7 Stock Valuation Models Investors Should Know Before Trusting a Fair Value Estimate
Stock valuation models can make a fair value estimate look like a fact when it is really the conclusion of a model.
The number may be precise to the cent. The chart may show a clean gap between market price and estimated value. The score may make comparison feel effortless. None of that tells you whether the model asked the right question for the business, whether its assumptions were defensible, or whether its output was even designed to represent intrinsic value.
Before trusting the estimate, an investor should identify the model that produced it. The first question is not simply, “What is fair value?” It is, “What kind of output am I looking at?”
StockGeniuses organizes seven approaches inside its Value Investing category: Discounted Cash Flow, Residual Income Valuation, Earnings Power Value, Buffett Intrinsic Value, Benjamin Graham Number, O’Shaughnessy Trending Value, and Joel Greenblatt’s Magic Formula. They belong together because each investigates value, but they do not all produce the same kind of answer.
Four estimate intrinsic value through different economic anchors. One provides a simplified defensive valuation screen. Two rank stocks relative to a defined universe rather than calculating fair value at all.
That distinction is the key to this article. The phrase “stock valuation models” is useful search language, but serious interpretation requires a more precise taxonomy.
Before the seven models: value is not one type of output
A strong stock analysis framework gives every model a defined job. It does not treat every number as if it carries the same meaning.
The seven StockGeniuses Value models fall into three practical groups:
| Model type | Models | Typical output | Main question |
|---|---|---|---|
| Intrinsic valuation | DCF, RIV, EPV, Buffett Intrinsic Value | Estimated intrinsic value under assumptions | What may the business or equity be worth? |
| Defensive valuation screen | Benjamin Graham Number | Formula-based valuation anchor | Do earnings and book value support a conservative price reference? |
| Relative factor ranking | O’Shaughnessy Trending Value, Magic Formula | Percentile or combined rank | How does this stock compare with the selected universe on value and another factor? |
The categories overlap philosophically, but their outputs are not interchangeable.
A DCF estimate near a market price and a top-decile Magic Formula rank do not confirm the same proposition. The DCF is an intrinsic-value argument built from future cash flows and discount rates. The Magic Formula rank says the stock combines favorable earnings yield and return on capital relative to other eligible stocks. One is an absolute estimate under assumptions; the other is a comparative position inside a universe.
Learning to read a stock analysis model therefore matters more than memorizing model names. The headline output sits downstream from the model’s philosophy, inputs, eligibility rules, and blind spots.
Model eligibility deserves as much attention as the final number. A model that declines an unsuitable business may protect the analysis better than one that always returns a precise answer. Availability is not applicability.
1. Discounted Cash Flow: valuing expected future cash generation
Discounted Cash Flow, or DCF, asks what a company’s expected future cash flows are worth today.
Inside StockGeniuses, the model projects Free Cash Flow to the Firm, discounts those cash flows using the Weighted Average Cost of Capital, estimates a terminal value, and moves from enterprise value to equity value through net debt and shares outstanding.
Its appeal is easy to understand. DCF forces the analyst to make important assumptions visible:
- the starting cash-flow base
- forecast growth
- reinvestment needs
- the discount rate
- terminal growth
- debt, cash, and share count
That visibility is also the source of its risk. A DCF can be mathematically correct while remaining economically fragile. Small changes in WACC, terminal growth, normalized cash flow, or the forecast path can move the estimated value materially. When terminal value represents most of the result, the apparent precision can hide how much of the estimate depends on the distant future.
The model is most useful when cash flows are positive or credibly normalizable, the business is understandable, and long-term assumptions can be defended without heroic optimism. It is less reliable for early-stage companies, structurally negative cash-flow businesses, and companies whose economics swing sharply across cycles.
The existing DCF valuation example uses Apple to show this directly. Filing data provides a real starting point, but the valuation still changes when growth, WACC, and terminal assumptions change. The estimate matters, but the more durable output is the map of assumptions carrying it.
2. Residual Income Valuation: valuing profit above the cost of equity
Residual Income Valuation, or RIV, starts somewhere DCF does not: current book value of equity.
It then asks whether the company earns more than shareholders require for supplying that equity capital. Residual income is net income after deducting an equity charge, which is beginning book value multiplied by the cost of equity. Intrinsic equity value is built from current book value plus the present value of expected future residual income.
The conceptual question is powerful:
Is the company merely reporting profit, or is it earning enough to create value after accounting for shareholders’ opportunity cost?
A company can report positive net income while earning less than its cost of equity. RIV makes that distinction explicit. It is particularly intuitive where book value is meaningful and accounting relationships are reasonably stable, including many financial institutions and other accounting-driven businesses.
A bank such as JPMorgan Chase is a recognizable model-fit illustration because book equity, net income, return on equity, and the cost of equity are central to understanding the business. That does not mean RIV automatically produces a trustworthy current valuation for JPMorgan. Accounting quality, changing capital requirements, credit conditions, and assumptions about how long excess returns persist still matter.
RIV becomes less natural when book value is negative, unstable, or disconnected from the assets that actually create economic value. An asset-light company built around internally developed software, brand, data, or network effects may have substantial economic value that accounting book equity captures poorly.
Its main advantage is not that it eliminates forecasting. It changes the anchor. DCF begins with cash flows; RIV begins with book equity and asks how much value future excess returns add beyond it.
3. Earnings Power Value: removing growth from the valuation story
Earnings Power Value, or EPV, asks a deliberately restrictive question:
What might the business be worth if it never grows, but continues producing a sustainable level of operating earnings?
The model normalizes operating earnings, applies taxes, capitalizes the resulting earnings using a required return such as WACC, then adjusts operating value for net debt, excess cash, and shares. Growth is not quietly buried inside the formula. It is intentionally excluded.
This makes EPV useful as a no-growth baseline. If an investment case looks compelling only when future expansion carries most of the value, EPV exposes how little of the thesis is supported by current earnings power. If current normalized earnings already support a substantial operating value, the analyst can separate that base from the more uncertain value assigned to growth.
The difficult part is normalization. A single strong year can overstate sustainable earnings. A recession year can understate them. Cyclical margins, unusual expenses, temporary pricing, acquisitions, and changing capital needs can all distort the base.
Procter & Gamble is a useful business-type illustration because a mature company with long operating history and recurring consumer demand gives the analyst more evidence for discussing sustainable earnings than an early-stage company still building its economics. This is not a current EPV conclusion about P&G. It simply shows why model fit begins with the stability of the earnings base.
The no-growth premise is easy to misread. As the Earnings Power Value explanation makes clear, it does not mean the company is expected to stop growing. It means the valuation refuses to pay for growth until the analyst examines that growth separately.
4. Buffett Intrinsic Value: thinking in owner earnings
Buffett Intrinsic Value approaches valuation from the perspective of a long-term owner of the whole business.
Its core concept is owner earnings: the cash the business can generate for owners after accounting for the capital required to maintain its competitive position. Reported net income is relevant, but it is not accepted as a complete description of owner economics. Depreciation, maintenance capital expenditures, working-capital requirements, business durability, and net debt all affect the interpretation.
The model’s strength is that it forces the investor to ask what part of reported earnings is genuinely distributable without weakening the business.
Its weakness is that maintenance capital expenditure is rarely reported as a clean line item. It has to be estimated. A business may also need continuing investment in technology, customer acquisition, product development, or physical assets simply to preserve its economics. Calling every such investment “growth” can overstate owner earnings; treating all capital spending as maintenance can understate them.
Buffett-style valuation therefore fits best when the business is understandable, owner earnings are positive and reasonably stable, and maintenance needs can be estimated with restraint. Durable economics matter because the approach is not merely discounting a number. It is making a judgment about what a long-term owner can reasonably expect the business to produce.
Apple serves as the real-company teaching example in the Buffett Intrinsic Value article. The analysis does not pronounce on Apple’s current price. It shows how owner earnings change the valuation conversation by separating reported profit from the cash economics of maintaining the business.
Comparing DCF and Buffett Intrinsic Value takes that distinction further. Both models estimate intrinsic value, but they can disagree because they define cash generation, reinvestment, discounting, and long-term durability differently.
5. Benjamin Graham Number: a defensive valuation anchor
The Benjamin Graham Number compresses valuation into a simple formula using earnings per share and book value per share.
Its familiar form is:
Graham Number = square root of (22.5 x EPS x BVPS)
The constant 22.5 reflects a combined ceiling derived from a price-to-earnings ratio of 15 and a price-to-book ratio of 1.5. The result is designed to create a conservative reference grounded in current earnings and balance-sheet backing.
That simplicity is useful, but it should not be mistaken for completeness.
The Graham Number does not forecast cash flow, estimate a discount rate, evaluate competitive durability, or account for intangible-heavy economics. It is invalid when EPS or book value per share is non-positive, and it can be a poor fit for businesses whose economic assets are not well represented on the balance sheet.
The Benjamin Graham Number guide uses Ford and Apple to illustrate model fit without declaring either stock cheap or expensive. Ford’s more asset-heavy accounting structure gives book value a different analytical role than it has for Apple, where internally created brand, software, ecosystem effects, and other intangibles are not fully represented by conventional book equity.
The Graham Number belongs in the Value category because it provides a defensive valuation screen. But it is not a full thesis and not a precision fair-value model. Its best use is often to create a question: does the formula appear favorable because the business is genuinely supported by earnings and assets, or because accounting measures are temporarily distorted?
6. O’Shaughnessy Trending Value: combining relative cheapness with momentum
O’Shaughnessy Trending Value is the point where the output type changes.
It does not estimate intrinsic value. It ranks stocks relative to a defined universe by combining a composite value measure with recent price momentum.
In the StockGeniuses implementation, the value component draws on several measures, including price-to-earnings, price-to-book, price-to-sales, price-to-cash-flow, enterprise value to EBITDA, and shareholder yield. Stocks are ranked on the composite. The relatively cheapest subset is then evaluated using six-month price momentum.
The model’s question is not:
What is this company worth?
It is:
Which eligible stocks combine stronger relative value characteristics with positive recent price behavior?
That is a legitimate Value-category question. It reflects a quantitative value philosophy that tries to avoid selecting statistically cheap stocks while price behavior continues deteriorating. Momentum acts as a second factor, not as proof that the business is sound or intrinsically undervalued.
The model also cannot be interpreted company by company without its universe. A percentile rank depends on which stocks were eligible, the data date, liquidity rules, and the distribution of factor values. Change the universe and the rank can change even if the company’s own financial statements do not.
A high Trending Value score should never be translated into a fair-value claim. It indicates relative factor alignment. It does not establish business quality, durable cash flow, or a margin of safety in the intrinsic-value sense.
7. Joel Greenblatt’s Magic Formula: combining cheapness with operating quality
Joel Greenblatt’s Magic Formula is also a relative ranking system, but it combines different factors.
The model ranks companies using:
- earnings yield, based on EBIT relative to enterprise value
- return on capital, based on EBIT relative to the operating capital required by the business
Earnings yield represents relative cheapness. Return on capital represents operating efficiency. The combined rank looks for businesses that appear both inexpensive and economically productive compared with the eligible universe.
That pairing explains why the Magic Formula belongs in the Value category even though it does not calculate intrinsic value. It is a systematic expression of a classic value-investing desire: avoid choosing between a cheap weak business and a strong business priced without restraint.
The tradeoff is compression. Return on capital cannot fully describe competitive durability, management quality, cyclicality, accounting risk, or the reinvestment runway. Earnings yield can be distorted by unusually high or low EBIT. The combined rank also depends on universe construction and can behave differently as factor regimes change.
Most importantly, a top rank is not a fair price. It says the company compares favorably on the two selected factors. It still needs business analysis, risk review, and interpretation of whether the reported EBIT and invested-capital base are economically representative.
The Magic Formula is valuable precisely because it is simple and systematic. It becomes dangerous only when simplicity is confused with completeness.
The seven models compared
The differences become clearer when the models are placed side by side.
| Model | Main anchor | Output type | Natural fit | Central limitation |
|---|---|---|---|---|
| DCF | Expected FCFF discounted at WACC | Intrinsic value estimate | Businesses with analyzable cash flows and defensible forecasts | Forecast, WACC, and terminal-value sensitivity |
| RIV | Book value plus future residual income | Intrinsic equity value estimate | Accounting-stable businesses with meaningful book equity | Accounting quality and excess-return persistence |
| EPV | Sustainable after-tax operating earnings | No-growth intrinsic-value baseline | Mature businesses with normalizable earnings | Normalization and changing business economics |
| Buffett Intrinsic Value | Owner earnings under conservative assumptions | Owner-oriented intrinsic value estimate | Durable, understandable businesses with estimable maintenance needs | Maintenance capex and owner-earnings judgment |
| Graham Number | EPS and book value per share | Defensive valuation screen | Traditional, profitable, asset-backed businesses | Ignores growth, cash-flow dynamics, and intangible value |
| O’Shaughnessy Trending Value | Composite relative value plus six-month momentum | Universe-dependent factor rank | Diversified quantitative screening | Universe dependence, factor cycles, and momentum reversal |
| Magic Formula | Earnings yield plus return on capital | Universe-dependent quality-value rank | Systematic search for relatively cheap, efficient businesses | Accounting distortion, universe dependence, and limited qualitative context |
The table should make one point unavoidable: comparing the final numbers without comparing the output types is a category error.
The raw evidence underneath these models also matters. Cash flow, EBIT, book equity, debt, share count, multiples, price history, and capital efficiency come from the broader layer of Core Metrics in stock analysis. The related guide to stock analysis metrics explains why those inputs need context before a model gives them a philosophical role.
Why reasonable models disagree
Different models can disagree even when each one is implemented correctly.
The reason is not mysterious. They begin from different anchors and place uncertainty in different locations.
DCF puts substantial weight on future cash flow, WACC, and terminal value. RIV puts weight on book equity, cost of equity, and the persistence of returns above that cost. EPV removes growth but depends on normalized current earnings. Buffett Intrinsic Value depends on owner earnings and maintenance needs. Graham emphasizes current earnings and book value. O’Shaughnessy and Magic Formula depend on relative ranks rather than absolute worth.
A growth-oriented, asset-light company might look more understandable through DCF than Graham. A bank may give RIV a more meaningful accounting anchor. A mature consumer business may be easier to examine through EPV or owner earnings. A company can rank well under Magic Formula while still producing an intrinsic-value estimate that demands aggressive assumptions.
Disagreement is therefore diagnostic information.
It can reveal that:
- future growth is carrying more value than current earnings power
- book equity does not represent the economic asset base well
- maintenance capital needs are uncertain
- a stock looks cheap only relative to an expensive universe
- price momentum and fundamentals are moving in different directions
- accounting profitability is strong while cash conversion is weak
Consider a company where DCF produces a much stronger result than EPV. The gap may show that forecast growth carries much of the DCF value while normalized current earnings support a lower no-growth baseline. If RIV is also restrained, the growth-heavy interpretation deserves closer scrutiny. The models have not voted against the company; they have located the thesis’s dependence on growth and persistence.
The wrong response is to select whichever output supports the preferred conclusion. The better response is to identify which assumption or model-fit difference created the conflict.
Real-company examples of model fit
Real companies make the model-fit question easier to see, even when no current valuation is calculated.
| Company illustration | Model questions that may be especially informative | Why this is only a fit illustration |
|---|---|---|
| Apple | DCF and Buffett Intrinsic Value | Cash generation and owner economics are relevant, but growth, maintenance needs, discount rates, and ecosystem durability still require judgment. |
| JPMorgan Chase | Residual Income Valuation | Book equity and returns relative to the cost of equity are meaningful, but credit conditions, regulation, and accounting quality affect persistence. |
| Procter & Gamble | EPV and owner-earnings analysis | Mature operating history can support normalization work, but it does not make any current output automatically correct. |
| Ford | Graham Number and normalized earnings approaches | Asset backing is more visible than in many intangible-heavy businesses, but cyclicality and capital intensity complicate a simple formula. |
O’Shaughnessy Trending Value and Magic Formula cannot be illustrated honestly by naming one company and inspecting it alone. Their outputs exist only relative to an eligible universe. A company can move in rank because its own inputs changed, because peers changed, or because the universe changed.
These examples are educational, not current assessments of the companies. No present fair value, score, rank, or investment conclusion is being stated.
How StockGeniuses reads the Value category
StockGeniuses does not use the seven models as seven votes on one identical question.
Each model preserves its own philosophy and output. The Value Overall Score then summarizes convergence across the valid model scores. At least four of the seven models must be available before the overall score is calculated. Missing models are excluded rather than treated as failures.
The resulting score is an orientation tool. It can show broad agreement, mixed evidence, or divergence across the Value category. It is not an intrinsic valuation, price target, recommendation, or substitute for opening the individual model outputs.
Coverage composition matters too. A score built mainly from DCF, EPV, and Buffett Intrinsic Value reflects more cash-flow-oriented evidence. A score built mainly from RIV, Graham, O’Shaughnessy, and Magic Formula is more accounting- and factor-oriented. Equal numerical coverage does not guarantee equal philosophical coverage.
Two companies can receive the same Value Overall Score through materially different evidence. One may have broad cash-flow-based support; the other may score well mainly through accounting anchors and relative factor ranks. The shared number does not make the cases equivalent. The coverage map explains what the score cannot.
The number alone is never enough. Ask which models were meaningful, which were unavailable, which model types dominated the coverage, and where the individual results disagreed.
That is the practical benefit of a multi-model system. It does not make uncertainty disappear. It makes the source of uncertainty easier to inspect.
A safer way to review a fair value estimate
When a platform, analyst, spreadsheet, or article presents a fair-value estimate, use this sequence before accepting the number:
- Identify the model. A DCF estimate, RIV estimate, Graham Number, and factor rank do not mean the same thing.
- Confirm the output type. Is it intrinsic value, a screen, a relative rank, or a blended score?
- Check model eligibility. Does the company have the cash flow, book value, earnings stability, or universe data the model requires?
- Trace the main assumptions. Find the growth, discount rate, normalization, maintenance capex, persistence, or peer-universe choices carrying the result.
- Check the business fit. Ask whether the model represents how this particular company creates value.
- Compare another lens. Use disagreement to identify hidden assumptions rather than averaging outputs automatically.
- Return to the underlying evidence. Review business quality, financial strength, cash conversion, debt, history, and market context.
- Keep the estimate inside the thesis. A valuation output can support or challenge a thesis; it cannot write the thesis for you.
This sequence belongs inside the broader work of learning to analyze a stock systematically. Valuation is one layer of the process. It becomes more useful after the business and its financial evidence are understood.
A mature dividend payer may also call for a model outside this Value category. The Gordon Growth Model, for example, sits in StockGeniuses’ Dividend & Income category because it values a stable dividend stream under perpetual-growth assumptions. Its absence from these seven models is an architectural choice, not a claim that it is irrelevant to valuation. The broader nine-model comparison shows how Value, Dividend, Growth, Momentum, Risk, and Sentiment models retain different analytical jobs.
The fair-value lesson
A fair-value estimate is not a fact waiting to be discovered. It is an argument built from a model, a set of inputs, and a view of what matters.
DCF argues from future cash flow. RIV argues from book equity and excess returns. EPV argues from sustainable current earnings without growth. Buffett Intrinsic Value argues from owner earnings and maintenance needs. Graham Number creates a defensive earnings-and-book-value anchor. O’Shaughnessy Trending Value ranks relative cheapness with momentum. Magic Formula ranks cheapness with operating quality.
All seven belong in the StockGeniuses Value category because they examine value through distinct philosophies. They should not be collapsed into seven versions of the same calculator.
The strongest investor habit is therefore not choosing one model and trusting it forever. It is knowing what question the model answered, why that question fits the business, what assumptions carried the output, and what evidence the model left outside its frame.
Once those distinctions are visible, fair value becomes more useful and less seductive. The estimate can inform judgment without impersonating certainty.