Sovereign Debt and Default: A Question of Ownership
Author: Nicolai Ellingsen, ESOP Scholarship Recipient 2014.
In this thesis I present a model featuring sovereign default in equilibrium. The government can choose a default rate between zero and one and the ownership structure of the debt matters for the decision. My main motivation is to propose an explanation of why the share of sovereign debt held by foreign investors has decreased in most countries in the European Monetary Union (EMU) since the beginning of the financial crisis. In the model decreasing foreign share is a rational response of expected utility maximizing investors when faced with increased risk of default. The model is also able to explain the negative correlation between bond yields and foreign share of bondholders found in the data. The empirical motivation for the thesis is the events in Portugal, Italy, Ireland, Greece and Spain (PIIGS), as those are the countries most affected by the crisis. From 2007 to 2011 the average foreign share of debt holdings in those countries decreased by 11 percentage points. The PIIGS countries also experienced the largest increases in interest rates on their sovereign debt. In light of the low and stable bond yields and increasing shares of foreign bond holders that characterized the beginning of the millennium, this is a dramatic change and an interesting topic for research.
Sovereign debt makes up almost one fifth of the world’s financial assets and has become an important interest rate benchmark and investment option. Before the financial crisis the trade of European sovereign debt across boarders increased, probably because of increased financial integration. Government debt obligations issued by advanced countries were viewed as a safe investment and credit ratings were strong. A lot has changed since 2007. European sovereign debt markets have been in turmoil for most of the time between 2008 and 2014. Yield spreads have been huge for many countries and credit ratings are poor. Debt obligations have changed hands at a rapid rate and are increasingly sold back to investors in the issuing country. It is this last development that is the main focus of my thesis. Little is known about the determinants of - and impact on - the ownership structure of sovereign debt. Some new empirical work show some facts and trends, but little theoretical work has been done on the subject. I suggest that the reduction in the share of foreign investors holding sovereign debt is a rational response to sovereign risk. The reduction is analyzed as a home bias in sovereign bonds only present when there is risk of default. Much of the home bias literature introduces trade costs or other market frictions to explain observations in the data. My home bias mechanism is not dependent on market failure or frictions, but is only applicable on sovereign debt. I build my model following a long tradition of modeling sovereign debt II as a non-enforceable asset with penalties for defaulting.
My model differs from the main literature in that it focuses on the ownership structure of the bonds. It also allows for partial default, where most models focus on the binary choice of default or full repayment. The model contains foreign and domestic investors in addition to a domestic government that issues debt. The home and foreign investors act as expected utility maximizers. All agents live for two periods and the second period income for the home investor is uncertain. The government’s debt is not enforceable and creditors risk default in equilibrium. The interest rate on government bonds is determined in the market and depends on expectations about government behavior. Debt is repaid with tax revenue in the second period and the government chooses the default rate by maximizing the utility of their population. Domestic investors incur a loss if the government chooses to default, the size of the loss depends on the default rate and the realization income. The decision is thus a trade-of between tax-gains and default-penalties. My analysis consists of varying the risk level by changing the income distribution and debt level and then comparing the results. Increased risk leads to higher interest rates and lower share of foreign bond holdings. Specifically I find that decreased expected GDP or increased debt has about the same effect. Both increase interest rates and decrease the foreign share by about the same magnitude. Increased probability of recession on the other hand has a much larger effect on interest rates and less effect on the foreign share than the first two. The mechanism at work is related to the government’s default decision. The government finances repayment of the bond by taxing their inhabitants.
Domestic investors observe that they get a lower tax in case of default and is thus partly "insured" against default risk. This effect is increasing in the risk level and induces domestic investors to take advantage of the increased interest rate to buy more of the bond. Foreign investors have income that is uncorrelated with the return on sovereign debt and thus only care about expected return and the risk level. The decline in foreign bond holdings is thus a result of increased risk. Following directly from this is also the proposal that there is no causal relationship between yields and the foreign share, both have a common driver; the default risk. I also propose mechanisms that might have contributed to the outbreak of the sovereign debt crisis. The model can describe a situation with multiple equilibria where the good, low risk, equilibrium is fragile both to expectations and the share of foreign bond holders. Expectations of default could increase interest rates and make default optimal for the government. Similarly, an increase in foreign bond holders makes default more likely, because the government only cares about domestic investors. Solution of the model and the analysis is carried out in Matlab.