Sovereign risk assessment: a practitioner's framework

Sovereign default happens when a government cannot pay, when it chooses not to pay, or (most commonly) when the political cost of paying exceeds the political cost of defaulting. Analysis must therefore address both capacity and willingness explicitly, rather than collapsing them into a single score.

Capacity is grounded in the fiscal accounts, external accounts and monetary policy space. Willingness is grounded in institutions, political economy and the sovereign's revealed preferences in past crises.

For capacity: gross public debt to GDP, gross financing needs to GDP, interest-to-revenue, average maturity, share of debt in foreign currency, external debt to exports, international reserves to short-term external debt, and the current account balance. For monetary space: inflation, the credibility of the inflation target, and the level of real rates.

For willingness: policy consistency across administrations, adherence to past IMF programmes, the credibility of the central bank's independence, and the history of restructurings or default. A single restructuring in the past 40 years is not disqualifying; a pattern of them is.

The output of the assessment should be comparable to market prices. Sovereign CDS spreads and eurobond yields imply a market-priced probability of default; if your assessment materially disagrees with that pricing, you have either a trade or a mistake in your model. Making the disagreement explicit forces the discipline to work out which.

Rating agencies provide a useful cross-check but a lagging one. A sovereign trading at a spread inconsistent with its rating is where opportunity (and risk) usually lives.

Translate the assessment into a portfolio position with explicit conviction, size, horizon and stop-out conditions. Sovereign trades are long-duration by nature; expect to hold the view for months or quarters and to revise it when specific data or political events change materially.

Keep a written record of the thesis, the key drivers you are watching, and the levels that would falsify it. Sovereign situations evolve slowly; that discipline is the difference between a repeatable process and a series of ex-post rationalisations.