Risk-taking and systemic banking crisis in Africa: do regulatory policy framework provide new insight in threshold models?

Daniel Ofori-Sasu, Emmanuel Sarpong-Kumankoma, Saint Kuttu, Elikplimi Komla Agbloyor, Joshua Yindenaba Abor

Research output: Contribution to journalArticlepeer-review

Abstract

This study examines how regulatory policy impacts the complex relationship between bank risk-taking and the predicted probability of a systemic banking crisis in 54 African countries for the period, 2004–2020. The empirical evidence is based on the instrumental variable probit panel regressions. The study found a non-linear U-shaped relationship between bank risk-taking and the probability of a systemic banking crisis. The study shows that a systemic banking crisis is likely to occur when the monotonically increasing levels of risk-taking of banks exceed thresholds of 0.015 and 0.79. The study also found that the thresholds of risk-taking in countries with stringent regulatory policies are relatively greater in countries operating in low regulatory policy regimes. In light of the conditional marginal effects, the study provides evidence to support that regulatory policy amplifies and augments the negative linear impact of risk-taking on the predicted probability of a systemic banking crisis. This is relevant to policymakers because the established conditional effects imply that regulatory policy is a sufficient complementary condition for shaping the negative effect of bank risk-taking and systemic banking crises.

Original languageEnglish
Article number10
JournalRisk Management
Volume26
Issue number2
DOIs
Publication statusPublished - May 2024

Keywords

  • Bank risk-taking
  • E5
  • E61
  • G2
  • G3
  • G32, E3
  • L10
  • L51
  • M21
  • M48
  • Regulatory policy
  • Systemic banking crisis
  • Thresholds impact

Fingerprint

Dive into the research topics of 'Risk-taking and systemic banking crisis in Africa: do regulatory policy framework provide new insight in threshold models?'. Together they form a unique fingerprint.

Cite this