Democratic dominant party systems hold fair but uncompetitive competitions. In a democratic country in which one political party wins national elections repeatedly and continuously, what determines whether it will become competitive, authoritarian, or if it will maintain both dominance and democracy? Prior work on this question generally has taken a view of party system evolution as a result either of top-down ruling elite factionalism or of grassroots-based opposition growth. I link the two processes, arguing that competitiveness emerges from a combination of a weakening dominant party and a growing opposition. The dominant party’s attitude to interparty and intraparty opposition, and the opposition’s viability as a credible alternative largely determine the evolution of a dominant party system. I use a multi-level approach, looking explicitly at the role of subnational party systems in a nationally dominant party system to gain analytical leverage. I evaluate both explicit efforts to spread subnational opposition success to other localities and to other levels of government, and the ruling party’s attempts to thwart these efforts. To support this theory, I draw extensively on literature about a diverse set of countries, including India, Italy, Japan, Mexico, and several states of sub-Saharan Africa. Then, I turn to an analysis of South African cities, using my own interviews with party elites, newspaper accounts, and archival sources to compare increasingly competitive Cape Town and increasingly African National Congress (ANC)-dominant Durban through the 2009 elections. This period coincides with a fraught succession fight within the ANC, culminating in severe factionalism with geographic bases and an organized defection to form a separate party. The weakened ANC created opportunities for opposition growth. I argue that opposition parties are most likely to find traction with strategies that, first, build their credibility through winning and governing subnational governments and, second, capitalize on voters’ doubts about the democratic credentials of the dominant party in order to capture some of the electoral center. However, democracy’s contested status may damage the quality of democracy and consolidation of the regime. Based on survey data drawn from the Afrobarometer and the Comparative National Election Project, South Africans are disillusioned with their democratic system in ways we expect and they support democracy in the abstract. However, there are reasons for concern about the public’s reaction if the dominant party were to lose national power. I conclude that the evolution of dominant party systems into competitive ones may be a double-edged sword; competitive parties should yield a better functioning democracy, but the process of achieving that competitiveness risks political instability.