Ponderosa Pine Trees
Several months ago I conceived the idea of how one might use Bullish Percent Indicator (BPI) data to actively manage portfolios. Readers of this blog who have been involved with the stock market for any significant period of time know just how difficult it is to outperform the S&P 500. Keep that thought in the forefront of your thinking as this idea is developed. We also know the S&P 500 is divided among eleven different sectors. In all the BPI posts, the second data table is the one that contains sector BPI information and it is that data we will use to manage Sector BPI portfolios.
The idea is to purchase a sector of the S&P 500 when the stocks that make up that sector are depressed and sell them when they are exuberant. That is the hypothesis behind this portfolio management experiment.
A hypothesis is an assumption before any research is done. That is the current condition of the Sector BPI Model, although we are launching the experimental research and hope to move this idea toward a theory by the end of 2023. It will take several Buy and Sell cycles before we know if this “Gedankenexperiment” is a profitable model.
The first move in this experiment was to convert the Carson LRPC into the Carson. Assets from the Carson HA and Carson BHS portfolios joined forces with the Carson LRPC to form the Carson. We are a little shy of two months working with the Carson. The next move was to shift the Franklin, a poor performing Dual Momentum™ portfolio, over to the Sector BPI model. Two more portfolios will make the shift over the next two weeks. More on this later.
Rules of the Sector BPI Model
- If the Bullish Percent Indicator (BPI) percentage dips to 30% bullish or lower, the sector is considered over-sold. When this condition exists, we purchase shares of the sector ETF. Each individual portfolio has maximum percentages that may be invested in a given sector. This information is provided in the Kipling spreadsheet. As you follow the reviews of the Carson and Franklin these limits will become quite clear.
- If the BPI for a particular sector rises to 70% or higher the sector is over-bought.
- If the percentage is in the 70s we set a 3% Trailing Stop Loss Order (TSLO).
- If the percentage is in the 80s we set a 2% TSLO.
- If the BPI percentage is in the 90s we set a 1% TSLO.
- If the BPI percentage for a sector lies between 30% and 70% bullish, we do nothing. Sector ETFs will spend the majority of their time in this zone.
- Available cash is held in a short-term treasury such as SHV or a TIP such as SCHP.
On occasion anomalies arise that falls outside the above investing rules. If and when these anomalies occur management decisions will resolve the problem.
Questions and Comments are most welcome. Post them in the Comment section provided below.
Discover more from ITA Wealth Management
Subscribe to get the latest posts sent to your email.
Hi Lowell,
This sounds like a promising idea. How will you handle the transition to this methodology within the existing portfolios?
Thank you,
Bill
Bill,
Good question. What I’m doing is placing TSLOs of 3% to 4% under existing holdings. With current cash I am placing limit orders for Energy, Technology, and Real Estate. These are the three sectors currently in the Buy zone.
I’m moving the Millikan and Gauss over to the Sector BPI Model as neither is performing all that well. This will give me a good base of portfolios to check the Sector BPI Model.
Lowell
Thanks for the quick and detailed reply.
Merry Christmas
Lowell, this sounds like a good hypothesis. Couldn’t this be backtested to speed up the validation of the hypothesis? I would volunteer but I am now so far out of the backtesting business it would take me too long to come back up to speed. Maybe John Dishman? David? With backtesting, you could also tweak the parameters for trailing stops and buy-in.
Happy Holidays to all and better returns in the future!
Ernie
Ernie,
There are sufficient BPI blog posts to run some back-testing, but I’m not up to it. As you know, I am somewhat of a skeptic when it comes to back-testing. Yet, it is about as reliable as one can get when it comes to checking out a model.
The Sector BPI model is quite simple in theory. Where I see problems is when nearly all the sectors dip into the Buy zone and there is insufficient cash to fill all the requests. That is when management decisions are required. I have a few ideas, but will test them when the situation arises – as I expect it will sometime in 2023.
I’m not asking anyone to run back-tests. The rules are sufficiently clear should anyone wish to delve into back-testing.
Ernie – If you or any other readers ever hear of a similar investing approach, let me know. I don’t want to take original credit if someone has already developed such a model. To my knowledge, the Sector BPI Model is unique to the ITA website.
Sending Holiday Greetings to all readers.
Lowell
Lowell, thanks. I share your skepticism about backtesting. After having fun doing years of backtesting, I finally realized the limitations of backtesting that are based on the fact that the economic and financial conditions of the market are fluid and ever-changing with time. But, I think for the same reason a real-time test of the hypothesis is fraught with the same issue. The conditions over the actual test period may not reflect the conditions going forward.
I am not aware of any similar approach, but I have not had my head in investment models for some time now. I’m not the right one to ask.
Regards,
Ernie
Ernie,
My goals for this new model are several.
1. I would like to outperform SPY. This is a tough assignment.
2. The Sector BPI model should outperform both the AOA index ETF and the Schrodinger portfolio.
3. The overall risk of the Sector BPI portfolios should be lower than the Schrodinger or computer managed Robo Advisor portfolio.
Lowell
Readers,
I should add another goal. The management rules are so simple anyone one can manage a portfolio if they follow the above guidelines.
Lowell
Lowell and Ernie,
Yes, I still do R-programming and back-testing using two models: 90 day momentum and 200 day simple moving average. I have several ETF portfolios that I test across the major economic sectors, in some cases up to 20 sectors. I could likely modify my code to do the BPI algorithm. However, I would need a formula for BPI to plug in. Is there someplace you could point me for the definition you are using?
Merry Christmas!
John
John,
Would the BPI addresses from Stock Charts for the 11 different sectors be what you are looking for?
Lowell
John,
The BPI spreadsheet should be in your mail box.
Lowell
LOWELL
https://stockcharts.com/c-sc/sc?s=%24BPENER&p=W&yr=3&mn=0&dy=0&i=t6595739872c&r=1671973918832
The above link is to a Stockcharts 3-year weekly BPI% graph for Energy sector. While this is not back testing it does show the pattern for 3years and when BPI% goes through the 30% or 80% limits.
Bob
Attention Readers: If you read the latest BPI blog post you will see I am refining the TSLO recommendations posted in this blog. One can only follow my latest recommendations if the broker permits one to set TSLOs to the nearest tenth of a percent. Schwab permits this, but TD Ameritrade does not. I don’t know the rules at Fidelity, Vanguard, etc.
Lowell