The SABR model is a mathematical model used to price derivatives on assets with uncertain future volatility, such as interest rate derivatives. It was developed by Patrick Hagan, Deep Kumar, Andrew Lesniewski, and Diana Woodward in 2002 and is named after their initials.
The SABR model assumes that the underlying asset follows a lognormal stochastic process, with volatility following a stochastic process driven by a stochastic volatility factor. The model is characterized by four parameters: the initial value of the underlying asset, the volatility of the underlying asset, the correlation between the underlying asset and the stochastic volatility factor, and the volatility of the stochastic volatility factor.
The SABR model is particularly well-suited for pricing interest rate derivatives, which often exhibit a "smile" or "skew" in their implied volatility curves. This occurs because the volatility of interest rates is itself a function of the level of interest rates, with higher levels of interest rates generally implying greater volatility. The SABR model allows for this non-linearity in the relationship between interest rates and volatility by allowing the volatility of the stochastic volatility factor to vary with the level of interest rates.
The SABR model is commonly used in interest rate derivative pricing, particularly for options on interest rate swaps, caps, and floors. It has been widely adopted by banks and financial institutions for its flexibility and accuracy in capturing the volatility smile features of interest rate derivatives. The model is also utilized in hedging and risk management scenarios, where traders or fund managers can use the model to estimate the Greeks or sensitivities for their portfolios.
For example, a bank with an interest rate derivatives portfolio can use the SABR model to estimate its exposure to changes in interest rates and volatilities, and adjust its hedging strategy accordingly. Additionally, the model can help traders identify trading opportunities in the market by comparing current market prices to the model’s implied prices.