Pricing Derivatives in Periods of Low or Negative Interest Rates

This study investigates which pricing model is more appropriate in a low or negative interest rate setting in terms of pricing accuracy. We compare six di_erent option pricing models where each model assume di_erent dynamics of the interest rate and the volatility. The models incorporate either a constant, Vasicek or CIR type dynamic for the interest rate and the models incorporate either a constant or Heston type dynamic for the volatilty. We test which pricing model has the better performance in either interest rate setting. Our simulation study shows that the pricing models that rely on a CIR type short-rate process exhibit abnormal parameter values and drastically underperform in terms of pricing accuracy in comparison with models based on a Vasicek type short-rate process. The models are calibrated on real world S&P 500 option data from the time period 2010 to 2015 and we _nd that the Heston-Vasicek model has the best pricing performance. In contrast, the Heston-CIR models shows a relatively poor out-of-sample pricing performance.