← Taiwan Stock Snapshot Database
Home / Indicator library / Three valuation models
Valuation

Three valuation models, every assumption on a slider

Discounted cash flow, Gordon dividend discount and the Peter Lynch fair value for any Taiwan-listed ticker, rebuilt from the official filings. Each one returns a pessimistic to optimistic range with the full assumption set attached, because a model valuation quoted as a single number is telling you about the analyst rather than the company.

Try it

Type a four-digit Taiwan ticker, pick a model and move the sliders. The demo ticker 2330 is free with all three models; other tickers are a paid tier. The chart is drawn from the same public endpoint shown below, and the sliders map directly onto its query parameters.

GET /v1/indicator/valuation-models/2330

A model valuation is not a price target. All three are dominated by their inputs — one or two points on the growth or discount rate moves the answer by tens of percent. Read the range, and move the sliders yourself to feel how wide it really is.
What we publish is data and tooling — not investment advice, not stock picks, and no promise of returns. Every figure is either an official public source reproduced as published, or a derivation we document well enough for you to recompute it. What you trade on it is your call. Model valuations are scenario ranges built on disclosed assumptions, not price targets and not a view on whether the shares are worth owning.

How it is built

Four steps. The fourth exists because one of the inputs is derived rather than announced, and saying so is cheaper than being found out.

01

Rebuild the inputs from the filings rather than take a vendor figure

Free cash flow is operating cash flow less capital expenditure, taken from the official quarterly cash flow statement, de-cumulated back to standalone quarters and summed over four. Earnings per share comes from the same filings on the same trailing basis. Share count is derived from share capital at the ten-dollar par value that Taiwan uses. Nothing here is a number somebody typed in.

02

Run three scenarios instead of one number

Every model is evaluated at a pessimistic, base and optimistic combination — low growth against a high discount rate, the defaults, and high growth against a low discount rate. The band you see on the chart is that spread swept across the whole discount rate axis, so you can read off how fast the answer moves when the assumption moves.

03

Disclose every assumption in the response

The assumptions block of the JSON carries each input the model used: the cash flow, the growth rate, the discount rates, the terminal growth, the horizon. Anyone can recompute the number from that block with a spreadsheet. A valuation you cannot audit is an opinion wearing a decimal point.

04

Say where the dividend figure came from

The dividend discount model needs a dividend per share. Where the ex-dividend announcement detail is not part of this dataset, the figure is implied from the official dividend yield and the closing price of the same day, taking the yearly median. That is a derived number rather than an announced one, and it is labelled as such in the assumptions block and on this page, because the alternative — printing it silently — is how a small approximation becomes a claim.

Terms

Four terms carry the page:

Discounted cash flowDCF
In short: add up the cash the business is expected to throw off in future years, each year shrunk by the discount rate to reflect that money later is worth less than money now. It is the most assumption-sensitive of the three: a two-point move in the discount rate can move the answer by half.
Dividend discount modelGordon growth
In short: value only the cash you actually receive, assuming the dividend grows at a steady rate forever. It suits stable payers and systematically undervalues a company that reinvests instead of paying out — that is a property of the model, not a verdict on the company.
Peter Lynch fair valuePEG rule of thumb
In short: a company growing earnings at twenty percent is worth roughly twenty times earnings. Fast arithmetic rather than a model, useful as a sanity check against the other two, and extremely sensitive to the growth estimate — which is why the growth input is clipped to a 5–25 band here.
Weighted average cost of capitalWACC
In short: the discount rate, meaning the return an investor would demand for taking this risk. It is an input rather than an observable, which is exactly why it is a slider on this page instead of a constant buried in the code.

Where the line is

The strong half: every input is a documented function of an official filing, the assumption set travels with the answer in the JSON response, and the whole calculation can be reproduced in a spreadsheet from that block alone. Nothing on this page is a number that only exists inside our system.

The weak half is the models themselves, and it is worth being blunt about it. Discounted cash flow on a cyclical semiconductor company with a ten-year explicit horizon is an exercise in compounding your own guess. The dividend model assumes a growth rate holds forever, which is false for every company that has ever existed, and it will undervalue any business that reinvests rather than pays out. The Lynch rule is a heuristic from a book, not a theory of value. These are shipped because comparing three wrong-in-different-ways answers against the market price is genuinely informative, and pretending any one of them is right is not. Where an input is derived rather than filed — the dividend per share, when the announcement detail is unavailable — the response labels it, and this page prints that label next to the number.

FAQ

Is this a price target?

No, and the design says so deliberately. Every model here is run at three assumption sets and the output is a range rather than a number. A single valuation figure implies a precision the underlying models do not have — move the discount rate by two points and the answer moves by tens of percent, which you can watch happen with the sliders.

Where do the inputs come from?

Official quarterly filings for cash flow, earnings and share capital, the official daily close for the current price, and the official dividend yield for the dividend model. Everything is recomputed here rather than copied from a data vendor, and the assumptions block in the JSON response carries every input so the number can be reproduced independently.

Why does the dividend model sometimes say there is not enough data?

It needs at least three years of dividend history to estimate a growth rate. A recently listed company, or one that has not paid a cash dividend, will not have it. The model is skipped in that case rather than run on a fabricated dividend, and the other two models still compute.

What does the shaded band on the chart mean?

The horizontal axis sweeps the discount rate — or the growth rate for the Lynch model — across a realistic range. The dark line is the valuation at the base assumption at each point on that axis, and the band is the pessimistic to optimistic spread. Where the current market price falls inside that range it is drawn as a dashed horizontal line, so the crossing point tells you what the market is implicitly assuming. Where the price sits far outside the range — which happens often, and is itself the finding — the line is omitted and a note under the chart says which side it is on.

Can I call this from code?

Yes. The endpoint is /v1/indicator/valuation-models/{stock_id} and it accepts growth, wacc and terminal_growth as query parameters, so a sensitivity grid is a loop rather than a click. The free tier covers the demo ticker 2330 with all three models; arbitrary tickers are a paid tier.

Run it yourself

The free tier needs no signup — call this feature's API straight away with the demo key. Unlock as-of history and full ticker coverage on a paid tier.

Try the demo key Plans & pricing