Naoise Metadjer

BBS, M.Econ.Sc.

PhD student NUI Galway

Naoise completed a Bachelor of Business Studies with French in the University of Limerick. He majored in economics and finance, graduating with a first class honours. He then completed a Master of Economic Science in International Finance at NUI Galway, graduating with a first class honours. It was while completing his master’s thesis that Naoise first developed an interest in his current research topic. While researching interactions between Eurozone sovereign bond and CDS markets Naoise found that the standard models of financial econometrics could not fully explain the dynamic behaviour of these markets. In 2010 he began a PhD in economics at NUI Galway; the main scope of his research being to examine the dynamic behaviour of sovereign bond and CDS markets during the Eurozone sovereign debt crisis using both linear and nonlinear methods. Naoise also has a strong interest in teaching. He has taught tutorials and lectures for macroeconomics, microeconomics, financial economics, econometrics and statistic at both undergraduate and master’s levels.

Dynamics of régime shifts in sovereign bond and CDS markets

The aim of this study is to examine the dynamics of the Eurozone sovereign debt crisis with an aim to better understand the large sudden increases in bond yields and CDS premia experienced by a number of Eurozone member states. The crisis is a funding crisis; investors viewed certain Eurozone countries as having an increased probability of defaulting on their debt and charged an increased risk premium for taking on new debt. In Greece, Ireland and Portugal this led to unsustainable interest rates and assistance from the Trioka of the ECB, the IMF and the European Commission. These increased risk premia captured in increased bond yields and CDS premia. When sovereign bond yields initially began to increase in mid-2009 to early 2010 the narrative was that speculation in the CDS market was causing an increase in bond yields. We therefore examine cointegration in between the markets using vector error correction models, and linear and nonlinear Granger causality testing. We find that the markets are strongly cointegrated in all cases with bi-directional causality indicating both markets were equally reacting to new information and developments in one market were not driving changes in the other.

There is a body of literature which shows that the standard tools of financial econometrics (such as ARMA, ARCH and GARCH models) do not capture the entire dynamical structure of financial time series. GARCH models in particular are very useful for capturing persistence in volatility, which is a stylised fact of financial time series. However, the sovereign debt crisis is in effect a régime shift in debt markets of the countries affected, from a situation where risk premia were low and stable to a sudden transition to very large and volatile risk premia. The GARCH class of model is unable to capture régime shifts and is not useful in predicting such crises.

Bond and CDS markets are forward looking in that their prices are formed based on investors’ views of the probability of a debt issuers defaulting in the future. It may therefore be possible to detect changes in the statistical properties of time series of risk premia which could provide early warnings of future regime shifts. There is a body of literature in the ecological sciences which models regime shifts as a catastrophic bifurcation. In this literature changes in the spectral properties of a time series can be an indication of critical slowing down, which is a decrease in the rate of return of a system to equilibrium after a small perturbations. Critical slowing down has been found in many real world systems prior to regime shifts. If it is detected prior to the sovereign debt crisis it will not only augment our understanding of how crises occur; it can also be used as an early warning signal for future crises. This methodology requires a long time series prior to the regime shift as critical slowing down typically begins far from the bifurcation point. CDS were not traded often enough before 2008 to provide us with a sufficiently long time series. We therefore use bond yields to search for evidence of critical slowing down. The strong cointegrating relationship between the premia in each market means that critical slowing down in the bond market may be used as an early warning signal for future developments in the CDS market.

Keywords financial crises, régime shifts, bond markets, credit default swaps, nonlinear dynamics, critical transitions


PhD research funding

  • IRC Government of Ireland PhD fellowship
  • NUI Galway Hardiman research scholarship

Head of Economics

Prof Alan Ahearne

Contact Economics

t:+353 91 492501/492177