Garbage in, Garbage Out – How to Make a Completely Useless Financial Model

As previously posted on Breaking Bay Street…

Among the holy trinity of valuation methods (comps, precedents & DCF), only one is bold enough to claim the ability to determine value formulaically. This is of course the discounted cash flow or DCF method, which essentially says that a company is worth the present value of all its future cash flows.  In other words, set up your formula, plug in the numbers and presto, you have the value of the company. Professionals call the DCF an “intrinsic,” or “fundamental” valuation technique, because value is determined solely by the attributes of the company itself, without looking to the values of other companies for guidance.

Comps and precedents on the other hand work in exactly the opposite way. These are “relative” valuation techniques, which determine value simply by making reference to the values of other companies. In other words, if your company is worth $100 and my company is just like yours, then my company is probably also worth $100. Understanding why your company is worth $100 isn’t part of the process. That’s the DCF’s job.

This special ability to determine value independently makes the DCF a little sexier and indeed it occupies a special place in the hearts of those who work in the finance industry. Unfortunately, in practice, the mythical powers of the DCF are often polluted in ways that despite being obvious are seldom recognized by those that use them.

Take for example the concept of terminal value. In theory, when you build a DCF you are supposed to forecast the company’s free cash flows for eternity. In practice of course, estimating a company’s free cash flow in the year 2200 is a bit of a challenge. Consequently, we compromise by limiting our forecast to a more reasonable time period (usually 5-10 years) and using a shortcut to estimate the value of the company at the end of the forecast period.

The two shortcuts used are called the “terminal multiple” and the “Gordon growth method.” Without digging too far into the details, the terminal multiple method basically requires you to decide how much each $1 of earnings, EBITDA or other profitability metric is worth in the last year of the forecast. If this sounds a little fishy to you, you’re on to something. A DCF is supposed to be intrinsic; determined without reference to the valuation of other companies and of course, it’s impossible to come up with this type of guess without referencing other companies. Indeed, typically when this method is used the analyst will just use the comps to come up with the appropriate multiple. Although the Gordon growth model appears a little more scientific, the comps are typically used to sanity check the results, meaning that in the end, both techniques lead to same problem.

So now our “fundamental” valuation has a “relative” valuation component. How big of a deal is this? Well, it’s fairly common for more than 50% of a company’s valuation to be attributable to the terminal value and a contribution of 60%, 70% or even higher is not uncommon, especially if the company is expected to grow quickly beyond the forecast period. In other words, we are often relying on relative valuation techniques to drive the majority of our supposedly fundamental valuations.

The major problem with this is that it becomes very easy to make a DCF model into a self-fulfilling prophecy with no predictive power. We let our understanding of the comps influence our design of the model to the extent that the DCF becomes merely an elaborate pseudo-scientific justification of the comps valuation.

I’m not anti-DCF, but I do think that people should be more aware of the consequences that result from the typical practices used when building DCFs. Whenever you’re confronted with a DCF, remember that the majority of the value will often be determined by a single assumption, based entirely on the judgment of a single human being, with absolutely no means to evaluate its reasonableness. Determining value is never an easy task.

2 thoughts on “Garbage in, Garbage Out – How to Make a Completely Useless Financial Model

  • Excellent article. I wish I had read this when I first got into Real Estate; this dissonance between objective and subjective inputs confused me for a long time. What’s more is that no one seemed to know where exactly my confusion was coming from! I just thought “well, maybe the know more than me”.

    Bottom line even if you aren’t comping returns on similar assets you have to determine desired rates of return from either lenders needs/wants, the WACC, or what buyers allegedly want as a return at a given point in time in a specific market. These are all assumptions that require subjective adjustments on the scale of comps.

Leave a Reply

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