Le résumé en français n'a pas été communiqué par l'auteur. / Throughout this thesis, I wish to study the political and economic determinants of the size and features of intergenerational public transfers, by which I mean all public policy programs that entail a transfer of public resources between different age cohorts. The first chapter of this thesis, jointly written with Georges Casamatta, reviews the latest developments in the political economy literature dedicated to the study of the consequences of population ageing, with a primary focus on the threat posed by ageing to the continued existence of public pension programs in developed countries. We review works that analyze how ageing might change the contribution rates, pension size or retirement age at the political equilibrium. We also present empirical assessments of the relationship between a population age structure and the size of its pension programs. Additionally, we give an account of both theoretical and empirical analyses of the impact of population ageing on the political support for other public programs, such as education. The second chapter of this thesis studies how demographic changes impact the public provision of social security and publicly-funded education, when these policies are determined as the outcome of a vote that involves both contributory and beneficiary generations. To this end, I set up a three-overlapping-generation model in which the two intergenerational transfers are funded through taxes on the working generation, and benefit respectively the young (for education) and the elderly (for pensions). Population ageing is shown to lead to a rising tax burden in the future, and is forecasted to lead to higher pension and education spending per recipient, under plausible parameter values. In the third chapter of this thesis, I study under which conditions political processes might lead to an efficient level of public subsidies to education acquisition, when the space of available policies also includes redistribution through distortive labor income taxation. To this end, I set up a two-period model where individuals differing in their intrinsic abilities on the labor market privately choose their levels of education expenditure and labor supply, and vote sequentially on a linear education subsidy and a linear labor income tax. When the elected government is unable to commit to a labor income tax rate before individuals select their level of education, I show that the electorate collectively selects too high a level of education subsidies in the first stage of the voting game, even when compared to the inefficiently high level of labor taxation selected in the second stage.