When in Investing, why worst is best

Originally written on 19th June 2013

Interesting chart on yearly and 10 year cumulative return delivered by different Asset Classes…..the exhibit underscores one often touted characteristic of the markets…That it is nearly impossible to predict the best performing or for that matter the worst performing asset from one year to next (Remember, safer and less volatile assets are usually at the bottom in return ranking, but that does not mean that you can predict their outperformance or underperformance relative to riskier asset classes from year to year). A equally weighted allocation across all asset classes as represented by the ‘Asset Allocation bloc (in Grey) is one best method to both remain reasonably well diversified and making sure your portfolio does not end up at the bottom (but will also preclude its possibility of ending at the top). A second observation, is that asset classes (particularly risk asset classes) that performs worst on relative basis in given year(s) usually has a higher probability of outperformance in subsequent period(s). Hence a strategy of buying losers and selling winners may result in long term outperformance.

Below i present a mechanical asset allocation strategy that:

1.   Invest in the subsequent year in the best performing asset of the previous year (eg. invest 100% in MSCI EME in 2004 after the asset class being best performing in 2003)

2.   Invest in the subsequent year in the worst performing asset of the previous year (eg. invest 100% in Cash in 2004 after Cash being the worst performing asset class in 2003)

3.   Invest in three best performing assets of any given year in subsequent year on an equally weighted basis (eg invests in MSCI EME, Russell 2000 and MSCI EAFE in 2004 on equally weighted basis)

4.   Invest in three worst performing assets of any given year in subsequent year on equally weighted basis (eg invests in Cash, Barclays Aggregate bond index and Market Neutral in 2004 on equally weighted basis)

The results of this exercise tabulated below are rather interesting, The contrarian strategy of holding worst performing assets in subsequent period is more than twice as profitable than the strategy of holding the best performing assets. Secondly, on risk adjusted basis, the sharpe ratio of the contrarian strategy is better than that of the momentum strategy. Thirdly, the portfolio model (three best and three worst) approach is more superior than holding 100% of single asset in both absolute return and risk adjusted return terms.

	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012	Cumulative Return	Std Deviation of Return	Maximum Drawdown	Sharpe Ratio
Best performing asset previous year		1.26	1.12	1.33	0.84	0.47	1.06	1.19	1.08	1.20	20.8%	26.2%	-53%	0.79
Worst performing asset previous year		1.01	1.03	1.04	1.16	0.62	1.79	1.00	1.05	1.19	74.8%	30.5%	-38%	2.45
Three best performing assets classes in previous year (Eq wt)		1.22	1.20	1.21	1.12	0.56	1.03	1.18	0.95	1.08	39.6%	20.8%	-44%	1.90
Three worst performing assets classes in previous year (Eq wt)		1.04	1.04	1.09	1.09	0.79	1.47	1.02	1.04	1.12	76.6%	17.4%	-21%	4.39

I have highlighted the year 2007, 2008 and 2009 to draw special attention to another interesting observation. As one would notice, the previous year best performing asset and equally weighted portfolio usually have a return in double digit whilst the worst performing asset and portfolio of worst performing assets has a low single digit return in most periods. So why does the worst performer come out best over the 10 year period?

The answer lies in 1) the sharp volatility of best performing but more riskier assets such as MSCI EME index (-compare Maximum Drawdown statistics). The best performing assets after three years of double digit return were cut into less than half in the financial crisis of 2008/09 while the worst performing asset of 2007 were trading at a depressed valuations going into 2008 and logically were relatively less impacted. Also the worst performing asset of 2008 were grossly undervalued  going into 2009 due to duress in the market at that time (having being cut in half or less in just one year time) and hence recovered sharply in the subsequent year.

So what lesson does the above study holds for an average investor?

1)   Not losing money is more important than making money due to the power of compounding (see the compounded performance of best and worst performing assets between 2004-2009). Every dollar invested in best performing asset and rebalanced annually was worth only 21 cents in 2009 despite three years of double digit growth. On the other hand, every dollar invested in the worst performing asset in 2004 was worth 41 cents in 2009 twice as much as its counterpart

2)  Dislocation in asset prices due to duress in the market can be good time to accumulative worst hit assets as prices are riddled for reason other than the long term earning capacity of the asset and investor outlook on valuations of such assets is at the nadir. Such dislocation provide opportunity to earn excess return in subsequent periods while valuation recover back to its historic average. This strategy is also safer as reflected in maximum drawdown and sharpe ratios.
---------------------------------END---------------------------------