Recent changes to clearing-house regulations have promoted exchange-traded products offering risk premia previously accessible only over-the-counter. Thus, as correlations increase between equity, bonds and commodities, a new strand of research questions the benefits of home-grown diversification using volatility products. First we ask: "What expected returns will induce equity and bond investors to perceive ex-ante diversification benefits from adding volatility?" We call this the optimal diversification threshold. We derive the theoretical thresholds for minimum-variance, mean-variance and Black-Litterman optimization. Empirical analysis of US and European markets shows that volatility diversification is frequently perceived to be optimal, ex-ante, but these apparent benefits are almost never realized, being eroded by high roll and transaction costs. Exchange-traded volatility only proved an effective diversifier during the banking crisis. At other times long equity and bond portfolios diversified with volatility futures have not performed as well as those without diversification, or even those diversified with commodities.JEL classification: G11, G15, G23