Emerging Market Sovereign Bond Spreads: Estimation and Back-testing
IMF Working Papers, August 1, 2012
Source details
- Canonical URL
- Emerging Market Sovereign Bond Spreads: Estimation and Back-testing
Other formats
Bibliographic details
- Authors: Fabio Comelli
- Published: August 1, 2012
- Series: IMF Working Papers
- DOI: https://doi.org/10.5089/9781475505627.001
Summary
- Study estimates sovereign bond spreads for 28 emerging economies over the period January 1998-December 2011.
- Tests the ability of the model to generate accurate in-sample predictions for emerging economies bond spreads.
- Key comparative methods: rolling regression routines, a linear prediction method, and a random walk (random guessing).
Methodology and scope
- Sample: 28 emerging economies; period: January 1998-December 2011.
- Frequency: monthly changes in bond spreads examined and predicted.
- Prediction approaches evaluated:
- Rolling regression routines (panel/rolling estimation).
- Linear prediction method.
- Random walk (benchmark for forecasts).
Major findings
- Heterogeneity of drivers:
- The impact and significance of country-specific and global explanatory variables on bond spreads varies across regions and economic periods.
- Crisis vs non-crisis differences:
- During crisis times, good macroeconomic fundamentals are helpful in containing bond spreads, but less than in non-crisis times, possibly reflecting the impact of extra-economic forces on bond spreads when a financial crisis occurs.
- Forecasting performance:
- For some emerging economies, in-sample predictions of the monthly changes in bond spreads obtained with rolling regression routines are significantly more accurate than forecasts obtained with a random walk.
- Rolling regression-based bond spread predictions appear to convey more information than those obtained with a linear prediction method.
- Bond spreads forecasts obtained with a linear prediction method are less accurate than those obtained with random guessing.
Implications for analysis and monitoring
- Model specification should allow for regional and period-specific variation in explanatory variable significance.
- Forecasting approaches:
- Rolling regression routines can provide meaningful improvement in in-sample prediction accuracy for some emerging economies relative to a random walk.
- Simple linear prediction methods may perform poorly and can be outperformed by random guessing for bond spread forecasts.
- Crisis-period analysis:
- Reliance on macroeconomic fundamentals to explain or contain bond spreads during crises may be limited; attention to extra-economic drivers is warranted.
Content in this bundle
- _wp12212 - Appendix Tables and References