Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply the Modigliani-Miller theorem to analyze capital structure decisions and identify when financing choices affect firm value. Use this skill when the user needs to evaluate debt-equity tradeoffs, assess the impact of leverage on firm value, understand tax shield benefits, or when they ask 'does capital structure matter', 'should we take on more debt', or 'what is the optimal leverage ratio'.
.claude/skills/asgard-ai-platform-grad-mm-theorem/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -6% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -66% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 22% | 0% |
The Modigliani-Miller theorem (1958) establishes that in perfect capital markets, firm value is independent of capital structure. This irrelevance result serves as the benchmark — every real-world reason capital structure matters is a violation of MM's assumptions.
IRON LAW: MM irrelevance holds ONLY in perfect markets — every
real-world deviation (taxes, bankruptcy costs, agency costs) makes
capital structure matter. MM is the null hypothesis, not the answer.Key assumptions (for irrelevance):
For each deviation, assess its magnitude:
Optimal capital structure balances marginal tax shield benefit against marginal bankruptcy and agency costs.
WACC = (E/V)Re + (D/V)Rd(1-Tc). Optimal structure minimizes WACC.
markdown## Capital Structure Analysis: [Firm] ### Current Structure | Metric | Value | |--------|-------| | Debt (D) | $X | | Equity (E) | $X | | D/E Ratio | x.xx | | WACC | x% | ### MM Imperfections Present | Imperfection | Magnitude | Direction | |-------------|-----------|-----------| | Tax shield | [high/medium/low] | Favors debt | | Bankruptcy costs | [high/medium/low] | Favors equity | | Agency costs | [high/medium/low] | [depends] | ### Recommendation - [Optimal direction of adjustment with reasoning]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,958 | 15,211 | -10% | 1 | 1 | 0% | 2,925 | 2,916 | -0% | 0 | 0 | — |
case-02 | fail→pass | 22,525 | 15,007 | -33% | 1 | 1 | 0% | 3,590 | 3,378 | -6% | 0 | 0 | — |
case-03 | fail→pass | 22,046 | 13,056 | -41% | 1 | 1 | 0% | 3,724 | 2,954 | -21% | 0 | 0 | — |
case-04 | fail→pass | 37,854 | 10,356 | -73% | 1 | 1 | 0% | 8,267 | 2,799 | -66% | 0 | 0 | — |
case-05 | fail→fail | 28,184 | 14,790 | -48% | 1 | 1 | 0% | 5,053 | 3,304 | -35% | 0 | 0 | — |
case-06 | fail→pass | 22,202 | 21,316 | -4% | 1 | 1 | 0% | 3,640 | 4,455 | +22% | 0 | 0 | — |
case-07 | fail→fail | 38,531 | 36,758 | -5% | 1 | 1 | 0% | 8,232 | 9,164 | +11% | 0 | 0 | — |
case-08 | pass→pass | 16,621 | 15,295 | -8% | 1 | 1 | 0% | 2,560 | 3,277 | +28% | 0 | 0 | — |
case-09 | pass→pass | 5,229 | 8,249 | +58% | 1 | 1 | 0% | 959 | 2,327 | +143% | 0 | 0 | — |
case-10 | pass→pass | 9,727 | 10,200 | +5% | 1 | 1 | 0% | 2,056 | 2,732 | +33% | 0 | 0 | — |
case-11 | pass→pass | 6,261 | 5,960 | -5% | 1 | 1 | 0% | 1,099 | 2,106 | +92% | 0 | 0 | — |
case-12 | pass→pass | 6,393 | 5,659 | -11% | 1 | 1 | 0% | 1,474 | 2,211 | +50% | 0 | 0 | — |
case-13 | pass→pass | 12,869 | 13,751 | +7% | 1 | 1 | 0% | 1,998 | 3,145 | +57% | 0 | 0 | — |
case-14 | pass→pass | 10,944 | 9,443 | -14% | 1 | 1 | 0% | 1,823 | 2,835 | +56% | 0 | 0 | — |
case-15 | pass→pass | 13,583 | 13,194 | -3% | 1 | 1 | 0% | 2,283 | 3,237 | +42% | 0 | 0 | — |
case-16 | pass→pass | 11,555 | 12,926 | +12% | 1 | 1 | 0% | 1,841 | 2,657 | +44% | 0 | 0 | — |
case-17 | pass→pass | 11,750 | 10,308 | -12% | 1 | 1 | 0% | 1,927 | 2,564 | +33% | 0 | 0 | — |
case-18 | pass→pass | 18,178 | 18,533 | +2% | 1 | 1 | 0% | 2,795 | 3,776 | +35% | 0 | 0 | — |
case-19 | fail→fail | 13,498 | 13,667 | +1% | 1 | 1 | 0% | 2,281 | 3,198 | +40% | 0 | 0 | — |
case-20 | pass→pass | 14,588 | 15,984 | +10% | 1 | 1 | 0% | 2,126 | 3,365 | +58% | 0 | 0 | — |
case-21 | pass→pass | 17,784 | 18,569 | +4% | 1 | 1 | 0% | 2,691 | 3,927 | +46% | 0 | 0 | — |
case-22 | pass→pass | 13,571 | 10,862 | -20% | 1 | 1 | 0% | 2,029 | 2,770 | +37% | 0 | 0 | — |
case-23 | pass→pass | 21,274 | 12,504 | -41% | 1 | 1 | 0% | 3,711 | 2,727 | -27% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted. The headline lift of +22 percentage points is the difference between those two pass rates over the 23 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.