Research

Working Papers

SOVEREIGN DEFAULT RISK CONTAGION IN THE EUROZONE AND THE EFFECTS OF A EUROBOND ON THE VOLATILITY OF SPREADS - Job Market Paper

ABSTRACT : Sovereign bond yields among countries in the European Union converged after 1999 along with lack of exchange rate risk, unified monetary policy, and a common currency. However, after the 2007 U.S. financial crisis, the sovereign bond spreads of Eurozone countries like Greece and Italy started to sharply diverge. This divergence was likely driven in part by differential default risk of some countries affecting them more adversely than other EU member nations. At the same time, capital flow reversal to safety lowered available capital in the regional banks of Portugal, Italy, Ireland, Greece, and Spain (PIIGS). It has been proposed that the resulting reductions in international liquidity also may have been a driving factor in the increased sovereign bond spreads for these nations leading up to the 2010 sovereign debt crisis.

I empirically document the effect of differences in EU nations' exposure to international capital flows at the onset of the U.S. financial crisis on their sovereign bond spreads. Specifically I show that, during the 2010-2013 sovereign debt crisis, higher claims of non-residents in the Euro area predict higher government bond spreads for PIIGS compared to other members. I next develop a quantitative sovereign default model to accommodate this phenomenon and explore whether introduction of a safe asset might mitigate the rise in sovereign bond spreads in times of global financial crisis. My sovereign default model allows a country borrowing from a financial intermediary to default on its outstanding debt obligations, and thus has an endogenous probability of default. The financial intermediary receives deposits from abroad that it can lend to the domestic government. However, the intermediary faces a capital requirement limiting the amount of such loans to the borrowing country. As a result, the loan discount factor on government debt is influenced by both default risk and liquidity risk, and either or both may rise in response to a real shock inside the country or a fall in international deposits.

My paper is the first to structurally quantify the joint roles of reduced liquidity and endogenous movements in default risk in exacerbating sovereign debt crises. I calibrate my model to match a core set of moments involving Italy. Next, I simulate business cycles in my quantitative model and confirm that it predicts a strong correlation between financial crisis abroad and sovereign bond spreads. Finally, I use it to conduct policy experiments, examining the effectiveness of alternative stabilization interventions. I find that a safe asset like a Eurobond can reduce mean and volatility of spreads during crises and incentive grants program can reduce default possibilities by encouraging repayment.

TRADE INTEGRATION AND SOVEREIGN DEFAULT RISK CONTAGION

ABSTRACT : Recent empirical work indicates that strong cross-country business cycle comovements are, to a large extent, attributable to strong trade linkages, suggesting that tighter trade ties might also imply higher sovereign default risk contagion. However, there is little existing evidence to confirm or disprove that relationship. I empirically test the extent to which higher trade intensity implies higher comovement in default risk by estimating a GMM-IV regression using trade data from the International Monetary Fund and spreads data from the OECD online database. I estimate a Frankel and Rose (1998) type regression with the sovereign credit spread correlation as the explained variable, interpreting sovereign debt spreads as an indicator of sovereign default risk. I find that a one percent increase in bilateral trade intensity increases the bilateral correlation of credit spreads by about 0.22 percentage points, raising it from 0.5 to 0.72.

Given my empirical finding, I next develop a structural model to explore whether the introduction of a safe asset could reduce sovereign default risk correlation. I allow trade intensities to vary with trade barriers, extending the workhorse Backus, Kehoe, Kydland (1994) model to include iceberg costs representing these barriers, and I introduce endogenous sovereign default risk. My international business cycle model studies two small open countries that each can borrow from a large country. The lending country is deep-pocketed, risk-neutral, and prices in default risk on its loans. The two small countries are directly linked through their trade in intermediate goods, and so are linked in their default risk. A default in one small country lowers the demand for the other's exports, thus reducing that country's output and raising its likelihood of default. This then creates a positive correlation of spreads across two small trade-linked countries. I calibrate my model to match a core set of moments involving Spain and Italy and examine its simulated business cycles. Given an empirically consistent degree of trade intensity, my model succeeds in generating the observed correlation of credit spreads between these two countries.