Equity Asset Allocation Didn't Work in 2008

The purpose of asset allocation is to own minimally correlated assets, which tends to reduce portfolio volatility while delivering the return of the blended classes.

Within the major equity category, asset classes are often thought of as US, non-US developed, and emerging country stocks.

Correlations between those major categories have converged somewhat in the past, but in 2008 they approached unity.

If equity class correlations equal “1″ there is no allocation benefit relating to diversifying the volatility within the equity class.

This correlation table for 1-year returns for various Vanguard mutual funds (with ETF proxies for those mutual funds) shows major and minor equity categories lining up in 2008 with correlations approaching “1″.

Securities listed in image:
VFINX, VTSMX, VFWIX, VDMIX, VEURX, VPACX, VEIEX,
SPY, VTI, VEU, VEA, VGK, VPL, VWO

Asset allocation with regard to equities will work better when correlations begin to diverge.

Comments

[donzelion:]

I wonder if, longitudinally, there has been a net acceleration in correlation among asset classes.

It would make sense that certain financial instruments and structures would increase correlation in ways that are less than obvious - and that correlations that are hidden would emerge during times of stress (e.g., when hedge funds liquidate certain companies, they sometimes must also liquidate non-correlated assets, and thereby, the "nocorrelated assets" become correlated as a result of market activity).

Indeed, once upon a time, there was a basic faith that the DOW, which represents uncorrelated industries, could never rise or fall, but in the long-run, it would eventually return to equilibrium (now a quaint notion long sense dispelled...)

Jan 08 07:00 AM

[User 283977:]

The reason everything is correlated is because there was a huge market move downward, and everything moved with it. That amount of shared variation would dominate the correlation coefficient. Portfolio diversification is designed to guard against negative sector moves not gross market moves. A point which I believe your article makes very clear.

As an investment professional, I wonder what your opinion is concerning degree of diversification. I know diversification seems to be a mantra with respect to risk reduction. However, I don't see too many people mentioning that blind diversification can limit upside potential. All sectors are not created equal with respect to their upside or downside potential. It seems that some sector performance would have a causal effect on other sector performance. I suspect Energy is one of these "first mover" segments. What do you think? Do you think that the sectors are inherently independent, or is performance in some sectors more likely to influence performance (positive or negative) in other sectors?

Jan 08 12:46 PM

[DavidT:]

Actually, the correlation between many of the equity asset classes is higher than people realize. VEU and SPY were highly correlated throughout 2007 as well. Even EEM (emerging markets) and SPY have been highly correlated (~0.8) for the past five years.

See www.assetcorrelation.c... for details

Jan 13 06:38 PM

[Atakan:]

Richard-
I have a question for you. If I want to select the world (free floating) as my portfolio benchmark. 5 years down the road when China makes up twice as large a part of the world market cap do I go back and readjust my benchmark or do I leave it where it started - frozen. We talk a lot about rebalancing without mentioning the fact that the world is dynamic and so are the benchmarks. How do we manage the inherent conflict between having to find a stable bencmark and one that reflects the world we live in?
May be I am missing sth. here...

Jan 15 01:47 PM