A Jump and Smile Ride : Jump and Variance Risk Premia in Option Pricing

We introduce a discrete-time model for log-return dynamics with observable volatility and jumps. Our proposal extends the class of Realized Volatility heterogeneous auto-regressive gamma (HARG) processes adding a jump component with time-varying intensity. The model is able to reproduce the temporary increase in the probability of occurrence of a jump immediately after an abrupt large movement of the asset price. Belonging to the class of exponentially affine models, the moment generating function under the physical measure is available in closed-form. Thanks to a flexible specification of the pricing kernel compensating for equity, volatility, and jump risks, the generating function under the risk-neutral measure inherits analytical tractability too. An application of the leveraged HARG model with dynamic jump intensity to the pricing of a large sample of S&P500 index options assesses its superior performances with respect to state-of-the-art benchmark models