New model revolutionizes forecasting of interest rates with Bayesian approach.
A statistical model was developed to adjust, interpolate, and forecast interest rates based on Diebold and Li's term structure model. The model uses Bayesian estimation with MCMC, allowing for flexible yield curve parametric forms and stochastic volatility. This method does not require pre-interpolation of the yield curve. The model was tested on daily data from SWAP DI-PRE contracts in Brazil, showing improved fitting and forecasting compared to other models.