SABR Model vs. Other Stochastic Volatility Models A Comparative Analysis
The Black Book of Day Trading Strategies
1,000 complete strategies · 31 chapters · Full trade plans
The SABR model is a effective and widely used tool for modeling stochastic volatility, but it is by no means the only game in town. The world of quantitative finance is populated by a diverse ecosystem of stochastic volatility models, each with its own strengths and weaknesses. In this article, we will conduct a comparative analysis of the SABR model against two other popular stochastic volatility models: the Heston model and the LMM-SABR model. We will highlight the key differences between these models in terms of their dynamics, calibration, and practical application, providing a nuanced perspective on their relative merits.
The Heston Model
The Heston model, developed by Steven Heston in 1993, is another classic stochastic volatility model. Like the SABR model, the Heston model assumes that the volatility of the underlying asset is a stochastic process. However, the Heston model differs from the SABR model in several key respects:
- Volatility Process: The Heston model assumes that the variance of the underlying asset follows a Cox-Ingersoll-Ross (CIR) process, which is a mean-reverting process. This is in contrast to the SABR model, which assumes that the volatility follows a geometric Brownian motion.
- Analytical Solution: The Heston model has the advantage of having a closed-form solution for European option prices, which is not the case for the SABR model. This makes the Heston model computationally more efficient for pricing vanilla options.
- Calibration: The Heston model can be more difficult to calibrate than the SABR model, as it has more parameters and the calibration problem is more prone to local minima.
The LMM-SABR Model
The LMM-SABR model is a hybrid model that combines the SABR model with the LIBOR Market Model (LMM). The LMM is a multi-factor model that is widely used for pricing complex interest rate derivatives, such as Bermudan swaptions. The LMM-SABR model extends the LMM by incorporating stochastic volatility at the level of the individual forward rates. This allows the model to capture the volatility smile for each forward rate, while also providing a consistent framework for pricing a wide range of interest rate derivatives.
A Comparative Table
The following table summarizes the key features of the SABR, Heston, and LMM-SABR models:
| Feature | SABR Model | Heston Model | LMM-SABR Model |
|---|---|---|---|
| Volatility Process | Geometric Brownian Motion | CIR Process | SABR on each forward rate |
| Analytical Solution | Asymptotic expansion | Closed-form | No closed-form |
| Calibration | Relatively easy | Can be difficult | Complex |
| Application | Vanilla and semi-exotic options | Vanilla options | Complex interest rate derivatives |
Conclusion
The SABR model, the Heston model, and the LMM-SABR model are all effective tools for modeling stochastic volatility. The choice of which model to use depends on the specific application. The SABR model is a good choice for pricing and hedging vanilla and semi-exotic options, as it is relatively easy to calibrate and implement. The Heston model is a good choice for pricing vanilla options, as it has a closed-form solution. The LMM-SABR model is the most sophisticated of the three models, and it is the best choice for pricing complex interest rate derivatives. In the next article, we will take a closer look at the LMM-SABR model and its applications.