COMPSTAT 2022: Start Registration
View Submission - COMPSTAT2022
Title: Exact likelihood for inverse gamma stochastic volatility models Authors:  Roberto Leon-Gonzalez - GRIPS (Japan)
Blessings Majoni - National graduate institute for policy studies (Japan) [presenting]
Abstract: A novel closed form expression of the likelihood for the inverse gamma stochastic volatility model is obtained. It is shown that by marginalizing out the volatilities the model that we obtain has the resemblance of a GARCH in the sense that the formulas that we get are similar, which simplifies computations significantly. We also obtain methods to draw the latent volatilities directly from their posterior distributions. Recent literature has also attempted to obtain closed-form solutions for the likelihood in stochastic volatility models. However, the literature has only obtained such a solution for non-stationary models, or for the stationary Gamma Stochastic Volatility model. We compare the empirical fit of our proposed model with the previous literature.