1Europe’s debt crises of 2009 onward raise at least two challenges for economics. First, the idea of a “graduation” to invulnerability to default1 looks false. Today’s advanced economies may default less frequently than during earlier centuries or than many developing countries today, but with large enough debt burdens, even they get into trouble.

2Fundamentals, in other words, still matter. Second, democratically elected governments seem extremely keen to avoid default, even if this implies considerable costs for their citizens.2 This observations points to politics as an important factor in determining whether a country that experiences difficulty in repaying its debt will actually default.

3Politics can only become relevant for the question of default because sovereign states are fundamentally different from private companies. Whereas creditors to private entities usually have legal means to enforce repayment, lenders to sovereign states cannot easily sue for compliance or seize the sovereign’s assets in case of default. In international law the principle of sovereign immunity prevents one state to be sued in the court of another3. This conundrum has inspired a large literature trying to explain the sustainability of lending to foreign sovereigns. Accounts have focused on different forms of economic sanctions that will be triggered in the case of default, with threats of sanctions preventing countries in difficulties from defaulting on their debt. Domestically, the legal means available to creditors of sovereign entities are likely even more limited. Sovereign states both make the law and enforce it and are thus, a priori, able to renegue on ex ante promises they have made by changing the law arbitrarily ex post. Therefore the repayment of sovereign debt depends not only on the ability of states to generate sufficient revenue to meet their obligations, but also crucially on the willingness to do so.4

4If, as the literature suggests, sanctions are supposed to ensure the willingness to repay, targeting them matters. Assuming people look out for their interests, punishment and threat of punishment will be most effective in deterring defaults if their intended target expects a large reduction in welfare in case the punishment is triggered. Sanction mechanisms will be less effective if they leave their intended target’s welfare essentially unchanged in expectation.

5Taking as an example the “threat of exclusion from international capital markets” argument, suppose countries as a whole value the ability to borrow because having this option is welfare-improving.5 Consequently, countries as a whole should be unwilling to default if the price of non-repayment is losing access to sovereign borrowing. Countries, however, do not make decisions. Government officials do. What might be useful for a country as a whole may or may not be so for the individuals making decisions.

6What then motivates governments to repay debt? Some authors posit politicians “idiosyncratic desire to repay” (McGillivray and Smith, 2003) or the weight a politician puts on social welfare (Borensztein and Panizza, 2008) as part of the story. This seems unsatisfactory. The natural assumption to make in an economic model is that self-interested politicians care about exercising power, i.e. staying in office as long as possible.6 Achieving this goal requires maintaining the loyalty of a sufficiently large group of political supporters. Whether threats of punishment will be effective in deterring default will therefore depend on whether the economic cost of lost borrowing opportunities or some other sanction will cost the government the loyalty of its supporters. If, following a default and the ensuing punishment, governments are certain to command the loyalty of a large enough group of supporters and therefore to remain in office, defaults should occur more frequently than if governments can be certain to be ousted.

7What is crucial therefore is whether and how economic sanctions following a default might cost the government the loyalty of its supporters. My hypothesis is that this hinges on a country’s institutions. Institutional mechanisms determine how economic sanctions that hit entire economies translate into political costs for countries’ leaders. The functioning of these institutional transmission mechanisms determines whether sanctions, such as investors shunning a country’s sovereign bonds, are effective in deterring frequent defaults. Analysing this transmission mechanism will therefore help explain why some countries default more frequently than others, why some countries default at debt levels that pose no apparent problem for others, and why some countries try to avoid default at all cost.

8To analyse the problem of incentive compatible sovereign debt repayment more formally, I draw on the Selectorate-theory, advanced by Bueno de Mesquita et al. (2003). Rather than conceptualizing governments in terms of categorical regime types such as “democracy” or “monarchy”, its authors define regimes along two institutional dimensions. The Selectorate{S} is the set of people who have at least a nominal say in choosing their leaders. In a democracy {S} coincides with the electorate. The most important aspect of belonging to {S} is that membership conveys the opportunity of belonging to the Winning Coalition{W}. Formally, {S} is defined as:

the set of people whose endowments include the qualities or characteristics institutionally required to choose the government’s leadership and necessary for gaining access to private benefits doled out by the government7

9The support of a subset, {W}, of {S} is needed to keep the incumbent in office. Again formally, {W} is:

a subset of {S} of sufficient size such that the subset’s support endows the leadership with political power over the remainder of the selectorate as well as over the disenfranchised members of society8

10With these definitions in place, I build a theoretical model of the political economy of sovereign borrowing. The central insight is that governments maintained in power by a small set of political supporters drawn from a large group are more likely to default on sovereign debt than governments dependent on a large Winning Coalition. Coalitions tend stay loyal to a defaulting government in the former, rigged-election autocracy, case. Coalitions in the latter, usually democratic, case tend to abandon their governments if they default. This mechanism keeps democratic governments “honest” and allows popularly elected autocrats the discretion of more frequent sovereign defaults.

11I then proceed to test these theoretical predictions using data assembled by Bueno de Mesquita et al. using a range of dependent variables and economic controls. Results are somewhat mixed, but point in the direction of the expected effect for Selectorate size, in particular for external debt. Using a threshold-estimator proposed by Hansen (1999), I find the expected effect of institutions on interest rates for countries below a 61% threshold of public debt to GDP.

12Overall, my findings support the contention that power politics plays a role in sovereign defaults. The rest of the thesis is organized as follows. Section below surveys the literature. Section 1 develops the theoretical model. Section 2 tests the model, and the last Section concludes.

Literature review

13Why do private investors lend to governments without binding laws that ensure repayment of sovereign debts? And why do governments repay? Eaton and Gersovitz (1981) model countries facing macroeconomic shocks and valuing access to the consumption smoothing possibilities of external borrowing. To maintain access to debt, countries must never default. If shocks are sufficiently large and borrowing opportunities compensate enough of this, sovereign debt is repaid. Bulow and Rogoff (1989) maintain that reputational concerns alone will not deter defaults, since countries would otherwise build up debts beyond the point where the cost of repayment exceeds discounted benefits from future borrowing and then default. Instead they argue that even if creditors of defaulting governments might not be able to seize assets in the sovereign’s own country, they might be able to do so abroad. As a further cost, creditors may try to cut off defaulting countries involved in international trade from the short-term financing needed for these transactions. The combined threat of these punishments then ensures repayment.

14Turning to political economy models of debt, Drazen (1997) considers a two-period endowment economy where individuals, who differ in wealth, aggregate their preferences to determine degrees of repudiation on domestic and foreign debt. Individuals with above average wealth prefer higher interest including repudiation on domestic debt and more domestic borrowing, since interest paid on domestic public debt is a tax-financed transfer to them. Poorer individuals prefer payment on domestic debt below the world interest rate and more foreign borrowing. Countries facing unexpected financing needs raise taxes and repudiate more domestic and foreign debt, with the weights given to each dependent on income distribution, mode of preference aggregation and punishments attached to repudiation of foreign debt.

15In the economic history of sovereign debt default, a key contribution is North and Weingast (1989). The authors interpret the Glorious Revolution of 1688 as a watershed for British government borrowing and propose that the political constraints established by constitutions make governments more credible as sovereign borrowers, since they allow credible commitments to repay. Stasavage (2007) argues that the true sea-change in British sovereign borrowing occurred not in 1688 but in 1715, at the onset of the parliamentary supremacy of the Whig party, which represented the interests of sovereign lenders in the legislature. Consequently, it were not constitutional constraints that enforced sovereign respect of property rights of lenders. Rather, a political coalition of creditors made commitment credible through effective exercise of voting power in parliament.

16Borensztein and Panizza (2008) propose a model in which the likelihood of defaults depends on the altruism of leaders. Politicians less concerned with public welfare are more likely to delay default, since defaults entail the possibility of losing office. In my model I will challenge their conclusions by arguing that the likelihood of deposition following a default depends on institutions, while leaders only care about power. The conclusion is quite different: leaders, who need not worry about losing political support and thus able to behave more selfishly, will default more often. Interestingly, these authors find that defaults carry, contrary to the surprising absence of large interest-premiums and prolonged periods of capital market exclusion, large political costs: Leaders and top bureaucrats tend to lose their positions following defaults. Defaults on bond debt appear to entail larger political costs in democracies and defaults on bank debts are more costly in dictatorships.

17The most direct inspiration for my thesis is McGillivray and Smith (2003). The authors build a political economy model of default and argue that leaders who are easily removed from office will default less often. They also draw on the data assembled Bueno de Mesquita et al. (2003) for the empirical portion of their paper. They do not, however, explicitly incorporate the Selectorate theory into their model to show why defaulting leaders facing certain institutions are more likely to be removed from office than others. This is what I attempt in the theoretical portion of my thesis. I show that the Selectorate yields a natural explanation of the political economy of default and thereby gives a direct justification for using Bueno de Mesquita et al.’s data to test the hypothesis that constraints on executive discretion influence sovereign borrowing behaviour.