Some interesting thoughts there.90% of the science in valuation...

Some interesting thoughts there.

90% of the science in valuation is in the forecast future free cash flows. That's what understanding a business is all about. The discount rate, in the traditional sense derived from the CAPM model, is almost purely mechanistic (there is some flex with how beta is calculated, depending on an investor's universe, but it's of minor consequence).

I'd point out a few things about Cam's 2-year DCF approach to valuing high growth companies:
- The NPV is going to be hugely sensitive to terminal value assumptions. If you are doing only a short-range (<3 years) as opposed to long-term forecast (7+ years), you have to get the terminal value as close to right as possible. The problem with the "traditional" TV formula Cam has used is that it assumes, in the case of ARB, that a 6% growth rate can be achieved into perpetuity without any reinvestment into the business - i can arbitrarily assume a higher growth rate and get a big pop in the TV (this is what sell-side research analysts do). This is impossible (ARB are selling to mature markets, where nominal GDP growth trends around 4%) - a 3-4% terminal growth rate is more realistic. The better formula to use is the McKinsey Value Driver terminal formula, which takes into account the fact that, if you want to assume a company earns well above its ROIC into perpetuity (a brave assumption itself - ordinarily the cash flow forecast should go out far enough to the point where the company's future RONIC is roughly equivalent to its WACC), *and* it will have a certain growth rate, the growth rate can only be achieved by additional reinvestment into the business at that RONIC.
- You shouldn't arbitrarily add or subtract from the discount rate. It's a mechanistic thing that "is", as derived from the CAPM - adding things into it like "liquidity discounts", "premium management", "additional return hurdle" etc. is, strictly speaking, incorrect. There is some leeway to suggest you can express very specific types of risks (i.e. sovereign risk) in the discount rate, but aside from these specific examples, all the "risk" is supposed to be expressed in the cash flows. The manner in which risks are expressed in the cash flows is to run a monte carlo analysis in which multiple cash flow forecast scenarios are run, and probability weighted accordingly. It's hard in practice to do this, but it's the more correct way.

For mine, i spend 90% of my time trying to get a handle on future free cash flows, and running multiple cash flow scenarios. At a minimum, to prevent almost the entire value being expressed as a terminal value, i'd try to forecast out 5 years, but normally 7-10. I don't spend much time bothering to figure out the discount rate - i just backsolve the discount rate such that the NPV of FCFs equals trading price, and determine if i am happy receiving that level of return.

Cheers