On the predictability of emerging market sovereign credit spreads

Abstract

This paper examines the quarter-ahead out-of-sample predictability of Brazil, Mexico, the Philippines and Turkey credit spreads before and after the Lehman Brothers' default. A model based on the country-specic credit spread curve factors predicts no better than the random walk and slope regression benchmarks. Model extensions with the global yield curve factors and with both global and domestic uncertainty indicators notably outperform both benchmarks post-Lehman. The nding that bond prices better reect fundamental information after the Lehman Brothers' failure indicates that this landmark of the recent global nancial crisis had wake-up call eects on emerging market bond investors.

A Coherent Framework for Predicting Emerging Market Credit Spreads with Support Vector Regression

“…In our predictive analysis we set τ = 5 years. The weekly frequency dataset of Audzeyeva and Fuertes (2018) contains the yields to maturity on zero-coupon U.S. Treasuries and U.S. dollar-denominated Eurobonds of each country we used to calculate the country-specific credit spreads.…”

“…The selection of predictive variables for the input vector in our analysis is based on prior research on predictability of sovereign emerging market spreads. We employ predictive variables, motivated by economic theory, that have been previously analyzed in Audzeyeva and Fuertes (2018). Accordingly, our Baseline model is rooted in the expectations theory of the term structure of interest rates of Sargent (1972) and Roll (1970).…”

“…The main idea is that investor expectations about future credit spreads, reflected in today's forward spreads, exploit all the available information. Consequently, Audzeyeva and Fuertes (2018) extract the credit curve level, slope and curvature factors from weekly cross-sections of country-specific bond prices; see their paper for details.…”

“…Furthermore, the few studies that attempt such forecasts document that credit spreads are difficult to predict out-of-sample (OOS). In particular, in a recent study Audzeyeva and Fuertes (2018) show that predictive models employing country-specific credit-spread-curve factors, namely, the level, slope and curvature, known to contain useful information for future yields in the context of riskless debt, cannot beat a random walk for emerging market bonds. They find that employing additional global and country-specific predictors improves the model predictive ability, but that even with these additional predictors the model-based-forecasts cannot always outperform a random walk.…”

“…All three sovereign borrowers are major emerging market economies and members of the G20. In our predictive analysis, we employ the data set of Audzeyeva and Fuertes (2018). Our analysis employs SVR with linear, sigmoid, RBF and polynomial kernels to generate credit spread forecasts using sets of input variables that are motivated by economic theory.…”