Answers are grounded in this paper and cite their sources.
This paper examines the term structure of interest rates in Bangladesh from April 2015 to January 2026 using the dynamic Nelson-Siegel (DNS) model, incorporating macroeconomic variables. The DNS model breaks down the yield curve into three latent variables: Level, Slope, and Curvature, and models the evolution of these latent factors within a Vector Autoregression (VAR) framework, including the policy rate and inflation. The DNS specification provides an excellent cross-sectional fit to Bangladesh government bond yields, with an average (𝑅2) of 0.991 and a root mean squared error (RMSE) of 24.7 basis points. Both the estimated Level and Slope factors show a high degree of persistence (0.905 and 0.956, respectively), whereas the estimated Curvature factor shows significantly less persistence. Impulse-response and forecast-error variance decomposition analyses suggest that monetary policy shocks have a small but growing impact on yield-curve dynamics at longer horizons. DNS-Macro-VAR yield curve forecasts remain upward-sloping and gradually converge toward a stable long–run configuration. These results show that the DNS framework is useful for modelling and forecasting the yield curve in the Bangladesh government bond market, offering valuable insights for monetary policy design, bond market development, and risk management.
No takes yet. Share an insight, caveat, or question.
Bulbul et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: