Thursday, July 23, 2009

Strength in Diversity

Jim Parker, Regional Director, DFA Australia Limited, July 2009

The post-mortem on the global financial crisis is raising the question in some minds about whether it is time to read the last rites for diversification — the principle that you reduce your investment risk by spreading it around.

Among the would-be mourners are critics who say the diversification principle failed spectacularly for investors just when they needed it the most — during the biggest market meltdown seen in generations.

After all, the idea behind diversification is that no two investments perform in exactly the same way at the same time. By mixing up the asset classes in your portfolio and diversifying within those asset classes, the idea is that you provide yourself with a cushion or shock absorber in the down times.

But then came the financial crisis of 2008 and suddenly there seemed to be no safety net. Just about every asset class — apart from government bonds and cash — got walloped. In the jargon of financial economists, those asset classes appeared to become much more "correlated".

But the appearance of rising correlations does not necessarily mean that diversification does not work. And you can prove that by considering what might have happened if investors had been less diversified than they could have been last year.

It's a point made by Dartmouth Tuck School of Business finance professor Ken French in a new video on the forum he co-hosts with Eugene Fama, a finance professor at the University of Chicago Booth School of Business.

"Diversification still works," French explains in the interview. "And they (investors) would have had even more uncertainty about the return on their portfolios if they had been poorly diversified."

French points out that diversification does not eliminate the volatility of theoverall market. What it does do is protect against the additional volatility arising from the characteristics of individual firms or asset classes.

"It's that extra volatility that you don't have to have if you diversify well."

On the question of rising correlations in highly volatile markets, French notes that the return on any asset comprises two components — the market return itself and the return attributable to the specifics of the individual asset.

During the extreme volatility of 2008, the market movement became proportionately much more influential so that it appeared that individual stocks and asset classes appeared to be much more lined up with each other.

"What matters when we think about diversification is the firm-specific pieces, not the market piece," French says. "And we know that in these volatile periods, the firm-specific pieces also get bigger. So while it may look like the benefits of diversification have gone down, they have actually gone up."

These differences in the firm-specific variation in returns are what financial economists call "cross-sectional dispersion". And it is well established that just as market volatility tends to spike after periods of negative performance, so do the cross-sectional dispersions.

This table below shows the performance of four Australian portfolios (cash, simple diversified, sub-asset class diversified and all equities) over three different periods - the peak months of the financial crisis (Nov 2007 to February 2009), the recent recovery (March 2009 to June 2009) and the past quarter century.

The simple diversified portfolio ("Diversified A") is just 70 per cent Australian equities, 25 per cent bonds and 5 per cent cash. The sub-asset class portfolio ("Diversified B") breaks up the Australian equity component into large, value and small components. Common indices are used.

You can see that even during the worst of the crisis, a diversified portfolio provided a better result (or at least a less bad result) than an all-stock portfolio. While cash was clearly the best result of the three in this period, look at contrast between the three portfolios in the subsequent, admittedly very short, recovery period. Note too, that the Diversified B portfolio has performed better in the rebound, reflecting the tendency of small and value stocks to do outperform around economic turning points.

However, the key message is in the longest time period here, covering early-1985 to mid-2009. This shows returns from the diversified portfolios were very similar to the all-stock portfolio, but with much lower volatility.

What this all means for investors is that the need for broad diversification both across and within asset classes becomes even more important at times of high volatility.

In the words of Mark Twain, reports of the death of diversification have been greatly exaggerated.

(Thanks to Rob Brown for compiling the data for this article)

Sources and Descriptions of Data:

All Cash:
UBS Warburg 90-Day Bank Bill Index: 100%
Diversified Portfolio A
S&P/ASX 500 Accumulation Index: 70%
UBS Warburg 90-Day Bank Bill Index: 5%
UBS Warburg Australia Composite Index All Maturities: 25%
Diversified Portfolio B (Sub Asset Class):
UBS Warburg 90-Day Bank Bill Index: 5%
MSCI Australia Value Index (net div.): 14%
UBS Warburg Australia Composite Index All Maturities: 25%
S&P/ASX 100 Accumulation Index: 35%
S&P/ASX Small Ordinaries Accumulation Index: 21%
All Stocks:
S&P/ASX 500 Accumulation Index: 100%
(All returns in $A)