Predicting cash flows

[oddballstocks]

In a previous thread someone linked to a presentation by some value investors who said they predict 10 years of cash flows and then invest at a discount to that.  Seems all well and good, sort of standard investment advice there.

 

So my question is this, does anyone look back at their cash flow predictions 10 years later and see how accurate they were?  I'd love to see the predictions these guys had 10 years ago and then compare them to how they really came out.

 

Does anyone do this?  If not then how do you know your predictions are accurate?  Or does it even matter?

[berkshire101]

I haven't been investing for 10 years so I can't say.  But when I made my first investments back around 5 years ago the cash flow is higher since my purchase.  It' actually higher than my estimates.  I used return on capital as a proxy for my estimates.  Then see if book value per share including dividends paid and share purchases grew at what rate.

[Morgan]

I haven't been investing for 10 years either, but I do something similar when buying a new apartment building. I always estimate FCF and how the purchase will change the balance sheet. I've looked at the estimates from the past and compared them to the outcome and I've gotten a bit better at estimating over time. By now, 3 solid years in, I can walk through a building and pretty accurately estimate the renovation costs, FCF and what is going to happen to the BS. Thinking about my strength for looking at real estate, makes my weakness for analyzing other businesses quite apparent. More filing to read!

 

Edit: As for looking at the cash flows 10 years out, I haven't modeled that far out. I see how much I'll make per year compared to how much cash I have to spend to buy the cash flows from the building.

[tede02]

Modeling 10 years out seems like a joke to me.  There is just way too much randomness to account for.  Go back to the year 2000, The US was projected to have paid off all of it's debt within the next decade, look what really happened.  At the same time, who was talking about Apple?  Now the company is the most valuable in the world (by market cap).  How many people had heard of Google in the year 2000?  I'd love to see the cash flow projections of retailers going back a decade. 

[jawn619]

Modeling 10 years out seems like a joke to me.  There is just way too much randomness to account for.  Go back to the year 2000, The US was projected to have paid off all of it's debt within the next decade, look what really happened.  At the same time, who was talking about Apple?  Now the company is the most valuable in the world (by market cap).  How many people had heard of Google in the year 2000?  I'd love to see the cash flow projections of retailers going back a decade.

 

agreed. 10 years is an insane amount of uncertainty to put any type of precision into. But i don't think you have to be precise, just right. Better to be roughly right than to be precisely wrong.

[racemize]

I used to use DCF/future earnings analysis.  A year or so ago I stopped and just went with the, "I better know it is cheap if I'm buying" it mentality. 

 

The false precision of long-term projections feels dangerous to me.

[Homestead31]

nate as i'm sure you are aware there are all sorts of studies that reveal how poorly people can predict the future, how bad analyst's track records are, and the fact that so called "experts" actually perform worse than layman b/c their high level of confidence in their abilities blinds them to their flaws.

 

10 years is just insane in my mind.  it is difficult to predict a 2 variable scenario out into the future.  a business has about a million different variables. i'd also be curious to see how these people think about recessions because in any ten year period there is a real good chance that cash flows are going to be seriously impaired at some point.

 

personally i try to think about cash flows 2-3 years out which is kinda crazy in and of itself, which is why i need to be able to drive a truck through a valuation gap caused by some obvious factor (non-economic selling, easily identifiable temporary problem, etc etc).  even when i can identify some obvious factor i still figure i'm probably at least half wrong about something.

 

i'm obviously talking about margin of safety here.

 

what is amazing to me is the amount of people that say, "buy XYZ, i think it is 25% under valued!" even on well respected sights like valueinvestorsclub.com. 25% might be acceptable for a truly great company, but in my mind it is no where near a big enough discount for an average company because 25% can vanish in a heart beat when you find out you were wrong about something (and you are definitely wrong about something).

 

part of this of course has to do with portfolio construction as well.  if you want to own 100 names youre gonna have to settle for some 25% off names.

 

if you want to own 10-15 names, they better all be potential doubles in 2-3 years in  my opinion.

[thepupil]

what is amazing to me is the amount of people that say, "buy XYZ, i think it is 25% under valued!" even on well respected sights like valueinvestorsclub.com. 25% might be acceptable for a truly great company, but in my mind it is no where near a big enough discount for an average company because 25% can vanish in a heart beat when you find out you were wrong about something (and you are definitely wrong about something).

 

maybe it's just me and that I've only been really investing since mid 2011 and am pretty US focused and not a spectacular investor, but I have never been able to put together a portfolio of  things I consider to be safe and at a 50% discount to fair value (ie a portfolio of 2X's)

 

I think I own a portfolio of growing 60-80 cent dollars (even some 100 cent dollars in there) and a few high risk multibagger payoffs with 100% downside, but If I could build a relatively diversified (say at least 6-10 names) portfolio of 50 cent dollars, I wouldn't be working for a living.

 

Maybe I'm just mediocre or set my expectations too low. Can someone walk me thru a portfolio of 10 names (just 2 bullet points) that they own and all of which are low-risk 2x's over the next 3-5 years?

[mjohn707]

maybe it's just me and that I've only been really investing since mid 2011 and am pretty US focused and not a spectacular investor, but I have never been able to put together a portfolio of  things I consider to be safe and at a 50% discount to fair value (ie a portfolio of 2X's)

 

I think I own a portfolio of growing 60-80 cent dollars (even some 100 cent dollars in there) and a few high risk multibagger payoffs with 100% downside, but If I could build a relatively diversified (say at least 6-10 names) portfolio of 50 cent dollars, I wouldn't be working for a living.

 

I must be an even worse investor than you, because I don't think I could put together a super safe portfolio of 5 names that were all at a 50% discount  ;)  I think it's possible to put together 30 relatively safe names at a 30-40% discount however

[berkshire101]

The way I see it, if it was easy to predict cash flow then everyone would be rich.  But it's very hard.  Looking from a company's management perspective, the good ones usually lay out a 5-10 year plan and execute them.  And that's the best we have to go with.  In the end, you're investing in a business.  And every business has a plan, it just depends if that plan is good or not.  I don't think anyone is going to go into business and say I'm starting a business so later on I can sell it for half the book value. 

 

There's no right way to invest.  I just think if I were to own a business, I would seek to grow it and that requires looking years out where the business will be.  I think that's why people are fascinated with future growth.  It's the entrepreneur inside all of us. 

 

Softbank has a crazy 300 year business plan.  Gotta admire the enthusiasm of  Masayoshi Son.

[SpecOps]

For the best investments you don't to do a DCF to know they're cheaper than their discounted cash flows. Unfortunately they don't come along too often.

 

I've not been investing for 10 years so unfortunately can't say

[rukawa]

It may be difficult to predict future cashflows but implicitly aren't we doing exactly that when we say stocks are cheap? You could proceed in the opposite direction and start with a stock that is cheap and ask yourself whether the cashflows needed to support the price you think it should be are reasonably attainable.

[anders]

Its an interesting topic indeed..  :)

 

I think there are so many different dimensions in a DCF that imho one is probably better off not relying too much on the concept and just keep focus on reading reports understanding the business... Further, imo its been developed into a tool to get as much pay as possible when selling an asset..

 

I have backtested my data, the only ones that I have been on track, are stalwarts that has a business structure that allow them to grow at a moderate rate every year... ie structures as walmart... but they always come with a premium..

 

I think the conversation below kind of illustrates my point on DCF..

 

Munger: But you say there is some vaguely established view in economics as to what is an optimal dividend policy or an optimal investment?

Bratton: I think we all know what an optimal investment is.

Munger: No, I do not. At least not as these people use the term.

Bratton: I don't know it when I see it... but in theory, if I knew it when I saw it this conference would be about me and not about Warren Buffett. (Laughter from the audience)

Munger: What is the break point where a business becomes suboptimal in an ordinary corporation or when an investment becomes suboptimal?

Bratton: When the return on the investment is lower than the cost of capital.

Munger: And what is the cost of capital?

Bratton: Well, that's a nice one (Laughter) and I would...

Munger: Well, it's only fair, if you're going to use the cost of capital, to say what it is.

Bratton: I would be interested in knowing, we're talking theoretically.

Munger: No, I want to know what the cost of capital is in the model.

Bratton: In the model? It will just be stated.

Munger: Where? Out of the forehead of Job or something?

Bratton: That is correct. (Laughter)

Munger: Well, some of us don't find this too satisfactory. (Laughter)

 

Best,