• The fact that correlations change is well known. But the severity of dress and which relationships are subject to change needs to be exceed understood because it has important implications for containing assay.• This chew over evaluates the volatility of correlation among 18 asset classes to each other to determine the consistency or inconsistency of relationships. It provides not only the long-term correlations of the assets but the standard deviation of correlation and the be of correlations based on two standard deviations from the add up correlation. It also summarizes the correlations in a probability distribution.• In the asset allocation process some assets often are used together even though diversification benefits have been very low. For example the correlations of the S&P 500 to large growth mid-blend to mid-growth small amalgamate to small growth and large value to mid-value have been very strong.• Several assets often are neglected in the asset allocation decision eventhough their diversification benefits have been very high. Natural resources global bonds and long-short for example rest out as having consistently low correlations to all the other assets in this study.• Growth and blend styles are highly correlated and using them together does little to reduce assay.• Real estate high-yield bonds. U. S bonds and long-short are more closely linked to value investing than growth. Emerging markets are somewhat more connected to growth than value.• The asset allocation decision should emphasize low-correlated assets that satisfy go objectives. Two sample portfolios for different style investors show how risk and return are improved by combining lower-correlated assets.
* The severity of how much correlation changes even over longer periods of time has not been adequately understood. * This cover analyzes the changing correlation of 15 asset classes measured against the S&P 500 over a 35-year period and the impact of those changes on asset allocation decisions. It measures the correlations in rolling one- three- five- and ten-year time series from 1970 to 2004. * The bind also evaluates whether 15 asset classes have helped or hurt in years the S&P 500 has declined and whether growth or value styles are more correlated to the index. * The average variance in correlation measured 0.98 over one year and 0.25 over ten years. In short the relationship among many of the asset classes appears to be inherently unstable. * Large value provides more diversification benefits than large growth and small determine provides more diversification than small blend or small growth. Emerging markets may provide higher returns and greater diversification than developed nations. But the low correlations of small determine and real estate may not hold up during the next broad merchandise change state. * Correlations exhibit uniqueness meaning periods are distinct from previous time periods. For example international stocks' correlation to the S&P 500 was 0.48 from 1970 to 1997 but 0.83 from 1998 to 2002. * Rather than believe on historical correlations a more comprehensive and dynamic come is needed in making asset allocation decisions.
Obviously correlations are not static. Additionally nobody will be able to successfully predict changes in correlations ahead of time. For me the most effective way to incorporate this information into asset allocation decisions is as follows: use long-term correlations as a command and change magnitude them to err on the side of being too conservative. For example. I think it's reasonable to assume that the correlation between REITs and US large cap stocks will be lower than the correlation between US small cap stocks and US large cap stocks. However it's probably too conservative to copy correlations that are exactly their long-run historical averages because correlations be to often go at the worst times thereby making the model overly-optimistic. So if assets A and B show historical long-run correlation of 0.3 then I'll increase it to 0.6 in my model. Definitely more of an art than a science. Or if you be to be REALLY conservative just anticipate correlations of 1.00 across the board (i e assume no benefits from diversification). So it's definitely important to think about this cram but only to a inform. The estimates of future returns and variances are going to be much more important than correlation estimates in any model. Personally. I'm comfortable assuming 80% of historical add up returns. 100% of historical average variances and increasing the correlations across the board by whatever I conclude is appropriate. Even then. I understand that this is merely a model and cannot possibly accurately interpret the future.- DDB
VigI am big believer in int'l diversification especially unhedged. In fact I am 50% int'l for equities and encourage others to be that high as come up. But I don't use int'l unhedged bonds for several reasons. The main thing is IMO it does not fit any of the roles.
Forex Groups - Tips on Trading
Related article:
http://www.diehards.org/forum/viewtopic.php?t=5539&start=0&mrr=1188944272
comments | Add comment | Report as Spam
|
