The effect of annual report narratives on the cost of capital in the Middle East and North Africa: a machine learning approach  

Gehan Mousa, Elsayed Elamir, Khaled Hussainey

Research output: Contribution to journalArticlepeer-review

Abstract

This paper contributes to accounting literature by reexamining the impact of the quantity and readability of annual report narratives on cost of capital. This study employs a machine learning technique, namely, the model-based (MOB) recursive partitioning, while the least absolute shrinkage and selection operator is used to select variables from a sample of 720 bank–year observations from eight Middle Eastern and North African countries between 2008 and 2019. The model-based (MOB) recursive partitioning works with local and global models to explore hidden information in the data that leads to better results in both linear and nonlinear relationships. Our analysis shows that, on one hand, the readability of annual report narratives has an insignificant impact on cost of capital. On the other hand, it shows that the greater the amount of narrative disclosure, the lower the cost of capital, a result that varies between countries and according to corporate profitability.
Original languageEnglish
JournalResearch in International Business and Finance
Publication statusAccepted for publication - 8 May 2022

Keywords

  • readability
  • corporate disclosure
  • annual report narratives
  • machine learning techniques
  • cost of capital
  • Middle Eastern and North African (MENA) countries

Fingerprint

Dive into the research topics of 'The effect of annual report narratives on the cost of capital in the Middle East and North Africa: a machine learning approach  '. Together they form a unique fingerprint.

Cite this