financial-frameworks

name	financial-frameworks
description	Deterministic financial tools — WACC, NPV, IRR, payback, EVA, real-options, Monte Carlo, sensitivity, covenant/liquidity checks. The Financial Hardcoding Directive.
allowed-tools	["Read","Write","Glob","Grep"]

Financial Frameworks — The Hardcoding Directive

Per the research doc (Section "Financial Framework Hardcoding Directive"), corporate finance discipline must be embedded as first-class deterministic tools and constraints, not informal guidance. Agents (cfo, mna-cockpit, capital-allocation, crisis-warroom) must invoke the relevant tool, log inputs and assumptions, and surface guardrail breaches before recommending action.

Each tool below has: Purpose · Inputs · Formula · When to use · Example · Pitfalls.

---

1. Weighted Average Cost of Capital (WACC)

• Purpose: discount rate for all cash-flow valuations
• Inputs: market value of equity (E), market value of debt (D), cost of equity Re, cost of debt Rd (pre-tax), marginal tax rate t
• Formula: WACC = (E/V)·Re + (D/V)·Rd·(1 − t) where V = E + D
• Re: typically CAPM: Re = Rf + β·(Rm − Rf) (Rf = risk-free; β = equity beta; Rm − Rf = market risk premium)
• When to use: any project / acquisition / asset valuation
• Example: E=$800M, D=$200M, Re=10%, Rd=5%, t=25% → WACC = 0.8·10% + 0.2·5%·0.75 = 8.75%
• Pitfalls: (1) using book values instead of market; (2) holding D/V constant when transaction changes capital structure (use target structure); (3) cross-border without country-risk adjustment; (4) negative-equity / pre-revenue companies — use comparable firm's WACC or sector cost-of-capital

2. Net Present Value (NPV)

• Purpose: aggregate value of an investment's discounted cash flows
• Inputs: cash flow stream CF_t for t = 0..n; discount rate r (typically WACC + risk premium)
• Formula: NPV = Σ(t=0..n) CF_t / (1+r)^t
• Decision rule: NPV > 0 accept; NPV < 0 reject; for mutually-exclusive, pick highest NPV
• When to use: any project with multi-period cash flows
• Example: I₀ = −$100, CF₁..₅ = $30 each, r = 10% → NPV = $13.7M
• Pitfalls: (1) using nominal cash flows with real discount rate (or vice versa); (2) ignoring terminal value when relevant; (3) double-counting financing costs (cash flows are unlevered FCF when WACC is the discount rate); (4) cash flows not risk-adjusted to discount rate

3. Internal Rate of Return (IRR)

• Purpose: the rate at which NPV = 0
• Inputs: cash flow stream
• Formula: solve for r such that Σ CF_t / (1+r)^t = 0
• Decision rule: IRR ≥ hurdle rate (typically WACC + risk premium)
• When to use: sanity-check returns; communicating to non-finance stakeholders
• Pitfalls: (1) multiple IRRs when cash flows change sign more than once — use NPV instead; (2) IRR alone misleads for mutually-exclusive projects of different scale — pick by NPV; (3) reinvestment assumption (MIRR fixes); (4) IRR is silent on size

4. Payback & Discounted Payback

• Purpose: time to recover initial investment
• Inputs: cash flow stream
• Formulas: payback = years until cumulative CF ≥ I₀; discounted payback uses discounted CF
• When to use: liquidity-constrained context, complementary to NPV (never primary)
• Pitfalls: (1) ignores cash flows after payback; (2) ignores time value (use discounted variant); (3) inadequate as sole decision criterion

5. Profitability Index (PI)

• Purpose: NPV per dollar invested — for capital rationing
• Formula: PI = (NPV + I₀) / I₀ (or equivalently PV(CF) / I₀)
• Decision rule: PI > 1 accept; for rationing, rank by PI
• When to use: when capital is constrained and projects are divisible / independent

6. Economic Value Added (EVA)

• Purpose: operating-period value creation above cost of capital
• Formula: EVA = NOPAT − (WACC × Invested Capital)
• NOPAT = EBIT × (1 − t)
• When to use: ongoing-business performance, executive compensation linkage
• Pitfalls: (1) invested-capital base must be consistent (book vs adjusted); (2) NOPAT volatility distorts year-over-year reads; (3) ignores future growth options

7. Real-Options Valuation

7a. Binomial Lattice (walked example)

• Purpose: value managerial flexibility (defer, expand, abandon, stage)
• Inputs: underlying value S₀; volatility σ; risk-free rate Rf; time T; exercise value X; up factor u; down factor d; risk-neutral probability p
• Formulas:
  • u = e^(σ√Δt), d = 1/u
  • p = (e^(Rf·Δt) − d) / (u − d)
  • At each node, option value = max(intrinsic, e^(−Rf·Δt) · [p·V_up + (1−p)·V_down])
• When to use: staged investments, abandonment options, expansion options on existing capacity
• Example: S₀ = $100M, σ = 30%, T = 2 yr, X = $80M (abandon for salvage), Rf = 4%; 2-step lattice yields option value substantially higher than naive NPV if downside is severe

7b. Black-Scholes Sketch (call-option analogy)

• C = S·N(d₁) − X·e^(−Rf·T)·N(d₂)
• d₁ = [ln(S/X) + (Rf + σ²/2)·T] / (σ√T), d₂ = d₁ − σ√T
• Useful for fast first-cut on growth-option value (e.g., R&D pipeline)

7c. Monte Carlo for Path-Dependent

When optimal-exercise depends on path (Least-Squares Monte Carlo, Longstaff-Schwartz), simulate underlying and regress to estimate continuation value.

• Pitfalls (all real-options methods): (1) volatility estimate is the dominant assumption — sensitivity-test it; (2) modeling managerial flexibility you won't actually exercise (over-optimism); (3) double-counting (don't add option value to a DCF that already implicitly captures flexibility)

8. Monte Carlo Scenario Engine

• Purpose: distribution of outcomes when key drivers are uncertain
• Inputs: distribution per driver (normal, lognormal, triangular, beta); correlations; 10k+ iterations
• When to use: any decision where downside tail matters or where multiple uncertain drivers interact
• Output: P5 / P25 / P50 / P75 / P95 of NPV (or any metric); plus probability of value-destroying outcome
• Pitfalls: (1) ignoring correlations (revenue and margin are not independent); (2) garbage-in / garbage-out on distributions — fit to history; (3) reporting only the mean (reading P5 matters more for guardrails)

9. Sensitivity Analysis & Tornado

• Purpose: rank assumptions by impact on outcome
• Method: one-variable-at-a-time swings (±10%, ±25%) around base; rank by NPV delta
• When to use: every material analysis — name the top 5 break-the-case variables
• Pitfalls: one-at-a-time misses interactions (Monte Carlo complements)

10. Capital-Budgeting Decision Tree

Strategic fit?
└── No → reject (archive with rationale)
└── Yes → Run NPV at WACC + risk premium
          ├── NPV < 0 → reject (unless real-option value justifies)
          └── NPV ≥ 0 → Check guardrails
                        ├── Any guardrail fail → restructure or board override
                        └── All guardrails pass → Run Monte Carlo
                                                  ├── P5 catastrophic → restructure / decline / hedge
                                                  └── P5 acceptable → Recommend (full / staged / conditional)

11. Capital-Rationing Prioritization

When capital is constrained: rank by Profitability Index (NPV per $ invested), subject to:

• Mandatory projects (compliance, safety) funded first
• Strategic enabler / option-value projects funded next (even at lower PI) when they unlock larger downstream optionality
• Discretionary funded by PI rank until capital exhausted

12. Covenant & Leverage Checker

Run before any material capital action:

Ratio	Formula	Typical covenant	Action if within 20% of limit
Net debt / LTM EBITDA	(Total Debt − Cash) / EBITDA	≤ 3.0–4.5x sector-dependent	Flag; alternative-action review
Interest coverage	EBITDA / Interest expense	≥ 3.0x	Flag
Fixed-charge coverage	(EBITDA − Capex) / (Interest + Lease)	≥ 1.5x	Flag
DSCR (debt service coverage)	NOI / Total debt service	≥ 1.25x	Flag

13. Liquidity Stress

Scenario	Revenue Δ	Action	Cash runway target
Base	0	Plan	≥ 12 mo
Mild	−10%	Operating actions only	≥ 9 mo
Moderate	−25%	Capex freeze + cost-out	≥ 6 mo
Severe	−50%	Working-capital tightening + hedge	≥ 3 mo
Catastrophic	−75% / event	Activate crisis-warroom	survive

14. Hard Guardrails Reference Table

Guardrail	Default	Override path
Hurdle rate	WACC + risk-class premium	CEO + Board with documented thesis
Net leverage	≤ 3.0x net debt/EBITDA	Board only
Covenant headroom	≥ 20% post-action	Board only
Cash runway	≥ 12 mo base / ≥ 6 mo stress	Board only
Counterparty concentration	< 10% recv/treasury	CFO sign-off
Sanctions / prohibited	Zero exposure	NEVER — non-overrideable
ESG / climate plan	No material breach	Board + Chief Sustainability sign-off

Required Logging Discipline

Every tool invocation logs: tool, inputs (named), assumptions, output, decision, time, author. This is non-negotiable for audit and AI-governance traceability (Research Doc Governance section + EU AI Act Article 9).