SA Financial Institution Systemic Risk Ranking

We rank South African financial institutions according to their contribution to systemic risk. How likely is it, that the default of this particular institution leads to the default of the financial system as a whole? Note that this is different from the risk that a particular institution defaults (which is extremely low).

We look at the 2000-2016 period. The ranking depicts the risk contribution at the last day of the period, and the graph below shows the risk contribution for each day in this period. All data originate from Bloomberg.

How do we do that?

We use two measures, an institution’s Marginal Expected Shortfall (MES), and its Systemic Risk Contribution (SRISK). MES is the average short run equity loss and SRISK is the expected capital shortfall of an institution conditional on a systemic event.

Generally, the more sensitive a financial institution is to aggregate market activity, the greater the it’s MES. As a result, these institutions are more likely to be systemically important, considering that their collapse may have a severe impact on the real economy. Similarly, greater SRISK indicates greater systemic importance. We rank financial institution according to how much each they contribute to the overall system wide undercapitalization.


We define the MES of an institution as the short-run expected equity loss conditional on the market taking a loss greater than a constant threshold C that defines a "systemic" tail event; we set this level at the 95\% VaR of daily index returns. Formally,

MES_{i,t-1}=\sigma_{i,t}\rho_{i,t}E_{t-1}\left( \varepsilon_{i,t}|\varepsilon_{m,t}<\dfrac{C}{\sigma_{m,t}}\right) + \sigma_{i,t}\sqrt{1-\rho_{i,t}^{2}}E_{t-1}\left( \xi_{i,t}|\varepsilon_{m,t}<\dfrac{C}{\sigma_{m,t}}\right)

where r_{m,t} and r_{i,t} are the returns on the market index and the equity of financial institution i respectively. \sigma_{m,t} and \sigma_{i,t} are the volatilities of the market and financial institution i at time t; \rho_{i,t} the correlation at time t between r_{m,t} and r_{i,t}. In this model, the disturbances \varepsilon_{m,t} and \x_{i,t} are assumed to be independently and identically distributed over time and have zero mean, unit variance and zero covariance under an unspecified distribution F. Conditional volatilities of the equity returns are modelled using an asymmetric GJR-GARCH specification.

SRISK is modelled as a function of the size of the financial institution, its degree of leverage, and its expected equity loss conditional on the market decline, the Long Run Marginal Expected Shortfall (LRMES). LRMES is a function of MES. Because systemic events tend to have persistent effects, LRMES is preferred to MES when computing SRISK. This is because LRMES allows us to capture expected losses over much longer periods of distress:

LMRES_{i,t} = 1-\exp(-18\times MES)

Assuming that debt cannot be renegotiated, we have

SRISK_{i,t} = k\textbf{L}_{i,t}- (1-k)\textbf{E}_{i,t}\left(1-LRMES_{i,t}\right)

where \textbf{E}_{i,t}is the market value of equity, \textbf{L}_{i,t} is the book value of debt and k is the prudential capital fraction. Ignoring capital surpluses, the contribution to systemic risk by any financial institution is given by

SRISK\%_{i,t} = \dfrac{ \left(SRISK_{i,t}\right)_{+}}{\sum_{i=1}^{N} \left(SRISK_{i,t}\right)_{+}}.


Our computations follow the approaches pioneered by Julien Idier, Gildas Laméa, and Jean-Stéphane Mésonnier (2014): "How useful is the Marginal Expected Shortfall for the measurement of systemic exposure? A practical assessment", Journal of Banking & Finance, 47, pp. 134–146 (DOI: 10.1016/j.jbankfin.2014.06.022) and Christian Brownlees and Robert F. Engle (2017): "SRISK: A Conditional Capital Shortfall Measure of Systemic Risk", The Review of Financial Studies, 30(1), pp. 48–79 (DOI: 10.1093/rfs/hhw060).

Financial institutions with the highest Systemic Risk Contribution

Rank Financial institution Contribution (%)
1 Standard Bank Group 21.56
2 Barclays Africa Group 13.00
3 FirstRand 12.94
4 Nedbank Group 12.61
5 MMI Holdings 10.70
6 Sanlam 10.34
7 Alexander Forbes Group Holdings 10.21
8 Old Mutual plc 2.77
9 Coronation Fund Managers 2.04
10 Sygnia 1.55
11 PSG Konsult 0.89
12 JSE 0.65
13 Investec 0.36
14 Investec plc 0.26
15 Prescient 0.09
16 Ecsponent 0.03

Table ranks financial institutions based in South Africa according to their SRISK Contribution. Contribution (%) is the SRISK defined on the last day of period, December 21st, 2016.

Top 7 risk contributors 2000-2016

Graph shows the SRISK Contribution for every day in the period 2000-2016. Use to your mouse to zoom in a selected rectangle. Double click to zoom out again. Some browsers may not work.

This is a project by Masters students at the University of Cape Town’s African Institute of Financial Markets and Risk Management (AIFMRM).

Both the source code and the data are on a public GitHub repository.


Qobolwakhe Dube (Student lead)

Qobolwakhe Dube is a PhD student at the African Institute of Financial Markets and Risk Management at University of Cape Town since January 2017. He works on market microstructure, focusing on the impact of FinTech in the banking industry.

For media inquiries, please contact Qobolwakhe.


Tresor Kaya (Student lead)

Tresor Kaya completed his Masters in Risk Management of Financial Markets with the African Institute of Financial Markets and Risk Management in 2016. He is currently interning with Prescient Securities and plans to study towards a PhD in Economics starting from 2017 focusing on macroeconomics, monetary economics and international finance.


Co-Pierre Georg (Academic lead)

Co-Pierre is a Senior Lecturer at the African Institute of Financial Markets and Risk Management at University of Cape Town and a Policy Associate at Economic Research Southern Africa. He is working on financial and economic interconnectedness, agent-based modelling, and microeconomics of banking. Co-Pierre is a frequent traveller and held visiting positions at Oxford, Princeton, and Columbia University. He has been a consultant for several central banks around the world on issues related to systemic risk and financial interconnectedness.


Michael E. Rose (Data visualization)

Michael E. Rose has been a PhD student at the African Institute of Financial Markets and Risk Management at University of Cape Town since April 2015. He works on social networks in financial and economic applications. He has been visiting De Nederlandsche Bank and Deutsche Bundesbank in 2016 and is also a Research Affiliate at Kiel Institute for the World Economy.