By assuming that insurgencies come 'from below' — i.e. are launched by exploited and deprived social groups —, existing theories are not suited for explaining insurgencies 'from above' — i.e. led by political, social, and economic elites. When most insurgencies come 'from above', why do insurgency levels vary across countries? In nineteenth-century Latin America insurgencies of this kind were very common. Therefore, I compare two countries with high insurgency (Colombia and Uruguay) and two with low insurgency (Chile and Costa Rica) during the century after independence (ca. 1820-1920). Three factors help explaining variations in insurgency levels. The first one is the strength of the ties between peasants and landowners. Strong ties allow landowners to use selective incentives for mobilizing their subordinates into rebel armies. Conversely, weak vertical ties reduce selective incentives and obstruct mobilization. This complements theories about insurgencies 'from below', which assume that weak vertical ties increase insurgency. The second factor is the timing of consolidation of the central state. Early state consolidation increases the costs of insurgency and leads government opponents to engage in other strategies (elections, informal agreements, and/or military coups). Conversely, late consolidation encourages an early resort to insurgency, which becomes self-reinforcing and persists even after the state consolidates. By emphasizing the timing of state consolidation I complement political opportunity and state breakdown theories, which overlook how past events shape outcomes across time. The third factor is the type of party system. Two-party systems simplify the process of blame attribution, allow the party in power to exclude its opponent, and encourage leaders to emphasize extreme positions for capturing the support of small and highly militant electorates. This increases polarization and boosts insurgency. Conversely, because in multi-party systems parties are unable to govern alone, they are encouraged to engage in flexible electoral and congressional alliances that decrease polarization and therefore insurgency. Consistent with this argument, in Colombia and Uruguay vertical ties were strong, central states consolidated late, and two-party systems polarized, leading to high insurgency. In Chile and Costa Rica vertical ties were weak, states consolidated early, and multi-party systems did not polarize, leading to low insurgency.