par Bagnoli, Lisa 
Président du jury Gassner, Marjorie
Promoteur Aldashev, Gani
Co-Promoteur Estache, Antonio
Publication Non publié, 2022-02-21
Président du jury Gassner, Marjorie
Promoteur Aldashev, Gani
Co-Promoteur Estache, Antonio
Publication Non publié, 2022-02-21
Thèse de doctorat
| Résumé : | This Ph.D. thesis focuses on the social and distributional effects of public policies, whether these are intended or not. To address the topic, I categorize public policies along two dimensions. First, I distinguish between policies with explicit social objectives and those without. Second, I distinguish between policies targeting specific vulnerable groups and those covering the whole population. The thesis consists of three empirical evaluations in three different contexts and sectors. The first chapter analyzes the introduction of a public policy with explicit social objectives and an explicit targeting of the most vulnerable. I focus on the impact of a social subsidy on electricity prices on the likelihood of energy poverty of vulnerable households in Spain. The second chapter assesses a public policy with social objectives but not directly targeted to the most vulnerable population. I investigate the heterogeneous health impact of the introduction in Ghana of free health insurance for all children. Finally, the third chapter analyzes the case of a policy without explicit social objectives or targeting. In this case I evaluate the unintended distributional effects of the introduction of a railroad in Kazakhstan through its effects on trade by the local population. |
| Chapter 1: How effective has the electricity social rate been in reducing energy poverty in Spain? This chapter analyzes whether the introduction of an electricity social rate was effective in reducing the likelihood of energy poverty. We focus on the Bono Social de Electricidad, introduced in 2009 in Spain's electricity market, a policy aimed at increasing the affordability of electricity by entailing a discount on prices for vulnerable consumers. Using data from the family budget surveys from 2006 to 2017, we rely on a difference-in-differences approach to measure its causal impact on energy poverty. We show that, on average, the introduction of the policy has reduced the likelihood of energy poverty. However, the magnitude of the effect is quite modest in practice as it takes only 59,000 households out of energy poverty out of the 2.8 million households still in that situation in 2018. It is noteworthy that the results also show that the lower effective prices do not lead an increase of the beneficiaries consumption of electricity. Overall this chapter provides new evidence on the institutional determinants of energy poverty and highlights the importance of assessing the magnitude of the effects as well as the targeting strategy of public policies to make the most of the resources allocated to social policies. | |
| Chapter 2: Does health insurance improve health for all? Heterogeneous effects on children in Ghana. This chapter investigates the extent to which the impact of a policy depends on the characteristics of the beneficiaries but also on the institutional capacity of the decentralized implementing agencies. To do so, I analyze the case of Ghana’s National Health Insurance Scheme (NHIS) that provides free coverage for all children. I exploit the Multiple Indicator Cluster Survey of 2011 and I use propensity score matching to account for selection in the scheme. The study finds that, even though the NHIS is successful on average in improving health outcomes among insured children in Ghana, gains are not shared equally across regions. The positive impact of health insurance is entirely concentrated among the lower-income households in regions with a high quality of public health care. This chapter sheds a new light on the mixed results of the literature on the impact of health insurance on health outcomes. It provides an understanding of the sources of the heterogeneous impact of a National Health Insurance Scheme and highlights the importance of context and implementation as drivers of its effectiveness. | |
| Chapter 3: Distributional effects of colonial railroads: Evidence from Kazakhstan. This chapter analyzes how the gains from a major public investment in railways undertaken without any explicit social objectives are shared among the population. More specifically, we study the distributional effect of Orenburg-Tashkent railroad constructed by the Russian Empire through Kazakh steppes in 1905-1906. The analysis is based on a unique historical dataset allowing us to provide evidence from a quasi-natural experiment to compare economic outcomes of households of varying wealth located in districts close and far from the railroad. We focus on the impact of the investment on the trading activities of households.We find that wealthier households benefited disproportionately more in trading activities from railroad closeness, whereas the trading activity of the poorer strata located closer to the railroad decreased. The colonial railroad appears affecting the indigenous population mostly through bringing in Russian peasant settlements rather than by reducing costs of long-distance trade. This paper provides novel evidence of the distributional effects of transportation infrastructure within regions. It further stresses the short and medium term social risks associated with failures to consider explicitly all the welfare impact of major public investments rather than just focusing on their long term growth payoffs. | |
| Conclusion: The analysis of these three case studies documents the diversity of social and distributional concerns arising from public policies. By studying different contexts with different methodological approaches, this thesis points to the importance of: (1) assessing ex-ante the size of the effects to reduce the risk of misuse of resources; (2) matching the targeting of beneficiaries with the stated policy objectives; (3) matching implementation strategies with institutional capacity; (4) considering the likely heterogeneity of direct impacts and their causes; and (5) considering explicitly the possibility of unintended or indirect social and distributional effects of policies to adopt mitigating solutions as needed. Overall, such evidence is relevant for both the evaluation and monitoring of existing policies as well as for the implementation of new policy reforms. |
