This article uses a multimethod research strategy to show that changes in electoral rules implemented in Colombia in 2003 did not have the expected effects in municipal council elections: party system fragmentation and personalistic intra-party competition increased. Moreover, while the effect of reform was more homogenous across municipalities with regard to intra-party competition, it was more heterogenous across these units with respect to party system fragmentation. Moreover, a higher average level of party system fragmentation does not necessarily reflect greater competition between parties in all municipalities. The use of multiple party labels and the ensuing increase in fragmentation can also be a strategy by hegemonic machines to control a greater number of seats in the new institutional context. In this way, this article does not only reveal the limits of institutional engineering mandated from the national level. It also shows how multimethod strategies in the context of a subnational research design (SNR) can provide more robust descriptions of local political processes, thus avoiding potentially partial interpretations derived from single-method designs.