Citizen Participation and Political Trust in Latin America and the Caribbean : AMachine Learning Approach

This paper advances the understanding of the linkages between trust in government and citizenparticipation in Latin America and the Caribbean, using machine learning techniques and Latinobarómetro 2020 data.Proponents of the concept of stealth democracy argue that an inverse relationship exists between political trust andcitizen participation, while deliberative democracy theorists claim the opposite. The paper estimates that trustin national governments or other governmental institutions plays neither a dominant nor consistent role in drivingpolitical participation. Instead, interest in politics, personal circumstances such as experience of crime anddiscrimination, and socioeconomic aspects appear to drive citizen participation much more strongly in the LatinAmerica and the Caribbean region. This is true across models imposing simple linear trends (logit and lasso) and othersallowing for nonlinear and complex relations (decision trees). The results vary across the type ofparticipation—signing a petition, participation in demonstrations, or involvement in a community issue—whichthe paper attributes to increasing net costs associated with participation.