CROIC ROIC Screen Strategy Backtest

March 4, 2010 | Comments (41)

Stock Screen Strategy and Backtest Series

Value Investing Screens

One of the most lacking aspects to all the widely available screens such as Yahoo, MSN, Morningstar etc are detailed screen criterias. They are all based on PE, PB, PS, ROE, BV and so on which don’t really help you find better investments than the next person.

This is where the Old School Value screeners try to fill this void. I’ve been working hard to come up with new value investing screen criterias and formulas to find the best strategies and opportunities.

One such strategy that I am very confident in is the CROIC screener.

CROIC – Cash Return on Invested Capital

As discussed in the previous CROIC and ROIC article, the formula for CROIC is shown below.

CROIC = FCF/Invested Capital

Invested Capital = Total Equity + Total Liabilities – Current Liabilities  – Excess Cash

Excess Cash = Total Cash – MAX(0,Current Liabilities-Current Assets)

CROIC and ROIC Above 10% Screen Strategy

In this test, I tried two different strategies.

The first one was where a portfolio of 15 stocks were selected based on the current year and prior 2 years CROIC and ROIC being above 10%. Anything less than 10% didn’t make the screen.

I always tend to try and find companies with CROIC in the sweet spot which is usually above 10% as it indicates a strong executive team and company.

In the NCAV screen and NNWC screen, I only selected companies that had trading volumes above 30k, but for CROIC, since many of the companies are not illiquid micro caps, I’ve lowered the volume criteria to 20k.

CROIC and ROIC Above 10% Criteria

  • CROIC and ROIC is greater than 10% for the past 3 years
  • No OTC and financial stocks
  • 15 stocks max in portfolio
  • Volume is greater than 20k
  • Slippage set to 1%
  • Rebalance frequency set to 6 months
  • Test period of 3 year time-frames

CROIC and ROIC Above 10% Results

The results show that companies able to produce high returns will beat or be consistent with the market. There is a good amount of volatility involved but the results are very respectable except for the 2007-2010 period where it just kept up with the market.

Over a 9-10 year period, the result is outperformance by a huge margin for the CROIC strategy but looking at the screen in different time frames, you wouldn’t have known.

Then it dawned on me that everyone looks for the above average companies. So I modified some more formulas and criterias to get the next screen.

CROIC and ROIC Increasing for 3 Years Screen

In the second screen, I removed the minimum CROIC and ROIC limit. This means that any company can make the screen, even if the CROIC and ROIC is negative. But my reasoning is that regardless of what the actual CROIC and ROIC numbers are, if the company has been increasing this metric for the past 3 years, the business is not only improving but management has done a good job of turning the boat around.

In other words, the screen should result in successful turnarounds which are often explosive opportunities.

CROIC and ROIC Increasing Screen Criteria

  • CROIC and ROIC has increased each year for the past 3 years
  • No OTC and financial stocks
  • 15 stocks max in portfolio
  • Volume is greater than 20k
  • Slippage set to 1%
  • Rebalance frequency set to 6 months
  • Test period of 3 year time-frames

CROIC and ROIC Increasing Screen Results

As you can see, this second screen completely annihilates the first screen.

Assuming you followed this strategy the only down period would have been from 2007-2010 but overall had you started in 2001, your $100 would have grown to $850 by 2010 using the CROIC method and $385 using the ROIC method.

One thing is for sure on both accounts. CROIC is a much better indication of performance and better metric even in a screener.

I also added another criteria where FCF is increasing each of the past 3 years which also yield great results which I won’t display here as the results are very similar to above.

I’ll be looking at FCF related screens and strategies in the next strategy review but first, here are the companies that showed up in the screen.

Disclosure

No holdings.

About Jae Jun


Jae Jun is the founder of Old School Value. He is on a mission to provide practical and actionable value investing tools, tutorials and educational material to help empower the individual investor. Keep in touch with Jae via any of the methods linked below.

Get Tips and Strategies to Achieve Higher Returns
Bonus: Get 9 FREE Investing Spreadsheet Calculators
  • Ozz

    Jae,

    Awesome work! One question; how do you go about generating the actual screen? Did you write code that grabs thousands of stocks and their raw data from a web database (MSN, Yahoo, etc), creates/populates fields normally not listed on screens (like CROIC), reduces the list somehow, and displays them in a screen? The screens themselves don’t have a long list of names, so I assume there may be filtering going on. I’d like to replicate it myself, just wanted to see if you could share your thought process

    Got the premium package btw, I like it a lot!

  • Eldinril

    I have noticed a couple of points in your work that have piqued my curiosity. If you don’t mind I would appreciate it if you would clarify these points for me.

    Your formula for invested capital is:

    Invested Capital = Total Equity + Total Liabilities – Current Liabilities – Excess Cash

    The first part of the formula is “Total Equity + Total Liabilities”. By definition, “Total Equity + Total Liabilities” equals Total Assets. Is there a reason why you did not simplify this?

    The next portion of the formula subtracts “Current Liabilities and Excess Cash”. Perhaps I am incorrect, but I have seen this also expressed as “non-interest bearing current liabilities minus excess cash”. I am not certain which is preferable, but would appreciate hearing why including short-term debt is preferable.

    Finally, my explorations regarding the calculation of excess cash have encountered multiple approaches. From the total cash on the balance sheet, some people simply subtract spontaneous liabilities, some subtract two percent of revenues, and some use the “current liabilities minus current assets” method that you include. For some reason, I find the first two more intuitive than the latter. Have you considered using the other methods?

    Best wishes,

    Jim

  • JJ

    Great posts! I’d like to save several of your posts but I do a print preview and it runs to 10 pages, wondering if you might be able to add a print button to put your posts in a more printable form. Thx.

  • http://www.oldschoolvalue.com Jae Jun

    @ Eldinril,
    By separating total assets, it helps most people to get a grasp of what invested capital is. If you just include total assets, it’s quite hard for new people to understand the concept.

    The correct formula would be to use “non-interest bearing current liabilities minus excess cash” but I include all of current liabilities because this is debt due within a 12 month period.

    A company won’t borrow short term debt to generate more money. Instead, a business would borrow for a few years and then use that cash to generate more capital.

    For the excess cash method, I use it in the EPV calculation but the shortfall is that they are both assumptions. 2% of sales is only a rule of thumb and a very rough one at that. I prefer to keep things safe and the excess cash formula I use suits me better than the other two.

    Things can only go wrong when you underestimate.

    @ JJ,
    You can send the article to your email using the “+Share” button on the blue toolbar you can bring up by click on the bottom right blueish tab. Then print from your email. Or I’ll put a print option very soon.

  • Alex MacKinnon

    Hey Jae,

    Great post ! Can I get the same e-mail that ‘Ozz’ got ?

    I would argue that you have surpassed fwallstreet as a site. No offense to Joe (he’s on to bigger and better things) but posts are almost daily on this site – YOU’RE A MACHINE !

    Cheers,
    Alex MacKinnon

  • Alan

    Jae, I am confused again. In your post, you say that DY and SPA both have increasing CROIC over the last 3 years, but when I use your spreadsheet and download the values, the CROIC on both are decreasing. (I am looking at the value your spreadsheet gives for CROIC on the DCF valuation page, under the “efficiency and profitability” heading.) If the source of information for the spreadsheet is different from the source you are using for your screens, which is correct? If both are the same, why are there such striking differences in the values for CROIC?

    Thanks,

    AB

  • http://chromainvesting.com Chroma

    Can I also get the Ozz email. I am interested in backtesting some other investment strategies.

    Thanks.
    .-= Chroma´s last blog ..Lessons Learned from Mike Burry =-.

  • psp627

    Jae,
    Excellent work. Could you provide more info on how you put the screens together? Also, how do you perform the backtests?

  • http://www.oldschoolvalue.com Jae Jun

    @ Alan,
    Remember that this is a screen so the data may be different when actually calculated by hand or the spreadsheet.
    You ask which one is correct if the source and spreadsheet is different, the correct one is manually checking to see whether the numbers are correct and match the SEC filings.

    But I would lean more towards the spreadsheet for accuracy as I have full control the data used and I am confident in those numbers.

  • http://MagicDiligence.com Steve

    Nice work. One thing I would mention with FCF is that I’ve found it is usually better to do something like:

    FCF = Cash From Operations – MAX(CapEx ; Depreciation)

    This tends to avoid penalizing businesses like expanding retailers where CapEx is high due to new locations. That should not really be a handicap – it is investment towards growth. By using Depreciation in these cases you are only penalizing free cash flow for upkeep on existing stores.

    Steve

  • Chuck

    Isnt the more conservative FCF equation something like:

    (NET INC + DEPRECIATION) – CAP EX?

  • Alan

    Thanks for the answer, Jae

  • http://www.oldschoolvalue.com Jae Jun

    @ Steve,

    That’s a very good idea. I’ll have to test out your suggested criteria. But depending on how depreciation is stated, couldn’t it overstate capex or depreciation which means you could end up with a fair aggressive number without realizing?

    @ Chuck,
    Your formula is a very basic version of owner earnings so of course you can use it.

  • texasyeehaw

    Jae could you send me the email you sent Oz? I would like to screen some stuff too if that’s cool with you. Thanks and the newest update is fantastic!

  • Jason

    Hi Jae, Nice work with this screen. Could you send me info so that I can perform some screens or replicate yours? Thanks.

  • http://n/a Duane

    This is great info. Much appreciated, Jae.

    Can I get the same info you sent Ozz?

    Thanks!

  • vernadel

    how is built croic strategy?

  • Ameet

    Jae,

    Is there a reason to rebalance every six months as opposed to a year and a day? I’m curious because of the bite short-term taxes would take out of such returns, even if good. Though if the strategy does worse on a year and a day rebalancing rather than a six month rebalancing, then that’s also worth noting.

    Thanks!

  • http://www.oldschoolvalue.com Jae Jun

    @ Ameet,

    I dont think 6 months is that bad. After all it’s about 60 trades a year max which is less activity than most active investors.
    I trade more than this considering I don’t trade often.

    Also keep in mind that these screens are for the idea purposes. To let you see how certain strategies work. I don’t advocate mechanical investing. Just a great way to find ideas and capitalize on the best ones.

  • k007

    in this type of screen, we are not just looking for stocks with good value but we are looking for good growth stocks. personally i think the vast majority of investors are young and still have annual income. i wonder what this strategy would show if instead of rebalancing the portfolio every 6 months, one were to simply buy 1 or 2 stocks every month from the list and hold them until a point where COIC were decreasing 2 years in a row or some other factor. this is prob difficult to backtest but just a thought.

    i thought of it mainly because many of these screens show great long-term returns but use much shorter strategies in terms of holding…very few people could hold after seeing a 75-80% drop in their portfolio and then imagine how fun it would be to rebalance at the bottom…maybe there is a way to screen for long-term holding of these companies

  • Owen

    Jae, when more than 15 stocks meet the screening criteria, how do you determine which 15 stocks to include in the portfolio? I guess what I’m really asking is if I do the same test, I may choose a different set of 15 stocks, and get totally different results. Correct?

  • Maxi

    Hi Jae,
    I would like to know if you run the backtest including companies that went bankrupt and are now delisted. What I mean is are your results favoured by not including the companies that we know have failed?

    thanks and keep the great work.

  • http://www.oldschoolvalue.com Jae Jun

    The backtest does include bankrupt companies. It’s far too difficult to exclude bankrupt companies as I search for really cheap stuff. However, if I use certain criterias to only test stocks above $5 as an examples, it should prevent bankrupt companies entering the backtest.

  • Joshua Chance

    Hi Jae Jun,

    I don’t think that addresses the survivorship bias issue. If your backtest only includes companies that are currently active and doesn’t include companies that were delisted (not necessarily bankrupt but most likely) then you might be overstating returns. The point of survivorship bias is that you’re testing a strategy on a set of companies that in 2010 we know survived, unless your database includes delisted companies. So there’s an element of “seeing into the future” which biases the test.

    I wouldn’t ignore this as the number of delisted companies in the past 10 years is quite large and most of them are the micro and small caps that your screens buy.

  • Joshua Chance

    Oh, I think I misread you. Your data does include delisted stocks? If you don’t mind me asking, what data source do you use?

  • http://www.oldschoolvalue.com Jae Jun

    Im using data from Reuters. Not the best in data integrity but it does its job.

  • Dan

    In your tests, do you account for look-ahead bias?

  • http://www.oldschoolvalue.com Jae Jun

    There is no look ahead bias in any of the tests.

  • Seth

    Following up on Dan’s question, how much do you time lag your data? I think the backtests in “What works on Wall St” use a lag of almost 1 year.

  • M. Kapman

    A lot of times when I enter a screen result into your dcf spreadsheet, the croic only appears to be increasing for 2 years. Why is that?

  • http://www.oldschoolvalue.com Jae Jun

    It is most likely that the screener and spreadsheet both use different data sources.
    Always a good idea to double check the screen results with the spreadsheet because it will be much more accurate.

  • http://www.sgValueInvestments.wordpress.com juno.tay

    Hi Jae Jun,

    Great blog you have. I have a query that I thought you might be able to help. What should we do if CROIC is extremely high? (40% to 50%). The company in question is a telecom company in a highly regulated industry (I live in Singapore – we have only three major telecoms, 2 big 1 small) so there is a really strong economic moat. However, if I were to use the given growth rate, wouldn’t the IV of the company calculated be extremely high? In your opinion, should there be a moderation of CROIC above a certain level?

    Thanks!

  • http://www.oldschoolvalue.com Jae Jun

    Hey Juno,

    If CROIC is very high and sustainable you have a good company on your hands. But to use it as a growth rate wouldn’t work. I find it best to use the owner earnings or FCF growth rate, whichever you prefer. After all the company grows at the rate it can grow cash.

  • DC Value Guy

    Jae:

    This is a very interesting post. I’ve been using Greenblatt’s “magic formula” for approximately 3 years and it has outperformed the S&P by ~ 7.5% per anum. I’m intrigued by the approach you’ve taken here. I have a couple of questions: (1) what computer package do you use to perform backtesting, (2) how would one (me) go about backtesting a strategy, and (3) given that you’re testing 15 stock portfolios, what do you do when > 15 stocks pass the screen (take the highest 15 as measured by FCF yield?)? Again, I thought this was a very interesting post. I enjoy reading your website. Best of luck and happy new year!

  • ValueInsider

    wow, @ DC Value Guy, I had exactly the same questions for Jae. Love this post, and have been using Joel’s magic formula with good success rate. Curious Jae, if you had combined this test with magic formula, and what software package would you recommend for backtesting?

  • http://www.oldschoolvalue.com Jae Jun

    I haven’t thought about running it along with the Greenblatt formula. I may just try. But what I find difficult is that I cannot seem to replicate Greenblatt’s formula. He definitely put in extra parameters other than what he mentioned in the book..

    im just using customized spreadsheets, my database and server, scripts and some other stuff.

    And yes I just rank them according to some metrics. I can’t remember which one for this test though..

  • http://www.buffettopia.com WestChi

    Any chance you could send me the email that ozz got? Thanks in advance

  • david

    Hi Jae Jun,

    can you recommened on stock screen that use CROIC ?

  • http://www.oldschoolvalue.com Jae Jun

    @ David,

    Do you mean recommend a stock from the CROIC list? My best rec would be to just go through a few that catches your eye. That way, you’ll learn about those companies and begin to find what you are looking for.

  • Rob

    Jae,

    I noticed a couple of things on the CROIC spreadsheet and hope you can comment on: Alot of the stocks seem to have recently reached an all time low in price. It’s as if they’re at the bottom of their charts. Is that indicative of stocks that show up on this screener? If so, why is that?

    Also…on the table of stocks for this screener, There appears to be 20 – 30 stocks or so. However, in your description of how you ran the initial screening tests, you say you chose no more than 15 stocks. So, why is there more than 15 stocks on the CROIC spreadsheet?

    Thanks – Rob

  • http://www.oldschoolvalue.com Jae Jun

    @ Rob,

    This screen is designed to find companies that were struggling but have started to turn the ship around. Since CROIC is increasing, management and the company went from losing money to making money, thus the type of companies that show up will likely be at their 52wk low stock prices.

    I only used 15 stocks for the backtest because it is realistic of what someone would be able to replicate and purchase. But the table has a max of 30 stocks to provide more options in case some junk stocks also shows up.

x
Receive Email Tips and Strategies on Achieving Higher Stock Returns

Bonus: FREE Set of 9 Investing Spreadsheets