Forecasting macroeconomic variables through the term-structure of interest rates in emerging countries

Details

Supervisors

Leonardo Iana

Faculty

Louvain School of Management

Degree label

Master [120] en sciences de gestion, à finalité spécialisée

Abstract

This paper assesses whether macroeconomic variables in emerging markets can be forecasted with three components -- the level, slope, and curvature -- of their national yield curve, as well as the euro area and USA curves. Unlike most papers in the literature who only use the slope, we employ all three factors of the yield curve. We then create three vector autoregression (VAR) models for each country that respectively includes the national, euro and USA curves. Almost all macroeconomic variables in every country can successfully be forecasted with varying horizons and degrees of certainty. All three factors of the curves contain predictive power. The euro and USA curves are particularly effective in neighbouring markets such as Russia or Mexico. Further analysis should be performed with more sophisticated non-linear techniques, and with a rolling time window to see if the forecasting power of the curves varies over time.