Skip to Main content Skip to Navigation

Sovereign risk exploration in times of crisis : a look at financial contagion

Abstract : Periods of deep risk aversion are usually marked by sizeable distortions in market prices, and substantial losses in portfolios. As observed during financial crises, a generalized debacle in financial markets is a very negative shock for the real economy. Against this backdrop, it looks relevant to explore how risk aversion tends to affect global market valuations, especially if this exercise helps make the promotion of more optimal portfolio rebalancing procedures.In this dissertation, we investigate different dimensions of risk aversion, with a focus on European Sovereign debt securities. For a given sovereign bond, the (quoted) yield to maturity has to reflect the underlying risk that the Treasury may default on its debt, before maturation of the bond. This is sovereign risk. Financial crises usually occasion an upward correction in bond yields. Since higher yields reflect larger sovereign risk and higher funding costs, national Treasuries are usually inclined to get a deeper understanding of how sovereign risk could evolve under the influence of fierce risk aversion. This is another objective of our research analysis.In Chapter I, we consider a probabilistic approach to sovereign risk exploration, with the main purpose of illustrating the non-linear reaction ensuing from a gradual deterioration in market sentiment. We consider heavy-tailed distributions, and we use the Generalised Autoregressive Score method as a means to capture the volatility momentum. The goodness of fit provided by Generalised Hyperbolic distributions is compelling, and results suggest that our approach is particularly relevant to fit periods or erratic volatility, typical of financial crises. As an attempt to simplify the model, we focus on an empirical formulation of the ‘untemporal’ volatility of each security. This estimator of the intrinsic volatility suggests that volatility tends to accelerate in a quadratic manner when it is expressed against the cumulative distribution function of the yield variations. In a second part, we extend this approach to a problem of larger dimension and we explore the dynamics of risk aversion from a bivariate point of view. Results look robust and illustrate multivariate correlations between sovereign securities. As a general conclusion, heavy-tailed distributions look remarkably efficient to replicate the distribution of times-series affected by distorted volatility and erratic price variations.Chapter II explores different ways to extract information from the model, about financial contagion and how it is supposed to propagate through sovereign securities. In particular, we explore the market reaction to a series of many shocks with gradual intensity. Results offer a high degree of granularity and we extrapolate empirical rules on the expected market dynamics, when risk aversion intensifies. Then we incorporate our estimators of volatility and market reaction (to shocks) into popular portfolio optimisation procedures and we see positive implications on the general resilience of these portfolios. Finally, we also design an in-house methodology for optimal portfolio rebalancing, based on mean reversion.In Chapter III, we explore how sovereign risk tends to affect the price of financial derivatives in a risk-off environment. We consider that risk aversion and the ensuing volatility now favour the emergence of sizeable discontinuities in market prices, that we model with stochastic jumps. The different approaches we investigate extensively rely on Hawkes processes. These stochastic processes seek to estimate the durable impact of risk aversion onto the dynamics of jumps, via the introduction of dedicated self-excited loops. We develop an original approach to the calibration, different from conventional procedures. In the end, the calculated implied volatility remains in the vicinity of the realised volatility and there is a visible capability to jump on any rise in risk aversion.
Document type :
Complete list of metadatas
Contributor : Abes Star :  Contact
Submitted on : Friday, April 27, 2018 - 12:33:16 PM
Last modification on : Monday, October 19, 2020 - 10:55:27 AM
Long-term archiving on: : Thursday, September 20, 2018 - 1:10:38 AM


Version validated by the jury (STAR)


  • HAL Id : tel-01780245, version 1


Marc-Henri Thoumin. Sovereign risk exploration in times of crisis : a look at financial contagion. Economics and Finance. Université Paris sciences et lettres, 2017. English. ⟨NNT : 2017PSLEM039⟩. ⟨tel-01780245⟩



Record views


Files downloads