Realizing Correlations Across Asset Classes

TY - JOUR

T1 - Realizing Correlations Across Asset Classes

AU - Grønborg, Niels Strange

AU - Lunde, Asger

AU - Olesen, Kasper Vinther

AU - Elst, Harry Vander

PY - 2022/6

Y1 - 2022/6

N2 - We introduce a simple and intuitive composite model for forecasting correlations for use in portfolio optimization. Each element of the composite model is based on a realized volatility model. To test our model, we consider an investor seeking to diversify an equity portfolio by including commodities. In a high-frequency setting, we demonstrate that significant economic gains can be achieved by basing portfolio decisions on our modeling framework. The gains depend on the quality of the chosen volatility model, and for our preferred model, they are economically significant despite the realistic constraints on short selling and portfolio turnover.

AB - We introduce a simple and intuitive composite model for forecasting correlations for use in portfolio optimization. Each element of the composite model is based on a realized volatility model. To test our model, we consider an investor seeking to diversify an equity portfolio by including commodities. In a high-frequency setting, we demonstrate that significant economic gains can be achieved by basing portfolio decisions on our modeling framework. The gains depend on the quality of the chosen volatility model, and for our preferred model, they are economically significant despite the realistic constraints on short selling and portfolio turnover.

KW - Commodities

KW - Futures markets

KW - Portfolio selection

KW - Realized beta GARCH

U2 - 10.1016/j.finmar.2022.100729

DO - 10.1016/j.finmar.2022.100729

M3 - Journal article

SN - 1386-4181

VL - 59

JO - Journal of Financial Markets

JF - Journal of Financial Markets

IS - Part A

M1 - 100729

ER -