Copenhagen 2012

1. The European Sovereign Debt Crisis: Background and Perspectives Copenhagen 2012 What is the Risk of European Sovereign Debt Defaults? Fiscal Space, CDS Spreads and Market Pricing of Risk by J. Aizenman, M. Hutchison and Y. Jinjarak Comments by Marcus Miller University of Warwick

2. What kind of market,friends, is this?Fundamentally driven; or fancy free?

3. AHJ analyse the sovereign debt component of the (largely OTC) CDS market This amounts to $2.5 trillion in 2010 , which exceeds the total of US government-issued international debt ($2.2 tr) and US GDP ($1.5 tr)! Regression analysis is used to relate sovereign spreads to fundamentals for 50 countries, over the period 2005-2011 for 3, 5 and 10 year CDS’ with the focus on the SWEAP* group in particular. The fundamentals include two measures of ‘fiscal space’: debt to tax base and deficit to tax base. *South-West Eurozone Periphery

4. All-too- brief summary of findings After a fascinatingh and well-documented empirical analysis, the authors conclude that: • ‘fiscal space is validated as an important determinant of market-based sovereign risk’, i.e. the fiscal fundamentals are statistically significant, but • there are large prediction errors in pricing SWEAP risk, with under prediction before the crisis and over prediction during the crisis. Two explanations of these prediction errors are suggested: • Market failure (‘excessive pessimism and overreaction’) • Expected future fundamentals.

5. 3 Comments • A. FF are treated as exogenous, but -if the deficit includes debt servicing costs-are likely to be endogenous. Is there not a risk of bias? • B. Is there not a possibility of Type II error? The Maintained hypothesis is ‘spreads reflect current fiscal fundamentals (FF)’: what of Alternative hypothesis that ‘there are multiple equilibria’, Calvo (1988)? • C. Need to add strategic aspects? • Games between financial intermediaries. • Interaction between markets and government.

6. Caveats • Definition of Fiscal Fundamentals (debt and deficits) not too clear from paper, in particular • Fiscal deficit: is this primary deficit, or does it include costs of debt service? (Note that text treats both fiscal fundamentals as ‘risk increasing’ (p.13); but in regressions it’s the fiscal surplus that is used, hence negative sign.) • No explicit account is taken of government action and reaction at national or supranational level. (If one is considering sovereign default, should variables not be included to reflect political economy considerations?).

7. SS Actual spread Is there a bias because actual market spreads affect fiscal fundamentals? Error vector Predicted spread, FF αFF SS= ∝FF+ ε If ε positively correlated with fundamentals, OLS estimate of ∝ will be biased upwards.

8. g+bRb x Endogenous Fiscal Fundamentals: Calvo’s self-fulfilling crisis for solvent sovereign (E0 -> E1) Partial default Full default x* g+αbRb E1 E0 No default g Government expenditure including cost of debt service at the safe rate R ‘Optimal’ level of tax chosen by govt Safe rate Risky rate Rb R Government expenditure including legal costs of default Market Rate, including sovereign spread

9. Θ Market and government: Expected and actual default - Calvo revisited 1/(1-α) 1 Full Partial default Default : actual and expected Government Reaction [Θ = rate of default] Θ1 R 1 No Rb 1 Market expectation 1 Θe

10. Spreads Alternative hypothesis - multiple equilibria (Calvo) Maintained and Alternative Hypothesis Maintained hypothesis (Aizenman et al.) β FF Good Bad - insolvent ME SS= ∝FF+ ε

11. Under the alternative hypothesis, there is of course the problem of ‘equilibrium selection’. Question: when will the market switch from the safe rate to one which reflects the partial default that it causes? One idea is ‘sun spots’, i.e. exogenous randomness. Another is that the switching reflects mixed strategies in a market game.

12. A CDS ‘market game’ with mixed strategy equilibrium

13. Frequency distribution of outcomes

14. Conclusion The authors are effectively testing the efficiency of the CDS market in measuring risk of sovereign default. The results for 2008-2010 imply that sovereign default risk was over-priced by a factor of 2.5 to 5 times (fundamentals greatly under-predict the actual spreads charged in the market). The authors suggest that market price risk ‘follows waves of contagion’. Are they not verifying Calvo’s model of multiple equilibria?

15. Appendix Numerical Example of Calvo Model • For the case of a country like Italy where the ratio of government debt to GDP is about 100%, we choose the parameters such that the critical value of market rates, above which the government will choose to default, is 7%, and the time consistent equilibrium with default is 20%. This implies α is about ¼ - i.e. You save ¾ by default! Parameters: b=1, g=0.38, R=1.03, α=0.235, c=0.684 • For a country like Germany, where the ratio of government debt to GDP is close to the Maastricht criteria of 60%, we assume that the interest rate is 3% (with no default). If so, it turns out that the government will not choose to default until market rates reach 10%; and the time consistent equilibrium with default turns out to be over 30%.

16. Current Situation: A schematic outline. Tax as fraction of GDP g+rb g+rb “Italy” 0.45 “Greece” 0.42 0.03 0.07 0.38 g b (debt/GDP) 1 1.2 1.6 2.3 Italy on the edge of default. Greece over the edge? Debt Write Down

17. What about a country like Greece where debt is 160% of GDP? If the tax take was as for Italy, namely 45% of GDP, then the sustainable level of debt turns out to be much higher, if only interest rates stayed at German levels, but Greece collects a lot less tax than other European countries. If the tax take goes down from 45% to 42% of GDP, it turns out that debt is unsustainable, even at German rates and a write down of about 30% to reduce debt to 120% GDP is needed for sustainability. (See diagram above)