How to Value a Stock: DCF, DDM, REIT, Airline & Every Method Explained

A note on authorship: The research, analysis, and opinions in this article are the author's own. Claude (Anthropic's AI) assisted with drafting and editing the prose.

The value of a business is the cash it will return to its owners over its life, expressed in today's money. That sentence is the whole idea behind intrinsic valuation. Everything else — the spreadsheets, the discount rates, the terminal values — is machinery for turning a stream of future cash into a single number you can compare against today's share price.

A single formula does not fit every business, though. A bank's "debt" is customer deposits, not a funding cost. A REIT's reported cash flow is distorted by depreciation that does not reflect real economic wear. An airline's earnings swing so hard with the fuel-price and travel-demand cycle that one trailing year says very little. KashVector's evaluation tool picks a model automatically based on what the company actually is — no manual toggle — and this article walks through every one of them: Discounted Cash Flow (DCF) for most businesses, the Dividend Discount Model (DDM) and Excess Return Model for banks/insurers and regulated utilities, dedicated REIT and Airline multiple models, and two adjustments layered on top of whichever model applies — a Cyclical normalization for commodity businesses and a Holding-Company discount for conglomerates.

How the tool chooses a method

Before any numbers are crunched, the tool decides which model applies — automatically, with no manual toggle. It checks the company's sector and industry against four special cases in order; whichever matches first wins, and anything left over gets ordinary DCF. Two more adjustments — for cyclical companies and conglomerates — are layered on afterward regardless of which model above was picked.

Method 1: Discounted Cash Flow

DCF answers one question: if I owned this whole business, what is the cash it will hand me worth today? It works in five steps, and the tool exposes the inputs for each.

1. Start from free cash flow. Free cash flow (FCF) is the cash left after the business has paid its running costs and reinvested to keep operating — the money genuinely available to owners. The tool seeds this from Yahoo Finance. Where reported FCF is missing, it derives it as operating cash flow minus capital expenditure (the real figure, not a rule of thumb); only if capital expenditure is also unavailable does it fall back to operating cash flow times 0.8 — a rough proxy that assumes capital spending is about 20% of operating cash flow, and which can overstate FCF for very capital-intensive businesses like miners and utilities. Whichever path is used is labelled next to the input, so you always know whether the figure is reported, derived, or a rough estimate you should replace.

2. Grow it over a forecast period. You set a growth rate, and the tool projects FCF forward over a projection window (five years by default, adjustable from three to ten). The growth rate is seeded from the company's recent revenue and earnings growth, but this is the first number you should sanity-check: no business compounds at 25% indefinitely.

3. Discount each year back to today. A dollar in five years is worth less than a dollar now, so each year's projected cash flow is divided by (1 + discount rate) raised to the power of the year. The discount rate here is WACC, covered in the next section.

4. Add a terminal value. The business does not stop at year five. Terminal value captures everything after the forecast as a perpetuity growing at a terminal growth rate — 2.5% by default. This single figure typically accounts for 60–80% of the total, which is why the terminal growth assumption deserves real scrutiny. It should never exceed long-run economic growth (roughly 2–3%); a business cannot outgrow the economy forever.

5. Adjust for debt, then divide by shares. Subtracting net debt (total debt minus cash) from the firm value gives the equity value; dividing by shares outstanding gives intrinsic value per share. That is the number you compare against today's price.

Cyclical companies: normalizing the base year

Step 1 above has a real weakness for mining, oil & gas, auto manufacturing, chemicals, and semiconductor companies: a single trailing year of free cash flow can sit at a cycle peak or trough, and projecting five years of growth off that one number either wildly overstates or understates the business. The tool detects these sectors and replaces the trailing-year FCF with a normalized figure instead: it averages the FCF margin (free cash flow ÷ revenue) across every real fiscal year Yahoo Finance returns — typically around four — and applies that average margin to trailing revenue. Everything downstream (growth, discount rate, terminal value) runs unchanged; only the starting FCF input is different.

This is a real, disclosed limitation, not a hidden one: a genuine commodity cycle usually runs five to ten years, and four years of history rarely spans a full cycle. When the tool applies this normalization it labels the input as such and states the number of years and the margin actually used, so you can judge for yourself whether that window looks representative.

The discount rate, and the numbers behind it

The discount rate is where most of the judgement lives. The tool builds it as a Weighted Average Cost of Capital (WACC) — a blend of what equity investors require and what lenders charge, weighted by how much of each the company uses.

Cost of equity comes from the Capital Asset Pricing Model (CAPM):

Cost of equity = Risk-free rate + Beta × Equity risk premium

The risk-free rate is what you would earn with near-certainty, so the tool uses the local 10-year government bond yield and fetches it live by market:

The equity risk premium is the extra return investors demand for holding shares over bonds. The 5.5% US default is the long-run historical figure from Professor Aswath Damodaran's data. Developed markets sit around 4–6%; emerging markets like India carry 7–9% for the additional political and currency risk. This input is more powerful than it looks: a one-percentage-point change in the premium can move the valuation by 15–25%.

Beta measures how much the stock moves relative to its market, scaling the risk premium. Here there is a data trap worth knowing. Yahoo Finance computes beta against the S&P 500, which is correct for US stocks but understates risk for everything else — an Australian or Indian company does not move in lockstep with the US market, so its measured beta comes out artificially low (a global insurer showing 0.17, for instance). To compensate, the tool applies a minimum beta when seeding the discount rate: 0.5 for ASX and most international stocks, 0.6 for India, and none for US stocks where Yahoo's figure is sound. It tells you when the floor has been applied, and you can override it with a local-market beta if you have one.

Cost of debt is estimated from a synthetic credit rating — the same shorthand professional analysts use when a company's bonds aren't liquid enough to have an observable market yield. The tool computes an interest coverage ratio (EBIT ÷ interest expense) and maps it to a rating band from AAA down to D, each carrying its own credit spread over the risk-free rate: a strongly covered borrower rated AAA adds only a small spread, a highly levered one adds much more. When EBIT isn't available, it falls back to a simpler book-yield estimate — interest expense divided by total debt, plus a small 0.5% spread to approximate today's borrowing cost rather than the rate on old debt — and finally to a flat default if neither figure exists. Tax comes from the income statement (defaulting to 25% if unavailable), because interest is tax-deductible and lowers the effective cost of debt. The two costs are then weighted by the company's equity and debt to give WACC.

A note for Australian investors: franking credits

Australia's dividend imputation system passes the company tax already paid to shareholders as franking credits, which reduce their personal tax. In a professional valuation this is captured by a variable called gamma, which the regulator (the ACCC) values at 0.4–0.5; including it lowers the effective tax rate and trims WACC by roughly 0.3–0.8%. For ASX stocks the tool explains this but does not apply it, staying on standard post-tax WACC — the conservative choice most individual investors make.

The assumptions that move the answer most

A DCF is only as good as its inputs, and three of them do most of the work:

Terminal growth rate. Because terminal value is the majority of the result, this is the single most sensitive input. Keep it at or below long-run GDP growth (2–3%). A terminal rate of 4% or 5% quietly assumes the company outgrows the entire economy forever, which inflates the valuation into fiction.

Equity risk premium. As above, a one-point change flows straight into the discount rate and swings the answer 15–25%. Match it to the market rather than leaving the US default on a foreign stock.

Growth rate. The forecast growth is seeded from history, but history is not destiny. Fast growth fades as companies scale; cyclical businesses show flattering growth at the top of their cycle. Adjust it to what you actually believe.

This sensitivity is the reason the tool runs three scenarios rather than printing one number — more on that below.

Method 2: The Dividend Discount Model

For most companies DCF works well. For banks and insurers it breaks down, and not because of a quirk you can patch — the breakage is structural.

For a bank, "debt" is mostly customer deposits, which are a raw material the business runs on rather than a funding cost you can subtract. There is no clean free cash flow, and the standard capital-structure WACC has no sensible meaning. For an insurer, reported cash flow is dominated by investment-portfolio movements and changes in reserves, not the owner earnings a DCF assumes. Running a DCF on either tends to produce numbers that are confidently wrong.

The Dividend Discount Model sidesteps the problem by valuing the business from what it actually pays its owners: dividends. The tool detects banks and insurers automatically and switches to it. The model is mechanically similar to DCF — project, discount, add a terminal value — but with two differences:

• It projects dividends per share rather than company-wide free cash flow, so the result is already per share — there is no net-debt or share-count step.
• It discounts at the cost of equity alone (the CAPM figure above), with no debt component, because the question is purely what the equity is worth to a shareholder.

The terminal value uses the Gordon Growth Model: the final dividend grown by a long-term rate, divided by the cost of equity minus that growth rate. Because that denominator can be small, the cost-of-equity and growth assumptions matter even more here than in a DCF — which is exactly why the non-US beta floor exists. Without it, a low Yahoo beta produces a cost of equity barely above the growth rate, and the model returns a wildly inflated value.

The tool also shows the metrics that actually matter for these businesses — price-to-book, return on equity, price-to-earnings, and dividend yield — as a cross-check. If a bank or insurer pays no dividend, the DDM has nothing to work with, so the tool shows those multiples and no intrinsic value rather than inventing one.

When DDM understates the value: the Excess Return Model

DDM only sees the dividend actually paid out. A bank that retains most of its earnings and reinvests them at a return above its cost of equity is creating real value DDM can't capture — it would show a low intrinsic value simply because the payout is low, even though the retained earnings are compounding at an attractive rate. For these companies the tool switches to the Excess Return Model instead:

Value of equity = Book value of equity + the present value of (ROE − Cost of equity) × book value, for each projected year

Each year's "excess return" — the return earned above what shareholders require — is discounted back at the cost of equity and added to today's book value. Deliberately, there is no terminal value: a permanent excess return would mean a bank earns above its cost of equity forever, which the model treats as an assumption to state explicitly for a handful of projected years, not to bake in indefinitely. The choice between DDM and Excess Return is made automatically by comparing the actual payout ratio against what would be sustainable given the bank's ROE and growth rate — a payout below roughly half of that sustainable level routes to Excess Return; anything higher stays on DDM.

The same formula, for the same underlying reason, is also how the tool values regulated utilities (electric, gas, water, and diversified utilities under direct rate regulation). A regulator sets how much a utility is allowed to earn on its asset base — the same "earn a return on book value" shape as a bank's economics — which makes ordinary FCF-based DCF a poor fit for the same reason it's a poor fit for banks. Merchant generators and renewable-only utilities, which sell into competitive markets rather than earning a regulated return, are explicitly excluded and stay on ordinary DCF.

Method 3: REIT — a Multiple, Not a Discounted Model

Property REITs (real estate investment trusts) have a specific accounting problem: GAAP depreciation is a large, non-cash expense that reduces reported net income and free cash flow even though a well-maintained building often holds or gains value over time. Running a standard DCF on a REIT's reported FCF systematically understates it.

The real-estate industry's own answer to this is FFO (Funds From Operations) and AFFO (Adjusted FFO), and the tool uses the same convention:

FFO = Net income + Depreciation & amortization
AFFO = FFO − Capital expenditure

AFFO per share is then multiplied directly by a sector-typical P/AFFO multiple — no discounting, no growth projection, no terminal value, the same shape as a P/E-based valuation. The multiple varies by property type, since different REIT sectors trade at structurally different multiples in practice:

Two things worth knowing: FFO deliberately omits gains on asset sales, because Yahoo's income statement doesn't break that figure out separately — a documented, usually-minor understatement, not a silent gap. And AFFO uses total reported capex as a proxy for "recurring" capex, since the recurring-vs-growth split is a REIT-specific disclosure no financial-data API exposes — this understates AFFO for a REIT that's actively expanding, a conservative bias rather than an inflated one. Mortgage REITs (which hold loans, not property) are excluded from this model entirely and valued as financials instead, via DDM or Excess Return.

Method 4: Airline — an EV/EBITDA Multiple

Airlines share REIT's problem in spirit if not in mechanism: extreme cyclicality and high operating and financial leverage make a single trailing year's free cash flow and growth rate an unreliable basis for a multi-year DCF projection. An airline's DCF can come back with no computable value at all in a weak year despite the business generating solid EBITDA, or show a materially different picture than a multiple-based read.

The tool values airlines the same non-discounted way it values REITs — a multiple applied to a single figure:

Enterprise value = EBITDA × 6  →  Equity value = Enterprise value − net debt  →  ÷ shares outstanding

The flat 6× multiple is a documented, explicitly uncalibrated mid-cycle placeholder — Yahoo tags every airline with the same single industry label, so there's no sub-classification to build a table from the way REIT's property types allow. Ideally this would be EV/EBITDAR, which adds back aircraft lease rent to normalize for owned-versus-leased fleet mix (a standard airline-analyst convention) — but that figure isn't reliably broken out in the free data feed this tool relies on, so plain EBITDA ships as a real, disclosed limitation rather than a hidden one.

The Holding-Company Discount

A conglomerate — a company operating several genuinely different businesses under one listing — creates a valuation problem none of the models above solve on their own: whichever model the ticker would otherwise get (almost always DCF) values it as if it were one coherent business, when in practice a sum-of-the-parts valuation would treat each segment separately.

A true segment-by-segment valuation isn't possible with this tool's data source, since it doesn't expose per-segment financials for a diversified holding company. Instead, once a conglomerate is detected, the tool applies a flat 15% discount to whatever value the underlying model already produced — a documented, deliberately uncalibrated placeholder reflecting the typically less efficient capital allocation and reduced transparency of diversified structures, the same "state the limitation plainly" posture the Airline model's multiple takes. This is not a per-segment sum-of-the-parts estimate, and it stacks on top of any of the models above — a utility-tagged or bank-tagged conglomerate still gets its normal model first, with the discount applied afterward.

Which method to use, and when

The tool makes the default choice for you, but it helps to understand the reasoning so you know when to trust the output and when to reach for something else.

If you want a ready-made starting list rather than valuing stocks one at a time, the automated monthly screens apply this same model dispatch — DCF, DDM, Excess Return, REIT, Airline, and the Cyclical/Holdco adjustments — across an entire index and rank every stock by margin of safety, useful for surfacing names worth investigating further:

• Most Undervalued ASX 300 Stocks → — ranked monthly by DCF margin of safety, AUD
• Most Undervalued S&P 500 Stocks → — ranked monthly by DCF margin of safety, USD
• Most Undervalued NIFTY 500 Stocks → — ranked monthly by DCF margin of safety, INR
• Most Undervalued KOSPI 200 Stocks → — ranked monthly by DCF margin of safety, KRW
• Most Undervalued Nikkei 225 Stocks → — ranked monthly by DCF margin of safety, JPY
• Most Undervalued European Stocks → — ranked monthly by DCF margin of safety, EUR

Sector reference ranges

A useful sanity check is whether your calculated discount rate lands near where the company's sector usually sits. The tool shows a reference range for the stock's sector and market; if your WACC is far outside it, an input is probably off. These are typical post-tax ranges, not rules.

United States: Regulated Utilities 5.5–6.5% · REITs and Energy 6.5–7.5% · Aerospace & Defense 7.0–8.0% · Retail 7.5–8.5% · Biotech & Pharma 8.0–10.0% · Technology & Software 8.5–10.5%.

Australia (ASX): Utilities & Infrastructure 6.5–8.0% · A-REITs 7.5–9.0% · Banks & Financials 8.5–9.5% · Healthcare & Biotech 8.5–10.0% · Mining & Resources 10.0–12.0%.

India (NSE/BSE): Utilities & Infrastructure 8.0–10.0% · FMCG 9.0–11.0% · Banking & Financials and Pharma 10.0–12.0% · Commodities 11.0–13.0% · IT & Software 11.0–14.0%.

The patterns are intuitive: regulated, debt-heavy businesses with stable cash flows sit at the low end; cyclical, equity-funded, high-growth businesses sit at the high end. Indian ranges sit above their US and Australian equivalents because the underlying risk-free rate and risk premium are both higher.

When cash flow is negative, or debt swallows the equity

Two situations break a naïve DCF, and the tool handles each explicitly rather than printing a nonsense number.

Negative or missing free cash flow. A fast-growing company reinvesting everything can report negative FCF, which a standard DCF cannot value. Instead of giving up, you normalise: estimate the FCF margin the business will earn once it matures (say 15% for software, 8% for retail) and apply it to current revenue. The tool seeds this target margin from the reported profit margin, runs the DCF on the normalised figure, and estimates roughly which year FCF should turn positive given your growth assumption. The result is clearly labelled as revenue-normalised, because it rests on an extra assumption you have made.

Net debt larger than firm value. For some heavily indebted or capital-hungry companies, the DCF firm value comes out below net debt, leaving negative equity. Rather than show an impossible negative share price, the tool states plainly that the equity is worth roughly zero on a DCF basis. That is not a glitch — it is the model telling you the market is pricing the stock on future growth or asset value that today's cash flow cannot justify. The honest response is to revisit your growth assumption, or to accept that a cash-flow model is the wrong lens for that particular business.

Bear, Base, and Bull — and the margin of safety

Because the output is so sensitive to assumptions, a single intrinsic value would imply a precision that does not exist. The tool runs three scenarios side by side instead. Base uses the seeded assumptions. Bear applies lower growth and a higher discount rate; Bull applies higher growth and a lower discount rate — for DCF, DDM, and Excess Return. REIT and Airline have no growth or discount rate to vary, so their Bear/Base/Bull instead move the multiple ±20% around its base value; the shape is the same three-scenario range either way.

Each scenario reports a margin of safety — the gap between intrinsic value and the current price. A positive margin means the stock trades below the model's estimate; a negative one means it trades above. The idea, which goes back to Benjamin Graham, is to buy with enough of a discount that you can be somewhat wrong and still not overpay. When the price sits below even the Bear value, that is a meaningful signal; when it sits above the Bull value, the stock looks expensive on these assumptions.

What the numbers cannot tell you

A valuation model is a disciplined way to convert assumptions into a number. It is not a crystal ball, and it is worth being clear about its limits.

The output is only as good as the inputs — change the growth rate or discount rate slightly and the answer moves a lot, which is a feature of the method, not a flaw in the tool. Every one of the sector-specific models above exists precisely because the plain version breaks somewhere, and each still carries its own real limitation: cyclical normalization is capped by however many years of history Yahoo actually returns, the Airline model's 6× multiple and the Holdco discount's 15% are both explicit, uncalibrated placeholders rather than calibrated estimates, and none of this is a substitute for reading the actual business. Early-stage companies with no cash-flow history at all and conglomerates priced heavily on future optionality remain the hardest cases regardless of which model applies. The data comes from Yahoo Finance, is roughly 15 minutes delayed, and carries no accuracy guarantee, so treat every figure as a starting point to verify rather than a fact.

Used well, the model is a way to make your assumptions explicit and test them. The number at the end is far less valuable than the discipline of having to state, out loud, what you believe about a company's future — and seeing how much the answer depends on it.