Post-Crisis Interest Rate Modelling

Specialeforsvar ved Simon Steinitz Andersen

Titel: Post-Crisis Interest Rate Modelling Modelling Roll-Over Risk

Abstract: Interest rate models specified within the stochastic model framework proposed by Alfeus, Grasselli and Schögl (2017)[2] are calibrated to swap market data covering the period from January 2013 to October 2016. This framework lays out a hybrid multi-curve interest rate model in that risk components associated with roll-over risk are modelled explicitly, that is, credit and funding liquidity risk, thereby creating a link between London Interbank Offered Rates (LIBORs) of different tenors resulting in only a parsimonious number of curves. The model framework exploits affine properties of the Cox-Ingersoll-Ross (CIR) process in deriving closed-form expressions for discount factors, the spot LIBOR and swap payment streams.
Using these closed-form expressions, a model is calibrated for 10 different specifications of the underlying CIR process, considering both 1-factor and 3-factor model effects alongside the effects of priori specified volatilities. It is shown that only discrepancies occur in the short end and when a high a priori volatility is specified in a 1-factor setup.
Vanilla and basis swap exposure profiles are generated from two of the calibrated interest rate models, revealing a sensitivity of the exposure profiles to the choice of model. In particular the underlying volatility affects the shape of these profiles through a diffusion effect. Again small irregularities are present in the short end. Finally, introducing a push to the discount curve is shown to have a very limited effect on the swap exposure.

Vejleder: David G. Skovmand
Censor:   Bjarne Astrup Jensen, CBS