DCF: Pro's/Con's

[SharperDingaan]

 

As there has been a lot of DCF discussion recently, for easy reference we thought we'd start a seperate string on the topic.

 

Some things that we've noticed:

 

The older you get the less you use it, & you use it differently.

- Right out of B-School you know squat, but you can do DCF! DCF crunching substitutes for experience.

- DCF is precise, & there is only 1 answer! Dogma can be blinding.

- There is only the AIMR/CFA way! Currency devaluation.

 

If you actually understand the business you're looking at, you will instantly recognize that much of its value is actually off BS. Market place reputation, network ability/access to deals, mgmts integrity/smarts, etc. None of it shows up in DCF, & you need maturity/experience to realize that.

 

Real businesses have monthly/quarterly ups/downs, but DCF projections dont. Wise men recognize that a small difference, early, will have a major impact on NPV (DCF end point) & valuation; & that because NPV calculation takes time, time starved market analysts are slow to recognize it.

 

The older you get the more you think like an owner.

- For most coy's/industries there is more than adequate published 'live time' information to give a reasonable sense as to whether the upcoming quarters earnings are likely to be better/worse than expected, & why. CF today & in the future.

- You also recognize that it is almost impossible to predict how much CF will occurr or when. You`re fairly sure its a good deal (ie: +NPV), but in practical terms CF determination is no more than your best reasonable guess. DCF as a range of values.

 

Hopefully, enough controversy to get things started!

 

Cheers

 

SD

 

 

 

 

 

 

 

 

 

[Partner24]

I have never used DCF to value businesses. In terms of valuation only, what I basicaly want is a high enough yield on free cash and profits generated by the company actually (on a trailing basis or on an normalized basis on insurance companies) and get a fairly high level of confidence that the cash generated will grow over a long enough period of time.

 

But I don't try to figure out what will be the cash that the business will generate in infinity. In infinity we're all dead and if we're not martians will attack us ;-)

 

And with Leucadia, how could I do that? So, in some cases, I use the equity that the business have, try to get it at a discount to a fair multiple to it and try to figure out by approximately how much the jockeys will be able to grow it over the next decade.

 

In each cases, I try to get a discount on the estimated fair value of the business and have some substainable growth prospects so that time will be my ally instead of being my enemy. With that, you get the best of both worlds: a return on narrowing the gap between price and fair value of the business AND a return on the increased fair value of the business.

 

 

[StubbleJumper]

If you are going to invest in anything, you really should make some effort to determine whether you are buying a 60-cent dollar or a 120-cent dollar.  To do this, you need to use some sort of valuation metric, be it a detailed DCF framework, or a more simple approaches like PE-normalized earnings framework, earnings yield+growth, price-to-book....whatever.  I have used all of these approaches in the past, and you'll note that I still haven't gotten rich.  ;)

 

DCF is a pleasant analytical activity if you've got lots of spare time to monkey around choosing growth rates, discount rates, terminal value and all of that great stuff.  However, in my experience you end up spending so much time fudging around with those parameters that you do not dedicate enough attention to thinking about the underlying business.  And, as we all know, the results can be so sensitive to the selection of those parameters that you can sit there farting around with your spreadsheet until you eventually convince yourself that there is value where none truly exists.

 

The other challenge with DCF is managing lumpy cashflows.  Just try doing a spreadsheet for FFH!  The folly of trying to select stable growth rates becomes immediately clear.

 

Nope, IMO, people are better off using rough metrics to assess value and spend more time thinking about the underlying business.

 

SJ

[scorpioncapital]

Can't you do DCF in your head, just use the rule of 72, that tells you what $1 in 3.5 years is worth today at 20% ($.50), it tells you how much a dollar today is worth in 3.5 years ($2) you can calculate anything pretty quickly, precision not necessary.

 

 

[calonego]

If one were a portfolio manager of a large fund, or due to their employment was required to have an opinion on or own a hundred securities or more, a DCF of model based structure may be the easiest way to get things done, with less risk of non-conformity (it's what the 'profession' of money management teaches). Similarly, one could rely on sell-side analysts and their models and "research" in a career managing a large mutual fund.

 

If one wants to make money, you should think like a businessperson, analyzing the business continually, based on the underlying economics and qualities. Purchases need to be made based on the economics, incentives within the organization, and valuation. If you can understand #1 & #2 and see it quickly, you can passively follow hundreds of good firms, #3 will jump out at you periodically (based on assets, CF generation, one-time issues, etc). You may also see a similar business that you don't follow, but you begin to understand it rather quickly due to your prior knowledge of similar firms. When it looks too good to be true and you start really trying to blow apart the seemingly good investment, but you can't, you gotta back up the truck.

 

I'm no pro, but that's my thoughts on the system Buffett et al seemingly use (tons of annuals get browsed).

 

There are (or 'were' I should say!) many people out there looking for a career in money management. Directing them to a life of entrepreneurial business ownership/investment may prove harmful as it can limit their employability. Few really good investors need employees in analysis, thats what they enjoy doing.

[Uccmal]

I see no value in this exercise whatsoever for what a value investor does.  I have tried it and decided it was a collosal waste of time. 

 

An easy challenge for DCF would be to get the 10ks of 5 "blue" chips from 10 years ago, which are available on the net, and calculate using forward assumptions.  Say GE, PFE, BAC, BRK, and KO.  Forward test if to see how it stacks up to the cash flow realities for 2008, or 2007. 

 

The irony is that doing it for FFH would reveal a fairly high return (from 1998) but no one could have stayed in the stock on the basis of DCF, given all that happened in between.  You had to thoroughly comprehend the companies' style to invest. 

 

The best way I have determined for investing so far is to get a general estimate of what a company is worth today, determine if it is significantly cheap by one metric or another, determine if it can remain a going concern, and get to know it as well as possible.

 

Getting to know a company well keeps me from getting sold out in downturns, or allows me to buy more.

 

Everyone always argues about whether WEB is a micro, macro, uses DCF, or whatever.  I think he is just very bright and loves business.  He has studied business 10 hours per day for 65 years and has a numerical aptitude to boot.  All of the sayings he has adopted or created such as looking for 7 footers on a basketball court, or looking for 1 foot hurdles in a footrace lead me to believe that his biggest skill lies in avoiding estimates such as DCF and sticking with facts and then adding in a requisite margin of safety. 

 

 

[calonego]

Well said Al!

The guy studies much more than 10hrs a day, seven days a week though ; )

 

To be clear on what I said earlier - if one wants a job in 'the biz', it's easier to get a job as an excel/DCF junkie, it's a career path.

If you wish to make money from investing (as Al and I do!) the best way is to think like a businessman.

[calonego]

One day I aspire to be a full member like you Al! My fluffy pointless posts help...

[arbitragr]

valuation depends on the company/industry.

many methods.

 

[calonego]

The largest problem is uncertainty.

A DCF is basically taking ownership of a company and comparing it to a bond CF. The problem is the CF isn't usually available, or it is lost and it clearly is not contractual (the case w/ a bond). If a company's equity is bond like, then maybe there is a case for a DCF, but that's only valid after a massive amount of analytical work on the business and and few businesses would qualify.

 

It's easier to argue the validity of a DCF on a lease or a contract.

Aggregated out (a business), it's rather difficult.

[SharperDingaan]

 

Thanks for your responses.

In fairness to the downtrodden... the case for DCF!

 

(1) It's a great training tool. It establishes a basic level of financial numeracy, & FS familiarity (you have to find the numbers in the BS and P&L). However ... the real lesson is recognizing its limitations, & why.

 

(2) It generates trading ideas

Extended periods of fixed growth assumptions exaggerate the peaks & troughs. They also collapse under their own weight, generating inflexion points. Guaranteed & repeatable volatility & momentum trading cycles.

 

(3) It makes all the $ look the same, & at least gives some sort of crude basis for comparisom. OK if applied only to the more extreme outliers, but don`t try to rank between close alternatives.

 

(4) It can expose the rough magnitude of a hidden Margin of Safety;

- FFH: FFH doesn`t discount its liabilities, but its competitors do. Recalculate the liabilities using the competitors discount rate, & the liability difference is a very sizeable off BS asset. Margin of Safety.

- Defined Benefit Plans: DB plans typically have ALM duration mismatches of between -11 & -16; ie an increase in interest rate will reduce the PV of the liability 11-16x faster than the asset value falls. Hence a DB plan, currently underwater, can be made whole again simply by increasing the discount rate - DCF gives you a way to calculate how many bp that increase needs to be. If you inflate the US economy by 350bp, does GM`s pension deficit suddenly become solvent again ... & all the fed has to do is print money.

 

On balance not a bad thing.

.... just keep in mind that it should be serving you, not the other way around!

 

Cheers

 

SD

 

 

[calonego]

Funny you mentioned that - there were recent posts about BMTC (CDN furniture), they have a DB plan and I recall them changing the '08 discount on the liabilities to 7% from 5% historically. I didn't mention it before, but it's not very cool (major difference).

 

The non-discounted liabilities on P&C are very attractive - seen some others like this, but economically they should be discounted, one should factor that in (it's either extra equity or the time value is extra free CF).