Reverse DCF

Everything so far runs the machine forward: assumptions in, value out, then compare with the price. The forward run has a known weakness. Your assumptions might be the wrong ones. And the argument "my DCF says ₹3,570, the market says ₹11,034, therefore the market is wrong by 3×" has a certain career-limiting arrogance to it. So flip the machine. That's expectations investing, Rappaport and Mauboussin's framing and the most useful idea in this phase. Stop treating the market price as a mystery to correct, and read it as a message to decode. Hold the machine fixed, make the market price the output you have to hit, and solve for the set of assumptions that produces it. Now the question changes. "What do I think it's worth?" is a question about your crystal ball. "Is the future the price requires more or less demanding than the future the evidence supports?" is a question about base rates, and that one you can actually answer. When you have an edge, it's in that gap.

How do you run a reverse DCF?

The five-step method. The Phase-8 deep dive uses this same sequence as Day 2, step 3:

1. Hold the machine. Freeze everything you have independent evidence for. That's the WACC with its audit trail, the tax rate, the terminal rules (g ≤ Rf, RR = g/ROIC_T, faded ROIC_T), and margins and sales-to-capital at delivered-history levels. You're granting the market every structural assumption, so the solve isolates the one that's contested.
2. Choose the solve lever. It's usually near-term revenue growth, the market's favorite dream. Sometimes it's margin, for turnaround stories. For quality compounders it's the duration of the fade, and you'll see why in a moment. The discount rate itself comes last, as a diagnostic.
3. Solve by bracketing. You don't need goal-seek, and you must be able to do this on paper so you aren't dependent on AI. Run the forward machine at a low guess and a high guess, then interpolate and refine. Value is monotonic in growth, so three or four iterations land within a decimal. In a spreadsheet, Goal Seek does it in one keystroke, once you can do it by hand.
4. Translate the solution into operational reality. "17% for ten years" is an abstraction. Compare it with "revenue of ₹1,14,000 cr by 2036, roughly 5.7× today, more than the entire current industry", which is a claim you can test. Convert implied growth into year-10 revenue, market share, store counts, subscribers: units someone could actually count.
5. Judge against base rates. Don't ask whether the story feels plausible. Every story feels plausible, since that's what stories are for. Ask instead: "of all companies that ever stood where this one stands, what fraction delivered what the price requires?" Mauboussin's The Base Rate Book (free PDF, Credit Suisse/Counterpoint Global) tabulates this for US companies. Among large companies, with revenue > $50bn in today's terms, the median 10-year real-terms sales CAGR is low-single-digit. Roughly one in twenty sustained >10%, and essentially none sustained >20%. Treat that as the illustrative shape, then pull the actual tables and quote their real figures. For India, build your own. Export 10-year revenue CAGRs for the NSE-500 from screener.in and count. You'll find nominal GDP at ~10 to 11% is the center of gravity, and the >15%-for-a-decade club is a short, famous list. The verdict has a name, the expectations gap: the difference between implied expectations and the ones base rates support. It can run in your favor, when the price implies less than the evidence supports, which makes it a potential buy. Or it can run against you, when the price implies more. Then you pass, or you hunt for what the market knows that you don't.

Two full reverse DCFs, one Indian and one US, follow in the worked examples. They're the centerpiece of the week, and the second act of each holds the deepest lesson. For premium franchises, the market's implied bet is usually about the duration of excess returns (the competitive-advantage period), not next year's growth. So crude "the P/E is high" analysis systematically misjudges quality companies, in both directions. Remember the steady-state anchor. A steady state supports ~15×, and everything above it is priced duration.

What growth does a multiple imply?

The quick version: inverting a multiple. For a teardown (question 11 of the twelve) you won't build ten rows. You'll invert the steady-state formula directly. Start from EV/NOPAT₁ = (1 − g/ROIC)/(WACC − g). Given the traded multiple and a defensible ROIC and WACC, you can solve it without a quadratic. Multiply out and isolate g. Take an Indian FMCG at 60× NOPAT with WACC 10.2%, and grant it ROIC = 25% forever: (1 − 4g)/(0.102 − g) = 60 → 1 − 4g = 6.12 − 60g → 56g = 5.12 → g ≈ 9.1%, forever. That's double the perpetuity cap. You can say the conclusion in one breath. A 60× multiple can't be a steady state under any defensible parameters, so the price must embed a long super-normal phase, and your job is to find out how long and test it against base rates. Sixty seconds, no spreadsheet, and you already know what the price believes. At 25×, 21g = 1.55, so the implied perpetual g ≈ 7.4%, still above the cap. Even "reasonable-looking" multiples in a 10%-WACC economy embed super-normal phases. Relative valuation trains that calibrated eye.

What explains a price your base case can't reach?

The three-lever diagnosis. If base-case growth can't reach a price, one of three levers explains it, or a blend of them, and the skill is naming which.

• (i) Growth or margins beyond base rates. The market sees an operational future you don't.
• (ii) Duration. The market grants a 15 to 25 year competitive-advantage period where your machine granted 10. This, far more than near-term growth, is usually what India's premium multiples and every "quality compounder" debate are about. It's a claim about moats, and Phase 4 is where you learn to audit it.
• (iii) Required return. The marginal buyer is happy with 6 to 7% where your WACC charges 10 to 11. That's common late in bull markets, in bond-substitute franchises, and in crowded quality baskets. It's a claim about the price of patience, and it mean-reverts.

The gift of the reverse DCF is that it turns a shouting match ("overvalued!" against "quality is priceless!") into a statement you can measure: this price = this growth × this duration ÷ this required return; pick your poison and defend it.

How should you use a reverse DCF in practice?

Process notes. Run the reverse DCF before you fall in love. It vaccinates you against plug-to-price, because you can't accidentally tune toward a target you've already decoded. Re-run it on every big price move. "What expectation just got repriced?" is a better question than "why is it down 9%?" And remember the asymmetric use Rappaport emphasized. To win, you don't need a better forecast than the market. You need to recognize when the implied forecast is far outside the base-rate distribution, in either direction. Most prices, most of the time, imply roughly reasonable things. So the tool's usual output is "fairly priced, pass," and that's a feature. It's how the tool tells you where you have no edge.