Sovereign Debt

This is a survey chapter, not an empirical paper: it takes one benchmark limited-commitment model of a small open economy and uses it as a common spine for the whole sovereign-debt literature, showing which branch of the literature each modification of the benchmark generates. The starting claim is that what distinguishes sovereign from private debt is not insolvency but enforcement – a firm is “at least technically, always subject to a legal authority,” a sovereign is not – so the sovereign’s option to walk away is modelled as a participation constraint that must hold at every history, and that constraint doubles as an endogenous borrowing limit. Before building the model the chapter assembles six empirical regularities it then holds the theory against: default recurs throughout history and in waves, and “graduation” to non-default status is extremely rare; default is more common in bad times but far from exclusively so (in Tomz and Wright’s sample of 175 countries output is on average 1.6 percentage points below trend at the start of a default, yet more than a third of their 169 episodes began with income at or above trend); creditor losses in restructurings are large and very heterogeneous (roughly 30 percent in Uruguay to over 60 percent for some Argentine and Russian bond series, averaging roughly 30-40 percent across the wider samples); renegotiation is slow, taking eight years on average across ninety episodes, with the median country leaving restructuring carrying a debt-to-GDP ratio 5 percent higher than at default; emerging-market spreads rise with maturity, co-move strongly with global factors, and the yield curve inverts while new issuance shortens during crises; and fast-growing economies are net exporters of capital, a pattern driven by government rather than private net foreign assets. Run through the benchmark, the model delivers a tight set of predictions: limited commitment impedes risk sharing and does so worst when debt is high; the efficient response is to back-load consumption, which means saving, so a patient sovereign eventually reaches full risk sharing while an impatient one never does; a large debt stock depresses and destabilises investment because capital makes walking away more attractive; and the participation constraint binds in high-endowment states, which the chapter is careful to say does not mean the model predicts “default happens in high-endowment states” – what it means is that borrowing is limited in bad times. Extensions then generate equilibrium default (add a shock to the outside option that lenders cannot see), costly delay in renegotiation (drop state-contingent assets, add hold-out incentives), self-fulfilling rollover crises (Proposition 1’s crisis zone, following Cole and Kehoe’s timing), and the quantitative Eaton-Gersovitz models of Aguiar-Gopinath and Arellano. The chapter’s own verdict on that quantitative literature is candid: it works, but “often relying on ad hoc assumptions that restrict equilibrium objects such as financial contracts and the output costs of default,” and it lacks both microfoundations and a coherent theory of equilibrium selection.

Mark Aguiar | Macro Paper Warehouse

Sovereign Debt