Читать книгу Asset Allocation - William Kinlaw, Mark P. Kritzman - Страница 16

 Investors commonly assume that standard deviations of asset returns scale with the square root of time and that correlations of returns between assets are invariant to the return interval from which they are estimated.

 Both beliefs rest on the same underlying assumption that returns are serially independent from one period to the next.

 However, this assumption is empirically false; standard deviations and correlations of longer-interval returns diverge substantially from the standard deviations and correlations estimated from shorter-interval returns.

 Investors usually attribute this divergence to non-normality of the returns, but instead it is usually driven by nonzero lagged autocorrelations and cross-correlations.

 The divergence of high- and low-frequency estimates of standard deviations and correlations has important implications for portfolio construction, performance measurement, and risk management.