Essays on Sovereign Risk

Abstract

This dissertation studies government debt and sovereign risk in emerging and developing markets. Chapter 1 studies local currency (LC) sovereign risk in emerging markets and finds that, unlike FC sovereign risk, the LC sovereign risk is insulated from global factors. Moreover, the discrepancy between FC and LC sovereign risk and the isolation of LC sovereign risk is not altered with higher foreign participation in the domestic market. An event study of the LC government bond inclusion into global indices suggests that higher foreign participation in the market suppresses the LC government bond yield and currency risk spreads, but LC sovereign risk stays invariant. Chapter 2 examines the currency choice of government bond issuance using a novel dataset containing bond issuance information. The empirical analysis, focused on 5 Latin American emerging economies, reveals a puzzling fact that governments issue more local currency (LC) debt when LC borrowing becomes more expensive than foreign currency (FC). To mitigate the endogeneity problem arising from currency choice and relative bond yields, the study employs an instrumental variable approach, using global risk aversion and interest rates as instruments, and conducts a two-stage least squares (2SLS) regression. Further analysis suggests that including the exchange rate risk cannot entangle this puzzle. Chapter 3 is coauthored with Ethan B. Kapstein, Jonah M. Rexer, and Adityamohan Tantravahi. We test for bias in sovereign credit ratings, considering various borrower-country characteristics and using machine learning models to estimate ratings. We find that even after accounting for economic, political, and borrower history fundamentals, rating agencies favor Western countries, including G7, EU, and OECD members, and penalize emerging Latin American and Asian nations. This bias results in coupon spreads between overrated G7 countries and underrated Southeast Asian states increasing by 62.7 basis points. The study also suggests that an unbiased portfolio of investment-grade bonds can generate risk-free excess returns, indicating persistent mispricing from credit rating bias.