The economics literature on sovereign debt builds on the seminal work of Eaton and Gersovitz (1981). The simplest formulation of their model has two key assumptions. First, the sovereign borrows from abroad to smooth exogenous income fluctuations. And second, sovereign borrowers cannot be forced to repay because of either sovereign immunity or the impossibility of credibly pledging collateral located within the sovereign’s own borders. The logical consequence of these two assumptions is that the sovereign will repay (and lenders will lend) only if the cost of default is higher than the value of the debt that needs to be repaid. Hence, the cost of default defines how much a country can borrow – its debt capacity.
The question then is: what is the cost of default that induces repayment? Eaton and Gersovitz’s original assumption was that countries repay to protect their reputation in international capital markets. But there are two flies in the buttermilk with this view. First, evidence suggests that reputational costs tend to be short-lived (Borenesztein and Panizza 2010, Gelpern and Panizza 2022, Mitchener and Trebesch 2022).Second, quantitative models of sovereign debt predict that reputation alone would imply debt limits close to zero (Uribe and Schmitt-Grohé 2017). One can try to reconcile the model with the data by introducing assumptions such as sovereign defaults causing exogenous ad hoc output losses (Arellano 2008).
There are other problems with this framing, such as the implicit irrelevance of law. Sovereigns can be, and are, sued and enforcement against them occurs (including in their own courts). And sovereign immunity has never been an insurmountable barrier (Schumacher et al. 2018). Defaults are lengthy and complex events that often lead to multiple restructurings (Trebesch et al. 2021) because of haircuts which are too small to restore debt sustainability (Reinhart and Trebesch 2014). Plus, perhaps paradoxically, conditional on a debt crisis, debt restructuring can be financially rewarding for bondholders (Schumacher and Andritzky 2021). Sure, it is difficult to sue and collect. But that is always the case, regardless of the defendant (Weidemaier and Gulati 2015).
These considerations led Rogoff (2022) to conclude that “[t]he most popular class of theoretical models, those building on Eaton and Gersovitz (1981) seminal reputation model of debt repayment, has limited practical relevance, despite decades of elegant generalisation and extensions”. In a new paper (Bolton et al. 2023), we focus on the disjunction between the worlds of sovereign debt economic theory and ground-level realities to identify more than 20 sovereign debt puzzles – some of them are well-known, some of them less so – and classify them into three categories: puzzles about sovereigns debt issuance, puzzles about the pricing of sovereign debt, and puzzles about sovereign debt restructurings.
The ‘too much debt’ puzzle
Seen through the lens of a model that determines the debt capacity of a country based on a willingness to pay constraint, many sovereigns borrow more than the canonical theory predicts. The ‘too much debt’ puzzle is actually two puzzles. The first is that countries borrow more than what is predicted by standard quantitative models of sovereign debt. As mentioned, reputation costs alone generate sustainable levels of debt that are close to zero. This is because, in a purely reputational model, the cost for the sovereign of losing its reputation is the inability to smooth consumption in the future through external borrowing and the welfare costs of consumption volatility are not very high (Lucas 1987). Even with ad hoc output cost of default, quantitative models tend to yield low levels of sustainable debt.
The second puzzle is that countries borrow more than is optimal from a long-term growth maximisation perspective. As Aguiar and Amador (2021) point out, higher levels of debt are associated with higher volatility and with debt crises, which contradicts the consumption-smoothing motive. Moreover, higher debt levels are not associated with higher growth (Gourinchas and Jeanne 2013), which contradicts the investment motive. Rather, the evidence is that countries characterised by high rates of GDP growth pay down their external public debt and accumulate foreign assets (Aguiar and Amador 2021).
So, why do countries borrow so much? The answer could be related to political economy considerations and to the political agency problems arising from the behaviour of self-interested politicians who are inclined to overborrow relative to what a social planner would do (Acharya et al. 2022, Collard et al. 2023).
Contract pricing puzzles
One of the most basic assumptions about efficient markets is that public information about financial deals will be priced in. One would expect, therefore, that differences in key contract terms would be priced. If investors worry about sovereign debtors too readily asking for the debt to be restructured, then bonds with a higher vote threshold should carry a higher price than those with a lower vote threshold.
This question of whether differences in contractual vote thresholds are priced by the market is perhaps the single most studied contract pricing question. But the results are perplexing. A number of papers find no pricing differentials, some find pricing differentials going one way for a subset of issuers and the other way for another subset, some other papers find the opposite, and yet other papers find the pricing effects seem to show up only very late in the game when a debtor is in deep distress and almost certain to need a restructuring (Becker et al. 2003, Eichengreen and Mody 2004, Bradley and Gulati 2013, Dottori and Bardozetti 2014, Carletti et al. 2016, Carletti et al. 2021, Chari and Leary 2021, Chung and Papaioannou 2021). When one looks beyond vote thresholds to the variation in terms such as trust versus fiscal agency structures, early redemption clauses, types of governing laws, whether one uses an arbitration clause, types of pari passu clauses and so on, the results of existing pricing studies get murkier still. Here, it is not even clear whether one gets any pricing of these contractual terms very close to default (Weidemaier 2008, Haseler 2012, Gulati and Scott 2013, Colla et al. 2017, Panizza and Gulati 2021;,Bradley et al. 2022).
If one talks to actual participants in the markets, the ones who draft the contracts and buy and sell the securities, the puzzle is magnified further. These participants, to put it mildly, scoff at the academic studies finding pricing impacts. What matters to them is whether the contract fits the ‘standard form’ or is ‘market’ (Gelpern et al. 2019). Small variations don’t matter for trading, except perhaps at the very end.
One answer, to go back to the Eaton and Gersovitz (1981) canonical model, might be that law does not matter – in which case, contract variation would not matter. But we know from real-world events that outcomes and payoffs are often impacted by contract terms.
Default variety
Most of the literature on sovereign debt assumes that just one type of default occurs: countries stop paying, and the haircut is assumed to be 100%. Until recently, this was also the case for the empirical literature, with defaults being classified as dichotomous events following a definition like that adopted by rating agencies. Consensus is growing that researchers should move away from a binary definition of default, as there are many different ways in which countries default and restructure their debts.
For instance, debt restructuring can take place after the default has happened or can be preemptive (Asonuma and Trebesch 2016). Restructurings also vary in their degree of coerciveness (Enderlein et al. 2012). Further, there are interim restructurings that fail to restore debt sustainability (and are thus part of longer default spells) and decisive restructurings that mark the end of a default spell. And then there are supposedly non-defaults, such as the case of Colombia in the 1980s, that actually had significant elements of reprofiling (Caselli et al. 2021).
Summing up
Global sovereign debt outstanding has surpassed the $70 trillion mark, with external sovereign borrowing by developing and emerging market countries amounting to more than $6 trillion and yet we still have large knowledge gaps in the way economists conceive of sovereign debt. To fill these gaps, we need to better ground economic theories in the institutional realities of how sovereign debt markets operate in practice. There are many areas where we have little more than anecdotes to go on. Important historical events have not been adequately investigated. Indeed, that holds true even when it comes to matters of recent history (e.g. the Brady deals, the European sovereign debt crisis, the IMF’s Sovereign Debt Restructuring Mechanism). Key players are retiring, and memories are fading. New paradigms will emerge from this closer institutional grounding, but this will require breaking out of artificial disciplinary silos and building theories that incorporate insights from historical, political, and sociological analyses of sovereign debt (e.g. Flandreau 2022). Marginal fixes to the Eaton and Gersovitz (1981) model have proved to be inadequate. It is time to abandon this failed hypothesis.
References
Aguiar, M, and M Amador (2021), The economics of sovereign debt and default, Princeton University Press.
Arellano, C (2008), "Default risk and income fluctuations in emerging economies", American Economic Review 98(3): 690–712.
Asonuma, T, and C Trebesch (2016), "Sovereign debt restructurings: preemptive or post-default", Journal of the European Economic Association 14(1): 175–214.
Becker, T, A Richards and Y Thaicharoen (2003), "Bond restructuring and moral hazard: Are collective action clauses costly?" Journal of International Economics 61(1): 127–61.
Bradley, M, and M Gulati (2013), "Collective action clauses for the Eurozone", Review of Finance 18(6): 2045–2102.
Bradley, M, I de Lira Salvatierra, M Weidemaier and M Gulati (2022), "A silver lining to Russia’s sanctions-busting clause?" Virginia Law Review 108: 326–42.
Bardozetti, A, and D Dottori (2014), "Collective action clauses: How do they affect borrowing costs?", Journal of International Economics 92(2): 286–303.
Bolton, P, U Panizza and M Gulati, (2023), "Sovereign Debt Puzzles", Annual Review of Financial Economics (forthcoming).
Borensztein, E, and U Panizza (2010), "The costs of sovereign default: Theory and reality", VoxEU.org, 6 May.
Carletti, E, P Colla, P, M Gulati and S Ongena (2016), "Pricing contract terms in a crisis: Venezuelan bonds in 2016", Capital Markets Law Journal 11(4): 540–55.
Carletti, E, P Colla, M Gulati and S Ongena (2021), "The price of law: the case of the Eurozone collective action clauses", Review of Financial Studies 34(12): 5933–76.
Caselli, F, M Faralli, P Manasse, and U Panizza (2021), "On the benefits of repaying", CEPR Discussion Paper 16539.
Chari, A, and R Leary (2021), "Contract provisions, default risk and bond prices: evidence from Puerto Rico", in P Penet and J F Zendejas (eds), Sovereign Debt Diplomacies, Oxford University Press.
Chung, K, and M Papaioannou (2021) , "Do enhanced collective action clauses affect sovereign borrowing costs?", Journal of Banking and Finance Economics 1(15): 59–87.
Colla, P, A Gelpern and M Gulati (2017), "The puzzle of PDVSA bond prices", Capital Markets Law Journal 12(1): 66-77.
Collard, F, M Habib, U Panizza and J C Rochet (2023) , "Assessing debt sustainability in the euro area", VoxEU.org, 27 Jan.
Eaton, J, and M Gersovitz (1981), "Debt with potential repudiation: theoretical and empirical analysis", Review of Economic Studies 48(2): 289–309.
Eichengreen, B, and A Mody (2004), "Do collective action clauses raise borrowing costs?", The Economic Journal 114(495): 247–64.
Enderlein, H, C Trebesch and L von Daniels (2012) , "Sovereign debt disputes: a database on government coerciveness during debt crises", Journal of International Money and Finance 31(2): 250–66.
Flandreau, M (2022), "Pari Passu lost and found: The origins of sovereign bankruptcy 1798-1873", INET Working Paper 186.
Gelpern, A, and U Panizza (2022), "Enough potential repudiation: economic and legal aspects of sovereign debt in the pandemic era", Annual Review of Economics 14: 545–70.
Gelpern, A, M Gulati and J Zettelmeyer (2019), "If boilerplate could talk: the work of standard terms in sovereign bond contracts", Law and Social Inquiry 44(3): 617–46.
Gourinchas, P-O, and O Jeanne (2013), "Capital flows to developing countries: the allocation puzzle", Review of Economic Studies 80(4): 1484–1515.
Gulati, M, and R Scott (2013), The Three and a half minute transaction: Boilerplate and the limits of contract design, Chicago: University of Chicago Press.
Haseler, S (2012),"Trustees versus fiscal agents and default risk in international sovereign bonds", European Journal of Law and Economics 34(3): 425–48.
Lucas, R (1987), Models of business cycles, Blackwell.
Mitchener, K J, and C Trebesch (2022), "Sovereign debt in the 21st century: Looking backward, looking forward", Journal of Economic Literature 60 (forthcoming).
Panizza, U, and M Gulati (2021), "Make-wholes in sovereign bonds", Capital Markets Law Journal 16(3): 267–99.
Reinhart, C, and C Trebesch (2014), "Sovereign-debt relief and its aftermath: The 1930s, the 1990s, the future?" VoxEU.org, 21 Oct.
Rogoff, K (2022), "Issues in the theory of sovereign debt and post-covid workouts", J Pol’y Mod. 44(4): 804-811.
Schumacher, J, J Andritzky (2021), "Bond returns in sovereign debt crises: The investors’ perspective," VoxEU.org, 18 January.
Schumacher, J, C Trebesch, H Enderlein (2018), "The legal cost of default: How creditor lawsuits are reshaping sovereign debt markets," VoxEU.org, 16 July.
Trebesch, C, J Meyer, C Reinhart, C von Luckner (2021), "External sovereign debt restructurings: Delay and replay," VoxEU.org, 30 March.
Uribe, M, S Schmitt-Grohé (2017), Open economy macroeconomics, Princeton Univ. Press.
Weidemaier, M, (2008) "Disputing boilerplate", Temple L. Rev. 82(1): 1–54.
Weidemaier, M, M Gulati (2015), "The relevance of law to sovereign debt", Annual Review of Law and Social Science 11: 395-408.
