We investigate the predictability of payoffs from selling variance swaps on the S&P500, US 10-year treasuries, gold, and crude oil. In-sample analysis shows that structural breaks are an important feature when modeling payoffs, and hence the ex post variance risk premium. Out-of-sample tests, on the other hand, reveal that structural break models do not improve forecast performance relative to simpler linear (or state invariant) models. We show that a host of variables that had previously been shown to forecast excess returns for the four asset classes, contain predictive power for ex post realizations of the respective variance risk premia as well. We also find that models fit directly to payoffs perform as well or better than models that combine the current variance swap rate with a realized variance forecast. These novel findings have important implications for variance swap sellers, and investors seeking to include volatility as an asset in their portfolio.
Dark, J., Gao, X., van der Heijden, T., & Nardari, F. (2022). Forecasting variance swap payoffs. Journal of Futures Markets. Doi:10.1002/fut.22371
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.