Using Ratios To Identify Stocks Set To Outperform Their Peers: Statistical Summary Of 2 Years Of Research

|
Includes: ABBV, ANDV, BUD, DEO, FEYE, FTNT, GILD, HEINY, JNJ, JNPR, MPC, MRK, PANW, PFE, RDS.A, S, STZ, SYMC, T, TAP, TMUS, USM, VLO, VZ, XOM
by: Stock Scrutiny
Summary

I update where my research is at after 2 years of data collection.

Presented is a table showing stocks' scores, ranks, and FY returns for each of the last two years.

From this, I unveil different trends and statistical summaries that show my research is still on the right track.

I also touch on the next steps I will be taking in moving this project forward.

Intro

For the past 2 years, I've been gathering data for a research project with hopes of supporting my hypothesis that stocks which are stronger financially than their peers will outperform similar stocks that don't have as strong of a balance sheet. To determine this financial strength, I use a wide variety of ratios and metrics aimed at measuring a company's debt, profitability, efficiency, and growth situations. If any new readers wish to see a deeper explanation of my process, my introductory article should answer any lingering questions. This week, my article's focus is to present different trends and summaries that I've pieced together from the data. There are some promising statistics I found, while there are also some developments that came as more of a surprise. Overall, I'm excited as ever about the future of the project and am encouraged by the results thus far.

Stocks used in this analysis are: Verizon (VZ), AT&T (T), T-Mobile (TMUS), Sprint (S), U.S. Cellular (USM), Johnson & Johnson (JNJ), Merck (MRK), Gilead (GILD), AbbVie (ABBV), Pfizer (PFE), Anheuser-Busch (BUD), Constellation Brands (STZ), Diageo (DEO), Molson Coors (TAP), Heineken (OTCQX:HEINY), Valero (VLO), Exxon (XOM), Andeavor (ANDV), Royal Dutch Shell (RDS.A), Marathon (MPC), Palo Alto Networks (PANW), Juniper (JNPR), Fortinet (FTNT), Symantec (SYMC), and FireEye (FEYE). All scores for companies were drawn from financials on E-Trade, and pricing data was gathered using Nasdaq.

My Data

Below, I will place a table summarizing scores for companies over the past 2 years. It may not seem very helpful at first, but when I present statistics later on in the article, it should become more clear.

2017 Score

2018 Score

2017 Rank

2018 Rank

2017 Return

2018 Return

Constellation Brands

2.38

2.12

2

1

35.68%

-22.4%

Diageo

2.18

2.36

1

2

19.87%

-1.1%

Heineken

3.47

3.5

4

3

39.9%

-15.6%

Anheuser-Busch

3.86

3.51

5

4

5.8%

-40.8%

Molson Coors

3.12

3.52

3

5

-15.7%

-31.6%

Valero

1.94

2.38

1

1

34.5%

-18.4%

Marathon

3.14

2.73

3

2

31.04%

-10.6%

RDS

4.11

2.82

5

3

22.7%

-12.7%

Exxon

3.49

3.14

4

4

-6.4%

-19.8%

Andeavor

2.32

3.95

2

5

29.77%

------------

Verizon

2.83

2.46

2

1

-1.5%

6.2%

AT&T

3.18

2.91

4

2

-9.3

-27%

T-Mobile

2.68

3

1

3

9.98%

.1%

Sprint

3.36

3.14

5

4

-46.6%

-------------

US Cellular

3.15

3.41

3

5

-13.5%

30%

Gilead

1.86

2.25

1

1

2.4%

-11%

AbbVie

2.88

2.47

3

2

52.8%

-4.7%

J&J

2.76

2.95

2

3

21.8%

-7.2%

Pfizer

3.88

3.17

5

4

9%

16.2%

Merck

3.61

4.05

4

5

-4.3%

33.4%

Fortinet

2.22

2.04

2

1

44.34%

61.2%

Juniper

2.21

2.64

1

2

.5%

-5.2%

Palo Alto Network

3

3.08

3

3

16%

50.4%

Symantec

3.71

3.44

5

4

-16.6%

-15.7%

FireEye

3.59

3.90

4

5

18%

9.9%

Some readers may notice I didn't include 2018 fiscal year return for either Andeavor or Sprint. This is because Andeavor was bought by Marathon and Sprint is awaiting approval to merge with T-Mobile, with each scenario causing upward price action for the stock not related to the fundamentals of the business. In the future, these companies will be replaced by others to fill the gap that was left (Chevron (NYSE:CVX) will replace Andeavor and CenturyLink (NYSE:CTL) will replace Sprint).

Statistical Summaries

There are some intriguing takeaways from my 2 years of data collection. Not all of it is good as I hoped, but with the small sample size of companies I've utilized thus far, there is a large amount of variation that causes my data to not be as efficient as it should. The only way to fix this is add more companies to the research and keep up the data collection for more time, as my goal for this was to be a long-term project anyway. Here are some of my findings that went as expected:

1. Stocks that scored a 1 or 2 in 2017 returned 19.73%, while stocks that scored a 4 or 5 returned just 1.22%. In 2018, there wasn't the alluring difference as a year before, as stocks that scored a 1 or 2 returned -3.3%, whereas those who scored 4 or 5 -2.3%. Combining these two years of data together shows that so far, stocks that score a 1 or 2 return 8.22% annually, on average. Interestingly, companies who have scored a 4 or 5 have returned an average of -1.08% per year. Since the goal of my research is to show that stocks who score higher than their peers (a 1 or 2) will return better than weaker companies (a 4 or 5) in the long run, these comparisons are what I'm most interested in. I'm sure the difference between the two numbers will decrease as more companies are added and as more years are logged in the study, but as of now, the difference supports my theory.

2. Stocks that have stayed in the top two of their respective groups have returned a total of 24.03% over the last 2 years, while those who have remained in the bottom two for each of the past 2 years have returned a meager -3.3%. This metric is useful because it helps me track how well consistent stocks are doing. For example, if my thesis holds any value, a company that stays as a 1 or 2 in their group for an extended period of time is establishing themselves as a constant outperformer in terms of financial strength, which should be rewarded with stellar returns in the market.

3. Over the past 2 years, stocks that have scored a 1 in their group have out-returned the number 5 stock in their group 66.67% of the time. Ideally, this number would be a bit higher, but nonetheless, it's still a pretty good percentage.

Not every trend or statistic in the data went as planned. For example:

4. Stocks that improved their score from 2017 returned -7.32% in 2018, while stocks' scores that worsened returned 3.54% in 2018. Going into this project, I thought that stocks that improved their scores from the year prior would appreciate more relative to those who dropped down. Perhaps analyzing a 1-year change in score is too narrow of a window and I should focus on how the score improves over a longer period of time. In other words, maybe comparing stocks whose scores are higher than they were 5 years ago to stocks whose scores are lower than they were 5 years ago would fit the long-term narrative of my research better. Unfortunately, this would mean I have to wait to obtain more years of data.

There are many other metrics that I would like to look at, but having just two years of data limits what can be drawn from the research. Once a third and fourth year of data is complete, I can start paying more attention to the many trends that would become available. One such trend would be how much stocks that improved their scores in each year of the study returned versus stocks whose score dropped in each year. One would think that constantly improving your average position against your competitors for a number of years would yield admirable returns. Another trend I look forward to seeing play out is how well the stocks that stay in the top 2 of their group perform over a period longer than just two years. Equally as interesting will be seeing if the stocks that remain in the bottom 2 of their group return as poorly as I predict (at least in relation to the higher scoring stocks).

Next Steps

There are a few major things I want to accomplish over the course of this year. The first thing is to add more groups of stocks to the research. Currently, there are just 5 groups totaling 25 individual companies. In order to add more validity to this whole process, I plan on adding more groups to the equation, such as airlines, auto-makers, etc. This will increase the workload quite a bit, but it's necessary to grow my research in order for it to accomplish what I want it to.

Another thing I want to tackle is finding a more efficient way to gather all of this data. It currently takes me about 6 hours to complete an analysis and write it up, so if I could cut down on some of the time it takes to gather data, it would allow me to finish more analyses. Finding a quicker way would also help me with backtesting newly introduced groups of stocks so that they have just as much data as groups that have been in the analyses for 2 years already. I'd like to think there's a way to program my computer to gather ratios from financial statements and insert them into an Excel spreadsheet, but I'd need someone who's more technologically savvy than I to confirm or deny this idea.

Lastly, I will obviously be updating the scores of groups that are already in this analysis. Companies whose fiscal years end on December 31st now have updated financial statements, so I can immediately start trying to complete those analyses. I'll have to wait to complete analyses for companies like Constellation Brands and Palo Alto Networks who have fiscal years that end later on in the calendar, but that will give me more time to add in new industries. As always, any suggestions or insights are more than welcome!

Disclosure: I am/we are long T, VZ, FTNT, JNJ. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.