The Magic Formula Investing Trading System
Joel Greenblatt is a system trader who considers himself to be a “value investor.” In the previous post on Defining a Trading System, I stated that every system answered certain questions, but some didn’t answer them very well. How does Magic Formula Investing answer the important five questions?
What does Magic Formula Investing trade?
This is a system for trading stocks that are listed on the U.S. markets.
What is the Magic Formula Investing set up?
The set up, as defined by the website and as intended for individual investors, looks at variables that are both FundaTechnical and fundamental. The website will sort a universe of stocks on two ratios, Price/Earnings and Return on Invested Capital (ROIC). The twenty-five, fifty, or one hundred stocks highest-rated stocks that meet a set market capitalization hurdle can be returned by the Magic Formula Investing website.
Numerous studies have been done on the Price/Book and Price/Earnings “anomalies” and at this point, only heathens who worship at the primitive idol of EMH aren’t believers in value investing as one of many strategies for long-term outperformance. The Magic Formula Investing website performs some adjustments for differing amounts of taxation and leverage, but these adjustments are in their book and aren’t beyond the ken of a brave man or woman with a spreadsheet.
Their measure of ROIC is also in the book, but the same description is applicable. It is similar to using ROA or ROE as a quality measurement.
In Joel’s hedge fund, I’m certain they also followed up with fundamental analysis like examination of the Q’s and K’s, and proxy statements to check for quality issues and risks that don’t necessarily appear in the balance sheet, but these aren’t mentioned in the Investing website’s FAQ. It is a good idea, however.
Another good idea with this type of set up is to either adjust the earnings for cyclicality, or to introduce a technical screen such as current price trading below the 200 dma. The major flaw of this type of set up is the tendency to catch peaking cyclical stocks – which would tend to have both low P/E and high ROIC. Joel’s hedge fund did an adjustment for smoothing earnings, but it’s unclear whether the website’s screener does this. I’m fairly certain Joel’s hedge fund didn’t use a technical screen.
In my opinion this is an excellent set up, and it would take monster mistakes on the other questions to make this a bad system. That doesn’t mean that good answers on the other questions couldn’t improve the system, however.
How much do I buy or sell?
This system specifies that all positions have equal dollar weight. While this seems extraordinarily unsophisticated, when viewed in the context of risk, it is subtly appropriate. The risk under this system, based on its entry and exit philosophy, is equal for every position, so equal position sizes make sense. This system would buy twenty to thirty stocks.
When do I enter a position?
This system enters a position when capital is available. Capital is available when added to the system or when a position is exited. The initial holdings are filled over the course of a year, with equal amounts of stocks added on a quarterly or monthly basis. Selection of screened stocks is suggested to be random. All positions are long – this system does not sell short.
When do I exit a position?
Positions are exited at approximately one year’s holding time. If the position is, at close to one year’s holding time, a capital loss, it is exited before one year passes. If the position is a capital gain at that time, it will be liquidated after a full year passes. This maximizes the tax efficiency of the system. At the trader’s discretion, if the stock is still on the eligible list at the time of potential liquidation, it may be “re-upped” for another one year holding time. There is no provision for exits other than at approximately one-year intervals.
Summary
The Magic Formula Investing trading system is atypical and controversial among “value investing” systems, in that it actually has an exit strategy – since most “value investing” systems do not. I believe the system “as is” has dramatic longer-term outperformance potential over the indices. Its strength is its set up, which could still be improved slightly. Its weaknesses are entries and exits, and some strategies to improve exits may force a modification of the position sizing algorithm.
Buy Joel Greenblatt’s “The Little Book That Beat The Market” at Amazon.com.
Visit the Magic Formula Investing website.
September 14th, 2006 at 8:50 pm
The follow-up to this specific post is Technical Analysis and the Magic Formula.
This post is part of a series. You can view it by clicking on the Systems category.
November 19th, 2006 at 6:48 pm
Bill,
Have you tried investing using the “magic formula?”
November 19th, 2006 at 8:09 pm
I haven’t specifically used MFI, but generally most “value” systems use substantially similar metrics. MFI uses earnings yield and ROIC, both smoothed, as their variables. My value screener uses PE (inverse of earnings yield), ROA (somewhat analogous to ROIC), D/E, Price/Book, and a technical screen.
I’ve found that there’s a lot of overlap between that and the MFI screen … maybe 25% to 50% depending on the luck of the draw and how you set their screener (Mkt Cap and 25-50-100 stocks).
January 1st, 2007 at 8:24 am
When people want to make a kind of trading system of the Magic Formula, possible the following is interested for them also:
Stocks which are bought based on the criteria (aka. Show up on the screen) but go down a specific amount, in a specific time frame, can be bought again.
This because when a stock could be bought in the first place (on fundamental criteria), and starts dropping, this makes this stock/company only cheaper.
Another trading ‘trick’ is to go with shorter holding periods, because on average the most undervalued stocks rise the most (see page 64 The little book that beats the market).
Personally I think that the ‘formula’ is a good screen, to START the selection process.
The results this year of the random selections are not that exiting. I believe this will continue in the coming years. This because of the –unexpected- rise in commodity prices over the last years.
By this many commodity companies have high returns on invested capital and high earnings yield, which make them to appear on the Greenblatt-screening, although they lack durable competitive advantages. As a starter for further research, I recommend screening on ROC and EY (like Greenblatt) and I also advice to read his book, The little book that beats the market.
Success in investing,
Hendrik Oude Nijhuis
http://www.magicformulastocks.com
May 5th, 2007 at 9:25 am
There is something that I do not understand or have a misconception about regarding the Magic Formula and would be grateful if I could get an answer.
I will use a company, Jackson Hewitt (JTX) to illustrate my question.
The Magic Formula website included JTX on April 23, 07 and contains the following information about the company:
Jackson Hewitt Tax Service Inc JTX 958.70 9% > 100% 04/22 01/31
The fact that JTX is on the “Magic” site means that it fits Greenblatt’s criteria.
Greenberg’s “Little Book That Beats the Market” indicates (page 136) that one could use PE and ROA as surrogates for Earning Yield and Return of Capital.
When I researched the stock on Yahoo on April 23, 07, it indicated that JTX has a PE of 18.91 and an ROA (ttm) 10.50%
If JTX has such a low ROA and a relatively high PE, why is this stock (and many others like it) on the Magic Formula screen?
What is it that I do not understand?
May 5th, 2007 at 11:30 am
Hi Dan, thanks for reading and thanks for commenting!
ROA over 10% is pretty darn high. The S&P 500 median is 3% on a TTM basis, according to my screener. The same screener shows a “market” median P/E of 20.1. I think the MF screener rank averages the two, so being very high in ROIC might make up for a pedestrian Earnings Yield.
Joel says one could use PE (ttm) and ROA as surrogates for his earnings yield and return on capital because his numbers (that he used at his hedge fund) are adjusted, and quite possibly the numbers on the website are adjusted, too. The earnings are smoothed, so the one-year snapshot most screeners will show is slightly off from Joel’s smoothed number, and his smoothed ROA would be different from a snapshot ROA, not to mention that ROA and ROIC might be slightly different, because not all assets might be counted as invested capital in the screeners.
All that being said, I think if you do a ttm P/E and ROA ranking, average the ranks, and sort the stocks, you would have a very high degree of overlap with what shows on his screen.
Feel free to leave another comment or email me (see the contact page) if you have other questions!
February 19th, 2008 at 8:33 am
Greenblatt did say that ROA and PE are good substitutes for ROTC and EY, but they are not perfect replacements.
ROA usually includes intangibles like goodwill, trademarks, and software. From my research, I’m pretty sure Greenblatt removes deferred tax accounting from his ROTC calculation as well. This makes sense as it’s not invested capital in the most fundamental sense.
PE is also not a perfect analog. EBIT/EV is a good valuation metric because it measures just the business. PE can be highly influenced by debt interest (either way), tax accounting, discontinued operations, etc.
Bill’s analysis is excellent and my website is dedicated to providing the due diligence on these stocks as he recommends:
http://www.magicdiligence.com
June 19th, 2008 at 1:26 pm
Bill, In your article you mention that it is up to the trader to keep stocks after one year that are still on the buy screens. can you elaborate on the decision parameters as well as the idea of cashing some of the earnings. many thanks
September 6th, 2008 at 7:46 am
Clarification on this old post - for quite a while I’ve been aware of research on the interaction of value and momentum, but it slipped my mind to add this comment.
“Peaking cyclicals” aren’t as much of a problem as value traps. The technical screen should most likely be for POSITIVE momentum.