Good In The Straightaways, Not So Good In The Turns

By: | Sun, Aug 24, 2008
Print Email

Rotational combines component rotation and asset class rotation to hold a small basket of ETFs or ETNs, selecting the handful with the most momentum from a representative sampling of classes and components. Throughout this article, when I refer to momentum, I am referring to an exponentially smoothed measure based solely on price movement.

Information is as of the close on August 22, 2008.

Model Allocation

Based on beginning with a $100,000 portfolio at inception, these are the current weights and holdings. The initial target was a buy of 10% weights per position. See my previous post on this system. Sort is alpha order by ticker and weights are rounded to the tenth of a percent.

Agriculturals (DBA) 10.2% weight
Gold (GLD) 8.8% weight
Oil Equip/Srvcs (IEZ) 10.4% weight
Natural Resources (IGE) 9.7% weight
Oil Services (OIH) 10% weight
Energy Exploration (PXE) 9.3% weight
Silver (SLV) 7.7% weight
Natural Gas (UNG) 9.1% weight
Oil (USO) 14.8% weight
Materials (XME) 9.8% weight
Cash 0.2% weight


Based on beginning with a $100,000 portfolio at inception.

Equity: $83,868.82
Gain, Past 4 Weeks: -6.75%
Gain, Year to Date: -17.46%
Gain, Since Inception on 11/19/2007: -16.13%

None of the ETFs in the Rotational portfolio paid dividends or distributions in the past four weeks.

Total dividends = $0.00 on the tracking portfolio. This amount is included in the returns shown above, and will remain in cash until needed for a new purchase. Note, commissions are expensed at $10.00 per trade when accounting for returns.

Changes To Model Allocation

Rotational screens for momentum inside a list of ETFs and ETNs by asset class category. The system is holding the top 10 issues, ranked by momentum, regardless of which asset class they are in or how much momentum they have.

If this system were to be initiated today, the target allocation would be a buy for 10% weight holdings of the ten issues highlighted in gold or green in the table below. Items highlighted in gray are "sells" from the existing model portfolio.

If the table is truncated in your browser, click on it to view it in its own pane. Depending on your browser, you may have to click again to view it in full size.


Everything except USO will be sold, market at open on Monday. The proceeds, plus cash, comprise 85.2% of portfolio weight, and will be used to buy shares as shown, based on the closing prices on August 22, at 9.5% weight each. I will round down any fractions in the share calculation.


Below, I present the change in rotational momentum from the last evaluation to the current one. It can be quite instructive.

Here is a table that shows the average momentum for the different issues in each asset class, at different evaluation dates from the inception of the program.

On a straight-weighted-average basis, all asset classes have negative momentum. This is an exceedingly rare occurrence.

Bonds are one of the two classes that gained momentum over the last month, although they're still in negative territory, as they have been since the June update. The U.S. Treasuries are positive - albeit barely - and corporate and emerging market debt is at the bottom of the momentum scale. This can only reflect a re-emerging flight to safety over the last month or two.

Commodities as a class have moved from having the most momentum, to having the second-lowest momentum, in just four weeks. At the last update, I spoke about them losing momentum rapidly over the past 16 weeks, but these recent four weeks have been a frightful collapse. The precious metal complex had briefly been a gainer - from a possible resurgence of "inflation fears" - but that was apparently just a fake-out, as they have fallen pretty badly in the last four weeks. This is the kind of momentum shift that kills trend-following systems like Rotational. The debate will rage about whether this is a correction in an uptrend, or the end of the bubble in commodities; my vote is that the bubble is popped, and the energy complex has seen a "blow-off top" that will constitute overhead resistance for any subsequent rallies.

In the April update, Rotational Portfolio is Overweight Commodities, and in my comment on that update, I detailed my reasoning for switching my personal trades from the Rotational system to one based solely on U.S.-traded stocks, Aggressive. My intuition was that the global macro trends would change and that other systems would start to outperform, since Rotational does poorly during shifts in trend. I had absolutely no idea it would shift so drastically, however. Over the next few months I'll be spending some testing time trying to mitigate the impact from future shifts of this nature. Note that my intuition about the shift influenced the timing of my changeover, but regardless of timing, I still think that Aggressive is a better long-term system for my account, based on size of assets managed and risk-tolerance.

Currencies competing against the dollar have fallen in momentum, moving into negative territory as a group for the first time since I've been tracking this model portfolio. Only one currency, the Mexican Peso, is in positive territory. The Aussie Dollar, being commodity-based, has fallen fastest, but as it's become obvious that the U.S. will rebound economically while malaise is still settling in across the pond, all of the European currencies are falling. The "carry trade" suffered this past month, as well, moving from positive to negative momentum.

The foreign stock markets fell out of the turnip truck in June; in July, the turnip truck shifted into reverse and backed over them; in August it looks as if the truck drove over them again on its way to market. Did I carry that metaphor too far? Probably. The foreign markets have the worst momentum of any group, and the worst momentum they've had since this tracking portfolio was launched. No foreign market that I track has positive momentum, and the U.S. market (represented by the SPY) has better momentum than any foreign market. This may bode well for my thesis that the U.S. markets will outperform from here. All markets except the U.S. and India lost momentum last month, and India's "recovery" may be just a bounce from a severely oversold low momentum score.

The domestic industry action is really entertaining. As an equal-weighted group, it was fairly stable since the last update, but was clearly bifurcated when viewed in detail. Energy and material industries were thrashed within an inch of their lives, and utilities fell; but all other groups gained in momentum. Banks and homebuilders gained from positions at the bottom of the momentum chart, but a large variety of other industries from the middle and top of the charts continued to rise in momentum. Gainers from the prior month continued to gain, with the exception of gold miners, who fell. Technology and industrial sectors had a marked rise from the middle of the charts. Technology was a riser two updates ago, with a falloff last update, so I would be watching them as a potential leading sector when the market calms down.

REITs are the other asset class, besides bonds, to have an increase in momentum this past month. Unlike the semi-negligible increase from the bond class, REITs gained pretty strongly, and though they're still in negative territory, momentum-wise, they are ahead of foreign markets, commodities, and the straight-weighted average of domestic industries. Only the international and mortgage-based REITs failed to gain momentum, and one category, the residential REITs, gained enough momentum to make them a "buy" for the system portfolio.

Rotational is a trend-following system that believes there's always a bull market somewhere, and in previous downturns this year, the portfolio has been able to ride it out OK because it was in "what was working." Not this time. This is the largest drawdown the system has generated in test or tracking. It's important to remember that, no matter how long any system's backtest, that its worst drawdown is always in front of it, just as its best performance is always in front of it. A backtest (or an actual trading return stream) is only a small sample from a larger heteroscedastic distribution, and as long as the recent results are within the same rough order of magnitude as the sample (they are), there's no evidence of "brokenness." In backtest, the system had a half-dozen drawdowns in the teens, and the current drawdown is 23.3% from peak equity, which is comparable to its backtested compounding rate and average annual return. It's potentially unnerving, yes, but not outside of the expected range of results. The system doesn't change, however. It just executes, regardless of circumstance, and while it does tend to to suck wind at turning points, it makes up for it in the straightaways - because there's almost always something trending. We'll see if the current leaders, stay leaders.

If you'd like to become of member of The Rempel Report, you can register here. Members receive email notification of new posts and can contribute to the site through comments. Registration is still free!



All Images, XHTML Renderings, and Source Code Copyright ©