Saturday, December 06, 2008

Fallacy of scruntinizing ratios for the past few years

Financial statements are very essential in fundamental analysis and value investing. To value investors, it's like soldier's M16, accountant's Excel, taxi driver's taxi, you get the idea. From the statements, we calculate the all important ratios. Basically it's one number divided by another and that's supposed to give you insights into the company's business operations, its edge or whatever. Something like adding apple juice to aloe vera and drinking the new juice will actually make you thin!

Actually around like 2% of the time, looking at ratios actually help a bit. The problem is we always only look at last yr's ratios. Or maybe some diligent analyst will look at ratios for the past 5 yrs. Not bad huh. 5 yrs quite long right? Presidential term only 4yrs. Alas, do they know on average how long is one economic cycle from peak to peak or trough to trough?

Well, its seven years on average. The recent one, 2001 to now and still going. The one before 1997 to 2001. 4 yrs, well sometimes it's shorter than average, obviously. So by looking at 5 yrs of ratios like ROE, operating margin, debt to equity ratio, can you really tell if the co. is really good? Esp if you are looking at 2002-2007 and the ratios just keep improving every yr? Like those we see in annual reports. Sooooo impressive. This co. is really something, we tell ourselves.

Our hero, Warren Buffett some yrs back started to ask co. owners to come forward to him if they intend to sell their stake to Berkshire. One of his criteria: GROWING earnings for the past 15-20 years. So seven years still quite amateur actually. Of course, most Sg co.s were not around 20 years ago. So we just have to make do with seven years lor.

So, that's the moral of the story: a co. with improving ratios for the past few years is not necessary a capable one. Rising tide lifts all boats. Keep the economic cycle in mind when looking at ratios for the past few years.


  1. Yes I agree with your article. A lot of emphasis is placed on the P/E ratio but not a lot of emphasis is placed on the E in the P/E ratio often. If we only use the E for the past 2 to 3 years, it is likely to give an inaccurate picture since like what you said, a rising boat lifts all tide. It is much better to look at the earnings history of the company.

  2. Agree with your point on Earnings and PE. Maybe the ratio emphasis should be on the balance sheet and cash flow such as current ratios and CF/interest expense which would be more meaningful. This is similar to the investment philosophy of Martin Whitman.

  3. "a co. with improving ratios for the past few years is not necessary a capable one. Rising tide lifts all boats. Keep the economic cycle in mind when looking at ratios for the past few years."

    True true.

    "One of his criteria: GROWING earnings for the past 15-20 years."

    From where? It is so Un-WB.

    Looking back 5 to 10 years is already very good enough if one is really scruntinising financial statements with good understanding of the company. Doing it back for 15-20 years don't make much diff. If so why not at least over 20 years of financial statements?

    But if it is just "scruntinising financial ratios" alone, doing it for 15-20 years will reveal a good company until NOW but it won't tell us the future.

  4. Hi Kay, btw we should always use E in the future when calculating PE ratios.

    Usually this is done by using the average or median earnings estimate by analysts. Or we can always try to predict that ourselves.

    Of course when we have earnings history of a co. for 20 yrs, and the E increases like 2% every year, then perhaps we can safely assume that it would be so next yr as well...

  5. Hi Donmihaihai,

    Buffett mentioned this a few times in his annual reports and a I read it from Essays of Warren Buffett, a compliation of articles from his annual reports.

    Berkshire started asking owners of co. who want to sell their stake for whatever reason (no successor, would like to cash out some of their success etc) to consider selling to Berkshire.

    Of course, he did not want just any guchi gurak firms, so he asked for growing earnings (not necessary straight line) for 15-20yrs. And of course other relevant financial details.

    Another criteria was that Berkshire cannot provide management, the seller needs to continue to operate the business.

    Business must be understandable to him etc.

    As to your qn about 5-10 yrs is enough? Why not 20 yrs?

    Let me try to answer. First the average cycle is 7 yrs, so I would say that 5 yrs is actually not enough. 10 yrs, well maybe.

    15-20 yrs you get to see 2.5 cycles, which is as practical as it can get. Asking for 30 yrs, which we can see 4.5 cycles is definitely better but may not be practical.

    It is also important to consider which industry the business is in. If it is tech, actually you only need 3 yrs, bcos the cycle is only 18 mths according to Morse Law. And technology just keep changing, and even Morse Law will become irrelevant soon.

    So 15-20yrs is only for those mundane industries, like drinks, shampoo, newspaper, shaver, chocolate etc. There will be fluctuations with the economic cycle but if some co. can demostrate good growth after 2 or 3 cycles, then perhaps it says that this co. is good enough for Buffett, or any other value investor.

    That would be my own intepretation as to why Buffett would want 15-20 yrs. Of course, it is definitely not a strict one nor would it be an exclusive one. He is known to make snap decisions as well.

  6. Hi 8percentpa.

    This time I don't like to go round and round. I asked from where in the 1st place is because I am extremely confident that this criteria has never been come out from WB mouth or hand.

    If it is really somewhere out there, please put a link for me.. I am always ready to take back my words.

    Lastly, I have read both chairman letters(all) and The essays of WB at least 2 times and some portions, many times.... But I never come across it. Anyway it is the idea behind it that is more important... From your article, it is so un-WB.

  7. Anyway since I am at this topic, I just point it out for you. Chairmain's letter 1991.

    We would love to see an intermediary earn its fee by thinking of us - and therefore repeat here what we're looking for:

    (1) Large purchases (at least $10 million of after-tax earnings),
    (2) Demonstrated consistent earning power (future projections are of little interest to us, nor are "turnaround" situations),
    (3) Businesses earning good returns on equity while employing little or no debt,
    (4) Management in place (we can't supply it),
    (5) Simple businesses (if there's lots of technology, we won't understand it),
    (6) An offering price (we don't want to waste our time or that of the seller by talking, even preliminarily, about a transaction when price is unknown).

  8. Sorry this is my last comment.

    As for the number of years, leaving those "standard" cycles taken from don't know where, I just like to say.....

    Have you invest in stock? I bet you do..So how many companies you have look back more than 10 years of their financial data?

    My questions are simple. To be good at something, practising is a must, so if you know that at least 10 years financial data is required, I bet you have been looking lot of companies with > 10 years of data

  9. Hi Dominhaihai,

    Thanks again for very valuable info posted. The criteria is in fact posted by yourself!

    (2) Demonstrated consistent earning power (future projections are of little interest to us, nor are "turnaround" situations)

    As for the 15-20yrs, I have to admit Buffett might not have explicitly written "15-20 yrs of earnings growth needed" anywhere.

    Below are some examples I found on the same Berkshire link that you gave as to why I think Buffett looks at businesses with long track record. No. of yrs that the company was in operations are in brackets.

    1992 Mrs B’s carpet business (9 yrs)
    1993 Dexter (37 yrs)
    1995 Helzberg's Diamond Shops (21 yrs)
    1996 Flight Safety (45 yrs)
    1997 International Dairy Queen (27 yrs)
    2000 Benjamin Moore Paint (117 yrs)
    2000 Johns Manville Corp (18 yrs, co. went bankrupted before that then restructured)
    2000 Ben Bridge Jeweler (89 yrs, Buffett mentioned how same store sales growth of seven yrs impressed him)

    Of course this does not imply 100% that Buffett die die wants 15-20 yrs of earnings growth and it has to be strict criteria before he says yes. The idea is truly in his own words “consistent earning power”.

    Of course, everyone has his own interpretation to what is consistent earning power. How long, how stable, etc.

  10. Hi again Dominhaihai,

    Just to answer the other parts of your qns.

    As for the average 7-yr cycle, you can check out this wiki page

    The Juglar Cycle is 7-11 yrs (an academic notion though)
    The actual average US economic cycle is about 5 plus yrs.
    Japan had a 13 yr down cycle 1990-2003
    China’s last 2 cycles were 6 and 8 yrs (1993-1998, 1999-2007)

    As to whether I invest in stocks. Yes, of course, since 1999 I think. How many co.s I looked at for more than 10 yrs? Geez, should be more than 100, I guess.

    If you have access the Bloomberg, you can use FA function and then change the no. of yrs to as long as you want. Usually it's about 15 yrs only though. For SG co.s shorter still.

    I have to admit it will be quite a hassle for retail investors with no access to Bloomberg to gather 10 yrs of financial data. So I wouldn’t be able to do that without technology.

    I have actually thought about such a business idea, providing Bloomberg use to retail investors. Say charging $2 for 10min use or something…

    Well in any case, hope this helps!

  11. Was looking at these past comments I found it amusing that I was so polite to someone so unruly.

    Just a further addition to the long term consistent earnings that Buffett wants before he invest.

    I re-read my post and realized I did not actually write anything about whether Buffett stated things in ink or said it out loud.

    Buffett has criteria for his investment. In written form, it's spelled out in one of the comments above. But in his mind, it has to be quantifiable.

    We do not know the exact no.s but judging from what he has done, he wants long term.

    In another comment that I posted above, he really looks at companies with a long history.

    Well, today, being more experienced, I would say he looks at things in at least 10 years perspective. 15-20 years would probably be the norm.

    Of course, there is no rule that one-size-fits-all. I would apply different time horizon to different co.s.

    It is said that we judge people by what they do more than what they say (or write).

    Buffett looked at IBM since 1991, but only bought it recently. That's 20 years.

    Excerpts from a recent WSJ article below


    Jeff Matthews, a money manager who holds Berkshire Hathaway stock and is a Buffett-watcher, sees similarities between Buffett’s investment in IBM and his longtime investment in Coca-Cola.

    “He had followed Coke for decades, and always said he should have done it sooner, and now he’s saying the same thing with IBM,” Matthew said.

    Another echo of the Coke investment, Matthews said, was the stock price — Buffett bought Coke at or near a high after the 1987 market crash, Matthews said. IBM has recently been trading near its alltime high, set a month ago.

    When Buffett invested in Coke, Matthews said, “People said, ‘what is he doing paying that high a price?’ It wasn’t a value stock, Buffett’s supposed to be a value guy.”

    In 1988, when Buffett first made a major investment in the drinks company, Coke shares were trading around $5 a share, adjusted for stock splits. Today, Coke shares are near $68 apiece.