Modelling systemic risk in the South African banking sector using CoVaR

ABSTRACT

In this paper, we model systemic risk by making use of the conditional quantile regression to identify the most systemically important and vulnerable banks in South Africa (SA). We measure the marginal contributions of each bank to systemic risk by computing the difference between system risk of individual banks when they are in a normal state and when they are in distress. Using daily stock market prices of six SA commercial banks¹ from June 2007 to April 2016; Our results show a considerable increase in the market risk of all the six banks during the 2008 global financial crisis with African Bank being the riskiest bank in the country. Our systemic risk ranking shows that First Rand Bank was the largest contributor to systemic risk followed by Standard Bank, Barclays Africa, Nedbank, Capitec and lastly African Bank. These results suggest that the largest banks pose a bigger threat to the banking system than the smaller banks. These findings clearly indicate that different banks pose different threats to the banking system and the economy at large. Hence, specific actions that go beyond limiting idiosyncratic risk are needed if stability is to be attained through macro prudential regulation.

KEYWORDS:

• Conditional quantile
• systemic risk
• conditional value at risk
• banking sector

JEL CLASSIFICATION:

• C13
• C22
• C58
• G01
• G21

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. The six South African commercial banks include Barclays Africa Limited, First Rand Bank Limited, Nedbank Limited, Capitec Bank, Standard Bank of South Africa and African Bank.

2. A previous version of this study was published as a working paper: Adrian and Brunnermeier (2008).

3. Who extended the CoVaR method to a multivariate GARCH setting?

4. Who studied systemic risk in sovereign debt markets?

5. Who investigated the effect corporate governance’s role on systemic risk?.

6. q represents a lower quantile for example 1% unless stated otherwise.