Abstract This paper shows that a high degree of intermediate imports increases the costs of sovereign debt defaults, as suggested by Mendoza and Yue (2012). Using an updated data set, it corroborates recent default cost estimates and demonstrates that countries with high shares of intermediate imports to GDP experience significantly larger GDP contractions following default episodes, compared to those with a low intermediate-import share. The paper augments the Inverse Propensity Score Weighted Regression Adjustment (IPSWRA) approach of Kuvshinov and Zimmermann (2019) with Smooth Transition Local Projections (STLP) in the style of Auerbach and Gorodnichenko (2012) to estimate state-dependent impulse responses, with the regimes determined by various trade-related variables. The results appear robust to various sensitivity checks and seem to persist and increase in the long run. This distinguishes the state-dependent results from the linear results, as the latter indicate that default costs are a short-run phenomenon.
Hinrichsen et al. (Wed,) studied this question.