Sovereign defaults that end up in a credit crunch are distinguished by a deep drop in output and protracted recovery. In these events, most of the public debt is issued through domestic markets and a significant share of domestic banks’ assets are government bonds. This paper studies how the exposure of the banking system to government debt accounts for the dynamics of investment and output during a default followed by a contraction in domestic credit. I develop a quantitative model that features capital accumulation, financial intermediation and endogenous sovereign default. Banks invest in capital and buy bonds issued by a benevolent government. They are financially constrained to issue deposits by the value of their net worth. In a sovereign default banks’ investment in capital drops as their net worth decreases. I calibrate the model to match the fraction of banks’ assets held as government bonds, the mean investment to GDP, and investment volatility for economies that experienced distress in domestic credit after a default. The model is able to reproduce the untargeted observed dynamics of output, investment, consumption, deposits, and bank’s assets around default events. I use the model to illustrate the trade-off that government debt held in the banking sector provides between the ex- ante incentives to default and ex-post cost of default. An increase of 50 % in the share of bonds to banks’ assets decreases the probability of default, but increases the volatility of investment to GDP by 23% and reduces the level of investment to GDP by 18%. With a lower capital stock, the ability to insure the economy against productivity shocks lessens, and the volatility of consumption relative to GDP increases by 13 %.
Contagion, sovereign debt, and non-fundamental risk
During the European sovereign debt crisis of 2011, several countries faced higher interest rate spreads, despite very different patterns of underlying fundamental shocks. In addition, governments’ interest rate spreads were highly correlated even when they faced different dynamics in their output. For instance, Italy had difficulty rolling over a large stock of debt, which was followed by sharp increases in interest rate spreads in Spain, Portugal and Ireland. Eventually, interest rate spreads decreased once the ECB intervened to alleviate Italy’s debt roll over problems. In this episode, Italy had problems rolling over its debt even when its stock of debt was sustainable. The episode showed that interest rate spreads are highly correlated and non-fundamental shocks can be transmitted across countries. In this paper, I build a quantitative model of sovereign default with two countries that share a risk averse investor. The assumption of a risk averse investor allows the existence of correlation between the interest rate spreads of different countries. In the model, an investor has a portfolio of sovereign debt in two countries. The value of the portfolio is negatively correlated with the price of the bond issued in each country. In equilibrium, given the risk aversion, the bond price of the marginal investor is positively correlated with the value of the investor’s portfolio. Therefore, bad states that increase the spread of government bonds in one country can affect the bond price of the other country. I calibrate the model to match the moments of Spain and Italy. The model is able to match 90 % of the untargeted correlation of the interest rate spreads observed between Italy and Spain. I extend the model to include non-fundamental risk to default. By introducing a probability of 1 % to states in which the government defaults if lenders are not willing to roll over its debt. In this extension I find that 5.1 % of the correlation reproduced by the model can be associated to non-fundamental risk.