According to market efficiency hypothesis, the markets are informally efficient. If investors have access to the same information, they cannot beat each other. There are three types of market efficiency: The weak form of efficiency suggests that prices reflect all publicly available information. Semi-strong form of efficiency suggests any new information is quickly reflected in the prices. Thus, unless you are a super-fast trader you cannot beat the market. The strong form of market efficiency suggests that the prices reflect all available and yet to be released information. Even an insider cannot beat the market.
I am not in favor of efficient market hypothesis. There are so many different types of investors in the stock market. They do differ in their primary investment philosophy. Some enjoy the swings in highly speculative stocks, some try to catch peaks and bottoms, and some just look for safe retirement ideas. Experience shows that those with long-term targets can beat the market with a large margin, using simple valuation metrics. One simple metrics I recently introduced is the T-Metrix.
I read about this valuation metrics in Prof. Jeremy Siegel’s book, "Stocks for the Long Run." T-Metrix uses simply three essential parameters:
A rule of thumb for stock valuation that is found on Wall Street is to calculate the sum of the growth rate of a stock's earnings plus its dividend yield and divide by its P-E ratio. The higher the ratio the better, and the famed money manager Peter Lynch recommends investors go for stocks with a ratio of two or higher, avoiding stocks with a ratio of one or less.
Dividend Yield: Higher is better.
EPS Growth: Higher is better.
P/E Ratio: Lower is better.
Here is the simple, yet powerful, formula, used by Wall Street investors:
T-Metrix = (Dividend Yield + EPS Growth) / (P/E Ratio)
Using, this formula we can determine the dividend picks for the next 5 years, as I did here. Many readers asked me about the success rate of the T-Metrix in picking the right stock. Therefore, I decided to go back to 2006, and calculated the T-Metrix scores in 2006. Next, I estimated the correlation between T-Metrix score with the annualized stock return performance. Data is from Morningstar, MSN Finance and Finviz. Here, is a back-test of T-Metrix model:
Company Name | Ticker | 2006 P/E | 2006 Yield | 5 Year EPS Growth | T-METRIX | Annualized Return | |
(A+) Average | 16,92% | ||||||
Petroleo B. (ADR) | 8,8 | 8,75 | 20,17 | 16,43 | 11,24% | ||
Vale (ADR) | 11 | 1,77 | 25,2 | 12,26 | 22,59% | ||
(A) Average | 16,61% | ||||||
BHP B. Ltd (ADR) | 9,7 | 1,81 | 16,98 | 9,69 | 18,30% | ||
Rio Tinto (ADR) | 9,6 | 3,6 | 13,76 | 9,04 | 7,91% | ||
BHP B. plc (ADR) | 10,8 | 1,94 | 16,98 | 8,76 | 13,39% | ||
CNOOC (ADR) | 10,1 | 2,99 | 14,68 | 8,75 | 26,84% | ||
(B) Average: | 16,00% | ||||||
Telefonica (ADR) | 16,6 | 3,24 | 19,77 | 6,93 | 14,38% | ||
IBM | 16 | 1,13 | 18,79 | 6,23 | 17,61% | ||
((NYSE:C)) Average: | 12,15% | ||||||
AT&T Inc. | 18,9 | 3,72 | 18,73 | 5,94 | 9,16% | ||
Chevron | 9,4 | 2,73 | 7,71 | 5,55 | 14,57% | ||
JPMorgan Chase | 12,6 | 2,82 | 10,72 | 5,37 | 2,63% | ||
McDonald's | 19,3 | 2,26 | 17,56 | 5,13 | 21,76% | ||
Coca-Cola | 22,3 | 2,57 | 19,92 | 5,04 | 11,72% | ||
China M. (ADR) | 20,5 | 1,9 | 16,88 | 4,58 | 13,05% | ||
(D) Average: | 7,57% | ||||||
Wall-Mart | 16,5 | 1,45 | 10,77 | 3,70 | 5,10% | ||
Oracle | 24,6 | 0 | 17,08 | 3,47 | 20,45% | ||
Siemens (ADR) | 21,7 | 1,59 | 12,91 | 3,34 | 9,82% | ||
Schlumberger | 21 | 0,79 | 13,18 | 3,33 | 6,22% | ||
PepsiCo. Inc. | 18,7 | 1,85 | 10,35 | 3,26 | 6,40% | ||
Novartis (ADR) | 19,4 | 1,56 | 10,21 | 3,03 | 4,41% | ||
Microsoft | 25,5 | 1,24 | 13,4 | 2,87 | 3,61% | ||
Johnson & J. | 17,7 | 2,2 | 6,68 | 2,51 | 4,55% | ||
Procter & Gamble | 22,9 | 1,88 | 9,09 | 2,40 | 6,87% | ||
Intel Corporation | 23,5 | 1,98 | 7,5 | 2,02 | 7,18% | ||
(F) Average: | 3,71% | ||||||
Exxon Mobil | 11,6 | 1,67 | 1,73 | 1,47 | 8,33% | ||
PetroChina (ADR) | 13,9 | 3,4 | 0,27 | 1,32 | 6,24% | ||
Wells Fargo | 14,3 | 3,04 | -0,36 | 0,94 | -0,65% | ||
Pfizer Inc. | 17 | 3,71 | -1,32 | 0,70 | 1,40% | ||
TOTAL SA (ADR) | 10,7 | 2,96 | -1,94 | 0,48 | 1,58% | ||
Royal D. (ADR) | 9 | 3,46 | -2,8 | 0,37 | 5,36% | ||
(Sub-F) Average: | -1,73% | ||||||
General Electric | 18,7 | 2,77 | -7,2 | -1,18 | -6,82% | ||
ConocoPhillips | 7,4 | 2 | -4,41 | -1,63 | 5,57% | ||
HSBC (ADR) | 13,2 | 4,15 | -11,81 | -2,90 | -5,62% | ||
Verizon | 19,8 | 4,19 | -19,43 | -3,85 | 10,56% | ||
GlaxoSmith (ADR) | 14,3 | 3,29 | -17,21 | -4,87 | -1,13% | ||
Merck & Co.. | 21,5 | 3,49 | -33,17 | -6,90 | 5,35% | ||
Toyota (ADR) | 16 | 1,34 | -28,41 | -8,46 | -4,35% | ||
Vodafone (ADR) | -6,3 | 14,63 | -0,18 | -11,47 | 10,55% | ||
Citigroup | 13,1 | 3,52 | -40,64 | -14,17 | -29,71% |
Peter Lynch suggests 2 as a perfect number. Therefore, I multiplied the scores by 5 in order to rank the stocks on a 10-point scale. The results show a stunningly high level of correlation between T-Metrix score and annualized total returns (including dividends). The following graph makes it easier to see the strong connection:
I also run a simple linear regression to estimate the whether the correlation between T-Metrix score and annualized return is strong enough to be statistically sufficient. The equation above shows that the variation in T-Metrix scores is strong enough to explain almost 50% of the variation in returns. The coefficient of 0.011 suggests that a one point increase the T-Metrix score implies 1.1% higher annualized return. Thus a stock that has a T-Metrix score of 10 is expected to beat another stock with a score 5 by 5.5% annualy (1.1% x [10-5] = 5.5%).
I also decided to include letter grades, as it will make things easier to understand. Here is the relationship between T-Metrix score and letter grades:
T-Metrix Score | Letter Grade |
T-Metrix > 10 | A+ |
10 > T-Metrix > 8 | A |
8 > T-Metrix > 6 | B |
6 > T-Metrix > 4 | C |
4 > T-Metrix > 2 | D |
2 > T-Metrix > 0 | F |
0 > T-Metrix | Sub-F |
If the regression above is too complicated for you, just go with the letter grades. The following graph shows the 5-year annualized return of stocks classified according to the letter grading system:
As the graph illustrates, companies with higher letter grades beat others that have lower grades. The formula above can be used to rank not only high dividend stocks, but low-dividend or growth stocks as well. As one might notice, according to T-Metrix grading system, growth is a perfect substitute for yield. Let’s explain this concept using two examples.
1) Apple (NASDAQ:AAPL)
P/E = 30.8 (in 2006)
Yield = 0 (in 2006)
Annualized EPS Growth = 57.56
T-Metrix Score = 9.34
Annualized Return = 39.04
2) Cisco (NASDAQ:CSCO)
P/E = 28.7 (in 2006)
Yield = 0 (in 2006)
Annualized EPS Growth = 8.86
T-Metrix Score = 1.54
Annualized Return = -4.47
The above examples show that in 2006, both Apple and Cisco was fairly priced for an annualized EPS growth rate of 30%. However, Apple beat the analyst estimates with a whopping growth rate of 57.56%, whereas Cisco disappointed the shareholders with a much lower EPS growth rate of 8.86%.
Result: Apple returned 39% annually, whereas Cisco’s annual return was -4.47%.
Note that all the data is used in a back-testing model, where I looked at the 2006 P/E ratios, yields, and actual EPS growth data. The T-Metrix model is a very simple model, but the data is dynamic in its nature. The earnings, prices, and yield are dynamic forms of data, which change frequently. We do not know much about the future, but we can make estimations. The P/E ratio and yield is already available. What we are left with is estimating the future EPS growth rate. That will be the real challenge, since the future has many uncertainties in it. Analysts make these estimations for us, which could be a good starting point. I will be on the hunt for dividend paying companies with high growth estimations, and excellent T-Metrix scores.
Disclosure: I am long INTC, C, MSFT.
Disclaimer: The article is a back-testing article which, shows that by the year 2011, stocks with higher T-Metrix scores in 2006 outperformed other stocks. I cross-checked the data from a variety of sources, but no data is perfect. Things get particularly complicated when there are mergers/break-ups involved. I am not a registered advisor [yet]. You should always do your own diligence before making any investment decision.