The structure of socio-economic transitions

Effective policy requires decision-relevant diagnostics, yet modeling choices that are not "fit-for-purpose" often miss the explanatory bottleneck, whether in measurement, mechanism, or intervention design, and obscure the very structure they aim to reveal. This thesis investigates the consequences of that mismatch and what becomes visible when it is corrected. Measurements of global poverty and development fail to capture temporal dynamics for the former and are blind to spatial distortions for the latter. By including these dimensions more appropriately, we contribute to settling debates on the existence of poverty traps and growth regimes. Similarly, social norms and decision processes are among the most complex processes in modern society, yet they are crucial for society's capacity to coordinate and solve problems. This thesis demonstrates that understanding their endogenous mechanisms yields clear strategic takeaways. Finally, policy design is constrained by political feasibility and shaped by technological developments. Governing technological transitions — be it green energy or artificial intelligence — requires investigating feasible and optimal policy responses under these constraints. Across these domains, this thesis shows that matching the model to the respective diagnostic bottleneck leads to decision-relevant insights.