financial-analysis-dcf

DCF Valuation

Estimate intrinsic value per share using a discounted cash flow model. Data from SEC EDGAR (financials), FRED (risk-free rate), and Yahoo Finance (price, beta).

Setup

No dependencies required. All scripts use Python standard library only.

Workflow

Step 1 — Run the DCF script

bash run.sh <SYMBOL>
# With custom assumptions:
bash run.sh <SYMBOL> --terminal-growth 3.0 --fcf-growth 15 --projection-years 7

The script:

1. Fetches 5 years of 10-K data from SEC EDGAR
2. Computes unlevered FCF = OCF − CapEx + InterestExpense × (1−T)
3. Derives historical FCF growth rate (or uses override)
4. Gets risk-free rate from FRED (10Y Treasury)
5. Gets beta and current price from Yahoo Finance
6. Calculates WACC = Ke × E/V + Kd × (1−T) × D/V
7. Projects FCFs, computes terminal value (Gordon Growth Model)
8. Generates 5×5 sensitivity table (WACC vs terminal growth)

Step 2 — Section 1: Summary

First output must be this summary box:

[TICKER] — DCF VALUATION

Intrinsic value:  $XXX.XX  |  Current price: $XXX.XX  |  ▲/▼ XX.X%
WACC:             X.X%     |  Terminal growth: X.X%   |  FCF base: $X.XB

Step 3 — Section 2: Sensitivity Table

Show the full 5×5 sensitivity table from the JSON output:

Terminal Growth \\ WACC  |  8.2%   |  9.2%   |  10.2%
------------------------|---------|---------|--------
1.5%                    | $198.4  | $176.2  | $158.1
2.5%                    | $221.3  | $194.5  | $172.6
3.5%                    | $251.7  | $218.0  | $191.4

Display N/A for null values (WACC ≤ terminal growth).

Step 4 — Section 3: Interpretation

Note key assumptions and their impact:

• Which inputs drive the most variance (usually WACC and growth rate)
• Whether the beta seems reasonable for the company
• How the historical FCF growth rate compares to analyst expectations
• Any red flags (negative FCF years, volatile cash flows, high debt)

Disclaimer: DCF output is highly sensitive to WACC and growth assumptions. It should be used as one input among many, not as a definitive price target. Always present the sensitivity table alongside the point estimate.

---

Model Assumptions

• ERP: 5.5% (Damodaran market average)
• Tax rate: derived from SEC filings (IncomeTaxExpense / PreTaxIncome); fallback 21%
• Cost of debt: derived from SEC filings (InterestExpense / LongTermDebt); fallback 4%
• FCF: Unlevered = OCF − CapEx + InterestExpense × (1−T) from most recent 10-K
• FCF growth: per-share UFCF CAGR from historical 10-K data (accounts for buybacks); fallback 3% if insufficient data
• Beta: OLS regression on daily returns vs SPY (~5 years); Blume adjustment applied (adjusted = 0.67 × raw + 0.33)
• Shares outstanding: from SEC EDGAR; fallback to market cap / price
• Mid-year convention: FCFs and terminal value both discounted at mid-year

Formatting Rules

• Intrinsic value and current price: "$XXX.XX"
• Upside/downside: "▲ +XX.X%" or "▼ −XX.X%"
• WACC and growth rates: one decimal, e.g. "9.2%"
• FCF: B or M, e.g. "$12.3B"

When NOT to Use

• Negative or volatile FCF: companies with negative free cash flow cannot be valued via DCF — the model requires a positive base FCF to project forward
• Banks and financial institutions: revenue and cash flow structures differ fundamentally (interest income vs. operating revenue); use price-to-book or dividend discount models instead
• Insurance companies: similar to banks — earnings driven by underwriting and investment income, not operating cash flow
• Pre-revenue or early-stage companies: no meaningful FCF history to extrapolate

Limitations

• US stocks only: SEC EDGAR data for US-listed equities
• Backward-looking: uses historical financials — future may differ significantly
• No analyst estimates: growth rates extrapolated from historical data unless user provides one