Evaluating efficiency determinants of rural bus routes in northern Taiwan: A network DEA with temporal data aggregation

Abstract

This study evaluates the production, service, and overall operational efficiency of 19 rural bus routes in northern Taiwan, focusing on the challenges of delivering transport services in remote and low-density areas. A two-stage analytical framework was adopted. In the first stage, multi-period Network Data Envelopment Analysis (NDEA) was employed to assess stage-specific efficiencies across time. Particular attention was given to ensuring consistency between data-level and efficiency-level aggregation, addressing a critical methodological issue in multiperiod efficiency measurement. In the second stage, Bootstrap Truncated Regression (BTR) was applied to examine the influence of external factors beyond conventional inputs and outputs. The findings identify route length, operating mode, and taxi service integration as significant determinants of efficiency. Further, quadrantbased importance-performance analysis reveals critical mismatches between route efficiency and strategic relevance, offering a basis for targeted improvement. This study provides actionable recommendations for policymakers and transit operators, including optimizing route design to reduce inefficiencies, integrating flexible transport modes, tailoring subsidy mechanisms, and institutionalizing multi-period evaluation frameworks to support data-driven governance in rural transportation systems.