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

This paper advances the understanding of the linkages between trust in government and citizen participation in Latin America and the Caribbean (LAC) using machine learning techniques and Latinobarómetro 2023 data. Empirically, we predict citizen participation based on trust levels, characteristics and circumstances of citizens. Proponents of the concept of stealth democracy argue that an inverse relationship exists between political trust and citizen participation, while deliberative democracy theorists claim the opposite. Based on our estimates, trust in national governments and other governmental institutions play neither a dominant nor consistent role in driving political participation. Interest in politics, personal circumstances such as experience of crime, and socioeconomic characteristics appear to drive citizen participation much more strongly in LAC. This is true across models imposing simple linear trends (Logit and Lasso) and those allowing for complex relations (decision trees). Results vary across types of participation—signing a petition, participation in demonstrations, or involvement in community issues. Ultimately, political trust can only influence political participation when certain other drivers are combined in some specific ways.