DEEP VALUE Supplementary Readings for Ch. 2, Deep Value


Below are supplementary readings and original sources for Chapter 2, Contrarians at the Gate, in Deep Value. This is an important chapter because we are introduced to the father of security analysis. He was the first person to systematize his analysis and separate the price of a security from its intrinsic value.

I realize that some may find Graham fusty and his prose turgid/boring, but Graham was a Renaissance man who had a razor-sharp intellect and integrity.  You can’t lose by reading his works.  Note his attitude, skepticism, and logic.

So far in the course I am surprised by the ironies and subtleties of Deep Value investing–picked up in the preface of the book, Deep Value.  Our Deep Value Group has close to 400 members, yet, currently (Jan. 17, 2015), the number of net/nets in the US market is negligible.  Financial assets (with a few exceptions) are sky-high in valuation due to Central Bank intervention, negative interest rates, and a six-year upward trend in US bonds and stocks.  I am surprised at the interest.

We invest today for the future, but the future is unknowable. Look at the predictive record of experts!  Investors tend to extrapolate the past into the future just in time for a reversal of fortune–reversion to the mean reverses the trend.

Deep value investing takes advantages of the cognitive biases of others to gain profits, but if we are human too then how do we avoid the same?  Cheapness or the discount from intrinsic value is the main determinant of margin of safety. The cheaper you buy the less risk and MORE reward. This smashes academic finance theory in the face.

Here I am quite surprised that the highest returns to the net/net strategy go to the MONEY-LOSING companies. The worst of the ugly gain the most. Perhaps because price drops the most due to fear and earnings trend extrapolation.  Like assuming the driver of the car will continue to motor off the cliff instead of turning or stopping the car.

A TEST: Let’s say a company earned $5 per share and it is growing at 5% per year, the cost of capital is 10%, and it is trading at $85 per share. Then the next year earnings drop to $1 per share and the following year earnings drop to $0.05 or five cents, the next year earnings go up to $3 and in the fifth year back to $5.25 and 5% growth. After ten years the company will continue to grow but only at 3%.  Guess the price drop quickly and write it down.  Now do the DCF and see where the intrinsic value is compared to your guess.

Last year $5.00

Year 1 $5.25

Year 2: $1.00

Year 3: $0.05

Year 4: $3.00

Year 5: $5.25

Year 6: $5.51 (5% growth)

…. next year, next year, etc.

After year 10, then terminal value 10% cost of capital with 3% perpetual growth. Can a financial wizard post in the comments section?  Was anyone surprised by your initial reaction the Intrinsic value result vs. their initial thoughts on how price would react?

Back to the post….

The higher probability (Montier, 5%) for each INDIVIDUAL company to go down 90% or more vs. 2% of non net/nets may cause other investors to shy away from investing. HOWEVER, the GROUP of net/nets STILL outperforms.   You have certain companies go to zero and some rebound multiple times but you don’t know which ones.  You have to deal with much uncertainty but believe you are playing the odds like an insurance company.

When net/nets are abundant like in 1932, the world appears to be ending. Investor fear is off the charts. The question is whether you will have the capital to buy and the courage to act. Again we come round and round to temperament or character or whatever you want to call acting in the face massive fear.

But knowing that a company is too cheap when it trades at less than 2/3rds to net asset value (Asset value is tenuous, liabilities are 100 cents on the dollar) does not seem difficult, but probably the surrounding circumstances for the company and/or market are ugly! “Don’t you read the papers!” an outsider might say if he or she learned of your purchase.

Remember 2009? Jim Cramer in a panic.   (just paste into your browser)

Those are my thoughts and questions so far as I keep reading.

Graham’s Writings and Testimony

Graham Testimony to Congress (note his remarks on the MYSTERY of price eventually closing the gap with intrinsic value)

Important writings on Liquidation Values during the 1930s

1932_American Corporations Worth More Dead than Alive 3 Parts by B Grahams  (Please read). You can’t understand the depths of despair in the financial markets (and thus the net/nets and prices below liquidation values) without understanding the preceding boom.

Historical perspective

A Study of Market History through Graham Babson Buffett and Others  The 1920s BOOM.

A Great Depression_Rothbard Obviously, you don’t have time to read this, but IF you do want to understand the causes of the biggest bear market in US history then this is the definitive work.  The book destroys the common wisdom that the depression was caused by the Fed’s tight monetary policy.

The Outperformance of Net/Nets

Net NEt Strategy in London


Benjamin-Graham-s-Net-Nets-Seventy-Five-Years-Old-and-Outperforming  I imagine that some net/nets are micro/nano-cap stocks under 50 million in market cap and with wide (5% to 10% bid/offer spreads) perhaps the studies do not deduct the spread?

There is a lot here so I will refrain from posting until the middle of the week. Take your time with Chapter 2.   I will be posting next on Klarman’s thoughts on liquidation and valuation.

Traditional finance savaged: The Dumbest Ideas in Finance_Montier

 How NOT to take this course

You don’t have to read EVERYTHING, but you do have a choice. Better to understand what you read.

24 responses to “DEEP VALUE Supplementary Readings for Ch. 2, Deep Value

  1. Accodring to my calcs, the company is worth $79.40 based on a a growth rate of 5% in perpetuity. It drops to $63.16 based on 3%.

    I figured that it would be worth about $80 on the 3% rate, so my initial guess was far off.

    I’m not sure if anyone else agrees with my figures – it is so easy to make off-by-one errors when doing DCF calcs. So we can thrash the numbers out if anyone disputes them.

  2. I found $105.00 for the intrinsic value of the 1st simulation (=5.25/(.10-.05)) and $63.16 for the 2nd simulation (DCF’ing it). Considered all the cash flows on the last day of each year, just for the sake of simplicity.

    There is a clear lesson here: the severe profitability erosion (-99% from 5.25 in Y01 to 0.05 two years later) didn’t cause as much value destruction as one would expect, because it was temporary.

    It looked as though the car was falling down the cliff but it didn’t. It may have suffered some heavy bumps, but definitely not total loss. People would probably overreact to an earnings tumble like that.

    Another conclusion is that the difference in the growth rate after Y10 (3% vs 5%) didn’t account for too much of a difference on the value.

    • You have to discount the terminal values back 10 years, though, and add on the cashflows for the first 10 years.

      I calculated that the PV of the 1st 10 years was $25.15. The bulk of the valuation comes from the terminal value – no surprises there.

      I calculate that the PV of the terminal value (discounted to present) for the 5% perpetuity is $54.25, whilst for the 3% perpetuity it is $38.01.

      So to get the total PV, you add the termianl values to the PV of the 1st decade. In other words:
      5% perp: 25.15 + 54.25 = 79.40
      3% perp: 25.15 + 38.01 = 63.16

      Agree, disagree?

      • No disagreement.

        I also found the $63.16 intrinsic value. The other intrinsic value I mentioned ($105.00) is another simulation concerning the value of the company if it kept growing 5% annually forever (with a 10% cost of capital). It would be, I guess, the expected value of a company which had been growing 5% since forever.

        P.S.: also agree with the $25.15 for the first 10 years.

        • gentlemen,
          can you please help me, My values are not eactly 63.16, where am i going wrong?

          Year 0 5
          Year 1 5.25
          Year 2 1
          Year 3 0.05
          Year 4 3
          Year 5 5.25
          Year 6 5.5125
          Year 7 5.788125
          Year 8 6.07753125
          Year 9 6.381407813
          Yer 10 6.700478203

          Terminal Value = 6.70/(0.1-0.03) = 95.72111719
          1. Present Value of TV = PV(0.1, 10, 0, 95.72) = $36.90
          2. PV of cash flows for 10 years = $25.15
          Sum of PVs (1+2) = $62.06


          • The terminal value should be 6.70*1.03/(0.1-0.03) = 98.59. The Gordon’s growth model states that
            V = D1 / (r-g), where D1 is the dividend at year “1” (although actually year 11 in our case).

            The dividend in year 11 is not 6.70, but 6.70*1.03. So in part 1 of your answer, you just need to take 36.90 * 1.03 = 38.01.

            If you make that adjustment, then our figures will agree.

            I bet that analysts make much more egregious errors than that on a regular basis, though!

          • @mcturra2000,
            brilliant! thank you. Now, I know where I went wrong. Ofcourse, I should have considered year 11. Prof Kaul (coursera, umich, introduction to finance) will be very unhappy with me because of this mistake.

            Thank you again.

    • # Terminal Value = 6.70/(0.1-0.03) = 95.72111719
      # 1. Present Value of TV = PV(0.1, 10, 0, 95.72) = $36.90

      “The terminal value should be 6.70*1.03/(0.1-0.03) = 98.59.”
      “you just need to take 36.90 * 1.03 = 38.01”

      Hey mcturra2000,
      Why is the 36.90 still used in the calculation to get PV of TV when the TV is not 95.72..
      Shouldn’t it be PV(0.1,1,10,98.59) which should give approx 39.13 and val of 25.14+39.13= 64.28 for 3%

      • Maybe I’m misunderstanding you, or have made an arithmetic error.

        98.95 is the terminal value of a 3% perpetuity as at the end of year 10. I need to discount that back 10 years: 98.95 / 1.1 ^ 10 = 38.01, as per my original calc. The Excel function is: PV(0.1,10, 0, 98.59), which gives the same answer.

        I used the $36.90 simply because it was out by a factor of 1.03. It was an arithmetical shortcut, and perhaps introduced confusion. I was working on the basis that multiplication is “commutative”, which means that:
        PV(0.1, 10, 0, 95.72) * 1.03 = PV(0.1, 10, 0, 95.72*1.03)

        Whilst aaruni was “one short” by a factor of 1.03, your calcs are “one too many”. What you did was:
        6.70 * 1.03 ^2 /(0.1 -0.03)/1.1 ^ 10 = 39.13
        but I’m saying it should be
        6.70 * 1.03 ^ 1 /(0.1 -0.03)/1.1 ^ 10 = 38.01

        Anyone else agree/disagree with this? DCF can be pesky at times!

  3. 1st simulation = the value of the company if it just kept increasing its $5 profit in the previous year by 5% every year until forever (discounted by an opportunity cost of 10%)

    2nd simulation = the one John propposed us to make

  4. John,your article about Graham on net-nets is very interesting, and explains something that I never knew about their existence: that the surpluses were the result of large capital infusions before the buble burst.

    This seems analogous to tech stocks in the early 2000’s. I was in the tech sector in the late 90’s, so I had a front row seat. Fortunately, I was sensible enough to realise that the sky-high valuations were a sign of a bubble.

    In 2002/3, th bubble had truly burst, and many companies had lost 90% or more of their value. I figured that despite such losses, they were still likely to be overvalued. But here, I think I missed a trick. A tech investor at the time said that people were arguing over whether high PE or low PE tech stocks represented the best value, and he turned around and said that it was the ones that were making a loss that would turn out to be the real winners.

    I don’t know if that was, in fact, the case, but I was probably worried about the fact that even if a company had a lot of cash, there was often significant cash burn. It would be interesting to see what loss-making tech net-nets there were at the time, and how they fared.

  5. This is a very interesting hypothetical question. Really put into perspective for me how temporary drops in earnings are a small part of a companies longterm PV.
    I must say though a drop from 85 to 63 is more than I would have expected from the calculation.
    However, on such a drop in profits I suspect its price would have fallen far further and it would have felt justified!

  6. I derived an intrinsic value of $123.47, based on the business earning $6.70/share at Yr 10, and growing in perpetuity at 3% ($98.33) + adding the discounted value of Yr 1 – 10 EPS stream ($25.15).

    mmm… my guess at the price drop would be in the order of 60% given that earnings all but disappeared. I’m not surprised… a business that can lose 90% of it’s earnings in 2 years should probably have a higher cost of capital

    • The purpose of this example is to show how:

      We tend to over-react due to recency bias. Since most of the value is in the terminal value, IF the business reverts back to normalized earnings, then two to four years of negligible earnings/cashflows will not effect intrinsic value as much as you would think.
      Also, when a company that you own has a blow-up you can go back and check your assumptions to determine whether the market over-reacted and thus you hold or buy more. Many times a stock may decline 50% but intrinsic value declines 20%–so unless there is a better opportunity at a wider discount–then selling would not make economic sense. Many investors neglect to remain steely eyed in the face of a collapsing stock price.

  7. Quick guess of price drop: $55

    Value of earnings in year 1-10: $25,15

    Terminal value (TV):
    – at 5% growth: $54,25
    – at 3% growth: $38,01

    Total value:
    – at 5% TV growth, $79,40 ($25,15+$54,25)
    – at 3% TV growth, $63,16 ($25,15+$38,01)

    Shrinkage in value from decline of TV growth rate from 5% to 3%: -16,24 (-20,45%), this from a decline in the terminal growth rate of 40%.

    Value of earnings growing at 5% in perpetuity: $100 ($5/(10%-5%)).

    My own initial guess was too low compared to the calculated values in the two scenarios.

  8. John this is a great example and your conclusion is borne out by crunching the numbers further. I know the intention of the course is to increase our thinking capacity over our ability to count beans.

    discounting the present value(PV) of a $5.25 per share earnings stream growing at a steady rate of 5% for 10 years and then at 3% into perpetuity yields an intrinsic value of $85. By coincidence the exact market price.

    Discounting your earnings as per your example gives a PV of $63.16

    My seat of the pants guess on price reaction to the cratering of earnings was the shares falling to $60 on the first miss and then to $28 when reporting $0.05 in year 3. This is my anticipation of the market reaction and would not be my gauge for valuation. There will always be players in the market that believe the fall in earnings is fundamental and that the business is doomed.

    If this is a one off event as is borne out by the comeback in earnings after year 3 then normalizing earnings over a longer period is the best approach and offers a plethora of opportunities. Not only would you not puke out your position below $63 but if Mr. Market drives the price lower you may consider adding to your holdings.

    Scenario 2 No comeback in earnings
    a) Crunching out the numbers on $1 per share earnings in year 2 with no growth into perpetuity gives a present value of $8.63

    b) $0.05 in year 3 with no growth into perpetuity gives a present value of $5 and 95% is the year one earnings of $5.25 which are never returning.

  9. At $85, the stock is trading at 17x earnings of $5.25, and if the eps falls to 1.00, then it would be easy for the market to price the stock at $20 or under, and when earnings drop to $.05 a share, the market may even price the security like it is going out of business. And you could literally get a $3-5 stock.
    I understand the premise of the question and to get us to think about reversion to the mean, but we have hindsight investing here. That is a tough call to make that a stock just had a blip and earnings will rebound. But even after the earnings revert higher, he market will price the stock at a lower multiple until eps goes back over peak numbers, $5.25. That is a major debacle to go from 5.25 to 1.00 to .05. Especially with the way earnings are typically adjusted adjusted, so those exclude one time items, bc who prices off of GAAP earnings these days (sarcasm, but true for majority of investors).
    That is just how I thought about the scenario, maybe wrongly or too pessimistic.

    • Good points. Don’t be hesitant to disagree, point out errors, provide alternative lessons. etc. We are all students of investing. Yes, it is easy to look at an example and play reversion to the mean vs. sitting with a stock down 50% or 60%? and showing $0.05 for the prior year’s earnings. The future looks bleak. Therein lies the opportunity if you can normalize earnings for a cyclical business. If you look hard enough you can find some pretty stable businesses.

      Back in 2009, Miller Industries, Inc. (MLR) was trading at $5, under my estimate of liquidation value plus it had earnings power value of about $1.00. The business was stable because of its dealer network–much of its recurring revenues came from parts and service of the winches. The company made tow-trucks. They sold the trucks at cost.

    • And to make a further point … let’s say that does play out the way you suggested … the stock goes to $5 on a EPS of 0.05.

      That gives it a PE of 100. And this is the awkward thing about Deep Value investing that I was kind-of arguing in my post about why I thought Greenblatt’s short of Carnival was a poor-quality idea: rather than being a lousy time to invest, a PE of 100 would in fact be a great time to invest. This is why you have to be careful about PEs, ROEs, and Magic Formulas.

      If we being “regular” value investors, we might have easily concluded that the company was heading to the tank. It is, of course, very easy to say what we should do when we have the future dividends laid out before us. In the hear of battle, though, where future dividends are not known, we would probably be inclined to conclude that the stock is too risky. Maybe we’d think we had another Munsingwear on our hands, and should avoid it like th plague.

      Very difficult, isn’t it?

  10. Question. Shouldnt the terminal value be discounted back 11 years instead of 10? Year 0 is last year, year 1 would be the end of the “current year” and year 11 the end of year 11.

    • I think the original calcs are correct.
      The Gordon’s growth formula states that:
      V = D1 / (r-g)
      where D1 is the the dividend at the end of year “1”, and V is the value “now”, where “now” is actually “in 10 year’s time” in our case. So you should discount back 10 years, rather than 11.

      I have carried out the calculations to Year 199 at the following spreadsheet:
      You can see that the numbers agree with my original calculations (phew!)

      I hope that helps.

  11. Damn, I think I always made that mistake before! ; P Thanks mcturra2000!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.