My previous article's conclusion was that the first benefit of market timing is to limit the drawdown and volatility, not to improve the return (click here to read). The portfolio was as simple as possible: only one position in SPY.
In this second part, I will consider two portfolios of ETFs (indexes and sectors) with three market timing strategies: an individual rule for each ETF, a global rule, and a combination of both.
In all cases, the signal is "on" when the 50-day simple moving average (SMA50) is strictly above the 200-day simple moving average (SMA200), and "off" when it falls below. There are better timing indicators. But you will see that even this one works decently.
The first portfolio contains the ETFs of five major U.S. stock indexes in equal weight: DIA, SPY, QQQ, MDY, IWM.
Simply holding them has outperformed the benchmark SPY from January 2000 to January 2013:
Let's apply the SMA timing rule in two ways:
1- Individually. We check the condition for each ETF once a week since 1/1/2000. If the signal is "off" (SMA50 <= SMA200), this position is sold or kept in cash. We buy again if the signal in "on" (SMA50>SMA200). Each position is limited to 20% of the portfolio value.
2- Globally. The timing rule is checked on the S&P 500 index. All positions are sold and bought together.
The next table gives the results of a 13-year simulation since 1/1/2000:
|5 indexes (MT: double SMA)||Total Return||CAGR||DDM|
|Buy & Hold||59.71||3.67||-54.23|
The global market timing shows better results in terms of return and drawdown.
The second idea is to build a portfolio with 11 sector ETFs from the iShares series:
|IYM||iShares Dow Jones US Basic Materials||June 2000|
|IYC||iShares Dow Jones US Consumer Services||June 2000|
|IYK||iShares Dow Jones US Consumer Goods||June 2000|
|IYE||iShares Dow Jones US Energy||June 2000|
|IYF||iShares Dow Jones US Financial Sector||May 2000|
|IYH||iShares Dow Jones US Healthcare||June 2000|
|IYJ||iShares Dow Jones US Industrial||June 2000|
|IYR||iShares Dow Jones US Real Estate||June 2000|
|IYW||iShares Dow Jones US Technology||May 2000|
|IYZ||iShares Dow Jones US Telecom||May 2000|
|IDU||iShares Dow Jones US Utilities||June 2000|
Taking into account the inception dates and the data required to calculate the indicators, the next simulations run from 6/1/2001 to 1/1/2013.
This time, the market timing rule is applied in three ways:
1-Individually. Same individual rule as the index portfolio, with a position limit of 1/11.
2-Globally. Same global rule as the index portfolio.
3-Combined. Individual and global rules are applied simultaneously, and the position limit is dropped. In other words: the portfolio is in cash if SMA50 <= SMA200 for the S&P500 index. Else, all ETFs with SMA50>SMA200 are equal-weight.
Here are the results, compared with the previous portfolios on the same period.
|SPY Buy & Hold||43.93||3.19||-55.42|
|5 indexes Global MT||137.64||7.75||-19.48|
|11 sectors Buy & Hold||82.09||5.31||-55.67|
|11 sectors Individual MT||104.22||6.35||-17.12|
|11 sectors Global MT||152.25||8.31||-17.3|
|11 sectors Combined MT||166.66||8.83||-17.34|
Once again, the global timing is better than the individual timing. The tiny additional return of the combined version is not really significant.
Sometimes, simple management rules give a higher return and a lower risk. This is the case here: timing the global index is better than timing all portfolio components.
I plan to write a 3rd article about timing global assets. Follow me if you don't want to miss it and Click here for more information about our research.