This study introduces a novel, replicable methodology for analyzing employment dynamics within public sector agencies, focusing on turnover and staff longevity. The methodology is designed to be generalizable and applicable to diverse national contexts where detailed administrative data is available. Using payroll data from over 325,000 Chilean civil servants (2006—2020), we apply mixed-effects Cox survival models and linear mixed models to examine patterns of employment stability across state agencies. By incorporating Propensity Score Matching, we further enhance the causal interpretation of turnover changes, especially in post-election years. Finally, we introduce two key metrics—Service Frailty and Relative Turnover Difference—to quantify long-term stability and short-term, post-electoral disruptions. Our findings highlight substantial differences in turnover patterns between regular and post-election years, as well as significant inter-agency heterogeneity in turnover and employee longevity, largely driven by latent agency characteristics. While major covariates like contract type and staff rank account for some variation, much of the disparity stems from agency-specific factors. This framework offers precise, cross-nationally comparable benchmarks for understanding public sector employment dynamics. Additionally, the methodology contributes to the literature by providing transparent and scalable tools for analyzing workforce stability across different contexts.