12/10/2012 Mousetrap

Small Portfolio XLF & IAU 17.10%
Sector XLF 23.67%
Secular IAU 10.53%

Large Portfolio Date Return Days
RIMM 7/16/2012 65.66% 147
SEAC 9/25/2012 26.07% 76
CAJ 9/25/2012 6.37% 76
DDAIF 9/25/2012 -2.99% 76
CFI 10/31/2012 14.50% 40
MO 11/8/2012 6.59% 32
EL 11/12/2012 5.53% 28
BOKF 11/19/2012 -0.34% 21
RE 11/26/2012 3.89% 14
GLW 12/3/2012 2.53% 7

S&P Annualized 3.51%
Small Portfolio Annualized 10.86%
Sector Model Annualized 14.89%
Large Portfolio Annualized 26.34%

*Returns on Friday, 12/7/2012.
**No rotation today (see below).
The other day I wrote something to a friend about compound returns that I quickly realized was wrong.
The point I was trying to make was essentially correct, but the number I gave for a compounded annualized return was wrong, and it reminded me that my model still used a placeholder linear equation that only loosely calculated an annualized rate of return, and not a more accurate exponential equation.
The good news is that the correction bumped the annualized return rate on my Large Portfolio up by a fraction of a percentage. The bad news was that my basic model may have been using the wrong rotation period.
Sure enough, the correct rotation rate on current data was 88 days, instead of 67 days.
Oh well.
My only fear with all of this is that I may discover one day that the model only works with bad math…
…we’ll find out…
Geeze, the things that keep me awake on a Friday night!
Come Saturday night I was plugging the new equations in all their proper places and realized that it would adjust my adaptive fundamental filter as well.
Come Sunday night I realized that I had no time to write a blog entry, but no need to do so, since the next rotation shouldn’t happen before Wednesday.
Come Monday, it’ll be alright… no wait, that’s a song…
Come Monday, I took a look at my revised rotation periods and fundamental filters and realized they make a lot more sense than they did last week.
So why all the fuss?
The fuss is about the difference between gross returns and net returns. Gross returns are what you think you have before trading costs and taxes leave you holding a shredded net.
Here’s how it works:

[ATTACH=CONFIG]16607[/ATTACH]

The blue line is the S&P 500 index from 1950 to the present. It’s what you would have in your account if you invested in the full index and held it for 62 years.
The red line is what you would get after you cashed out.
The green line is what you would get if you rotated your stocks once a year and paid long term capital gains each year.
Short term capital gains is even worse:

[ATTACH=CONFIG]16606[/ATTACH]

Instead of 1413.94, you’d have 1204.35 if you cashed out today and paid taxes (Bush rates), or 1078.59 if you cashed out in January (Obama rates). But if you had been actively trading the entire time you wouldn’t even have that. You’d have 800.43 if you traded once every 367 days (paying long term capital gains each year), or 355.28 if you rotated once every 364 days (paying short term capital gains each year).
355.28??????????????
Yep. Uncle Sam would have the rest, you patriot.
Whatever you pay in taxes in 1951 you can’t re-invest in 1952, and so on. The taxes keep killing your base and you can’t grow your retirement income at nearly the rate you need to survive.
Back to the only Buffet rule that we should care about: the more you trade, the less you have.
So why trade at all?
Short answer is that most people shouldn’t.
No, really. They shouldn’t trade at all. They should deposit it into SPY and walk away and NEVER look at the account. This compounding problem is also the reason that the vast majority of mutual funds underperform the market averages – they trade too much. Statistically, men trade more often than women do, and women make better returns than men do. It’s an insidious cycle in which you work harder and harder for less and less, when you would have done far better by doing nothing at all.
For 95% of investors, SPY is your friend.
Still, being a male of the species I’m doomed to trade more often, so I might as well make the best of it. Hence the periodicity chart for my model to determine the optimal holding period after trading costs and taxes are accounted for:

[ATTACH=CONFIG]16605[/ATTACH]

The original fundamental filter for the model was based on the work of Joel Greenblatt, which is designed for a one year holding period (that spike you see at the one year mark). It’s not designed for shorter holding periods, and I adjusted my model to use a more complex fundamental filter designed after the work of Benjamin Graham (that second top shows a clear outperformance over Greenblatt).
Finally, I created a self-adaptive fundamental filter that evolves with, and for, the Mousetrap. That’s the third spike on the left hand side.
The blue line is the net return for any given holding period on my model, after you remove taxes and trading costs.
The red line is the after tax net you would have from SPY (assuming you always held longer than a year and only cashed out with long term rates).
The green line is the difference – and the highest point on the green line is the optimal holding period for my model.
Of course, as the fundamentals self-adjust to the holding period, the holding period will also self-adjust to the fundamentals, until they finally reach a long term equilibrium.
Incidentally, that red line should eventually become flat. It only dips at the beginning because I started the model on 5/31/2011, just before a nasty drop in the S&P. That drop will be diluted as more trades and more data are entered.
And, as the rules change in Washington D.C. the chart will change as well. I can’t force them to make good decisions, but I can certainly try to navigate around the chaos they create.