28 January 2013

Bundesbank: CDS spreads and systemic risk – A spatial econometric approach

This study applies a novel way of measuring, quantifying and modelling the systemic risk within the financial system. The magnitude of risk spillover effects is gauged by introducing a specific weighting scheme.

This approach originally stems from spatial econometrics. The methodology allows for a decomposition of the credit spread into a systemic, systematic and idiosyncratic risk premium. Bundesbank  identifies considerable risk spill overs due to the interconnectedness of the financial institutes in the sample. In stress tests, up to one fifth of the CDS spread changes are owing to financial contagion. These results also give an alternative explanation for the nonlinear relationship between a debtor's theoretical probability of default and the observed credit spreads  known as the "credit spread puzzle".

This study explicitly models the degree of proximity between financial institutions, approximated by the equity correlation between two firms. The framework of spatial econometrics offers an efficient way of introducing banks' interconnectedness in terms of equity correlation into a regression model. Moreover, the spatial econometric approach allows for a decomposition of the variance of banks' CDS premiums into a systemic, a systematic and an idiosyncratic risk component.

The results indicate that the systemic risk component in the CDS market is important. The extent of spillover effects,  measured by the spatial autoregressive parameter, and the magnitude of transmission summarised as direct as well as indirect effects are found to be significant and considerable. These effects are still present even when macro-economic variables such as the Libor-OIS spread or proxies for risk mitigation are introduced into the model.

The Bundesbank finds that the systemic risk charge varies by geographical region as well as by time. While European and US institutions are strongly affected by financial contagion, Asian banks are found to be rather independent. CDS spreads to be up to a tenth higher than a model without spill over effects would predict. For first differences of CDS spreads, the degree of influence by systemic components is even higher.

The presence of indirect effects offers a new perspective on the so-called credit spread puzzle. While the systemic risk charge plays a rather substantial role for smaller credit spreads, the relative impact on CDS spreads decreases with the absolute level. Consequently, the answer to the credit spread puzzle might lie in the non-diversifiable risk stemming from the financial system.

The additional findings correspond to other analyses. CDS spreads not only depend on a firm's leverage and equity volatility but also on its other characteristics and the market conditions. The Bundesbank finds a rebate for the size of a financial institution which may be due to diversification effects, economies of scale or the fact that certain firms are simply too big to fail. Furthermore, market liquidity measured by the bid-ask spread plays an important role. Changes in spreads are largely driven by the abnormal return of companies' stock and, again, market liquidity.

The magnitude of risk propagation not only stresses the role of contagion as a determinant of CDS premiums but also highlights the need for macro-prudential supervision, since the failure of a single entity may threaten the whole system. Historically, financial regulation has concentrated on ensuring the stability of each individual financial institution and neglected the risk stemming from the financial system as a whole. Recent advances in the regulatory framework (Basel III, Solvency II) already address the issue of countercyclical buffers counteracting the overdrawn market movements arising from both negative and positive financial spill overs.

This analysis shows that systemic risk is an important risk factor in today's economic and financial system, arguing for a further step towards macroprudential regulation. The findings also serve as a warning to national and regional authorities which have not yet geared their regulation towards a macro perspective. In this kind of a regulatory framework, financial institutions may possibly over-invest, which is then particularly painful in the event of a crisis.

This study focuses on the threats of connectivity patterns between financial institutions, yet neglects the potential implications of the concentration of risks stemming from massively writing CDS.

Full paper

© Deutsche Bundesbank