How to value and price dirty CSAs from a funding perspective? Developing a pricing and valuation model for ING Bank
Files
JORIS_75492400_2026.pdf
Closed access - Adobe PDF
- 4.52 MB
Details
- Supervisors
- Faculty
- Degree label
- Abstract
- Since the 2008 financial crisis, collateral has become a central element of over-the-counter derivatives markets. More recently, the increasing use of dirty Credit Support Annexes (CSAs), allowing collateral to be posted in the form of securities rather than cash, has created new pricing challenges for financial institutions. Unlike cash collateral, securities must be financed through the repo market, making the repo curve the relevant discounting curve and introducing the need to compute the Collateral Valuation Adjustment (ColVA). However, repo transactions are concentrated at short maturities, while derivatives often extend over several decades, requiring the long end of the repo curve to be extrapolated. This thesis develops a methodology to value and price dirty CSAs from a funding perspective in collaboration with ING Bank. Three approaches are investigated to extrapolate the repo spread curve: the Nelson–Siegel model, the Smith–Wilson method and a Vasicek short-rate model. These models are calibrated on historical UK government bond repo spread data and compared through their impact on ColVA. The results show that both the Nelson–Siegel and Smith–Wilson approaches provide satisfactory repo curve extrapolations and can be used for ColVA estimation. However, because market data are only available for short maturities, the extrapolated part of the curve inevitably depends on modelling assumptions, leading to different but equally plausible ColVA values. In contrast, the Vasicek model proved unsuitable for modelling repo spreads in this context. Overall, this thesis provides a first methodology that can support the pricing of dirty CSAs at ING Bank while highlighting the sensitivity of ColVA to the choice of repo curve extrapolation method.