically, the degree to which political institutions are inclusive versus exclusive has been shown to explain an important part of countries’ long-run development. Although the literature has found that the inclusiveness of political institutions is positively related to economic growth and development, little attention is being paid to the inclusiveness of political institutions with regard to WPR. This is a significant shortcoming as democratic systems per se are no guarantee for the inclusion of women in political decision-making, with many democracies featuring less than 30 percent of women represented in parliaments.
Empirically, most indicators of Women’s Political Representation (WPR) are positively associated with health outcomes at the country level. However, existing studies on the relation between WPR and population health are analytically weak and inconclusive due to their: 1) almost exclusively observational nature, 2) reliance on aggregate-level, or 3) cross-sectional data. As a result, an open question remains whether greater WPR is causally related to population and child health, or whether the causality is in fact reverse, as countries with higher levels of development may simply be more likely to have higher levels of WPR. Furthermore, the reliance of aggregate-level data (mostly on the country-level) has hampered the identification of concrete mechanisms linking WPR with (child) health.
By addressing these shortcomings of previous studies, the EPOCH project makes an important contribution to the scholarship on the primary explanations for population health and the role of political institutions as a determinant. Robust evidence on the relation between WPR and (child) health has the potential to further question the dominant view that economic growth is the most important driver of population health – a key dimension of human development. From a policy perspective, the expected results of this study are highly relevant for setting policy priorities in LMIC as increasing WPR, if proven causal, could represent a very cost-effective way for health promotion in LMIC and for achieving Sustainable Development Goal (SDG) target #3.2.
The project derives from the hypothesis that the quality of a country’s political institutions is the primary explanation for countries’ economic and health progress. Specifically, the degree to which political institutions are inclusive versus exclusive has been shown to explain an important part of countries’ long-run development. Although the literature has found that the inclusiveness of political institutions is positively related to economic growth and development, little attention is being paid to the inclusiveness of political institutions with regard to WPR. This is a significant shortcoming as democratic systems per se are no guarantee for the inclusion of women in political decision-making, with many democracies featuring less than 30 percent of women represented in parliaments.
Empirically, most indicators of Women’s Political Representation (WPR) are positively associated with health outcomes at the country level. However, existing studies on the relation between WPR and population health are analytically weak and inconclusive due to their: 1) almost exclusively observational nature, 2) reliance on aggregate-level, or 3) cross-sectional data. As a result, an open question remains whether greater WPR is causally related to population and child health, or whether the causality is in fact reverse, as countries with higher levels of development may simply be more likely to have higher levels of WPR. Furthermore, the reliance of aggregate-level data (mostly on the country-level) has hampered the identification of concrete mechanisms linking WPR with (child) health.
By addressing these shortcomings of previous studies, the EPOCH project makes an important contribution to the scholarship on the primary explanations for population health and the role of political institutions as a determinant. Robust evidence on the relation between WPR and (child) health has the potential to further question the dominant view that economic growth is the most important driver of population health – a key dimension of human development. From a policy perspective, the expected results of this study are highly relevant for setting policy priorities in LMIC as increasing WPR, if proven causal, could represent a very cost-effective way for health promotion in LMIC and for achieving Sustainable Development Goal (SDG) target #3.2.
The work during the first year of the project has been focused on Objective #1: Estimating the effect of WPR on child health in LMIC: This aim will be pursued by using Regression Discontinuity Designs (RDD) in combination with cross-national longitudinal and cross-sectional data on child health as well as detailed information on sub-national electoral outcomes in LMIC. Specifically, the work addressed three related research questions: A) What is the impact of WPR on child health and mortality in LMIC? B) Are there specific ‘sensitive’ periods, e.g. infancy or the transition to primary school, during which the effect of being exposed to WPR matters most for child health? C) To what extent does the effect of WPR on child health depend on household, maternal or child characteristics?
As proposed in the original project proposal, to empirically address these questions, I accessed and compile data from Mexico and Brazil as well as the Demographic and Health Surveys (DHS). The data mere merged with detailed data on elecion outcomes on the level of municipalities for different years. The latter allows identifying whether a child, at a given year of age and in a given place, was exposed to a female or male head of government. Using a so-called regression-discontinuity design that exploits close elections between a male and female candidate in electoral districts, I am able to estimate the causal effect of the existence of a female head of government on child health in these different datasets.
In terms of concrete research outputs, based on these data, two articles have been written and are currently under review with journals. Earlier versions of the paper were presented at research seminars of different acadeic institutions (University of Lausanne, University of Geneva, Barcelona Institute of Global Health, Social Science Research Center Berlin, French Institute of Demographic Studies) and I benefited greatly from the feedback of colleagues from these institutions.
The work on Objective #2 (Establishing and quantifying the mechanisms that explain the relation between WPR on child health in LMICs), planned for the second year, was not started due to an early termination of the project after the first year.
As proposed in the original project proposal, to empirically address these questions, I accessed and compile data from Mexico and Brazil as well as the Demographic and Health Surveys (DHS). The data mere merged with detailed data on elecion outcomes on the level of municipalities for different years. The latter allows identifying whether a child, at a given year of age and in a given place, was exposed to a female or male head of government. Using a so-called regression-discontinuity design that exploits close elections between a male and female candidate in electoral districts, I am able to estimate the causal effect of the existence of a female head of government on child health in these different datasets.
In terms of concrete research outputs, based on these data, two articles have been written and are currently under review with journals. Earlier versions of the paper were presented at research seminars of different acadeic institutions (University of Lausanne, University of Geneva, Barcelona Institute of Global Health, Social Science Research Center Berlin, French Institute of Demographic Studies) and I benefited greatly from the feedback of colleagues from these institutions.
The work on Objective #2 (Establishing and quantifying the mechanisms that explain the relation between WPR on child health in LMICs), planned for the second year, was not started due to an early termination of the project after the first year.
The first paper, which was submitted to the "Journal of Health Economics" but rejected and therefore being currently revised for re-submission to another journal, assesses whether there is a long-run effect of growing up in a municipality governed by a woman, and also whether this effect may be channeled trough increased investments into conditional cash trasnfers (CCT), i.e. the Progresa program. Results reveal that, overall, there is no significant effect of exposure to femal municipality mayors during school-going ages on adult mental health. We do however, find evidence that the Progresa program, which has been consitently associated with better child health in other studies, was implemented faster in municipalities with a female mayor than in municipalities with a male mayor. However, we find strong effect heterogeneity. Being exposed to a female mayor in schooling ages implies a decrease in self-reported mild anxiety among male and urban respondents during adulthood. Exposure to female mayors in childhood is also associated with more years of schooling, greater employability, and better health.
The second paper, under review with the "Journal of Development Economics", finds that electing a woman as mayor has no statistically significant effect on child death overall However, electing a female mayor decreases child mortality in municipalities with a higher share of legislative power held by women, and where most of the population is illiterate. The findings of this paper suggest that electing women politicians, in this case as mayors of municipalities in Brazil, does not necessarily lead to improvements in child health in all instances. This is in contrast with much of the existing literature which has shown repeatedly that the presence of female politicians is strongly associated with lower mortality rates among children, including in Brazil. However, previous evidence has been based almost exclusively on correlations and the causality behind these findings has always been in doubt. The results of my paper thus put into questions whether these associations are causal. Despite the finding that overall there seems to be no significant effect of female mayors on child health in Brazil, the findings do show that electing female mayors can significantly reduce child mortality in municipanities that are the leaast developed and that already have a higher share of female politicians represented in the municipal council.
The second paper, under review with the "Journal of Development Economics", finds that electing a woman as mayor has no statistically significant effect on child death overall However, electing a female mayor decreases child mortality in municipalities with a higher share of legislative power held by women, and where most of the population is illiterate. The findings of this paper suggest that electing women politicians, in this case as mayors of municipalities in Brazil, does not necessarily lead to improvements in child health in all instances. This is in contrast with much of the existing literature which has shown repeatedly that the presence of female politicians is strongly associated with lower mortality rates among children, including in Brazil. However, previous evidence has been based almost exclusively on correlations and the causality behind these findings has always been in doubt. The results of my paper thus put into questions whether these associations are causal. Despite the finding that overall there seems to be no significant effect of female mayors on child health in Brazil, the findings do show that electing female mayors can significantly reduce child mortality in municipanities that are the leaast developed and that already have a higher share of female politicians represented in the municipal council.