Scaling portfolio volatility and calculating risk contributions in the presence of serial cross-correlations

摘要

In practice daily volatility of portfolio returns is transformed to longerholding periods by multiplying by the square-root of time which assumes thatreturns are not serially correlated. Under this assumption this procedure ofscaling can also be applied to contributions to volatility of the assets in theportfolio. Close prices are often used to calculate the profit and loss of aportfolio. Trading at exchanges located in distant time zones this can lead tosignificant serial cross-correlations of the closing-time returns of the assetsin the portfolio. These serial correlations cause the square-root-of-time ruleto fail. Moreover volatility contributions in this setting turn out to bemisleading due to non-synchronous correlations. We address this issue andprovide alternative procedures for scaling volatility and calculating riskcontributions for arbitrary holding periods.