Market Timing: A Strategy Worth 25% A Year For A Decade (Part 3)

by: Fred Piard

This is the third article of a series about Market Timing. In the first one (click here to read) we have seen that the first benefit of market timing is to limit the drawdown and volatility, not to improve the return. In the second part (click here to read), we have seen that a simple management rule sometimes gives a higher return and a lower risk than a complicated one.

Now I will consider a portfolio invested in less correlated asset classes:
- Treasury Bonds
- Worldwide Equities
- Real Estate

Here is a set of seven ETFs I have chosen for this portfolio:

Ticker Name Inception
IEF iShares Barclays 7-10 Year Treasury Bond July 2002
TLT iShares Barclays 20+ Year Treasury Bond July 2002
SPY SPDR S&P 500 Jan 1993
IEV iShares S&P Europe 350 Index July 2000
ILF iShares S&P Latin America 40 Index Oct 2001
EPP iShares MSCI Pacific ex-Japan Oct 2001
ICF iShares Cohen & Steers Realty Majors Jan 2001

Because of the inception dates and our data requirements for indicators, the simulations are performed from 6/1/2003 to 1/1/2013.

I want to evaluate this portfolio with two variants of a market timing rule. In both cases we apply the individual timing of my previous article. For each ETF, if the 50-day moving average is strictly above the 200-day moving average, the ETF is in the portfolio, else it is out of it. The portfolio is rebalanced once a week, positions are equal-weight. The variants have the same timing, but a different money management:

1) Every position is limited to 1/7 of the portfolio (14.3%).
2) No limit.

The next table shows the results for a "buy-and-hold" strategy, the two variants of market timing, and the previous article's strategies on the same time interval.

CAGR: Compound average growth return (%).

DDM: Maximum Drawdown (%).

  Total Return CAGR DDM
SPY Buy & Hold 80.59 6.36 -55.42
SPY MT 111 8.1 -17.18
5 indexes Global MT 127.32 8.94 -19.48
11 sectors Combined MT 154.17 10.22 -17.34
7 classes Buy & Hold 194.95 11.94 -45.82
7 classes MT limited 180.09 11.34 -11.38
7 classes MT unlimited 272.11 14.69 -17.88

Even the buy-and-hold version has a better return that the previous strategies, but the drawdown is still high. The market timing with limited position keeps the return in the same range and divides by four the maximum drawdown.
The unlimited version improves the average return at the price of a higher, but acceptable, drawdown.

The following table is a more realistic and detailed comparison of both variants, including leveraged versions. A 0.1% rate is applied for the trading fees and a 2% rate for borrowing the leveraged part.

Some definitions:

"Sortino"= Sortino ratio, a Sharpe-like ratio. This is a risk-adjusted performance indicator. The higher, the better.

"Kelly"= Kelly criterion, a money-management indicator that can be used as a probabilistic robustness indicator. The higher, the better.

"Sterling"= Sterling ratio, an indicator of the return strength against the drawdown. The higher, the better.

"DD-MaxDuration"= Maximum duration in drawdown. The lower, the better.

7 classes MT Total Return CAGR DDM Sortino Kelly Sterling DD-MaxDuration
limited 176.63 11.2 -11.39 0.84 0.2 0.52 90 weeks
unlimited 265.64 14.48 -18.04 0.88 0.19 0.52 88 weeks
limited leveraged x2 522.47 21.01 -21.93 0.96 0.19 0.66 91 weeks
unlimited leveraged x2 811.32 25.92 -34.5 0.9 0.18 0.58 88 weeks

Three technical remarks:

- The Kelly criterion is theoretically an invariant by leveraging. But it is not after applying trading fees and borrowing rates.

- There is a higher penalty on the return to leverage the unlimited version. In the limited version, the portfolio is sometimes partly in cash. In this case, the borrowing fees and the beta-slippage are lower.

- The drawdowns of the 2x leveraged strategies are lower than twice the non-leveraged drawdowns, in spite of the borrowing rate. Here, the beta-slippage plays in our favor.

The most aggressive investors might be seduced by the unlimited leveraged version, which has the highest return: over 25% a year. The most conservative ones might prefer the limited non-leveraged version, which has the highest Kelly ratio and the lowest drawdown.
Objectively, the best risk-adjusted performance is the limited leveraged, which has the highest Sortino and Sterling ratios.

You may want to Click here for more information about our research.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.