Beyond Rankings: Bayesian stochastic frontier analysis of the determinants of university efficiency

University rankings predominantly focus on outputs while neglecting the efficiency with which institutions convert resources into outcomes. We contribute to addressing this limitation by analyzing the determinants of university efficiency using a Bayesian stochastic ray frontier model applied to 47 Spanish public universities over the 2016–2021 period. Unlike traditional approaches, our methodology jointly estimates efficiency and its determinants in a single stage. We adopt a multi-output framework encompassing the three university missions: teaching, research, and knowledge transfer. Using backward stepwise selection with the deviance information criterion, we identify key efficiency determinants including the average department size, number of campuses, academic staff characteristics, and multi-province location. Results reveal substantial efficiency variations across universities, with approximately half showing positive efficiency changes over the period. The Bayesian approach provides full efficiency distributions rather than point estimates, enabling robust statistical comparisons. Our findings offer valuable insights for university managers and policymakers seeking to enhance institutional performance beyond traditional output-based rankings.