This dissertation explores the diverging social spending patterns in labour-abundant regimes in the Middle East and North Africa (MENA). It is motivated by two main research questions: 1. Why have social spending levels and social policy trajectories writ large diverged so drastically across labour-abundant MENA regimes? 2. How can we explain the market persistence of spending levels after divergence? To answer the first question, this study develops a theory about the emergence of authoritarian welfare states. It argues that autocratic leaders need both the incentives and the abilities to distribute welfare for authoritarian welfare states to emerge. The former are shaped by coalition building dynamics at the onset of regime formation while the latter are conditioned by the external environment. At the level of incentives, broad coalitions emerge in the presence of intra-elite conflict and the absence of salient communal cleavages and, if present jointly, provide a strong incentive for welfare provision. Conversely, a cohesive elite or salient communal division entail small coalitions with few incentives to distribute welfare broadly. At the level of abilities, a strong external threat to regime survival is expected to undermine the ability to provide social welfare in broad coalitions. Facing a 'butter or guns' trade-off, elites shift priority to security expenditures and the population accepts that because no alternative regime could credibly commit to neglecting external defence in the presence of external threats. Only an abundant resource endowment can provide the necessary resources to avert this trade-off. To answer the second question, I rely on two important mechanisms in the welfare state literature to explain path dependance. The first one can broadly be summarised as 'constituency politics' in that beneficiaries of social policies succesfully avert deviations from the spending path in the form of systemic reforms or large-scale spending cuts. Mobilisation of these constituencies should be particularly vigourous if initial advantages conferred to these groups habe been reinforced over time, for instance, because these groups grew in size or got entrenched in the state administration. The second mechanisms are spill-over effects to unintended beneficiaries who can over time become important gatekeepers against path divergence. Methodologically, the study is characterised by a mixed-methods approach which combines quantitative tests with the analysis of qualitative evidence in the form of arhcival material, newspapers, and field interviews. Moreover, the study also follows a multi-level approach in that the viability of the argument is tested comparatively at the cross-country level and process-traced at the micro-level in two in-depth case studies of Tunisia and Egypt.