The DCF is the Randy Watson of Valuation

I’m not very well-rounded.  I studied finance in college, and finance in business school.  And man could I turn out a perfect DCF.  I could make that model sing.  After joining Fidelity, I remember ramping up on my first sector, and after finishing, I showed one of the senior analysts my fancy-schmancy DCFs.

“Drew, what in the hell are you doing?”

I was crushed.  Here I had been staying up all night for weeks modelling every aspect of the P&L, cash flow statement, and balance sheet over the next 15 years (for all of my companies) - arriving at what must be the fair value of the stock - and this very smart, well-accomplished, senior guy (hi Sanjeev) basically tells me I am an idiot. [i]

Now, you have a choice here to react as I did initially.  You can think to yourself “oh, he only said that because he knows you won’t have time to do all that work” or “yeah, big institutional shops don’t do proper analysis” or “he just knows the PMs won’t have time to review your model and won’t be able to build any conviction in your ideas.”  Sure, some of those points may be true at some shops, but I finally figured out the bigger picture:

DCFs don’t work because they are laced with bias.

Don’t get me wrong, the DCF is the answer to the “what is it worth” question.  The DCF is perfect that way.  But the DCF is both good and terrible.  We are humans, humans build DCFs, and we have biases.   These biases manifest in the DCF, which itself is full of more moving parts than a steam locomotive.  If we get just one of those parts wrong, the error can compound on itself, and before we know it, our thesis has derailed.

If we have an inkling of excitement about a company or a management team (or indeed, an inkling of resentment), then the DCF is the ultimate camouflage machine; burying our biases and emotion behind annual changes in working capital and terminal growth rates.   As I mention above, I could make a DCF model sing; in fact, I could make it sing any tune I subconsciously wanted it to sing.

And even if we do it objectively (as I am sure nearly all of you DCFers think you are) we are only answering the value question.  The DCF doesn’t care about Mr. Market.  I suppose for some that has its advantages, but not for us.  We are public market investors.  We want to seize on opportunities where Mr. Market is making a mistake of overreaction or underreaction that will fix itself anytime within the next three years.  The DCF doesn’t help us do that.  

Additionally, to the creator of the DCF, it can never be wrong, and that is dangerous.  The commander of a DCF sits in the captain’s chair, presses auto-pilot, puts on the blindfolds, and falls asleep, hoping to wake up at the right destination.  Our view is in some cases, that blindfolded DCF captain leads us straight into the Bermuda Triangle of overconfidence, confirmation bias, and value traps.

Of course there is bias in any valuation methodology, and this will be controversial, but we contend that – because of these points above – a forward (not trailing, for goodness sakes) fundamental multiple (e.g. P/E, EV/EBITDA) or an SoP model may sometimes be a better proxy for an honest DCF valuation than a DCF is itself. [1]

And if the stock performs well over the long term, it will be because the company beat fundamental expectations.  That’s it.  

Moreover, and finally, the DCF can be just as inappropriate for the high flyers as it is for the potential value traps.  

In 2007, the most offensive, egregiously wild Amazon bull estimated that 13 years from then (in 2020) Amazon was going to print nearly $40 billion of revenues. [2]  

Today, 12 years later, the average sell-side analyst thinks Amazon will do over $320 billion of revenues in 2020.  

Amazon has been a great stock, and a big winner for the bulls; but not because some analyst and his or her DCF figured all this out 12 years ago.  Amazon won because they crushed expectations.

On the other end of the spectrum, potential value traps end up not becoming value traps because they turned the fundamentals around.  They started beating expectations.  The narrative changed.  

Fundamentals matter.  Narratives matter.  DCFs don’t help with either, and can lead us astray in our pursuit of the truth - or at least the truth in all of our time horizons.