ETF Update: On the Couch with Dr. Brett

Our ETF Update typically takes a sector-based look at the market, and spotlights a particular choice.  This week is a bit different.  Last week's signal to go short was a big loser.  It calls for deeper consideration.  We have whimsically called this a trip to the couch with Dr. Brett Steenbarger.  (While we enjoy an occasional meeting for coffee with Brett, the advice given here comes from his blog, not a specific consultation:)


Each week we take a look at the ETF universe, emphasizing sectors rather than the overall market or individual stocks.  Our ratings reflect sector Trends, but also pick up some Cyclical behavior.  Finally we add a touch of Anticipation, so we call it the TCA-ETF Model.  (The complete current rankings are at the end of the article, along with an explanation of our methodology).

The model has a one-month time horizon, but we generate new ratings at least twice a day.  Can the one month forecast change from day to day?  Of course.  Events change, and we react.

Last week we wrote that there were no attractive sectors, and we took positions in the inverse ETF's, leaving us 40% to 50% net short at the start of the week.  This was an instant loser, and got worse as the week progressed.

A Fictional Conversation with Dr. Brett

(Jeff and Vince, the modeling guru, have a fictional chat with Dr. Brett Steenbarger , noted author, trading consultant, and featured site for our readers.  BTW, it could be Dr. Jeff and Dr. Vince, since we have a group of PhD's.  We are invoking the psychological professional sense here, since we are the patients!)

Jeff:  Vince, what went wrong last week?

Vince:  This has been a news-driven market.  None of the data-based models can deal with this.

Jeff:  Markets are always driven by news.  You must mean something more….

Vince:  Yes.

Jeff:  There is news that is "in the market" somehow.  Perceptions change before stocks move.  Insiders act.  Markets signal what is going to happen.

Vince:  Yes.  When that happens, there is a message from the market.  Our method is first-rate at picking up that message.

Jeff:  Sometimes there is a news "shock" if you will.  Something happens that is totally unanticipated — a natural disaster  — there are few genuine examples.  Even wars generate market anticipation.  Does Meredith Whitney qualify?

Vince:  Obviously, her positive call on Goldman was not anticipated.  It also came at a crucial time, a make-or-break point for many amateur technicians.

Jeff:  Perhaps we should have a "regime switching" model, reflecting the possibility of a trading range.  We have discussed it.  Dr. Brett (and some others) highlighted this possibility.

Dr. Brett:  The source of the market message was determined bearishness by the most active traders.  Here was Dr. Brett's take from mid week:

"How could we go higher when the
economy is so weak?" was a common response. The key to the replies,
however, was the emotional tone of indignation–almost as if I had
insulted their family members.

traders had a *need* to believe in a bearish thesis; their beliefs were
not grounded in how the market was actually trading.

Jeff:  Some of our accounts have a different time frame.  The long-term positions did very well–better than the averages.

Vince:  No system will be right all of the time and certainly not every week.  Your "fundamental" picks did better this week, but my model will prevail another time.

Dr. BrettIt's vitally important that the time
horizon of your analyses fit the time horizon of your trading and
investing. Mismatches will take you out of good investments on
short-term weakness, and they will keep you out of short-term rallies
on longer-term pessimism.

Jeff:  Those actually trading a model and following the day-to-day performance are focused on  short-term performance.  We compare results to our simulations.   Last week was not at the extreme for total losses, but it was at the limit for comparison to the benchmark.

Vince:  This is why we develop and test models over long time horizons and varying market types.  Effective simulations are better than real-time results, since they cover differing market circumstances.

Jeff:  Some developers pollute their out-0f-sample data by using it too much.  Dr. Brett wrote about this problem.

Dr. Brettit's possible to run so many tests and so reduce your sample size that you wind up with false certainty.

also possible to find results specific to the historical period that
you are testing–dangerous if regimes are in the process of shifting.

Used properly, however, the combination of careful observation and testing can yield very fertile hypotheses for traders.

Jeff:  While I understand and appreciate the theoretical strength of testing methods — and Vince is the best — it is not a satisfactory answer to a week like this.  We have been talking about expanding the ETF universe…

Vince:  It is always good to have more choices.

Jeff:  It would not have changed much last week.  We are going to use a larger list and will show the top twenty in this report.

Weekly TCA-ETF Rankings

Our performance for last week was down about 2.5%, while the market was up about 7%.  This wiped out our recent gains versus the benchmark, and a bit more.

There is still very narrow leadership in our sectors, but the inverse ETF's have dropped into the penalty box.  We have shifted to a neutral position in the the Ticker Sense Blogger Sentiment poll.

Here is the table for the most recent trades and current ratings as of Thursday's close:


Bonus Rankings

Here are the top sectors from an expanded universe of ETF's.

Extended ETF improved

Note for New Readers

Our weekly ETF Update is designed to assist both investors and
traders interested in ETF's and Sector Rotation.  Before turning to the
current rankings, let us undertake a review for readers new to this

Our Method.  In this past article,
we described our basic methodology and why we believe the rankings are
useful for fundamental traders and technical traders alike.  While we
urge readers to check out the entire article, the key point is that
ETF's pose challenges and opportunities different from investment in
individual stocks.  The fundamentals may be more difficult to assess. 
Even with a good grasp on fundamental trends, there is a lot of
technically-based trading in ETF's.  This means that those trading with a fundamental approach (and we do this as well) want to monitor the "hot money" moves.  Here is an article on that point.

The system synopsis.
We look at Trending sectors, Cyclical Sectors, and build in an element
of Anticipation for both entry and exit — thus the name of the model,
TCA-ETF.  While we do not reveal the exact methodology for spotting
trends and cycles, the system is not a "black box."  The basic elements
are used by many, and widely reported.  We even discuss the need for human analysis as opposed to black box trading.

We report the rankings
each week, now on the weekend with a one-day delay, using the Thursday
output from the model.  We monitor and trade this daily, and offer a
free report (request via the email address on the top left of the site)
for those interested in our weekly trading program.

You may also like


  • Kenboe July 20, 2009  

    Hi Jeff,
    I know you were not happy with your ETF model last week. Many technicians were caught “short” including technical sage, John Murphy of I have a couple of questions: 1. Do you always let the model play out for the week or do you ever acknowledge it’s wrong and change course midweek? 2. You state this week that “the inverse ETF’s have dropped into the penalty box.” But only PSQ is in the box, not SH or DOG according to your 7/19/09 ETF Update. What’s up with the discrepency between the Sector Report Chart and your narrative?
    Thanks for responding. Kenboe

  • Jeff Miller July 21, 2009  

    Kenboe — We run the model at least twice each day. Circumstances change during the week. In the “partnership account” which trades daily, we make immediate adjustments. For those participating in the “weekly program” we make mid-week adjustments when the general tenor of the model recommendations change.
    As to the discrepancy in the narrative, you made an astute observation. I was trying to suggest that additional data were getting us involved in new sectors — a hint about the Friday ratings and what I saw coming.
    Thanks for a helpful question!