DCF in spirit, not letter

"It was one of those cases where you approve the broad, general principle of an idea but can't help being in a bit of a twitter at the prospect of putting it into practical effect. I explained this to Jeeves, and he said much the same thing had bothered Hamlet." - P.G. Wodehouse

I don’t DCF. Buggy humans, including yours truly, are disastrous forecasters. So, I can’t populate numerator. Reducing risk to a number is an abomination. So, denominator is also out. I’m left with a row of fraction bars separated by plus signs (or lately, minus signs).

Sure, there are variants. Reverse DCF tries to figure out what’s baked into the price. When such an exercise is attempted, answer is usually known beforehand (“hope”). Sensitivity analysis can map a range of outcomes to assumptions. Again, range is guessable without fancy XL (“3 scenarios: bull-case, anything goes, Tiger Global”). This sort of exercise only confirms DCF’s real purpose as rationalizing, not reasoning. And I haven’t even got to the two funniest words in English language: terminal value. DCF is self-delusion with decimal points.

Now, here’s the problem. Those who DCF look down on people like me who don’t. It’s CBSE looking down at state-board all over again. Articles abound on pitfalls of PE multiples or how EBITDA is for L-boards (latter is true, though). While I am used to being state-board, I cringe upon seeing nonsense like 8.97% cost of capital. This sort of thing reveals more about author than about topic. In messy world, with others’ hard-earned money at stake, fake precision can do real damage. As this essay series is about a way of thinking, I wanted to articulate how I think about DCF, in spirit and in letter.

In spirit, I buy into DCF. More cash a business is likely to generate, more it’s worth. Lower the riskiness of that cash, more it’s worth. More conviction I can build around cash and risk, more it’s worth. Someone acquiring an entire business is likely to think on similar lines. So, the general idea of linking valuation to predictability, cash and risk is fine.

Problem arises in going from general to specific. There’s the practical problem of populating a model with unknowable parameters and fiddling with assumptions till it produces a convenient result. There’s also a philosophical problem. Science, even faltu social science, has to be empirical. I can’t stop at saying DCF is theoretically correct. DCF-produced valuations have to be consistent with observed valuations, across decades, markets and companies, at least on average. Inputs into DCF have to be mapped with observed outputs, showing consistent patterns.

Such mapping of theory to practice hasn’t been done till date, in any robust sense. I don’t know if it’s doable. DCF lives in one world, and empirical valuation references in another. In fact, one crucial DCF input (Beta) has been proven to be empirically invalid. Not one soul has generated superior returns by buying more squiggly stocks in history of humanity. Many have gone broke doing so. Should projected cashflow in year-14 be discounted precisely by some number raised to 14th power? If you stop to think, this sort of thing is a leap of faith. Time value of money hasn’t been quantified in any reliable sense. All we know is that buggy humans are notoriously inconsistent about this sort of thing. On any first-principles basis, DCF’s edifice is flimsy. DCF is the sort of thing that is accepted because it’s easier to pretend than question.

Where does that leave me? I focus on principles behind DCF rather than mechanics of it. Essence over XL. Spirit over letter. While I don’t do DCF, I obsess over three aspects that underpin any DCF: predictability, cash, risk. Price I pay in every investment is entirely dependent on these DCF ingredients, even as I pour scorn on DCF charade.

Predictability

World is messy but not uniformly so. There are pockets that show a sense of steadiness, if not stability. Industry structures hold. Pace of change is modest. Competitive advantages are real. Competitive positions stay stable. Basis of competition is established and rational. Above-par economics sustain, for reasons that seem understandable. While there will be ebbs and flows, general trajectory of business is decipherable. Certain businesses are highly likely to emerge stronger and bigger over time, even if I can’t be sure of exactly how big. Rather than explicitly predict future, I implicitly limit myself to these relatively predictable pockets. A general sense of such businesses continuing to do fine offers better odds of getting the future right than (decimal) point forecasts.

Return on capital = cash

In my early investing years, I focused on nonsensical metrics: growth, margins, TAM, credentials. Hard knocks and an apprenticeship later, I reformed. If there’s one metric I obsess over, it’s Return on Capital. Its main use is as a foolproof measure of good but fuzzy attributes: management quality, industry attractiveness, company track record, competitive position. A side benefit is that algebra of return on capital helps solve our DCF conundrum. A business with 30% ROE growing 10% a year is left with two-thirds of profit as surplus cash. A 15% ROE business is only left with one third. If I choose former over latter, I get 2X cash flow per unit of profit. Coincidentally, that’s the exact comparison between businesses I own and average business in market. PE-multiple method may be state-board but my E is ISCE. Take that, CBSE. PE-multiple discussions focus on the multiple without fully appreciating that some earnings are better than others, on both predictability and cash flow.

(As an aside, I only use real, delivered earnings that are representative of earnings power over a cycle. No imaginary FY30E fantasy earnings that’s worse than DCF. )

Risk, risk, risk

Regular readers would be sick of my lectures on risk. Risk first. Risk is everything. Avoid ruin. Most businesses are buy-at-no-price. Control risk, return is a mere consequence. Margin of safety as an overarching mindset, not a valuation check. I won’t repeat that entire spiel. End-goal of risk obsession is to ensure that a business that is good stays good long after I invest. Industry nature, competitive position, balance sheet strength, business economics and managerial trust should hold over time. At the least, very few in the portfolio should turn crappy. While there’s no guarantee how future will pan out, hit rate on this front has been decent so far.

Bringing it all together

In valuation, like in everything else, I want to achieve two goals. (1) Be roughly right. (2) Err on safe side. If I manage (1) and (2) often enough, I’ll end up with sound decisions and acceptable outcomes. I’ll have mistakes but not disasters. Between messy world and my frailty, that’s as good as it gets.

Above approach, of adopting DCF in spirit, is my surest path to achieving these goals. I have predictability of cashflows, without futility of predictions. I minimize risk of deterioration of cashflows by ensuring safety at many levels. Even as I use valuation methods that tie into earnings, I don’t lose sight of cashflow. In fact, I ensure superior cashflow per unit of earnings. While I don’t reduce risk to a number, I have a construct that deserves a lower discount rate than most. Every ingredient of a DCF is in place.

So, why not just do a DCF? Being scientific is about valuing observation over theory. This is all the more crucial in social sciences that aren’t underpinned by natural law or universal axioms. There is no theory. Only broad patterns inferred from carefully observing real world. 8.97% cost of capital for some iffy business has no basis in reality. However, there is empirical basis for how a wide swathe of businesses have been valued over decades across markets. It is possible to check how an average business is valued over time. Or what kinds of businesses tend to be valued higher. Or what sort of valuations are untenable even for better businesses. What strategic acquirers pay. Multiples (not necessarily on earnings) are merely a way to covert this historical encyclopaedia on business valuations into a usable form.

In my Sound Judgment series, I wrote about importance of a historically-grounded frame of reference. Using multiples over DCF is a manifestation of this principle. Aim is to be empirically sound, even if it appears theoretically unsound. This frame of reference also helps with my goal of playing it safe. Despite sticking to above-par businesses (on predictability, safety, cashflow), I stay within a range that’s justified by long history of all businesses. I try to buy 90th percentile businesses way below 90th percentile valuations. I define red lines, to protect me from worst of myself. I can stay internally consistent across businesses and over time. After a while, I can introspect to tweak my rules of thumb.

Unlike in textbooks, real-world methods are about practicality and efficacy. Soundness of decisions matter, not elegance of method. An empirically-grounded, conservative, multiple-based approach has worked well over the years. While I haven’t ever done a DCF, incorporating key ingredients of DCF into each of my investment decisions has made them sounder. I adhere to DCF in spirit, not in letter.

I started with a quote from one genius. Let me end with an apocryphal quote, attributed to another.

“In theory, theory and practice are the same. In practice, they are not.” – Albert Einstein