Abstract:
This paper deals with the problem of interpolation of discount factors between time buckets. The problem occurs when price and interest rate data of a market segment are assigned to discrete time buckets. A simple criterion is developed in order to identify arbitrage-free robust interpolation methods. Methods closely examined include linear, exponential and weighted exponential interpolation. Weighted exponential interpolation, a method still preferred by some banks and also offered by commercial software vendors, creates several problems and therefore makes simple exponential interpolation a more logical choice. Linear interpolation provides a good approximation of exponential interpolation for a sufficiently dense time grid.