Bank and Sovereign Risk: The Case of European Economic and Monetary Union

Abstract

This thesis consists of four self-contained but related papers trying to uncover different aspects of banking and sovereign risk in the member countries of European Economic and Monetary Union (EMU). From a methodological point of view, they all have in common the contingent claims model from the theory of finance, which is used to value call options on a stock.

The first paper, “Bank risk behavior and connectedness in EMU countries”, studies the structural differences in banking sector and financial regulations at country level to measure and analyze the banking sector risk behavior. Deviating from the current view, which in our opinion is excessively focused on Systemically Important Financial Institutions (SIFIs), we introduce a micro approach to emphasize the role of smaller financial institutions in build-up of risk. The paper starts with a discussion of the reasons that are needed to consider this choice. Contingent claims analysis model is employed to calculate the risk of individual banks which is then aggregated at country level. The remaining of the paper tries to highlight the information content of country level banking risk indices. It is shown that if banking sector risk is calculated at country level using a bigger sample of banks, it can provide a simple, convenient and intuitive forward looking risk measure. The risk measures differentiate countries based on the structural differences in their financial sectors and show strong correlations with national and regional market sentiment indicators. They outperform the regulatory risk measures based at the European level and the causal linkages run from them to the latter indicators, suggesting better information content. And even though they have high correlations, causality and connectedness tests reveal no systemic component.

The second paper, “Sovereigns and banks in the euro area: a tale of two crises”, attempts to quantify the directional intensity of sovereign-bank linkages in the euro area countries. To this end, we borrow the indicator of banking sector risk in each country from the first paper, and use a traditional measure of sovereign risk (10-year government yield spreads over Germany). The paper starts with the review of channels via which banks and sovereigns are linked in a vicious cycle. We apply a dynamic approach to testing for Granger causality between the two measures of risk in each country, allowing us to check for episodes of significant and abrupt increase in short-run causal linkages. The empirical results indicate that episodes of causality intensification vary considerably in both directions over time and across the different EMU countries. The directionality suggests the presence of causality intensification, mainly from banks to sovereigns, in the crisis periods. Our findings also present empirical evidence about the existence of an adverse feedback loop between sovereigns and banks in some euro-area countries.

The third paper, “Incorporating creditors' seniority into contingent claim models: Application to peripheral euro area countries”, develops and uses a seniority structure of sovereign's creditors to analyze the impact of sectoral distribution of debt on the sovereign credit risk. Specifically, this paper highlights the role of multilateral creditors (i.e., the ECB, IMF, ESM etc.) and their preferred creditor status in explaining the sovereign default risk of peripheral euro area (EA) countries. Incorporating lessons from sovereign debt crises in general, and from the Greek debt restructuring in particular, we define the priority structure of sovereigns' creditors that is most relevant for peripheral EA countries in severe crisis episodes. This new priority structure of creditors, together with the contingent claims methodology, is then used to derive a set of sovereign credit risk indicators. In particular, the sovereign distance-to-default indicator, proposed in this paper (which includes both accounting metrics and market-based measures) aims to isolate sovereign credit risk by using information from the public sector balance sheets to build it up. Analyzing and comparing it with traditional market-based measures of sovereign risk suggests that the measurement and predictive ability of credit risk measures can be vastly improved if we account for the changing composition of sovereigns' balance sheet risk based on creditors' seniority.

In the last paper, “Revisiting the sovereign-bank linkages: Evidence from contingent claims analysis”, we reconsider the sovereign-bank nexus as discussed in the second paper to check the robustness of our findings. Using the banking sector risk indicator developed in our first paper, together with the sovereign risk index build in the third paper we re-inspect the bank-sovereign linkages. We use three different statistical measures of interconnection based on principal components analysis, Granger causality network and Diebold-Yilmaz's connectedness index. We also compare our results with alternative specifications using existing market-based indicators of banking and sovereign risk. Our results suggest strong connectedness and co-movement between country-level banking and sovereign risk indicators. We also find evidence of an increasing role of idiosyncratic risk factors driving the evolution of all risk indices in the post-crisis period, thus supporting the “wake-up call hypothesis” that the sensitivity of financial market participants to fundamental differences increased during the crisis. Country-wise analysis of time-varying bi-directional linkages using dynamic Granger-causality suggests the development of a bank-sovereign doom loop in Spain corroborating for this country the findings of our second paper. Connectedness analysis also suggest that increasingly the risk is being driven away from market-based uncertainty to the idiosyncratic risk factors, which are better captured by the contingent claim based indices.