I know I’ll probably sound a bit defensive here, but let me explain why I wrote such a long post.
-
I am uncomfortable with “pick a hurdle rate and just go with it.” For me, I feel a required return must vary with the amount of risk the investment presents, or it’s just a purely subjective or random number thrown into a calcuation and people start thinking “that’s how you do valuation.” And at worst, it can lure you into buying risky stuff (priced for a high rate of return) that is overvalued (because it passes the hurdle rate, but ought to be priced to deliver even more).
-
What I like about the index model (which produces the same security market line as CAPM) is that it has simple, believable asusmptions (unlike CAPM) and adjusts for (admittedly, only one kind, but still a major kind) of risk. The market factor, even if not perfect, tends to explain more variation in returns than almost anything else. So even if it’s not complete in and of itself, you lose a lot by not including it, I think.
-
Most people who love the single hurdle rate approach see that the index model looks like CAPM and fire back - “yeah, but CAPM doesn’t work,” and call it a day. My take is that CAPM doesn’t work perfectly, but at least it varies with the amount of market risk, so you can compare whether you did better with your stock pick than you would by levering up or down a passive market portfolio, which is simpler. A single hurdle rate is an even noiser estimate of required return than one that uses a regression line, unless the regression is not significant (i.e. F-test not significant).
-
Palantir did point out that you can do the risk adjustment by using very conservative cash flow estimates that account for risk there, and then say you need to pass a hurdle rate after that. Although I don’t do it like that personally because of the kind of data I work with, there is risk adjusting going on in an early phase of valuation, and so that’s a reasonable way to do it that does allow a single hurdle rate, because it addresses risk adjustments.
What’s happening in Palantir’s version (consciously or not) is that you’re trying to standardize the cash flows so that they reflect the same amount of risk (i.e. cash flow estimates for different companies analyzed are equally likely to be met or exceeded, no matter what company it is). Then you can use a single discount rate reflecting how much compensation you need to assume that amount of risk. You can interpret that required return as either a risk premium or a required margin of safety.