HOME Recent Events RMI in the News ALUMNI
  Issue 17 | Archive November 2013

Cascading Defaults and Systemic Risk of a Banking Network

A paper by Jin-Chuan Duan (National University of Singapore) and Changhao Zhang (National University of Singapore)

In a recent research paper entitled “Cascading Defaults and Systemic Risk of a Banking Network”, Prof. Duan Jin-Chuan, RMI Director, and his PhD student from the Department of Finance, Zhang Changhao, have developed a model to measure systemic risk. Using actual data for 15 UK banks, they investigated interesting characteristics of systemic risk.

Systemic risk has often been confused with systematic risk. The former arises from cascading defaults due to interbank linkages, whereas the latter arises from exposures to common risk factors. By and large, literature on systemic risk is based on correlation studies that are unable to distinguish the two. The authors overcame this limitation by characterising the banking system as a network of banks, of which the interconnectedness is represented by an interbank exposure matrix. The systemic cascade effect can then be analysed when the network is subjected to a systematic shock.

Banks are set up heterogeneously, each with their own assets and liabilities, which respond to a set of common risk factors. Specifically, the authors calibrated the response of each bank’s assets against the UK FTSE 100 Index, a Trade-Weighted GBP Index (by Deutsche), the GBP Libor Spread (12m-1m), and an orthogonal latent factor.

A key challenge lies in obtaining the market value of assets. Traditional methods, such as KMV’s, are unsuitable for banks because deposits are ignored, even though they make up roughly half of the bank’s total liabilities. Therefore, the authors used implied asset values produced by RMI’s Credit Research Initiative (“RMI-CRI”). RMI-CRI’s methodology for computing implied asset values is particularly apt for banks as it captures an appropriate proportion of bank deposits in its measurement.

The key advantage of such rigorous modelling is that the system may be evolved dynamically. In the UK example, the authors conditioned the UK FTSE 100 to fall by at least 40% over a 6-month period, and evolved the asset-liability dynamics day-by-day. The default of a bank will impact the balance sheet of other banks, and thereby induce systemic effects.

Two concepts of systemic risk are formalised: (1) Systemic Exposure, which measures expected uncovered losses under the stress event; and (2) systemic fragility, which measures the pervasiveness of defaults under the same.

Their measurement is responsive and confirms a spike in systemic risk during the credit crisis. They found that systematic shocks are more likely to drive systemic risk, as opposed to banks idiosyncratic elements. In turn, systematic shocks are positively related to the performance of market-wide risk factors.

The model is also suited for ranking banks by “systemic-importance”, as two important elements identified by the FSB are captured – systemic interconnectedness and loss absorption capacity.

Back To Newsletter

Published quarterly by Risk Management Institute, NUS
Editor: Ivy Wang (rmiwy@nus.edu.sg)