The Discounted Cash Flow & Margin of Safety
And how I screwed it up
In a previous note, I described how I mis-used the Margin of Safety (MOS) concept in the past. Today, I’d like to talk about a common value framework and how I’ve botched using it over the years:
The Discounted Cash Flow (DCF)
John Burr Williams in his book The Theory of Investment Value is often credited as the originator of the concept of intrinsic value of a company. The intrinsic value is the cash flows that can be taken out of a business over its remaining life, discounted at an appropriate rate. Thus was born the DCF or Discounted Cash Flow analysis. For some, this is the Holy Grail of investing truth. For others it is a complete waste of time. For me… let’s say the results have been mixed.
In theory, it all makes sense. How much cash will this black box spit out over time? And how much should I pay for it? Of course, we should also discount what we would pay to obtain a MOS. For years I slavishly built cash flow models of companies to tease out the magic cash flows and determine an appropriate intrinsic value based on those cash flows. It certainly helped me learn more about what makes companies tick in general. I’m not sure it provided the investment insight that many claim. Here is a non-exhaustive list of the short comings of this approach:
What discount rate? This is the first hurdle you encounter. Should you have a higher discount rate than the risk-free rate (US Treasuries)? Should it be higher for big companies vs. small companies? Should it be higher for risky companies v. stable companies? There are many, many debates in financial circles on appropriate discount rates.
What are the company’s cash flows? If you look at a given company’s past cash flows, you will see wide variation – depending of course on the company. Or maybe it is a young company, and it doesn’t have any history. Many people develop detailed spreadsheets with revenue growth, variable and fixed expense categories, tax rates, etc. to get to a company’s operating or free cash flow. These are assumptions on top of assumptions multiplied by assumptions. And if you tweak any given assumption, you will get a different answer. Sometimes dramatically different.
What is the terminal value? A common way to approach this task is to model a “two stage” DCF. You have a period where you estimate the explicit cash flows. Then, you have a “terminal value” or a fixed value for the remainder of the company’s history. Based on some simplifying assumptions like a multiple of some imagined steady state cash flow. This terminal value often ends up as a large percentage of the intrinsic value (depending on yet other assumptions like discount rate!)
When I first started investing this way, I made detailed models – sometimes even by quarter. I mean, it’s just more columns in the spreadsheet, right? Why not!
In retrospect, the models were crap. Many things I invested in with a DCF analysis continued to drop in price despite my prowess in Excel. I remember one of my attempts to use a DCF approach was with a company called Coach. They make handbags and accessories for women and men (now they are part of a company called Tapestry). I dutifully projected out margins, tax rates, revenue growth (base on even more assumptions around store openings) and share count growth over time. By quarter. For 10 years. I found that the current price offered a MARGIN OF SAFETY, so I bought some. The next quarter, they were wildly off my model’s assumption for cash flow. Then the next. And the next. And not just a little bit, but 20-50%. It wasn’t any better by year. I think I owned it for like 3 years and I doubt that any of the yearly cash flows were closer than +/- 30% of my original estimates!
The funny thing is, Coach ended up being a successful investment for me. Let’s be clear, a successful investment for me is one where I have a positive return. Did I beat the market with Coach? I have no recollection, but I made a bit of money, and I learned a lot about the utility of discounted cash flow analysis.
My solution: Use a fixed 10% discount rate for all investments regardless of size or risk.
Your discount rate should equal your minimum hurdle rate for what you want to earn as an investor.
The only form of DCF I'd use is the simple Gordon Growth Model: Price = Dividend Yield / (Discount Rate % - Growth Rate %)
Or you can swap out Dividend Yield for Earnings Yield.
The fewer the assumptions the better.
I 100% agree with you. DCF has so many flaws and is in many ways a guessing game. Why should your projections be the ones that are right, and not those of the guy selling to you? It also has the tendency to compound errors as you go further into the future.
What’s more, a 1% difference in the discount and growth rate can lead to values that differ by as much as 100%!!! (search up Greenwald’s example).
I think that’s what makes investing beautiful — the idea that there is no perfect formula or math equation that you can use to get your answer. The future is, by definition, uncertain so trying to project cash flows 10 or even just 5 years out is a highly speculative process. Thanks for the piece!