The Wrong Kind of Learning: The Case of Subnational Recall Diffusion
P10-S250-1
Presented by: Mariana Alvarado
The diffusion literature has tended to see learning as a process through which efficient or successful policies expand. While the possibility that negative practices may diffuse has sometimes been acknowledged, it has never been the object of serious study. This paper uses the diffusion of subnational recall elections in Peru (the most intensive user of the recall worldwide) to conduct the first study of negative or perverse diffusion dynamics. It exploits the availability of a large number of observations (1,634 districts) and a relatively long time series (1996-2014), to leverage variation in the presence of recall elections within districts over time and uncover the effect of neighbours’ behaviours and the dynamics of political competition on the probability a recall takes place. The paper shows how recalls diffuse spatially through a process of social learning in which information about the political viability of this institution is transmitted to the local politicians who are responsible for its repeated activation. These politicians learn to (ab)use participatory institutions to re-do elections regardless of voter's preferences.
Keywords: diffusion, recall elections, direct democracy