Equities

JoelLewis/finance_skills/plugins/wealth-management/skills/equities

作者 JoelLewis5c498eacf7057e31238c4c5a8012a1afe9ec7c8a無授權條款收錄於 2026年10月9日更新於 2026年10月9日

Analyze equity securities, factor models, and equity portfolio construction. Use when the user asks about stocks, equity valuation ratios, index construction methods, or style analysis. Also trigger when users mention 'P/E ratio', 'growth vs value', 'market cap weighting', 'sector allocation', 'GICS classification', 'earnings per share', 'Fama-French factors', 'CAPM', 'dividend yield', 'PEG ratio', 'EV/EBITDA', or ask which factors explain equity returns.

AI 產生的概覽

指導股票分析:選擇估值指標、指數編製方法以及因子與風格背景。

功能
此技能提供一套股票證券分析的決策流程,涵蓋哪種估值指標適合哪類產業與資本結構、哪種指數編製方法適合哪種投資授權,以及分步的證券分析順序。它包含 EV/EBITDA、PEG、戈登成長模型和 CAPM 等關鍵公式,並附有一個工業公司估值的完整範例和常見陷阱。隨附的 Python 指令碼會輸出估值指標、以合成資料進行的因子迴歸和產業集中度分析,並可對照範例驗證示範輸出。
適用情境
當使用者詢問股票、股票估值比率、指數編製方法或風格分析時使用。也適用於提到本益比、成長與價值、市值加權、產業配置、GICS 分類、每股盈餘、Fama-French 因子、CAPM、股息殖利率、PEG 比率、EV/EBITDA,或詢問哪些因子能解釋股票報酬的問題。
執行需求
需要 Python 和 uv 執行器(或已安裝 numpy 的 python3)來執行隨附指令碼 scripts/equities.py;該指令碼使用 PEP 723 內嵌相依套件。未說明需要憑證或網路存取。

Equities

This skill is a decision procedure: which valuation metric to use for which company, which index methodology fits which mandate, and the order of operations for analyzing a stock. It assumes the user can look up definitions; the value here is choosing the right tool.

Core Concepts

Choosing the Valuation Metric

Match the metric to the sector and capital structure — using the wrong one is the most common equity-analysis error.

SituationUseAvoidWhy
Financials (banks, insurers)P/B, P/TBV, ROE vs P/BEV/EBITDADebt is raw material, not financing — EV and EBITDA are meaningless; book value is marked closer to fair value
Capital-intensive (industrials, telecom, energy)EV/EBITDA, EV/EBITP/E aloneNeutralizes depreciation policy and leverage differences across peers
Mature dividend payers (utilities, staples)Dividend yield + payout sustainability, P/EPEGGrowth is low and stable; income and coverage matter most
High-growth, low/no earningsEV/Sales, PEG (if earnings exist), unit economicsP/E, P/BEarnings are depressed by reinvestment; book value is mostly intangibles
Cyclicals (autos, semis, materials)Mid-cycle or normalized P/E, P/B at troughSpot P/EP/E is lowest at the cycle peak and highest at the trough — spot P/E inverts the buy/sell signal
Negative earnings, positive cash flowEV/EBITDA, P/FCFP/E, earnings yieldRatio is undefined or misleading with negative denominator
REITs and listed real estateP/FFO, P/AFFO, NAVP/EGAAP depreciation distorts earnings for property — handled in detail by the real-assets skill
Cross-border / different leverageEV-based multiplesEquity multiplesEnterprise value normalizes for capital structure

Cross-checks that apply everywhere:

  • Use forward (next-12-month) estimates for the numerator decision when the business is changing; trailing figures when estimate quality is poor.
  • Compare against the company's own history and a true peer set, not the whole market.
  • Translate any multiple into its implied assumptions (growth, margin, required return) before declaring cheap/expensive — a low multiple usually encodes a real problem.

Choosing the Index Methodology

MandateMethodologyTrade-off to flag
Cheap, tax-efficient market exposureCap-weighted (S&P 500, total market)Momentum-chasing by construction; concentration in mega-caps — a single sector can exceed 30%
Reduce concentration / small-cap tiltEqual-weightedHigher turnover and rebalancing cost; structural size and contrarian tilt
Break the price-weight linkFundamental-weighted (revenue, earnings, book)Effectively a value tilt with extra steps; compare cost vs an explicit value fund
Explicit factor exposureFactor/style index (value, momentum, quality, low vol)Verify the factor definition and rebalance rules; factor timing rarely works
AvoidPrice-weighted (DJIA-style)Weight proportional to share price is economically arbitrary — legacy only

Selection rules: default to cap-weighted for core beta; add equal- or fundamental-weighted only when the user explicitly wants the embedded tilt and accepts the turnover; treat any "smart beta" product as a factor portfolio and evaluate its factor loadings, not its marketing name.

Security Analysis Sequence

  1. Classify the business — sector (GICS or equivalent), cyclical vs defensive, capital intensity, leverage. This determines the valuation toolkit (table above).
  2. Quality screen — revenue trend, margin trend, ROIC vs cost of capital, balance-sheet risk (net debt/EBITDA, interest coverage), share count trajectory (dilution vs buybacks).
  3. Earnings basis — pick trailing vs forward EPS, check for one-offs, use diluted share count. For cyclicals, normalize to mid-cycle.
  4. Value with the matched metric — primary multiple from the table, one cross-check multiple, and where dividends are central a dividend-based check (Gordon growth: P = D1 / (r - g), valid only when g < r).
  5. Factor and style context — regress (or eyeball) exposures to market beta, size, value, momentum, quality. Distinguish stock-specific thesis from a factor bet you could buy more cheaply via an index.
  6. Portfolio fit — marginal effect on sector concentration and factor tilts; total return (price + dividends) is the comparison basis, never price return alone.

Key Formulas

FormulaExpressionUse Case
EV/EBITDA(Market Cap + Debt - Cash) / EBITDACapital-structure-neutral valuation
Earnings YieldEPS / PriceCompare equity vs bond yields
PEG(P/E) / Earnings Growth Rate (in %)Growth-adjusted valuation
Gordon GrowthP = D1 / (r - g)Dividend-based intrinsic value
CAPME(R) = R_f + beta × (E(R_m) - R_f)Required return input for valuation
Total ReturnPrice Return + Dividend ReturnPerformance comparison basis

Worked Examples

Metric Selection and Valuation

Given: An industrial company with market cap $500M, total debt $100M, cash $50M, EBITDA $75M, EPS $7.50, price $150. Decide and calculate:

  1. Capital-intensive industrial → primary metric is EV/EBITDA (table above), with P/E as cross-check.
  2. EV = $500M + $100M - $50M = $550M. EV/EBITDA = $550M / $75M = 7.33x.
  3. Cross-check: P/E = $150 / $7.50 = 20.0x; earnings yield = 7.50 / 150 = 5.0%.
  4. Interpretation: 7.33x EV/EBITDA is modest for an industrial if margins are stable — compare against the peer set and the company's own 5-10 year range. The 20x P/E looks richer than the EV multiple because the company carries little net debt; the EV multiple is the better cross-peer comparison.

Common Pitfalls

  • Applying EV/EBITDA to banks or P/E to REITs — metric/sector mismatch is the dominant error this skill exists to prevent
  • Buying cyclicals on low trailing P/E at the cycle peak (the "value trap" inversion)
  • Treating a fundamental-weighted or smart-beta index as alpha rather than a packaged factor tilt
  • Confusing price return with total return — dividends compound to a large share of long-run equity returns
  • Survivorship bias in backtested factor or screen results

Cross-References

  • historical-risk (wealth-management plugin): volatility and drawdown measurement for equity return series
  • statistics-fundamentals (core plugin): beta estimation via CAPM regression
  • performance-metrics (wealth-management plugin): Sharpe ratio and related risk-adjusted return measures
  • fund-vehicles (wealth-management plugin): equity fund selection (ETFs, mutual funds, SMAs)
  • currencies-and-fx (wealth-management plugin): international equity currency effects
  • asset-allocation (wealth-management plugin): equity allocation within multi-asset portfolios
  • real-assets (wealth-management plugin): REIT valuation (P/FFO, NAV) is owned by that skill
  • qualitative-valuation (wealth-management plugin) and quantitative-valuation (wealth-management plugin): deeper single-company valuation workflows
  • financial-statements (wealth-management plugin): EBITDA, free cash flow, ROIC, and margin analysis underpinning fundamental stock selection
  • equity-compensation (wealth-management plugin): employer stock acquired through RSUs, options, and ESPPs carries equity risk plus tax and insider-trading constraints
  • factor-investing (wealth-management plugin): the factor-loading evaluation of smart-beta and style products prescribed above lives in that skill

Running the Script

bash
uv run scripts/equities.py            # run the demo (uses PEP 723 inline deps)uv run scripts/equities.py --verify   # check demo outputs against the worked example (exit 1 on mismatch)python3 scripts/equities.py            # alternative (requires: pip install numpy)

The demo prints valuation metrics (including the worked example's EV/EBITDA and earnings yield), a factor regression on synthetic data, and sector concentration analysis. Run --help for a list of the classes and functions. For programmatic use, import the module rather than running it — the demo only executes under python equities.py.

來源與署名

來源:JoelLewis/finance_skills位於plugins/wealth-management/skills/equities提交5c498ea

授權條款: 無授權條款

內容歸原作者所有。SourceWeft 從公開儲存庫中收錄這些內容。

檢舉或申請下架