TL;DR
This paper introduces a new model for pricing European quanto options that accounts for "fat tails" (extreme events) and "asymmetric dependence" (uneven relationships) in financial markets. By using a multivariate normal tempered stable (NTS) process, the authors demonstrate that their model provides more realistic pricing than traditional Black-Scholes models, especially when dealing with market conditions prone to extreme events.
What Problem Does This Paper Address?
Quanto options are financial derivatives that offer exposure to an asset in a foreign currency without the exchange rate risk. Their valuation is highly sensitive to the dependence structure between the underlying asset and the exchange rate. Traditional models, like the Black-Scholes model, often assume that asset returns follow a normal distribution and that the dependence is linear. However, empirical evidence shows that financial asset returns exhibit "fat tails" (meaning extreme events are more common than a normal distribution would predict) and "asymmetric dependence" (the relationship between assets can change depending on market conditions, e.g., becoming stronger during downturns). These deviations from normality can lead to significant mispricing of options. This paper aims to address this by developing a more sophisticated pricing model that can capture these observed market characteristics.
Methods
The core of this paper's contribution lies in its use of the multivariate normal tempered stable (NTS) process to model the underlying asset and exchange rate dynamics. This process is a generalization of the multivariate normal distribution that allows for:
- Fat-tailedness: The NTS distribution can exhibit heavier tails than the normal distribution, better reflecting the observed frequency of extreme market movements.
- Asymmetric Dependence: It can capture more complex relationships between variables than simple linear correlation.
The authors derive a closed-form solution for pricing European quanto options under this NTS framework. They also compare their results to the standard Black-Scholes model, which serves as a benchmark. The methodology involves:
- Defining the NTS process: The paper formally defines the N-dimensional NTS process and its properties.
- Change of Measure: They utilize Girsanov's theorem to transition from the physical (market) measure to a risk-neutral measure, which is essential for option pricing.
- Deriving Option Pricing Formulas: They develop specific formulas for quanto call and put options under both the NTS and Black-Scholes models.
- Empirical Application: They apply their model to real-world data of the Nikkei 225 index and the JPY-USD exchange rate.
Datasets
The paper uses historical daily log-return data for the USD-valued Nikkei 225 index and the JPY-USD exchange rate from January 4, 2000, to June 21, 2013. This dataset is used to estimate the market parameters for both the Black-Scholes and NTS models.
Key Insights and Ideas
- NTS is a better fit for financial data: The empirical analysis strongly suggests that the NTS model provides a significantly better fit to the observed financial data (Nikkei 225 and JPY-USD exchange rates) compared to the traditional Gaussian (Black-Scholes) model. This is evidenced by goodness-of-fit tests.
- Fat tails and asymmetry matter for pricing: The NTS model's ability to capture fat tails and asymmetric dependence leads to substantially different option prices compared to the Black-Scholes model.
- Underestimation by Black-Scholes: The Black-Scholes model tends to underestimate the risk associated with extreme events, leading to lower quanto option prices.
- NTS prices are generally higher: The NTS model, by accounting for higher probabilities of tail events, generally yields higher prices for quanto options, especially for out-of-the-money options.
- Asymmetry impacts pricing: The paper demonstrates that asymmetries in the underlying returns (like the negative skewness of Nikkei returns) affect option pricing, with out-of-the-money put options being valued higher than out-of-the-money call options.
Key Results
- Model rejection: The Black-Scholes model was rejected by the data based on goodness-of-fit tests (Kolmogorov-Smirnov and Anderson-Darling statistics), while the NTS model was not rejected at any reasonable confidence level.
- Price differences: Quanto call and put prices derived from the NTS model differ substantially from those derived from the Black-Scholes model.
- Higher NTS prices: The NTS model generally results in higher option prices, reflecting its better capture of tail behavior.
- Figure 3 and 4: These figures visually demonstrate the price differences between the two models and the impact of moneyness and asymmetry on option pricing. The NTS model's prices are consistently higher, and the deviation increases for out-of-the-money options.
Limitations and Future Work
- Future Research: The authors suggest that future research should focus on:
- Empirical analyses using observed quanto option transaction data.
- Comparisons with alternative quanto pricing strategies.
- Further investigation into how asymmetries specifically affect NTS option pricing.
- Model Assumptions: While the NTS model is an improvement, it still relies on specific assumptions about the underlying processes. The paper does not explicitly detail limitations of the NTS model itself, but future work could explore even more complex dependence structures or incorporate other market frictions.
Summary updated