Methodology — Greenwald Ladder (Reproduction Value → EPV → Franchise)

Three questions, stacked from most to least certain: what would the assets cost to replicate, what are the current earnings worth with zero growth, and how much of the price is franchise and growth on top? Display-only — the ladder never feeds a fair value, the screener upside, or any other served number (the cardinal rule holds by construction).

This is the public methodology. The approach and its source are named in full; the exact replication recoveries and normalization windows are proprietary calibration and deliberately not enumerated. Each leg in the product is labeled with what was replicated — and a suppressed leg always says why it was withheld.

The idea

Source. Bruce C. N. Greenwald et al., Value Investing: From Graham to Buffett and Beyond (Columbia Business School's value-investing canon). Greenwald's insight: a DCF stakes most of its answer on the least knowable inputs (growth and the terminal value). His ladder re-orders the valuation by reliability instead:

1. Reproduction value of the assets. The balance sheet restated at what a competent competitor would pay to replicate it: cash at face, receivables and inventory below face (collection risk; replication buys inputs at cost), plant at depreciated book, plus a capitalized research asset for the trailing R&D it would take to re-create the research program — minus all liabilities. No forecast anywhere in it.
2. Earnings-power value (EPV). Normalized sustainable operating profit — a multi-year median margin applied to today's revenue, so the cycle is smoothed but the scale is current — taxed at the trailing effective rate and capitalized as a zero-growth perpetuity at the same cost of capital our DCF uses, then bridged to equity by netting debt against cash. One assumption (sustainability), no growth story.
3. Franchise value — what the earnings-power value adds over the reproduction value. Positive only when the business earns more than its assets alone would — which requires something that stops competitors: a moat. A negative rung is reported as-is (the business earns less than its assets should).

On top of the ladder we report the growth premium: what our served headline fair value adds over the zero-growth EPV. That splits any price into asset value + franchise + growth — and tells you exactly which storey of the building is load-bearing.

Honesty rules

• Missing lines suppress, never zero-fill. A filer with no stored inventory or receivables line gets the reproduction leg suppressed with the missing line named — "not tagged" and "zero" are different facts. An untagged R&D line, by contrast, is a genuine zero (no research program was expensed), and is said so in a note.
• Financials suppress entirely. For a bank or insurer the balance sheet is the business; "replicating" a loan book with haircuts reads nothing. Detected by the same industry routing the fair-value engine uses.
• Same discount-rate world as the DCF. The cost of capital and effective tax reuse the engine's own treatment (including the live per-currency risk-free rate), so ladder and DCF disagree only where they should — on growth.
• Understatement is disclosed, not repaired. If fewer R&D years are stored than the research life, the research asset is understated and the payload says so; we never extrapolate a missing vintage.
• Display-only. Like the base-rate panel, the ladder is a reading of the same inputs — it cannot inflate a fair value because it never touches one.

Honest limits

• Reproduction value uses book-derived replication proxies, not appraisals — intangibles beyond R&D (brands, networks, licenses) are not replicated, so asset-light franchises will show large franchise rungs partly by construction.
• EPV assumes the current earnings power is sustainable; a structurally declining business is worth less than its EPV, and no zero-growth read can see that.
• The ladder is annual-filing granular and inherits every coverage limit of the underlying store (foreign filers, partial tagging).