Sector Rotation and ETF's: The Need for a System

We are attempting to comment weekly, each Thursday night, on the subject of system trading.  We are using our own TCA-ETF model as an example, mostly because things are easier to see when the illustrations are clear.

We expect the content of this weekly series to include topics related to System Trading, ETF Matters, Sector Rotation, and our own Weekly Update.  Reader suggestions, as always, are most welcome.  Covering these topics carefully involves a weekly review of developments on all fronts.  Readers might well choose to view this as a "linkfest" of thinking on important topics illuminated by the sector ratings.

System Trading

Most individual investors and traders fail.  Human instincts --perhaps even survival instincts -- may teach the wrong lessons for investing.  Two recent posts from Barry Ritholtz illustrate the problem.  One shows the success of email spam ventures.  (Guess what?  The pros are not buying these.)  The other links emotion and trading, citing a book that we join Barry in recommending and one that we now have on our reading list).

As usual, Dr. Brett Steenbarger writes authoritatively on this topic.  He shows how emotions interfere with the normal process of causal reasoning.  Let us consider an example we can document from our own trading.

A system can be based upon fundamental analysis.  For those of us holding Apple Computer, Inc. (AAPL) fifteen months ago, there was a great challenge related here and here.  Since we were confident of our fundamental analysis, we ignored extreme volatility in the stock.  Trying to exit and get back in was almost impossible, unless one is capable of selling and then buying a gap to a higher level.

A quantitative, computer-based system helps with this.  If you are confident in your method, you can ride out adversity.  Getting that confidence requires proper understanding and testing of the model.  (There can be exceptions to the black box approach, but that is a future topic.)

For now, let us merely note that selling Apple (which we still hold) because it looked "toppy" or got "oversold" would have been a major mistake.   A system can help with this temptation.

ETF Matters

International ETF's are extremely overbought, we learn from the astute technical analysts at Bespoke Investment Group and TickerSense.

Part of the rise has been based upon the falling dollar.  The decline of the dollar is just a question of recency, according to former Fed President Bill McTeer.  Market participants are looking at the wrong index, says Bill Rempel.  An orderly decline of the dollar is a good thing, argues Karl Smith, a young professor who has quickly earned our respect through his writings.

Since many of our positions relate to the dollar, as we noted last week, this bears careful watching.

Sector Rotation

Our TCA-ETF method is based upon sector rotation, a concept we have implemented successfully for nearly ten  years.  We were therefore surprised to learn (thanks to Abnormal Returns--always a great source for new sites) that there is a strong argument against this approach from an authoritative, scholarly source.  Why the difference in conclusions?  While details are beyond the scope of this update, let us merely say that there have been major changes in trading sector rotation.  These are largely due to the trend, described by Tom Lydon, away from mutual funds and toward ETF's.  Those who have adapted their methods can continue to succeed.  We shall pursue this important subject further.

Weekly Update

Our own sector ratings have changed little from last week.  Here are the ratings and results as of Wednesday's close.

The energy  position replaced a former holding that was up 5.91%, but was a loser over the course of last week.

Conclusion

A disciplined system can keep one on the right side of the market.  Nearly all of the rated sectors are in the "buy" range, reflecting overall market strength.  Once again, Dr. Brett shows why this is important.  In answering a reader question he illustrates how fundamental analysis of earnings supports the technical picture.  Readers should look at his chart and read the entire analysis.

This type of information is in sharp contrast to the arguments from the bearish punditry, a team on a mission.

Markets described as "overbought" become so on the way to new highs.  Since making new highs is the most bullish event, they get even more "overbought."  There will be future weeks where these sectors decline, get stopped out, and rotation occurs.  One of the themes of this series is that we must understand the inevitability of losses and drawdowns.  Handling these problems is just as important as capitalizing on winners.  That is why we invite readers to look over our shoulder.

Comments

[Jason Ruspini]

I'm not sure those papers are much of a challenge to what I think your style of trading is, Jeff. The Stangl/Jacobsen/Visaltanachoti paper examines rotating pre-set groups of industries by business cycle, and it actually finds that such a strategy does improve returns - before transaction fees and it is not possible to execute without the benefit of hindsight. Either way it is pretty moot with respect to your strategy.

The Cremers/Petajisto paper is more complicated, but using it as an argument against sector trading might be misleading. First, high tracking error + "low active share" is a blunt way to describe funds that rotate sectors. For one thing it's possible that some managers ran non-rotating "closet sector funds", especially during the late 90s in technology, and the study ended in 2003 which would penalize such managers. Second, the gross return statistics (8A) for high tracking error + low active share funds in the paper are not significant. Third, the "4-factor alpha" gross and net statistics for those funds are significant, but the 4th factor (in addition to the usual three) is momentum. If you are running your own ETF momentum strategy, aside from evaluating the quality of your momentum logic, you are probably most interested in comparing its performance to gross returns without a momentum component.

Most importantly, the study covers mutual funds that are not allowed to sell short. Maybe my reading of the paper is colored by the recent past and the ease - even without hindsight - of running a 130/30 portfolio with a mix of housing, consumer disc. and finance on the short side. To me that is more of a modern sector rotation portfolio.

[Bill aka NO DooDahs!]

I'm actually long several of those you mentioned, based on *trend-following asset allocation models.* I don't personally play the macro "what if" game; my statement is aimed at those who do, and do so incorrectly.

Here's the consensus, bearish, and incorrect view, based on the dollar hitting "all time lows" - imminent "recession" in the U.S. and a "slowdown in world economic growth," as incorrectly measured by economists. The contrary macro-based play is "soft landing." What does well in each scenario? You do the math.

Me, I'll stick with more mechanical methods. On top of being fools for macro modeling on the wrong data series, I think it's foolish to try and build a trading model based on another model of macroeconomics. 2 models to get to 1 result means twice as many chances for error.

[Mike C]

"Macro models based on the trading index are flawed. To the extent that less thoughtful or less knowledgeable market participants make decisions based on such models, it give the smarter participants a contrary opportunity."

Just curious, how would one implement this knowledge in terms of actual investing/trading decisions.

It seems the less knowledgeable market participants are piling into anti-dollar trades like long gold, commodities, international stocks, emerging stocks, etc.

Would your position be to short these areas against being long U.S. stocks and U.S. bonds since the U.S. dollar is actually much stronger on the alternative measure?

What exactly is the contrary opportunity here based on ignoring the conventional measure?

[Bill aka NO DooDahs!]

Thnx for the link love.

In recent years, both the trading and the macro dollar indices have been down. As traders, this needs to be accounted for.

My contention is that those who use the trading index to infer macroeconomic conclusions are ignorant, because the trading index is not representative of economic reality. Vis a vis U.S. trading partners, the dollar is nowhere near its all time lows.

Macro models based on the trading index are flawed. To the extent that less thoughtful or less knowledgeable market participants make decisions based on such models, it give the smarter participants a contrary opportunity.