— —
— —- EPS近四季
- —
- 每股淨值最近一季
- —
- 每股自由現金流最近年度
- —
- 現金股利近四季
- —
- 本益比依你的股價
- —
- 股價淨值比依你的股價
- —
- 股價營收比依你的股價
- —
- 現金殖利率依你的股價
- —
- ROE近四季
- —
- 市值依你的股價
- —
- 在外流通股數
- —
- 財報基準最新一期
- —
獲利品質
毛利率與營業利益率的走勢。這不是估價指標，不納入上方彙總 —— 它看的是各公式分母的那個「盈餘」品質如何。 Gross and operating margin over time. Not a valuation input and not included in the summary above — it shows the quality of the earnings the formulas divide by.計算值彙總
下方各公式在你設定的參數下算出的數值，做敘述統計。 本站不對這些數值作評價，也不提供買賣建議。勾選卡片可調整納入哪些公式。
Descriptive statistics over the values the formulas below produce under the assumptions you set. This site does not evaluate these numbers and gives no buy or sell advice. Use the checkboxes to choose which formulas count.
統計說明
—
計算參數
所有假設都由你設定，改動會立即重算
自訂每股盈餘（EPS）
填入數字後，所有以「預估 EPS」為基礎的公式（年化本益比、本益成長比、 葛拉漢公式）一律改用你輸入的值，卡片會標示「你輸入的 EPS」。 留空則沿用本站依 SEC 季報算出的數字。
Enter a number and every formula built on an estimated EPS (forward P/E, PEG, Graham) switches to your figure, and the card says so. Leave it blank to keep the value derived from SEC quarterly filings.
本益比倍數（年化版與近四季版共用）
預設值高於台股版（10 / 15 / 20），因為美股整體本益比水準本來就比台股高， 但這只是一組起點，請自行依產業調整。
The defaults are higher than the Taiwan edition's (10 / 15 / 20) because US P/E levels simply run higher. This is only a starting point — adjust it for the industry.
股價淨值比倍數
股價自由現金流比（P/FCF）
資本支出重的產業（半導體、電信、能源）自由現金流波動大， 單一年度的數字容易失真，建議搭配多年趨勢一起看。
Capital-intensive industries (semiconductors, telecom, energy) have volatile free cash flow, so a single year is easily distorted. Read it alongside a multi-year trend.
目標現金殖利率
殖利率與價格成反比，殖利率設得越高、算出的對應價越低。 美股整體殖利率低於台股（大量以庫藏股取代配息），預設值因此較低。
Yield moves inversely to price: the higher the target yield, the lower the resulting price. US yields run below Taiwan's because so many companies return cash through buybacks instead of dividends, hence the lower defaults.
股價營收比倍數（P/S）
合理倍數因產業差異極大 —— 軟體與雲端常在 5~10 倍以上， 零售與通路低於 1 倍。務必依同業水準調整。
Reasonable multiples vary enormously by industry — software and cloud often run above 5–10x, retail and distribution below 1x. Always adjust to the peer group.
本益成長比（PEG）
PEG = 本益比 ÷ 盈餘成長率。文獻上常以 1 作為對照基準，這裡的三組值 由你自行設定。成長率上限用來避免基期過低時算出極端的倍數。
PEG = P/E ÷ earnings growth rate (%). The literature often uses 1 as a reference point; the three values here are yours to set. The growth cap keeps a very low prior-year base from producing an absurd multiple.
葛拉漢公式 V = EPS × (8.5 + 2g)
8.5 是葛拉漢 1962 年提出的「零成長公司」基準本益比。g 由你自行判斷輸入， 本站不替任何個股預設成長率；敏感度用來同時看 g ± 若干百分點的結果。
8.5 is the base P/E Graham observed for a no-growth company in 1962. You supply g — this site never presets a growth rate for any stock. The sensitivity setting shows g ± a few percentage points side by side.
ROE 法
g 必須小於 r，否則模型無解。
公式怎麼算？
七種公式與台股版相同；月營收動能法因為美國沒有月營收而換成自由現金流， 本益成長比則只有美股頁有（台股的開放資料沒有去年全年每股盈餘）。 下面每一段都寫明這個公式在美股的資料來源與邊界。
1. 本益比法 —— 年化 EPS（Forward P/E）
近四季 EPS 反映過去一整年；獲利正在成長或衰退的公司，近四季會落後現況。 這個方法改用本會計年度已公布的季報年化：
年化 EPS = 本年度累計 EPS ÷ 已公布季數 × 4 對應價 = 年化 EPS × 你設定的本益比倍數
1 季 ×4、2 季 ×2、3 季 ×4/3、4 季就是全年。季數越少，外推成分越重， 卡片會標明用幾季推算。注意這裡的「年度」是各公司自己的會計年度 —— 蘋果 9 月底結束、輝達 1 月、微軟 6 月，所以每檔的期間標示都不一樣。
剛結束一個會計年度、還沒送出新季報的公司（例如年報剛發布時）， 沒有可年化的對象，這個方法會標示不適用 —— 此時近四季用的就是同一份年報， 兩種算法會得到相同的數字，不必重複列出。
2. 本益比法 —— 近四季 EPS（Trailing P/E）
對應價 = 近四季 EPS × 你設定的本益比倍數
用已實現的一整年獲利，不做任何外推，是上面年化版的保守對照組。
近四季怎麼來的：XBRL 有個結構性缺口 —— 公司的會計年度 第四季通常不會單獨標記（年報只揭露全年數字，Q4 要用 「全年 − 前三季」才推得出來），直接把四個日曆季相加，大部分公司會少一季。 所以本站改用會計年度重建：
近四季 = 最近完整會計年度
+ 該年度結束後的累計
− 去年同期累計
個股卡上會列出實際用到的年度與季別。若去年同期的季別資料不齊、無法完成扣減， 會退回直接採用該會計年度的全年數字，並在卡上標示。
3. 本益成長比（PEG Ratio）
PEG = 本益比 ÷ 盈餘成長率(%) 對應價 = 預估 EPS × 成長率(%) × 你設定的目標 PEG
成長率取「預估 EPS ÷ 前一個會計年度 EPS − 1」，並設上限（預設 40%） 避免基期過低時算出極端的倍數。去年虧損或今年衰退時不適用， 景氣循環股也容易在高峰期被高估。
4. 自由現金流法（P/FCF）
自由現金流 = 營業活動現金流 − 資本支出 每股自由現金流 = 自由現金流 ÷ 股數 對應價 = 每股自由現金流 × 你設定的 P/FCF 倍數
盈餘含折舊攤銷與各種應計項目，可以在不動用現金的情況下被調整； 自由現金流是真的收進來、扣掉維持營運所需投資之後剩下的錢。 兩者長期背離通常值得追究，這是前面幾種獲利類公式看不到的角度。
兩個限制講在前面。其一，現金流量表在美國是以 年初至今累計申報，半年與九個月的累計值不在 SEC 的期間彙總表內， 所以近四季無法穩定重建 —— 本站取的是最近一個完整會計年度的數字， 比其他公式舊，卡片上有標期間。其二，銀行與保險公司的現金流量表結構不同， 沒有可比的資本支出科目，會標示不適用。大幅擴廠或處分資產的年度也會失真。
5. 葛拉漢公式（Graham Formula）
V = EPS × (8.5 + 2g)
8.5 是葛拉漢 1962 年觀察到「零成長公司」的基準本益比， g 為預期年成長率（%）。本站不替任何個股預設 g —— 它完全由你在參數區輸入，卡片會同時列出 g ± 敏感度的三組結果。
此式是 1960 年代的經驗公式，未考慮利率環境（原著另有以 AAA 公司債殖利率 調整的版本），對高成長股會給出很高的倍數。
6. ROE 法（股東權益報酬率法）
ROE = 近四季淨利 ÷ 股東權益 模型股價淨值比 = (ROE − g) ÷ (r − g) 對應價 = 每股淨值 × 模型股價淨值比 上下限 = 對應價 × (1 ∓ 你設定的安全邊際)
r 是你要求的年報酬率、g 是長期成長率，兩者都由你輸入。 ROE 必須大於 g 才有意義。
美股要特別注意庫藏股。買回的股份直接沖減股東權益， 長期大量買回的公司帳面淨值偏低，ROE 因而被墊高 —— 那不代表它真的 比同業會賺錢，只代表它的分母被買小了。極端者股東權益為負，模型直接無解。
7. 股價淨值比法（P/B Ratio）
每股淨值 = 股東權益 ÷ 股數 對應價 = 每股淨值 × 你選定的股價淨值比倍數
適合資產型、獲利波動大的公司（銀行、保險、房地產、景氣循環股）。 帳面淨值不含自行發展的品牌、專利與軟體，所以輕資產公司的 P/B 天生就高； 加上前述的庫藏股效應，美股的 P/B 比台股更難跨產業比較。
8. 股價營收比（P/S Ratio）
每股營收 = 近四季營收 ÷ 股數 對應價 = 每股營收 × 你選定的股價營收比倍數
營收恆為正，所以獲利尚未轉正的公司也估得出來 —— 這是其他獲利類方法完全無解的族群，美股的軟體與生技業有大量這種公司。
本法預設不納入彙總。合理倍數因產業差異極大， 沒有一組通用預設值。請先確認同業水準再勾選納入。它完全不看獲利能力， 營收高但長年虧損的公司會被高估。
9. 股利法（現金股利折現）
對應價 = 每股現金股利 ÷ 你設定的目標殖利率
殖利率與價格成反比，最高的目標殖利率對應最低的價格。 適合配息穩定的成熟公司。
美股有兩點和台股不同：一是大量成長股完全不配息， 改以庫藏股回饋股東，這個公式對它們沒有意義；二是美國對外國投資人 預扣 30% 股利稅，實收金額低於本頁顯示的帳面殖利率。
數字怎麼來的
所有數字都取自美國證券交易委員會（SEC）：財報來自 EDGAR 的 XBRL 彙總表（frames API）， 股票代號來自 SEC 官方代號表。美國聯邦政府的著作屬公共領域，可自由商業利用。 本站不提供股價，比率類數字以你輸入的股價計算：
股數 = 公司申報封面的在外流通股數（SEC） 每股淨值 = 股東權益 ÷ 股數 每股營收 = 近四季營收 ÷ 股數 EPS = SEC 申報的稀釋後每股盈餘（近四季重建） ROE = 近四季淨利 ÷ 股東權益
雙重股權（如波克夏 A／B 股、Alphabet GOOGL／GOOG）共用同一份申報， 每股數字只對應其中一個股別。若申報的每股盈餘與在外股數對不起來（差距超過兩倍）， 本站改用「近四季淨利 ÷ 每股盈餘」推得的股數，讓每股淨值、每股營收與 EPS 同一口徑； 這類公司的卡片會提醒你確認輸入的股價是同一個股別。
Seven formulas match the Taiwan edition. Monthly revenue momentum is replaced by free cash flow because the US has no monthly revenue, and PEG appears only here (Taiwan's open data has no prior-year full-year EPS). Each section below states the data source and the limits of that formula in the US market.
1. P/E method — annualized EPS (Forward P/E)
Trailing twelve-month EPS reflects the past year; for a company whose earnings are growing or shrinking, it lags reality. This method annualizes the quarters already filed in the current fiscal year:
annualized EPS = year-to-date EPS ÷ quarters filed × 4 price = annualized EPS × your P/E multiple
1 quarter ×4, 2 ×2, 3 ×4/3, 4 is the full year. The fewer the quarters, the more extrapolation, and the card says how many were used. Note that "year" means each company's own fiscal year — Apple's ends in late September, NVIDIA's in January, Microsoft's in June — so the period label differs by stock.
A company that has just closed a fiscal year and not yet filed a new quarter has nothing to annualize, so this method reports that it does not apply. In that situation the trailing method is using that same annual report, and both would return the same number — no point listing it twice.
2. P/E method — trailing twelve months (Trailing P/E)
price = TTM EPS × your P/E multiple
A full year of realised earnings with no extrapolation — the conservative counterpart to the forward version above.
How TTM is derived: XBRL has a structural gap — a company's fiscal fourth quarter is usually not tagged separately (the annual report discloses only the full year, so Q4 has to be inferred as "full year minus the first three quarters"). Add four calendar quarters together and most companies come up one quarter short. So this site rebuilds it from the fiscal year:
TTM = latest complete fiscal year
+ year-to-date since that year ended
− the same period one year earlier
Each stock card lists the actual year and quarters used. If the prior-year quarters are incomplete and the subtraction cannot be done, it falls back to the fiscal-year figure and says so on the card.
3. PEG ratio
PEG = P/E ÷ earnings growth rate (%) price = estimated EPS × growth rate (%) × your target PEG
The growth rate is "estimated EPS ÷ prior fiscal-year EPS − 1", capped (40% by default) so that a very low base cannot produce an absurd multiple. It does not apply when last year was a loss or this year is shrinking, and cyclical stocks are easily overvalued at the peak.
4. Price / free cash flow (P/FCF)
free cash flow = cash flow from operations − capital expenditure free cash flow/share = free cash flow ÷ shares price = free cash flow per share × your P/FCF multiple
Earnings contain depreciation, amortisation and every kind of accrual, and can be adjusted without any cash moving. Free cash flow is money actually received, after the investment needed to keep the business running. A persistent gap between the two is usually worth investigating — an angle none of the earnings-based formulas can see.
Two limits, stated up front. First, US cash flow statements are filed year-to-date, and the SEC's period frames only cover 3-month and 12-month spans, so six- and nine-month cumulative figures appear in no frame at all and TTM cannot be reconstructed reliably — this site uses the latest complete fiscal year, which is older than the other formulas, and the card shows the period. Second, banks and insurers structure their cash flow statements differently and have no comparable capital-expenditure line, so the method reports that it does not apply. A year with a major plant build-out or asset disposal will also distort it.
5. Graham formula
V = EPS × (8.5 + 2g)
8.5 is the base P/E Graham observed for a no-growth company in 1962; g is the expected annual growth rate (%). This site never presets g for any stock — you enter it under "Assumptions", and the card shows g ± your sensitivity setting side by side.
This is an empirical formula from the 1960s that ignores the interest-rate environment (the original also has a version adjusted by AAA corporate bond yields) and produces very high multiples for high-growth companies.
6. Return on equity method
ROE = TTM net income ÷ shareholders' equity model P/B = (ROE − g) ÷ (r − g) price = book value per share × model P/B bounds = price × (1 ∓ your margin of safety)
r is your required annual return and g the long-run growth rate; both are yours to set. ROE must exceed g for the model to mean anything.
Watch out for buybacks in the US. Repurchased shares are charged directly against shareholders' equity, so companies that have bought back stock for years carry a low book value and a correspondingly inflated ROE — that does not mean they earn more than their peers, only that the denominator has been shrunk. In extreme cases equity is negative and the model has no solution at all.
7. Price / book (P/B)
book value per share = shareholders' equity ÷ shares price = book value per share × your P/B multiple
Suited to asset-heavy companies and those with volatile earnings (banks, insurers, real estate, cyclicals). Book value excludes internally developed brands, patents and software, so asset-light companies carry a naturally high P/B; combined with the buyback effect above, US P/B is even harder to compare across industries than Taiwan's.
8. Price / sales (P/S)
revenue per share = TTM revenue ÷ shares price = revenue per share × your P/S multiple
Revenue is always positive, so companies not yet profitable can still be valued — a group none of the earnings-based methods can handle, and a large one in US software and biotech.
Excluded from the summary by default. Reasonable multiples vary enormously by industry and there is no universal default. Check the peer group before ticking it in. It ignores profitability entirely, so a company with high revenue and years of losses will be overvalued.
9. Dividend method
price = dividend per share ÷ your target yield
Yield moves inversely to price, so the highest target yield gives the lowest price. Suited to mature companies with stable payouts.
Two US-specific points: many growth companies pay no dividend at all, returning cash through buybacks instead, so the formula is meaningless for them; and the US withholds 30% of dividends paid to non-resident investors, so what you receive is less than the headline yield shown here.
Where the numbers come from
Every figure comes from the US Securities and Exchange Commission: financials from the EDGAR XBRL frames API, tickers from the SEC's official ticker list. Works of the US federal government are in the public domain and free to use commercially. This site does not provide prices; ratios are computed from the price you enter:
shares = shares outstanding from the filing's cover page (SEC) book value per share = shareholders' equity ÷ shares revenue per share = TTM revenue ÷ shares EPS = diluted EPS as filed with the SEC (TTM, reconstructed) ROE = TTM net income ÷ shareholders' equity
Dual-class companies (Berkshire A/B, Alphabet GOOGL/GOOG) share one filing, so per-share figures match only one class. If reported EPS does not reconcile with the share count (off by more than 2x), the site derives shares as "TTM net income ÷ EPS" so that book value and revenue per share use the same basis as EPS, and the card reminds you to enter the price of the matching class.
🇺🇸
輸入美股代號或公司名稱開始計算
收錄紐約證交所、那斯達克與 CBOE 掛牌的普通股， 財報與股數來自美國 SEC 的公共領域資料，股價由你自行輸入
Covers common stocks listed on the NYSE, Nasdaq and CBOE. Financials and share counts from US SEC public-domain data; you enter the price.
美股版和台股版哪裡不一樣？
公式本身是同一套財務教科書內容，兩地通用。真正的差別在於 兩個市場公開的東西不一樣，加上本站只用可合法商業利用的資料， 所以有幾處和台股版不同：
| 每股數字的來源 | △ 方向相反 |
| 股價 | ✗ 不提供，由你自行輸入 |
| 倍數的參考基準 | ✗ 沒有歷史區間，使用你自訂的固定倍數 |
| 月營收動能法 | ✗ 美國沒有月營收，改放自由現金流 |
| 本益成長比（PEG） | ✓ 美股才有（SEC 財報有去年全年每股盈餘，台股開放資料沒有） |
| 其他七種公式 | ✓ 完全相同 |
一、每股數字的方向相反
台灣的證交所與櫃買中心每天直接公布每一檔的本益比、股價淨值比 與現金殖利率，所以台股版是「由比率反推每股數字」：
EPS = 收盤價 ÷ 本益比 每股淨值 = 收盤價 ÷ 股價淨值比
美國沒有這種官方統計 —— SEC 只收財報，交易所只有價格。所以美股版反過來， 先從 SEC 的 XBRL 財報取得每股盈餘、股東權益、營收與現金流，再除以股價算出比率：
本益比 = 股價 ÷ 近四季 EPS 股價淨值比 = 股價 ÷ 每股淨值 每股淨值 = 股東權益 ÷ 股數 股數 = SEC 申報的在外流通股數
結果是一樣的四個數字，但美股版多了一層自己算的成分，因此 每張卡片都把中間過程完整列出來，你可以自己核對。
二、不提供股價，也沒有歷史區間
美股的財報可以合法取得（SEC 屬美國聯邦政府，資料為公共領域），但股價沒有可合法 商業利用的免費來源：交易所行情需要另簽授權，而先前使用的那斯達克官網資料， 其使用條款限「個人、非商業用途」並禁止以程式擷取，所以已經停用並刪除。
因此這一頁的股價由你自行輸入（可以照你看到的報價填）。沒有股價時， 各公式照樣算得出對應價，只是無法跟現價比較；填了之後，本益比、股價淨值比、殖利率、 市值與彙總比較才會出現。輸入的股價只存在你的瀏覽器裡。
少了股價，也就無法建立「同業分布」或「歷史區間」這類需要大量股價的倍數基準， 三組倍數一律使用你在「計算參數」中設定的固定值。
三、月營收動能法換成自由現金流法
台灣規定上市櫃公司每月 10 日前公布上月營收，比季報快約 35 天， 是台股最即時的基本面資料 —— 台股版的月營收動能法整個價值就建立在這個「快」上。 美國沒有月營收制度，只有季報。
硬做一個「季營收動能法」不是不行，但推導後會發現它和上面的年化 EPS 完全相同（季數在分子分母消掉了），等於同一個數字算兩次：
推估年營收 × (累計EPS ÷ 累計營收) = (累計營收 ÷ n季 × 4) × (累計EPS ÷ 累計營收) = 累計EPS ÷ n季 × 4 ← 就是年化 EPS
所以那一格改放美股拿得到、而且其他八種公式都看不到的東西 —— 自由現金流：
自由現金流 = 營業活動現金流 − 資本支出
盈餘含折舊攤銷與各種應計項目，現金流是真的收到的錢；兩者長期背離通常值得追究。 這是美股慣用而台股資料較難取得的指標。
還有幾件事要先知道
- 外國公司的 ADR 沒有資料。台積電（TSM）、艾司摩爾（ASML）、 豐田（TM）這類外國企業以美國存託憑證掛牌，依 20-F 用 IFRS 及當地幣別申報，不在 SEC 的 us-gaap 彙總表內。要換算還需要匯率與 ADR 兌換比例（1 單位 ADR 不等於 1 股原股），這兩項沒有可靠的免費來源 —— 與其算出一個看起來合理卻是錯的數字，本站選擇不算， 這類股票不在收錄範圍內。台積電的原股在台股掛牌， 可以到台股頁查 2330。
- 近四季（TTM）是重建出來的。XBRL 有個結構性缺口：公司的 會計年度第四季通常不會單獨標記，年報只揭露全年數字。 直接把四個日曆季相加，大部分公司會少一季。本站改用 「最近完整會計年度 ＋ 該年度結束後的累計 − 去年同期累計」重建， 每一檔的個股卡上都會列出實際用到的期間。
- 庫藏股會壓低淨值。美股大量實施庫藏股，買回的股份直接沖減 股東權益，長期買回的公司帳面淨值偏低、股價淨值比與 ROE 因而偏高， 極端者（如麥當勞、星巴克）股東權益甚至為負，股價淨值比法與 ROE 法 會直接標示不適用。這不是資料錯誤，是這兩個公式碰到了它們的邊界。
- 會計年度不是日曆年。蘋果的年度在 9 月底結束、輝達在 1 月、 微軟在 6 月。卡片上的期間標示是各公司自己的會計期間， 跨公司比較時要留意基準日不同。
- 股利預扣稅。美國對外國投資人的股利預扣 30%， 實收金額低於本頁顯示的帳面殖利率。本頁不做稅務計算。
The formulas themselves are standard finance-textbook material and travel between markets unchanged. The real difference is that the two markets disclose different things, and this site only uses data that can lawfully be used commercially — so a few things differ from the Taiwan edition:
| Where per-share figures come from | △ Derived in the opposite direction |
| Price | ✗ Not provided — you enter it |
| Reference basis for the multiples | ✗ No historical range — your fixed multiples |
| Monthly revenue momentum | ✗ No monthly revenue in the US — replaced by free cash flow |
| PEG ratio | ✓ US only (SEC filings include prior-year EPS; Taiwan open data does not) |
| The other seven formulas | ✓ Identical |
1. Per-share figures are derived in the opposite direction
The Taiwan Stock Exchange and TPEx publish P/E, P/B and dividend yield for every stock, every day, so the Taiwan edition works backwards from those ratios to the per-share figures:
EPS = close ÷ P/E book value per share = close ÷ P/B
The US has no such official statistic — the SEC collects filings, the exchanges have prices, and nobody publishes the ratio. So this edition runs the other way: pull diluted EPS, shareholders' equity, revenue and cash flow from SEC XBRL filings, then divide by price to get the ratios:
P/E = price ÷ TTM EPS P/B = price ÷ book value per share book value per share = shareholders' equity ÷ shares shares = shares outstanding as filed with the SEC
The four numbers end up the same, but this edition computes more of them itself, so every card lays out the full derivation for you to check.
2. No price, and no historical range
US financials can be used lawfully — the SEC is a federal agency and its data is in the public domain — but there is no free price source that may be used commercially: exchange quotes require a separate licence, and the Nasdaq.com data used previously is licensed "solely for your personal, non-commercial use" with automated capture prohibited, so it has been dropped and deleted.
So on this page you enter the price yourself (from whatever quote you see). Without one, all the formulas still produce their prices; you just cannot compare against the market. Once you enter one, P/E, P/B, yield, market cap and the summary comparison appear. The price you type stays in your browser.
Without prices there is also no way to build a peer distribution or a historical range, so the three multiples always use the fixed values you set under "Assumptions".
3. Monthly revenue momentum becomes free cash flow
Taiwan requires listed companies to publish monthly revenue by the 10th of the following month — about 35 days ahead of the quarterly report, and the most timely fundamental data available there. The whole value of the Taiwan edition's monthly-revenue method rests on that speed. The US has no monthly revenue requirement, only quarterly filings.
You could build a "quarterly revenue momentum" method, but work through the algebra and it turns out identical to the forward P/E above — the quarter count cancels, so it is the same number computed twice:
annualized revenue × (cumulative EPS ÷ cumulative revenue) = (cumulative revenue ÷ n quarters × 4) × (cumulative EPS ÷ cumulative revenue) = cumulative EPS ÷ n quarters × 4 ← that is the forward EPS
So that slot holds something the US does disclose and the other eight formulas cannot see — free cash flow:
free cash flow = cash flow from operations − capital expenditure
Earnings contain depreciation, amortisation and every kind of accrual; cash flow is money actually received. A persistent gap between the two is usually worth investigating. This is a standard US metric and one that is harder to assemble for Taiwan.
A few things to know first
- Foreign companies' ADRs have no data. TSMC (TSM), ASML and Toyota (TM) list as American Depositary Receipts and file on Form 20-F under IFRS in their own currency, so they are absent from the SEC's us-gaap frames. Converting would also require an exchange rate and the ADR ratio (one ADR is not one ordinary share), and neither has a reliable free source — rather than produce a plausible-looking but wrong number, this site declines to compute one; such stocks are not included. TSMC's ordinary shares trade in Taipei, so you can look up 2330 on the Taiwan page.
- Trailing twelve months is reconstructed. XBRL has a structural gap: a company's fiscal fourth quarter is usually not tagged on its own, because the annual report only discloses the full year. Add four calendar quarters together and most companies come up one quarter short. This site rebuilds it as "latest complete fiscal year + year-to-date since it ended − the same period a year earlier", and each stock card lists the exact periods used.
- Buybacks depress book value. US companies repurchase shares heavily, and repurchases are charged directly against shareholders' equity. Companies that have bought back stock for years carry a low book value, which inflates P/B and ROE; in extreme cases (McDonald's, Starbucks) equity is negative and the P/B and ROE methods simply report that they do not apply. That is not a data error — it is those two formulas hitting their limits.
- Fiscal years are not calendar years. Apple's ends in late September, NVIDIA's in January, Microsoft's in June. The period labels on each card refer to that company's own fiscal periods, so watch the reference dates when comparing across companies.
- Dividend withholding. The US withholds 30% of dividends paid to non-resident investors, so what you actually receive is less than the headline yield shown here. No tax calculation is performed on this page.