The Bergomi model is an advanced stochastic volatility model that addresses some limitations of the earlier stochastic volatility models, such as the Heston model or the SABR model. In a nutshell, the Bergomi model is a two-factor model that aims to capture the forward-looking behavior of the implied volatility, which is quite relevant in options pricing and risk management. The model is named after its creator, Lorenzo Bergomi.
Let’s first discuss the limitations of the traditional stochastic volatility models and then see how the Bergomi model improves on them.
Limitations of traditional stochastic volatility models:
1. Traditional stochastic volatility models typically assume that the volatility process is a one-factor model, which means they cannot capture the full dynamics of the volatility surface. This limits their ability to fit the term structure of the implied volatility skew and its dynamics, leading to potential mispricing of options.
2. Most stochastic volatility models tend to use a risk-neutral approach to price derivatives, which does not give insights into the underlying risk factors of the market. This limits their ability to provide accurate forecasts for hedging and risk management purposes.
Now, let’s see how the Bergomi model addresses these limitations.
1. Two-Factor Model: The Bergomi model assumes a two-factor structure for the instantaneous volatility. This allows the model to capture the term structure of the implied volatility skew more accurately. Specifically, the Bergomi model can incorporate the negative correlation between the forward variance and the spot volatility, which is observed empirically. This property is crucial for fitting the steep skews observed in equity markets. The model is defined by the following system of SDEs:
$$dS_t = S_t\sqrt{v_t}\,dW_t^S$$
$$dv_t = \kappa(v_t)(\theta - v_t)dt + \eta\sqrt{v_t}\,dW_t^v$$
dκt(v) = λ(κ − κt(v))dt + σκt(v) dWtκ
where St is the spot price of the underlying, vt is the spot variance, κt(v) is the mean-reversion level of the variance process, and WS, Wv, and Wκ are standard Brownian motions. Note that the correlation between WS and Wv is assumed to be constant and negative.
2. Forward-Looking Volatility: The Bergomi model considers the full forward variance curve as a state variable, unlike traditional stochastic volatility models, which only include the spot variance as a state variable. This approach helps capture the forward-looking feature of the implied volatility dynamics, leading to better predictions of the option prices and improved insights for risk management.
Applications in Quantitative Finance:
1. Options Pricing: The Bergomi model has been extensively used for more accurate pricing of equity, index, and FX options. The superior ability to fit the implied volatility skew helps reduce pricing errors, especially for options with longer maturities.
2. Risk Management: Thanks to its forward-looking approach, the Bergomi model provides improved risk metrics for managing option portfolios, such as the Greeks (Delta, Gamma, Vega, etc.). This leads to more effective hedging strategies, ultimately minimizing the risk exposure for market practitioners.
3. Model Calibration: The forward variance structure allows the practitioner to use market data, such as the implicit variance curve, or other non-parametric procedures to calibrate the model parameters. This enables more accurate computational results and specific scenario analyses, which are crucial for risk management and trading.
In conclusion, the Bergomi model is a significant improvement over traditional stochastic volatility models due to its two-factor structure and its ability to capture the forward-looking behavior of the implied volatility. These features enable more accurate option pricing, better risk management insights, and greater flexibility in model calibration, making it a valuable tool in quantitative finance.