Examining the determinants of bank efficiency in transition: empirical evidence from Ghana

Nathaniel Blankson, Ebenezer Bugri Anarfo, Godfred Amewu, Louis Doabil

Research output: Contribution to journalArticlepeer-review

11 Citations (Scopus)

Abstract

This paper measures and evaluates the relative pure technical efficiency and cost-efficiency of Ghanaian banks over the period 2008–2019 using a nonparametric DEA technique. The study also examines the factors that determines bank efficiency in Ghana using both static panel and dynamic panel regression estimators. The results show that the overall average bank efficiency (pure-technical and cost) levels in Ghana are relatively low compared to the benchmark “best-practice” efficiency level of 1. However, the results indicate that there are remarkable improvements in both the pure technical efficiency and cost-efficiency levels since the coming into effect of the new banking Act in 2016. Further, the study depicts bank size, GDP growth rate and inflation to be the most important factors that must be considered in the determination of bank efficiency in Ghana. Also, bank capitalization is depicted to have a significant impact on bank cost-efficiency. In contrast, the findings reveal that return on assets (ROA), liquidity and loan loss provision are, however, not important factors in the determination of both banks pure technical and cost-efficiency. Bank capital is further depicted to have insignificant impact in the determination of bank pure technical efficiency in Ghana.

Original languageEnglish
Article numbere10156
JournalHeliyon
Volume8
Issue number8
DOIs
Publication statusPublished - Aug 2022
Externally publishedYes

Keywords

  • Banking reforms
  • Cost-efficiency
  • Data envelope analysis
  • Dynamic panel regression analysis
  • Pure technical efficiency
  • Static panel regression analysis

Fingerprint

Dive into the research topics of 'Examining the determinants of bank efficiency in transition: empirical evidence from Ghana'. Together they form a unique fingerprint.

Cite this