1. I develop a two-city model to explain a new stylized fact: high-income cities specialize in income-elastic sectors. The model has heterogeneous income elasticities and mobile agents, and either heterogeneous fundamental productivities or heterogeneous amenities generate the specialization pattern through the home market effect. The model also implies that tradable sector shares affect income inequality across locations by driving the home market effect. I provide empirical evidence of this with U.S. CBSA-level data. Finally, I theoretically and empirically suggest that, through this effect by tradable sector shares, the disproportionate trade cost reduction of business services has been supporting a stable city-size wage premium since 1980.2. I develop a model that demonstrates how the reputation of a military can generate different paths of political transitions. When a military has a reputation for not holding on to power for a long time, citizens can tolerate frequent coups. A military with such a reputation voluntarily democratizes to maintain the reputation. This equilibrium replicates the characteristics of the political transition paths of countries like Thailand. When there is no such reputation, the citizens resist a coup to avoid a non-democratic regime. Once the military seizes power, it will never voluntarily democratize, and the non-democratic regime becomes persistent. I show that citizens can choose the combination of frequent coups and voluntary democratizations over buying out their military.