Abstract

Chile was long considered a beacon of stability in a turbulent region, until the 2019 social outburst and two failed constitutional processes prompted a reassessment. We argue that the reliance of much of the literature on national aggregates obscured underlying dynamics that could have raised institutional alerts, and perhaps even diverted the course of the crisis. To test this claim, we focus on the Chilean party system as it has been tied to the fateful events and ask whether inferences would have been different had alternative data been considered. In this effort, we introduce the Chilean Political Landscape Dataset (CPLD), a standardized longitudinal dataset of district-level electoral outcomes and indicators from 1989 to 2024 that includes more than 5300 observations across legislative, regional, and municipal contests. The dataset allows us to highlight meaningful gaps between prevailing explanations and observed patterns, showing why district-level evidence is necessary to identify the sequence, scope, and ideological contours of change. We show that more detailed data suggest that fragmentation began well before 2015, that the party system was drifting leftward over time, and that fragmentation occurred asymmetrically, impacting parties on the left more than those on the right. We argue that this had previously gone untested and unfound because of the lack of data and conclude by illustrating its relevance by showing additional applications of the CPLD, including assessments of social structures and institutional design.