Probabilistic modeling of retirement financial preparedness among higher education staff: A bayesian regression perspective

This study addresses the problem of low retirement financial preparedness among higher education staff in Indonesia, a critical issue as demographic, behavioral, and institutional uncertainties affect post-employment well-being. To respond to this challenge, the research applies a Bayesian probabilistic modeling framework to capture the stochastic nature of retirement readiness and propose data-driven solutions for financial policy and education. The research contribution is the integration of behavioral and institutional determinants into a unified Bayesian model that quantifies uncertainty and identifies probabilistic predictors of retirement preparedness among academic personnel. The study utilized survey data from 110 respondents, both civil servants (ASN) and non-civil servants, and employed Bayesian logistic regression with Markov Chain Monte Carlo (MCMC) estimation. Posterior diagnostics confirmed model convergence (̂R < 1.01; ESS > 5,000) and predictive adequacy (WAIC = 0.41; LOOIC = 0.63). The results indicate that behavioral engagement and institutional access strongly influence readiness: investment planning, pension benefit availability, and financial strategy implementation significantly raise preparedness probabilities. Income adequacy interacts positively with strategic financial behavior, while age and education show weaker effects. In conclusion, retirement readiness in Indonesian academia remains low yet improvable through behavioral interventions and expanded institutional support, providing empirical evidence for targeted financial literacy and pension reform initiatives.