Account Information

  • My Account

    Manage all your subscriptions, update your address, email preferences and change your password.

  • Help Center

    Get answers to common service questions, ask the analyst or contact our customer service department.

  • My Stock Talk Profile

    Update your stock talk name and/or picture.



Close

“What Works on Wall Street” and Trending Value: Best Stock Screen of All Time!

By Jim Fink on January 20, 2012

Print Friendly

I’ve been a fan of James O’Shaughnessy ever since his investment classic What Works on Wall Street was first published in 1996. As I wrote in Mechanical Investing and Fundamental Indexing: Be a Quant!, what I love about quantitative stock screens is that they can be backtested over many years and compared to find those that really outperform the overall stock market over the long term. I want proof, not anecdote or conjecture.

Investing is like baseball – Moneyball to be exact – in that the best criteria for picking outperforming stocks or players are often very different from what most people think the best criteria are. For example, investors often search out companies with large annual sales gains because they “think” that the stocks of such high growth companies generate investment returns that outperform the overall market. In fact, O’Shaughnessy discovered that companies generating the highest annual sales growth severely underperform the overall market! Only through data crunching and historical analysis can the truth be uncovered and shibboleths refuted. Investors with this “variant perception” are the big winners.

“Cornerstone Growth” Stock Screen is Good . . .

In the third edition of O’Shaughnessy’s What Works on Wall Street, the best growth-style stock screen was called “Cornerstone Growth.”

The criteria were as follows:
  • Market capitalization greater or equal to $225 million
  • 3-month average daily share volume greater or equal to 100,000
  • Price-to-sales ratio less than or equal to 1.5
  • 3-month total return greater than the median return of Russell 3000 companies
  • 6-month total return greater than the median return of Russell 3000 companies
  • Trailing 12-month earnings per share is greater than zero
  • Rank by highest 12-month total return

Between January 1, 1964 and December 31, 2009, the 50-stock Cornerstone Growth portfolio (rebalanced annually) returned 17.00% per year compared to only 11.22% for the overall stock market. Not bad, but there’s a new king in town . . .

But “Trending Value” Stock Screen Generates the Absolute Best Returns

Say hello to a growth-style stock screen called “Trending Value” which is the star attraction of the fourth edition of What Works on Wall Street released in November 2011.  Its annualized return over the 46-year period between 1964 and 2009 is a whopping 21.19%, virtually doubling the return of the overall stock market and beating the Cornerstone Growth strategy by more than four percentage points per year. Four percentage points may not sound like much, but over a 46-year period a strategy with a four-percentage point annual edge results in an account value six times larger.

Value Composite is the New Innovation

The main difference between this new Trending Value screen and the old Cornerstone Growth screen is the use of a “value composite” to measure undervaluation rather using the single valuation ratio of price-to-sales. As O’Shaughnessy explains in the video embedded in this Forbes article (4:19 minute mark), the problem with single-factor valuation ratios is that they move “in and out of favor” and can significantly underperform the overall market over any given 10-year period despite their long-term outperformance.

In contrast, a valuation composite consisting of six different value factors “smoothes out” performance in the short term because at least a few of the valuation factors are always in favor and balance out those factors that are out of favor. Using a value composite is another version of diversification, which is probably the most important investment concept in all of finance. O’Shaughnessy found that stocks selected based on the value composite outperformed stocks scoring highest on any single value factor 82% of the time in all 10-year rolling periods between 1964 and 2009.

The value composite is composed of six value factors:

  1. Price to book value
  2. Price to sales
  3. Earnings before interest, taxes, depreciation and amortization (EBITDA) to Enterprise value (EV)
  4. Price to free cash flow
  5. Price to earnings
  6. Shareholder yield

O’Shaughnessy explains how the Trending Value stock screen filters stocks based on the value composite:

Each stock in the universe gets a score of 1 to 100 for each of these factors. The final value score is an average of these scores. The Trending Value portfolio narrows the investable universe to the 10% of stocks with the best score based on the value composite, and then selects a concentrated portfolio of 25 stocks based on trailing six-month momentum.

For example, if a stock has a P/E ratio that is in the lowest one percent of the investable stock universe, it gets a P/E value rank of 100. A P/E ratio in the lowest four percent gets a P/E value rank of 97, etc. Do a similar analysis for each of the five other value factors. Add up the six value ranks, divide by six, and you get the value composite rank. Do this for all stocks in the investable universe and then isolate the 10% of stocks with the highest value composite rank. Lastly, sort the stocks with the top 10% value composite rank by six-month relative price strength (i.e., the trending part of the screen) and select the 25 stocks with the highest six-month relative price strength for the Trending Value portfolio.

Can a Stock Screen Be Truly Best if it Can’t Be Reproduced?

That’s it – the road to riches in mechanical stock investing. Or is it? I see one big problem with the Trending Value stock screen – besides the ever-present caveat that “past performance does not guarantee future results.” Namely, how does the average investor get access to this screen? None of the stock screeners that I highlighted in my article Best Stock Screening Tools on the Web have the capability of running a screen with this complex set of criteria. Not even my trusty Bloomberg terminal – which costs thousands of dollars per year – can do this screen.

Based on the methodology O’Shaughnessy outlines in an appendix to the book, it appears that the only way to run this screen is by subscribing to several incredibly expensive databases (Standard & Poor’s Compustat and the Center for Research in Security Prices (CRSP) equity database) and software packages (FactSet and SAS).  Absent buying these services, the only way to get exposure to a Trending Value-like portfolio is to sign up with O’Shaughnessy Asset Management (how financially convenient for O’Shaughnessy!).

Similar But Not Identical Stock Screen May Generate Completely Different Backtested Returns

The closest I can get to the Trending Value screen on Bloomberg is by screening for stocks that satisfy the following criteria:

  • Market cap greater or equal to $225 million
  • 6-month total return greater than median 6-month return of Russell 3000 index
  • Top 40% percentile rank of price to book value (lower value is better)
  • Top 40% percentile rank of price to sales (lower value is better)
  • Top 40% percentile rank of EV to EBITDA (lower value is better)
  • Top 40% percentile rank of price to free cash flow (lower value is better)
  • Top 40% percentile rank of price to earnings (lower value is better)
  • Shareholder yield greater or equal to 2%

Even though I upped the percentile rank to 40% from O’Shaughnessy’s 10% percentile limit for the value composite, I still came up with only seven stocks! The reason is that my Bloomberg screen requires that a stock be in the top 40% percentile of each value factor, whereas O’Shaughnessy’s screen sets no percentile limit on any individual value factor but only focuses on the 10% percentile limit of the value composite.

For example, a stock that is in the best 1% percentile for all of the value factors except the P/E ratio, for which it places in the 65% percentile would be excluded from my screen but would probably make the cutoff of O’Shaughnessy’s Trending Value screen because its average rank would still be very low: (1+1+1+1+1+65)/6 = 11.7. Similarly, a stock with individual rankings of the 39% percentile for all value factors would make my screen but the 39 average rank [(39*6)/6] could very well be too high to make O’Shaughnessy’s cutoff.

Just the fact that my Bloomberg screen only spits out seven names total and O’Shaughnessy’s Trending Value strategy is based on a 25-stock portfolio is a very good indication that something is amiss.

Bottom line: My Bloomberg screen may be similar to O’Shaughnessy’s Trending Value, but it probably generates a completely different list of stocks which means that it has very little chance of matching the excellent backtested returns of the actual Trending Value stock screen. Bummer.

With that caveat out of the way, here are the top 5 stocks that made the cut of my Bloomberg screen. Over the past six months, the return of the Russell 3000 index has been -4.1%. To save space, I’ll just list the EV to EBITDA value with its percentile rank in parenthesis, but keep in mind that these five stocks are highly ranked in all six valuation factors:

Company

6-Month Total Return

EV to EBITDA Ratio

Shareholder Yield

Industry

Universal Corp. (NYSE: UVV)

27.5%

6.8 (71)

4.3%

Tobacco

Tech Data (NasdaqGS: TECD)

12.4%

4.2 (91)

11.5%

Computer products

Gamestop (NYSE: GME)

5.5%

3.7 (94)

8.6%

Video game retailer

Tesoro (NYSE: TSO)

2.0%

2.3 (98)

2.6%

Oil refining

Wellpoint (NYSE: WLP)

-1.8%

2.8 (98)

12.9%

Health insurer

Single-Factor EV/EBITDA Valuation Stock Screen is Much Easier to Reproduce

I’m much more confident about reproducing one of O’Shaughnessy’s single-factor value screens. Another interesting change to O’Shaughnessy’s thinking in the fourth edition of What Works on Wall Street  is which single-factor valuation measure is the best. In the first three editions of the book, O’Shaughnessy stated that the price-to-sales ratio was the single best valuation measure. But in the fourth edition, O’Shaughnessy crowns a new single-factor value champion: EV to EBITDA, which has been the primary valuation tool used by investment banks and private equity firms for many years. And a recent Drexel University study has also anointed EV to EBITDA as the best valuation factor for predicting future equity returns. 

According to O’Shaughnessy’s research, the annual returns from buying the stocks with the lowest price-to-sales ratios is now only the seventh-best single-factor valuation measure between 1964 and 2009:

Valuation Factor

Annualized Return

EV/EBITDA

16.58%

Price to Earnings

16.25%

Price to Operating Cash Flow

16.25%

Buyback Yield

15.81%

Shareholder Yield

15.56%

Price to Book Value

14.53%

Price to Sales

14.49%

Dividend Yield

13.30%

 

If you’re wondering why the price-to-sales measure is shown to have had an annualized return of only 14.49% when I previously mentioned that the Cornerstone Growth screen (which is based on price to sales) had an annualized return of 17.00%, I think the reason is that Cornerstone Growth added a momentum criteria (i.e., relative price strength) that improved returns over and above a screen based solely on valuation.

Consequently, I believe you can improve upon the 16.58% annualized return generated by the best valuation-only screen based on EV to EBITDA by adding the Cornerstone Growth momentum criteria. Specifically, by replacing the criterion in O’Shaughnessy’s Cornerstone Growth screen of “price-to-sales ratio lower than or equal to 1.5” with the new criterion “EV to EBITDA ratio in the top 10% percentile (lower is better)” – and ranking by 6-month relative price strength — my Bloomberg terminal came up with the following five names:

Company

6-Month Total Return

EV to EBITDA Ratio

Industry

ChinaCast Education (NasdaqGS: CAST)

26.7%

3.6 (94)

Education services

LyondellBasell Industries NV (NYSE: LYB)

23.8%

4.2 (92)

Chemicals

Lexmark International (NYSE: LXK)

23.0%

3.2 (96)

Computer printers and peripherals

Rent-a-Center (NasdaqGS: RCII)

19.7%

3.0 (97)

Household appliance leasing

Net1 UEPS Technologies (NasdaqGS: UEPS)

18.8%

3.3 (95)

Payment processing services

 

Stock Talk — Post a comment Comment Guidelines

Our Stock Talk section is reserved for productive dialogue pertaining to the content and portfolio recommendations of this service. We reserve the right to remove any comments we feel do not benefit other readers. If you have a general investment comment not related to this article, please post to our Stock Talk page. If you have a personal question about your subscription or need technical help, please contact our customer service team. And if you have any success stories to share with our analysts, they’re always happy to hear them. Note that we may use your kind words in our promotional materials. Thank you.

You must be logged in to post to Stock Talk OR create an account.

Create a new Investing Daily account

  • Use Social Connect
  • - OR -

* Investing Daily will use any information you provide in a manner consistent with our Privacy Policy. Your email address is used for account verification and will remain private.

Stock Talk

  1. avatar
    Philip Reply February 2, 2013 at 12:00 PM EST

    The Value factor one screen is available on value-investing.eu for Europe, USA, Australia and Japan

  2. avatar
    Johnny Wad Reply August 4, 2012 at 3:21 AM EST

    Hello investors!

    So, after discovering James O’Shaughnessy’s Trending Value screen (and being a huge fan of his work), I went on a mission to find stocks that passed his rigorous tests. After spending many hours on various free stock screeners (e.g. FINVIZ, Google, Motley Fool, Scottrade and Yahoo stock screeners), I came up with a series of thresholds that led to 44 passing stocks. As a disclaimer, I set the minimum market capitalization at 50 million, but even for the little guys, I required at least $100,000 trading dollars per day (average volume). This extra measure required a little math, but it did not take too long. Additionally, I did not calculate Shareholder Yield since I was fuzzy on how to do it manually (not to mention, that would require reading many annual reports, something that Mr. O’Shaughnessy frowns upon).

    To note, of the 44 stocks that passed the composite screening, 21 of them were micro caps and the remaining 23 had a market cap of 300 million or more.
    Of course, by having micro caps in your portfolio, you increase risk and volatility, but your reward over the long-term, as pointed out by O’Shaughnessy, will be significantly higher. However, unlike his aggressive Tiny Titans strategy, the Trending Value stocks have met much more criteria, so the downside risk is reduced. I will be tracking a 25-stock Trending Value portfolio for the rest of the year to see how it performs; my guess is that it will do extremely well considering I think the 2nd half of this year will be fantastic for equities. I think I will put real money to work for 2013 and beyond.

    Lastly, for the small investor, trading liquidity is not as important. Money managers who are working with millions, if not billions of dollars are forced to invest in large companies by default. But, for the average investor, you can incorporate micro and small cap stocks, like the ones that showed up in the Trending Value portfolio, and take advantage of the best-performing strategy of all-time!

    Not to leave anyone hanging, here are the top 25 stocks that passed the composite tests coupled with highest 26-week price appreciation:

    1. HCII 14. GASS
    2. CKEC 15. SEM
    3. HRG 16. SAIA
    4. JBSS 17. CSV
    5. BXG 18. HCOM
    6. YTEC 19. VECO
    7. DK 20. SIM
    8. BSET 21. SSI
    9. TCX 22. ITIC
    10. FLXS 23. TSO
    11. SYPR 24. SBS
    12. DDS 25. UFI
    13. WWIN

    The thresholds I used in my tests are as follows:
    ~Price to Sales ratio below 2 (most fell below 1)
    ~Price to Book ratio below 2 (most fell below 1.5)
    ~Price to Cash Flow ratio below 15 (only 2 stocks were above 10; most below 7)
    ~Enterprise Value/EBITDA below 10 (most fell in the 3-6 range)
    ~P/E ratio below 20 (most fell in the 8-13 range)

    It bothered me that I could not perform the last step in the composite process, Shareholder Yield. Who knows, maybe some of these stocks would have been in the 90th percentile, I have no idea. I considered adding one of the criteria from Cornerstone Growth, annual earnings higher in the past 12 months, to the mix as a substitute for the missing link. Why not, considering it is used by the previous champ! When I added that screening measure, only 29 stocks remained – still enough for the recommended 25-stock portfolio. Food for thought to those who can’t screen for Shareholder Yield either.

    Furthermore, I am curious how Mr. O’Shaughnessy himself would feel about this list. Although I don’t have access to COMPUSTAT data, I think I did well given what most investors have at their disposals. The portfolio was up over 2 percent in the first trading day (8/3/12); so far, so good!

    Happy trading…

  3. avatar
    Jack Goodman Reply March 25, 2012 at 10:19 PM EST

    Ronald – curious to know if you have been able to come up with anything on the AAII STPro tool?

  4. avatar
    the cowboy Reply February 1, 2012 at 6:10 PM EST

    Well, I’ve been following Mr. Oshaughnessys work since the beginning and have always been impressed. That being said, if he has based all this on some unattainable data base I will be much less impressed. Hope this hasn’t devolved into a pitch for crsp/compustat and his investment fund. Theres really no need to write a book about this if there is no cost effective way to run the numbers. This makes me long for the old Value Line days ( still using some screens from them and still quite profitable) where the data is available and understandable.
    Hope we’re all wrong about this – I thought more of the author than this foolishness.

  5. avatar
    ronald ferrill Reply January 28, 2012 at 12:37 PM EST

    Excellent review. First, I’ll grab the new addition. Second – gonna research YOU more… Third, using AAII’s Stock Investor Pro screening tool, I’ll mess around with these screens and see what I come up with.

    If OK with you guys/gals I’ll even share any interesting findings right here.

    Cheers!

  6. avatar
    Bill B Reply January 22, 2012 at 11:58 AM EST

    Jim, good info. I too was attempting to reproduce the information in the book and discovered the same – I can not reproduce the exact screens for the the composite factors, which so provide such a consistent positive edge. For example, my normal Fidelity screen doesn’t have EBITA, 6 month price performance and BBY which are key factors. I have found “pieces” of the EBITA and 6 month price performance in other screener, but not the BBY. I considered substituting FCF for EBITA, and 3 mo for 6 mo price performance (in conjuntion with 12 mo momentum), and going through the tedius process of deriving the quantitative ranking via iterative analysis of the other factors to eliminate the problem you mentioned above. Based on your analysis, and the fact that BBY is a bit hard to come by, I think I will postpone this rigorous process.
    Thanks for your work!