Tasks, cognitive agents, and KB-DSS in workflow and process management

The purpose of this paper is to propose a nonparametric interest rate term structure model and investigate its implications on term structure dynamics and prices of interest rate derivative securities. The nonparametric spot interest rate process is estimated from the observed short-term interest rates following a robust estimation procedure and the market price of interest rate risk is estimated as implied from the historical term structure data. That is, instead of imposing a priori restrictions on the model, data are allowed to speak for themselves, and at the same time the model retains a parsimonious structure and the computational tractability. The model is implemented using historical Canadian interest rate term structure data. The parametric models with closed form solutions for bond and bond option prices, namely the Vasicek (1977) and CIR (1985) models, are also estimated for comparison purpose. The empirical results not only provide strong evidence that the traditional spot interest rate models and market prices of interest rate risk are severely misspecified but also suggest that different model specifications have significant impact on term structure dynamics and prices of interest rate derivative securities.