Embedded Betas and Better Bets: Factor Investing in Emerging Market Bonds

Vol. 17, No. 3, 2019

Johnny Kang, Kevin So and Thomas Tziortziotis
Our paper documents novel empirical insights driving the pricing of sovereign external emerging market bonds (EMB) using a factor pricing framework. While similar to traditional fixed-income instruments driven primarily by interest rate and credit risk factors, the unique dynamics of emerging markets suggest that additional factors may be necessary to understand this asset class. We therefore use the EMB universe as an empirical setting to explore the ability of macro factors to explain the time series of returns and the power of style factors in forecasting the cross-section of returns.

We begin by highlighting the historical returns of the value-weighted EMB portfolio and showing that its impressive performance can be explained by its embedded betas to a diversified set of macro factors (rates, credit, currency, and equity). Not only does the benchmark portfolio have significant exposure to all four factors, but the relative risk contribution from each factor also appears to be well balanced over the full sample period. Repeating the same analysis on a rolling basis reveals that these risk contributions are not static, but in fact vary significantly over time.

Next we construct value and momentum style factors that help explain the cross-section of country expected returns. Our value measure, which we call default-adjusted spread, identifies issuers trading cheap or rich relative to expected sovereign default risk. Our momentum measure uses a cross-asset insight from currency markets to identify financial conditions that may hinder an issuer’s ability to pay. We find that these two style factors are highly complementary, given the value factor’s risk- seeking profile and the momentum factor’s defensive nature. Based on this result, we introduce a risk-on versus risk-off framework to characterize the correlation structure spanning our macro and style factors. Finally, we incorporate our factor insights into a long-only optimized portfolio with practical investment constraints. Our results show that this multi-factor portfolio outperforms the value-weighted benchmark by 60 basis points per annum net of estimated transaction costs.