Analysis of Technical, Pure Technical and Scale Efficiencies of Pakistan Railways Using Data Envelopment Analysis and Tobit Regression Model

Khalid Mehmood Alam*, Li Xuemei, Saranjam Baig, Li Yadong, Akber Aman Shah

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)

Abstract

Efficient railway transportation systems are very important for economic development of countries. Although extensive research has been conducted on railway efficiency in developed countries, however, such studies for developing economies in general and Pakistan is lacking. Here, we used data envelopment analysis (DEA) to evaluate the technical, pure technical and scale efficiencies of Pakistan railways from 1950 to 2016 and super efficiency model is employed to rank the efficient DMUs. This study uses seven internal factors and three externals factors to analyse efficiency. The empirical result indicates that there are five DMUs with an overall technical efficiency value of 1. The Pakistan railways operate at 86.8% level of overall technical efficiency i.e., inputs could be decreased by 13.2% without sacrificing output if all DMUs were efficient. Furthermore, the contribution of scale inefficiency in overall technical inefficiency has been observed to be smaller than what has been observed due to pure technical inefficiency (managerial inefficiency). The results of the Tobit regression model suggest that economic development and trade have positive effects on technical efficiency while the growth of highways affects it negatively as external factors.

Original languageEnglish
Pages (from-to)989-1014
Number of pages26
JournalNetworks and Spatial Economics
Volume20
Issue number4
DOIs
Publication statusPublished - Dec 1 2020

Keywords

  • Data envelopment analysis
  • Overall technical efficiency
  • Pakistan Railways
  • Pure technical efficiency
  • Scale efficiency
  • Tobit regression model

ASJC Scopus subject areas

  • Software
  • Computer Networks and Communications
  • Artificial Intelligence

Cite this