ETF Complete Portfolio Strategy Insights: Crossing the Zero Divide (Part 3)

James Kimball | July 27, 2015

The ETF Complete portfolio closed the week down -1.6%, outperforming the SPY which closed down -2.13%. The lag in performance was somewhat expected given the nature of our positions and the amount of cash in the portfolio.

The week saw a slow march down for all the indexes. After mostly putting in new recent highs on Monday, all four indexes sold off back towards the middle of their ranges and towards their 200 DMA’s. There was no great explanation for the markets. Earnings were mixed, in addition, there were some concerns about China and the upcoming FED meeting next week.
We have one position change to start out next week:

Country: Sell EWL at the market on the open 7/27/15
Country: Buy SSO at the market on the open 7/27/15



Please use the link below to see the detailed breakdown of all the ETF Complete Portfolio’s current positions and past performance:

If you have any questions about getting started please drop us an email at:

This Week’s Strategy Lesson: Crossing the Zero Divide (Part 3)

With the major markets still range-bound in this large sideways trend over the last six-months, we have been dealing with some of the higher turnover that comes with low overall TSI scores and the fact that the average TSI score for the whole investible universe remains negative.

Because the model will only hold positive TSI ETFs, these low overall TSI scores can cause situations where we have a significantly higher turnover. The turnover in the portfolio is a good thing if the new positions better reflect where the market is going from here (like getting short before a continued market drop). Unfortunately, when we make these transitions, we can never be sure which way the market will end up going.

Last week we looked at the summer of 2011 when the TSI went negative and we had a small degree of follow-thru, enough to reach some targets. This week, we will briefly look at most infamous market dive in our sample, the fall of 2008.

The Fall of 2008
It’s easy to look back on this period now with calmer eyes, but at the time, it often looked and felt like the market was in 1928 crash mode and the United States might be headed for the next great depression—(and depending on what metric you look at, the crash and long recovery time had many similarities).

The SPY put its (at the time) all-time new high in October of 2007 and spent the next year limping down before things exploded down in September and October. The average TSI score had already crossed negative a few times, but in July 2008, it made another decisive negative move and remained negative for the next 9-10 months.


We are going to take a look at both how the Sector Moderate Model fared relative to the SPY over this period and what position rotations were driving this performance difference.


The chart above tracks the overall equity of a portfolio in SPY and one in the Sector Moderate model. We can see above that during the first few months of this period, the model lagged behind the SPY. This is somewhat expected. When the market is switching between a positive and negative bias, our position switches can often lag by a small margin. And when this occurs back-and-forth, it can widen this performance gap. It is not unlike what we have been seeing over the last few months where the market has been trending sideways, occasionally looking like it was going to try to breakout in either direction before stubbornly retracing its steps.

But in the example above, the market did eventually and decisively (for our purposes) break out, this time to the downside. And it was during this transitory period that the model rotated into some alternative holdings that allowed it to actually put in a positive return over this period.


In the chart above, we have “zoomed” into the model and are looking at each individual portfolio slot. One of our money management rules is that when we rotate out of a position, we rotate into the new one using the position sizing of the old position at the time of exit and the only time money switches between the portfolio slots is in January and July, when we rebalance the three portfolio slots.

The green portfolio slot labeled above was the lucky winner of the three. It was actually in SDS for most of this period save for a short stint in XRT in early September. But it rotated back into SDS and caught most of the move up in that ETF. We actually left a little money on the table in that we moved to cash before that ETF peaked, but the net effect of taking profits was better because SDS put in its peak in November and proceeded to fall well below our exit price.
The blue portfolio slot above performed between the other two, ending the period positive, but spending most of that time in the red. The rotation out of IBB into TLT proved positive, but it was a little slow and missed some of that move. Still, to be up about 10% over this period versus the market down over -30% is a great outcome.

Similar to what we saw last week, one of the portfolio slots lagged by a greater degree. This slot did not have access to a “pure” short like SDS. Instead it was in a few of the long-oriented sectors that were holding up on a relative basis. However these too eventually sold off and were stopped out into cash. After getting stopped out this position remained in cash over most of this period as there were only two ETFs with positive TSI during most of October and November. It was only in late December that GDX went positive and this portfolio slot established a position in that ETF.

This is a good illustrative example of how our money management rules, market conditions, and the investible universe all work together to determine how the model performs over a given period.

With these ETF models, we are not trying to predict where the market will go next, rather we are creating rules and situations that will have us in the right ETFs for whatever trend the market is showing us and if conditions change, our holdings will change with them. These transitionary periods can cause a lot of “noise”, but based on how we have seen the model behave in the past, we believe that the “signal” will eventually be found.

Next week we wrap up this series looking at one final sample period where the market correction was considerably shallower and what we can learn from it.

The Current Condition of the Model

The Complete portfolio is 8 out of 9 possible positions.

Buy Zone: XHB, EWQ, PBS, TBT

Stay tuned to the daily emails for any position changes and updates.

Best wishes for your trading,

James Kimball