“There have always been risk spillovers in Europe”

Loriana Pelizzon is SAFE Professor of Law and Finance at Goethe University Frankfurt since October 2013. Previously, she was at the University of Venice, MIT Sloan School of Management and London Business School where she also earned her doctorate in finance in 2002. Her current research focuses mainly on issues related to the financial crisis, such as systemic risk, risk management, credit derivatives and contagion risk.  

You have conducted a lot of research on contagion effects among Eurozone countries. What are your latest findings?

I have a number of recent papers on this issue, each looking at different aspects. For instance, in Caporin et al. (2015), we look at the period from 2003 to 2013 and investigate whether the shock transmission between Eurozone states changed over this period. Clearly, in a monetary union with a single market, countries are heavily interlinked. When one has an infection, this will, most probably, propagate to the rest of the system. We show that there have always been risk spillovers in Europe. Sovereign bond yields were highly interconnected before the financial crisis. Markets were not really distinguishing between German or Greek government bonds. However, we observe a structural break in 2008, which changed the way through which risk is spread. This means that serious contagion effects did not just start with crisis developments in Greece, Portugal, Spain or Italy, but already with the Lehman default in 2008.

Can you explain whether contagion is based on fundamental economic relations between countries or more due to psychological effects?

We did not find evidence for an overreaction of the market. The spillovers we observe are caused mainly by the fact that some countries became more volatile. So, the relationship between countries did not change substantially, but the amount of risk that was spread became larger, due to the larger volatility in the countries where risks originated. In fact, the economic interlinkages, for instance between Greece and Germany, have declined rather than increased since the crisis. This was definitely beneficial. The amount of risk generated, especially by Greece, was so large that the overall damage would have been much larger if interconnections had been at pre-crisis levels.

In a related paper (Aït-Sahalia et al., 2014) you looked more into the details of how risk spreads among countries.

In this paper we analyze how much of a risk jump in a certain country will be propagated to other countries. We designed a model to capture the dependencies between countries and then studied whether contracts for credit default swaps (CDS) were priced consistently with the predictions of the model. As CDSs provide insurance against the risk of default of the debt issuer, their spreads provide key information about the market's view on the arrival rate of default events and, as such, about default probability itself and cross-excitation. The broad empirical implications of the model are reassuring due to their plausibility and realism: Ireland, Portugal, Spain and Italy are the main countries affected by events in Greece, with results varying in strength across the countries, France and Germany are affected to a lesser extent (see figure).

Impulse-response analysis: Estimated increase in annualized probability of default in several countries due to an initial exogenous shock in Greece. The shorter and thicker the arrows, the more exposed is a country to excitation by Greece.

In your paper, you suggest to use the knowledge about the different strengths of contagion effects to make policy interventions more effective. 

From an economic point of view, policy makers should intervene in those parts of the system that send out the largest shock waves. Our model provides policy makers with the information they need to design effective stabilization policies. It offers information on the extent to which the probability of bond defaults is reduced by preventing a risk jump in a certain country. Conversely, it also tells you how much risk you are taking on when you do not intervene. Thus, it can help to decide how to allocate limited resources for policy measures. With respect to the recent crisis, we can say that the funds for Greece were definitely well placed because the majority of risk jumps came from there. In comparison, the jumps in Italy or Spain were quite small. So, costs would have certainly been higher, if the money had not have been used for Greece. If anything, the measures taken may not have been sufficient.

In a third paper (Pelizzon et al., 2015), you look at the impact of several interventions by the European Central Bank (ECB). Which was the most effective?

In the period from 2010 to 2012, the ECB intervened in several different ways to stabilize markets and provide liquidity to the system: they bought government and corporate bonds (SMP), they gave money directly to banks (LTRO), they even tried to persuade clearing houses to reduce their margins. In fact, this was a huge problem. On top of overall rising default risks, clearing houses all around Europe increased their margins three- to fourfold during this time. Clearly, they wanted to be safe but, by doing so, they dried out market liquidity.

We analyzed which of these interventions was the most effective in terms of stabilizing market liquidity and reducing the connection between default risk and liquidity. We found that there was a structural break in 2012 when belief in the market came back. In our view, this was due to LTRO, which not only provided banks with liquidity but also affected demand and supply on repo markets. Banks no longer needed to go to the repo markets to get funds, they could directly go to the ECB. Thus, clearing houses had to take fewer risks and could lower their margins. 

What can you say on the state of markets today? 

Government bond markets have restarted and are doing well. However, we have a new phenomenon today with quantitative easing (QE). The ECB is buying bonds for 60 billion euros every month, the majority of which are sovereign bonds. This time, the objective is to restart the economy. But, of course, there are secondary effects. QE may create problems by reducing supply in the sovereign bond market and by providing too much liquidity to the equity market. If market participants turn to more risky assets than bonds, this could create a serious bubble. So, my current work is to include QE in these types of models, in order to see how this is affecting market liquidity and equity prices.

Questions by Muriel Büsser.

References

Aït-Sahalia, Y., Laeven, R., Pelizzon, L (2014)
“Mutual Excitation in Eurozone Sovereign CDS”,
Journal of Econometrics, Vol. 183, Issue 2, pp. 151-167 (SAFE Working Paper No. 51). 

Caporin, M, Pelizzon, L., Ravazzolo, F., Rigobon, R. (2015)
“Measuring Sovereign Contagion in Europe”,
SAFE Working Paper No. 103. 

Pelizzon, L., Subrahmanyam, M. G., Tomio, D., Uno, J. (2015)
“Sovereign Credit Risk, Liquidity, and ECB Intervention: Deus Ex Machina?”,
SAFE Working Paper No. 95.