Ranking My Tracking

A week ago, we had a discussion on my blog about alternative ways to identify the "best" stocks on the MFI lists.  I have shown that since 2006, you have been much better off going with dividend stocks (yields of 2.6% or greater) and that you have been better off staying away from the small stuff, say $800 million and lower in market cap.

The question was asked about ranking.  I do create a top 200 list every month or so and I do rank the top 50 official stocks when I do that.  So if you picked the top 5 or top 10 from my ranking would you do "better"?

Difficult question to answer.  I think the biggest issue is the data.  I only have data going back to July 2011.  It has been a steady upmarket since then and MFI has done well. In addition, in the entire market the Russell 2000 Index (smaller stocks) has done very well. So I would argue the data may have some bias.  Then it is always a bit messy to do back tests.  I have some decent tools, but I always worry if I am capturing special dividends, stock splits, name changes, buy outs etc etc etc.  I do the best I can, but I do have a day job.

So with those caveats, here is a table I created.  It has about 23 or 24 portfolios going back to July 2011:

Rank	Pct Change	Rolling Five	Rolling Ten
1	15.4%
2	17.0%
3	25.1%
4	12.6%
5	16.0%	17.2%
6	26.7%	19.5%
7	20.2%	20.1%
8	17.9%	18.7%
9	40.2%	24.2%
10	41.2%	29.3%	23.2%
11	32.5%	30.4%	24.9%
12	52.3%	36.8%	28.5%
13	12.6%	35.8%	27.2%
14	35.0%	34.7%	29.5%
15	17.2%	29.9%	29.6%
16	33.5%	30.1%	30.3%
17	22.1%	24.1%	30.5%
18	28.4%	27.2%	31.5%
19	20.1%	24.3%	29.5%
20	19.4%	24.7%	27.3%
21	36.0%	25.2%	27.6%
22	32.0%	27.2%	25.6%
23	40.5%	29.6%	28.4%
24	9.3%	27.4%	25.8%
25	31.6%	29.9%	27.3%
26	9.9%	24.7%	24.9%
27	22.1%	22.7%	24.9%
28	9.8%	16.5%	23.1%
29	17.6%	18.2%	22.8%
30	22.1%	16.3%	23.1%
31	33.1%	20.9%	22.8%
32	25.3%	21.6%	22.1%
33	29.1%	25.4%	21.0%
34	19.4%	25.8%	22.0%
35	19.3%	25.2%	20.8%
36	12.3%	21.0%	21.0%
37	20.2%	20.0%	20.8%
38	28.7%	20.0%	22.7%
39	12.2%	18.5%	22.2%
40	18.4%	18.4%	21.8%
41	11.6%	18.2%	19.6%
42	14.1%	17.0%	18.5%
43	24.5%	16.2%	18.1%
44	22.1%	18.2%	18.3%
45	18.4%	18.2%	18.3%
46	13.1%	18.4%	18.3%
47	26.1%	20.8%	18.9%
48	21.4%	20.2%	18.2%
49	17.1%	19.2%	18.7%
50	11.0%	17.7%	17.9%
Grand Total	22.7%	22.7%	22.7%

Now you can see that the average stocks has gone up 22.7% (generally a year).  As I emntioned, this has been a real bull stretch. For whatever reason, the #9, #10 and #12 slots have done very well.  Do I think there is something magical about being #12 on the list?  No.  Here are all the #12 stocks

Date	Stock	Initial Price	End Price	Percent Change	Mkt Cap	Rank	EY	ROIC
8/16/2013	UIS	24.25	33.06	36%	1,060	17	18.3%	123.7%
7/12/2013	HLF	48.36	66.16	37%	5,050	16	12.6%	267.1%
6/28/2013	AUXL	16.63	29.72	79%	820	17	12.6%	244.0%
5/31/2013	DLX	36.78	52.25	42%	1,896	16	12.5%	307.1%
4/19/2013	CPLA	30.18	65.43	117%	376	20	21.3%	131.7%
3/11/2013	RTN	54.61	100.73	84%	18,378	23	15.2%	151.9%
2/1/2013	AAPL	429.1	500.6	17%	425,937	18	13.7%	357.4%
12/7/2012	LPS	24.11	35.11	46%	2,066	21	14.4%	390.9%
9/21/2012	WCRX	12.56	22.93	83%	3,256	20	13.9%	421.0%
9/21/2012	VG	2.4	3.01	25%	544	21	16.0%	162.1%
8/17/2012	GME	17.61	47.05	167%	2,312	22	27.3%	105.1%
7/21/2011	SCEI	2.27	1.02	-55%	53	15	--	159.2%
7/9/2012	UIS	18.22	24.02	32%	856	19	34.4%	179.7%
5/4/2012	CPLA	32.98	36	9%	411	20	27.2%	159.7%
3/23/2012	VECO	30.35	38.41	27%	1,137	21	42.9%	150.3%
2/24/2012	ASTX	1.96	3.1	58%	189	24	21.8%	184.3%
1/29/2012	CHKE	10.28	14	36%	123	21	16.4%	377.4%
12/21/2011	UIS	20.27	17.3	-15%	14,887	20	26.7%	167.9%
12/8/2011	ACAD	1.05	4.55	333%	53	21	15.7%	2177.3%
11/3/2011	CPLA	36.62	31.74	-13%	446	22	23.8%	189.6%
8/26/2011	CPLA	30.17	35.32	17%	453	19	31.3%	209.6%
8/26/2011	MNTA	16.35	14.41	-12%	832	18	32.0%	212.0%

So you can see that the results here were a bit distorted by ACAD running from $1.05 to $4.55 (I looked and it is $26 now!).  My view in looking at the data overall is that there does seem to be a case made for staying away from #41 to #50.  But the first 40, I do not see a pattern.

By Earning Yield

My second cut of the data was by earnings yield.  We just showed that overall rank within the top 50 did not seem to matter, except in the bottom 10 (perhaps).  How about if you are one of the higher Earnings Yield stock (ignoring return on capital)?  Recall earnings yield is kind of the inverse of price to earnings ratio.

Earning Yield Decile	Avg % Chg
1	27%
2	41%
3	32%
4	30%
5	23%
6	19%
7	19%
8	14%
9	15%
10	11%
Grand Total	23%

This is a bit more interesting.  This seems to be a descending pattern.  A "10" are the 10% of stocks with the highest earnings yields.  So it appears that these may be the classic value traps from time to time.  Meanwhile, the lower earnings yields (recall they are still good) actually do better.  The #10 slot had stocks like ASYS, CECO, ESI and NSU that dampened their results.  So the moral here is that we do want to buy "cheap" companies, but if they are the cheapest of the cheap - we may need to kick the tires a little harder.

By ROIC

Same exercise, now looking at ROIC by decile.

ROIC Decile	Avg % Chg
1	14%
2	10%
3	41%
4	32%
5	29%
6	27%
7	0%
8	24%
9	21%
10	33%
Grand Total	23%

This one seemed a little more random.  Not sure I see anything.  Obviously the #7 stands out. For whatever reason, that decile got a lot of the APOL and ESI types that did poorly for a chunk of the period we're looking at.

Conclusion

While I did have two "findings": the bottom 10 in the top 50 appear to be the worst performers and that the very highest earning yield stocks appear to under perform as well;  I would call these directional at this point.  After another year or two, I can follow up with a little more data and we will see what we will see.