This thesis contributes to the public economics and economics of ageing literature. It focuses on heterogeneity in pension policy. In the first chapter, the heterogeneity relates to the age of individuals at the time of the reform. It analyses the distributional consequences of gradual pension reforms required to balance the intertemporal budget constraint, by adjusting either the accrual or the indexation rate. It shows that, based on employment and demographic forecasts for Belgium, 80% of the voting population would prefer the accrual reform to the indexation reform; implying that the youngest half of the population would bear 85% of the costs. So, there is a tradeoff between democracy and equality. The second chapter examines how pensions could account for occupation arduousness heterogeneity. It shows, using SHARE data, the possibility to infer a ranking of occupation arduousness from the health at old age. Then, it computes the adjustment of career length required to compensate for occupation arduousness. It concludes with caution by showing that the impact of occupation on health is dwarfed by that of childhood conditions. The third chapter considers lifespan heterogeneity. Life expectancy differs across income groups, and there is a growing consensus that this matters for pension policy. This chapter argues that, in addition to life expectancy, lifespan heterogeneity should also be looked at. It shows, using lifespan simulations based on US data from Chetty et al. (2016), that indexing retirement age to life expectancy differences could be misleading due to the large lifespan variance within income groups and the overlap of the lifespan distribution across them.