The econometric modelling of banks nonperforming loans at the country-level and the bank-level

Okpala, Ginikachukwu (2023) The econometric modelling of banks nonperforming loans at the country-level and the bank-level. Doctoral thesis, University of East Anglia.

Preview

PDF Download (2MB) | Preview

Abstract

There has been a lot of debate regarding the many determinants of banks’ non-performing loans across several regions of the world. This debate between several researchers and policy makers has been ongoing since the 2008 global financial crisis and even intensified with the emergence of the sovereign debt crisis in the Eurozone in 2010. Due to the rapid increase of non-performing loans (NPLs) immediately after the global financial crisis that led to the folding up of a significant number of banks, several studies have been done to investigate the determinants of banks’ non-performing loans. This thesis, amongst other determinants, focuses on the effect of banking competition on NPLs at the country-level and at the bank-level and also investigates the relationship between NPLs and banks credit ratings. In chapter 1, I provide a general introduction to the thesis.

In chapter 2, we identify banking competition as a key determinant of NPLs and investigate the relationship between them at the country level. Using a panel data set covering a total of 105 countries for the period 2004-2016 and employing different panel data models, focusing on the fixed-effects estimator which takes into account country-level unobserved heterogeneity, this chapter tests the two views in the literature, i.e. the “competition-fragility”hypothesis and the“competition-stability” hypothesis). Contrary to previous studies that find evidence that supports only one hypothesis, this study provides evidence that supports both hypotheses. The results show that overall, a U-shaped relationship exists between NPLs and banking competition, with negative relationship at low competition levels, while positive for higher levels of competition. We go further to test whether the country’s level of development in which banks operate significantly influences the relationship between banking competition and NPLs. The results further show that, when the country’s level of development is accounted for a U-shaped relationship still holds for low and medium developed countries while for highly developed countries, the result suggests a less quadratic relationship. Results from this chapter contribute to the existing literature by using more recent data which covers a bigger sample size and by providing evidence that shows that overall, there is an optimal level of banking competition at which predicted NPLs are at their minimum. This chapter also contributes to the existing literature by investigating this relationship before and after the global financial crisis, employing different measures of banking competition and accounting for the potential issue of endogeneity of the banking competition variables through the application of the Instrumental Variable technique.

In chapter 3, using a combination of bank-level data and country-level data, we go further to test this relationship at the bank-level. Covering a total of 706 banks operating in 85 countries for the period 2004-2016, this chapter employs the use of a multi-level model approach due to the unique structure of our dataset which exhibits crossed random effects. The results suggest that a U-shaped relationship still holds for low and medium developed countries while for highly developed countries, the result suggests a less quadratic relationship as observed in low and medium developed countries. These results confirm that even at the bank-level, the country’s level of development in which banks operate has an impact on the relationship between NPLs and banking competition.

In chapter 4, we investigate the relationship between NPLs and banks credit ratings. Using a total of 145 banks for the period 2004-2016, this chapter employs the use of the Panel Vector Autoregressive Model as well as the Panel Granger causality test to examine whether a two-way relationship exists between bank credit ratings and NPLs. The results reveal that, not only do NPLs affect banks credit ratings, but banks credit ratings also – through lending channels - affect NPLs.

Item Type:	Thesis (Doctoral)
Faculty \ School:	Faculty of Social Sciences > School of Economics
Depositing User:	Chris White
Date Deposited:	19 Dec 2024 12:55
Last Modified:	19 Dec 2024 12:55
URI:	https://ueaeprints.uea.ac.uk/id/eprint/98035
DOI:

Downloads

Actions (login required)

View Item