White Lies, Lies and Statistics

As my readers know, i have been creating a MFI tracking portfolio every month since January 2006.  As we approach 2014, I thought it would be fun to look at some statistics from the overall database of over 4,500 stock years.  I almost always take the top 50 stocks over $100 million from the screen,  There have been a few times (maybe ten) where I did not follow that rule.

Performance by Market Cap Decile

This table looks at the 4,500+ stock years and aggregates them into 10 decile buckets by market cap and time of buying.  At some point, I suppose I should market adjust this or inflation adjust it, but that seems like a lot of effort.  It is extremely interesting that larger cap stocks have clearly done better.

Decile	Min Mkt Cap	Max Mkt Cap	Avg Pct Chg
1	8,335	500,025	16.0%
2	2,931	8,327	14.1%
3	1,581	2,925	15.2%
4	1,013	1,579	16.2%
5	744	1,010	5.7%
6	507	743	6.7%
7	361	507	8.0%
8	251	360	7.8%
9	167	251	4.1%
10	-	167	6.1%
Overall	-		10.0%

Performance by Purchase Month

One would not expect seasonality, as you hold the stock for a year.  I do not see much. Just eye balling the quarters, it seems you are best off buying in 3rd month of the quarter.  I guess most companies report in the second month of the quarter, so earnings have been digested and perhaps stocks have drifted based on other news or fear.

Decile	Avg Pct Chg
1	8.6%
2	15.3%
3	12.7%
4	9.5%
5	11.1%
6	16.6%
7	1.5%
8	6.6%
9	8.7%
10	14.6%
11	5.3%
12	15.7%

Performance by Number of Letters

I know this one seems stupid, but it actually seems to have predictive power.

Letters	Avg Pct Chg	Count
1	34.4%	1
2	19.3%	176
3	16.6%	1,748
4	6.4%	2,721
5	-0.4%	21
7	-53.6%	37
8	-66.1%	13

Stocks with 3 letters or fewer have done much better than stocks with 4 or more letters in their ticker.  Looking at my 6 tranches since I rebooted MFI, stocks with 3 or fewer letters are up 49.2% vs 19.3% for their longer counterparts.  Very Interesting.

Performance by Yield

I have shown numerous times that stocks in my tracking portfolios with yields greater than 2.6% have done very well.  I have tried to show that here, but I only captured yield on a regular basis since January 2010.  And even then, keep in mind that I generally used Yahoo Finance, which is not the most accurate data source on the web.  But directionally, this is still helpful I believe.

Yield	Avg Pct Chg	Count
Under 1%	7.9%	1,436
1 to 2%	19.7%	143
2 to 3%	20.9%	177
3 to 4%	35.7%	167
4 to 5%	29.1%	95
5% or higher	18.1%	330
Grand Total	13.9%	2,348

You can see the real sweet spot has been in that 3 to 4% range.  If I looked at over 26.6% and under 2.6% I would get this:

Yield	Avg Pct Chg	Count
Under 2.6%	9.4%	1,676
2.6% and higher	24.9%	672
Grand Total	13.9%	2,348

I guess we already knew this, but it is interesting to see the confirmation.