08/17/2014 Madness to my Method

Style Model	Small Value
Sector Model	XLF	-0.70%
Large Portfolio	Date	Return	Days
TIVO	4/23/2014	11.18%	116
SHOO	4/28/2014	-5.35%	111
PM	5/27/2014	-1.37%	82
SR	6/2/2014	16.70%	76
CFI	6/9/2014	-3.63%	69
FRAN	6/16/2014	-12.72%	62
RRD	7/21/2014	6.58%	27
CHFC	7/28/2014	-1.21%	20
ESI	8/4/2014	-41.30%	13
BSET	8/11/2014	-2.08%	6
(Since 5/31/2011)
S&P	Annualized	12.34%
Sector Model	Annualized	25.17%
Large Portfolio	Annualized	23.03%

 

Rotation: selling PM; buying STRA.

This is the second purchase in the latest iteration of the fundamental selection process.  The original goal of the Full Model was to treat the Sector Model as a benchmark, and to find a way to combine fundamental and technical screens into a single model.

The first iteration used a simple Greenblatt sort for Earnings Yield and Return on Total Capital.  In the first year of trading, that method soundly beat the Sector Model.

Then I got creative.

The British would call this kind of creativity “too clever by half.”

I could claim that there was a method to my madness, while it may be closer to say there is madness to my method.

In any case, it’s useful to state my current goal:

I want to combine the returns of my Sector and Style models, in a fully scalable model that can handle both small individual funds (like my own) and large institutional ones.

If I wanted to simply make 25%, I’d simply buy the largest stocks in those industries that fell in the buy zone on my full model, and call it a day.  That would provide both scalability and duplication of the Sector Model’s returns.

But it’s that third aspect to my goal that has proven elusive thus far: to add the potential of the Style Model to my Full Model’s returns.

To give an idea of how that looks, here are the sector and style models, with the combined goal projected as an overlay:

So far I haven’t succeeded.  Although 23% annualized returns for the full model might look interesting, it’s not what I’m after.

The key is to find fundamentals that reflect each “style” as a market “condition.”  So, for today the style is “Small Value.”  Recently I tried to simply look for small caps with high earnings yield, but that proved particularly disastrous in the selection of ESI.

It was a useful failure.

Mapping out correlations is time consuming, but this is what I’m currently looking for:

Low dividend growth and low long term returns, combined with high sales growth and earnings yield.

The low dividend growth seems counter intuitive for a “value” stock, but it has a strong correlation with good returns against my model.

Time will tell.

I’d write more, but I have two toddlers to play with.

Have a great week.

Tim