The ability to create an accurate and reliable discounted cash flow (DCF) model was once the hallmark of a powerful and well-paid financial analyst. Now anyone who brags about their ability to create a DCF model is laughed out the room.
What changed?
In a word, computers. With automation and better tools, the difficulties in the quantitative aspects of DCFs have disappeared, and the qualitative backdrops of DCFs, which were ignored for so very long in financial history, have become the defining characteristic that makes one analyst more valuable than another.
To understand this change and what it means for the skills you need to focus on, let’s talk briefly about the way DCF models are constructed.
To create a DCF model, an analyst needs to begin with the Net Present Value (NPV) of the company’s unlevered free cash flow. To do this, we first need to look at the company’s earnings before interest and tax (EBIT), its capital expenditures, depreciation costs, and changes in working capital. If we subtract those expenses to the EBIT, we get the unlevered free cash flow, which can then be used to project future earnings. Note that all of these metrics can easily be found in a company’s balance sheet, and public companies report these figures on their 10-Qs and 10-Ks.
With this, we then want to forecast expenses by either assuming a rate of revenue growth in the future without varying our expectations for expenses, or we want to assume a rate of earnings growth based on changes to both revenue and expenses.
Which one is the “right” approach? The answer: it depends. There are a lot of evaluative criteria that can be behind the choice to go with revenue or earnings as your model’s guiding approach, and the keen analyst will know when to use one and when to use the other. They will also think out fully the implications of which choice they make before deciding on which approach for their DCF.
Once these judgments have been made, things get very mechanical. One then takes a look at capital assets and structure to compute the terminal value of your DCF model. This is important, because you need to set a goal at some point in the future which correlates to the end of your model; in other words, your DCF has an end point, and the terminal value tells anyone seeing the model that that is the point when you have stopped effectively analyzing the company.
Following the terminal value, one can then use WACC to create the discount rate by which you will discount the cash flow—that gets you to the enterprise value of the company, from which your future values are determined.
Note how much of this process depends on an algorithmic approach to putting in numbers and spitting out results—but it all is based on a major qualitative assumption: the rate of revenue/expenses growth. The end result of this process can vary radically based on that assumption. It’s the assumption that truly matters above all else.
While financial professionals need to know the mechanical process of creating a DCF, the value of that skill has been deprecated substantially in the last few decades. There are only DCF calculators and templates that you can download for free. They’re all very similar, which is unsurprising; it’s a very mechanical process, so there isn’t much room for differing opinions.
So what makes the analyst special? It isn’t her knowledge of the DCF model or his ability to make one. It’s her ability to create accurate, predictive models by accurately predicting future revenue, expenses, and/or earnings growth. When analyzing sell-side analysts, many funds will focus on their ability to predict revenue while wholly ignoring the DCFs themselves since, well, any computer can recreate a DCF. That’s not what’s important. What’s really important is: how did the analyst predict growth and how right or wrong was the analyst’s predictions?
So ask yourself: how can you predict growth better than anyone else?