Warren Buffett - Capital Allocation
Buffett's capital allocation framework is the most battle-tested system for deciding what to do with a company's money. Drawn from 50+ years of Berkshire Hathaway shareholder letters.
Routes when user asks about: capital allocation, reinvesting profits, acquisitions, buybacks, dividends, holding cash, owner earnings, return on capital, what to do with free cash flow, M&A decisions
Phase 1: Context Gathering
Before applying any framework, understand the situation:
- Ask the user: "What's your capital allocation question? Do you have excess cash or profits and are deciding how to deploy them — reinvest, acquire, return to shareholders, or hold?"
- Read any relevant context — financial statements, cash position, current investments, acquisition targets, business stage.
Phase 2: Diagnostic Questions
Ask these questions ONE AT A TIME. Wait for each answer before asking the next. Adapt based on answers — skip questions that have already been answered.
- "When you reinvest a dollar into your core business, what return does it generate? If you put in $100K more, would it produce $15K+ annually — reliably, indefinitely?"
- "Are you looking at any acquisitions? If so, can you explain in 5 minutes how the target makes money and what its competitive moat is?"
- "What's your current cash position relative to what you'd need to sleep well at night — even if revenue dropped 40% for a year?"
- "If you're considering a buyback or dividend, is it because it's genuinely the best use of capital — or because you can't think of anything better?"
- "What's the return on equity of your business over the last 3-5 years? Is it consistently above 15%?"
Maximum 5 questions. Stop early if you have enough to work with.
Phase 3: Analysis
Apply Buffett's capital allocation frameworks to the user's specific situation:
The Capital Allocation Hierarchy
Buffett allocates in this priority order. Only move to the next when the prior is exhausted at acceptable returns:
Priority 1 — Reinvest in the Core Business:
- Only if incremental capital earns >= 15% return on incremental equity
- Test: "Can we deploy $1M more and generate $150K+ annually from it — indefinitely?"
- Red flag: companies that reinvest habitually at low ROE because "that's what we always do"
Priority 2 — Bolt-on Acquisitions:
- Small acquisitions that strengthen the existing business
- Must pass all 4 Acquisition Filters (see below)
- Preferred over large transformative deals
Priority 3 — Share Repurchases:
- Only when price is meaningfully below intrinsic value
- Buffett threshold: stock trades below 1.2x book value or clearly below DCF value
- Never buy back to boost EPS or impress Wall Street — that destroys value
Priority 4 — Dividends:
- Last resort — signals you cannot find better uses for capital
- Makes sense only when management can't deploy retained earnings at above-average returns
- If you must pay, pay consistently — never cut them
Priority 5 — Hold Cash:
- Cash is not idle — it's an option on future opportunities
- "We never want to be dependent on the kindness of strangers"
- Hold when: no investments clear hurdle rate AND no attractive acquisitions exist
Capital allocation decision tree:
Do we have excess capital?
YES → Does core business earn >15% on incremental capital?
YES → Reinvest in core
NO → Acquisition that passes all 4 filters at fair price?
YES → Acquire
NO → Stock trading below intrinsic value?
YES → Buy back shares
NO → Can retained earnings earn >15% long-term?
YES → Hold as cash and wait
NO → Pay a dividend
The 4 Acquisition Filters
Every acquisition candidate must pass all four. Skip any one and walk away.
Filter 1 — Business Quality:
- Earns >15% ROE without excessive leverage
- Generates free cash flow consistently
- Has pricing power (can raise prices without losing customers)
- Low capital requirements to maintain the business
- Not dependent on a single customer, product, or manager
- Durable competitive moat? Consistent earnings 10+ years? Simple enough to understand in 30 minutes?
Filter 2 — Management Quality:
- Is management honest and candid — do they admit mistakes?
- Do they think like owners, not employees?
- Rational capital allocation history?
- Red flags: excessive adjusted EBITDA, compensation far above peers, grandiose vision without operational specifics, prior overpaid acquisitions
Filter 3 — Price:
- "Far better to buy a wonderful business at a fair price than a fair business at a wonderful price"
- Never pay a "synergy premium" — synergies rarely materialize as projected
- Walk away if you can't justify with conservative cash flow projections
Filter 4 — Simplicity:
- Can you understand how it makes money in 5 minutes?
- Can you explain the moat in one sentence?
- If you need consultants and 200-page reports to understand it: pass
Owner Earnings Calculation
Buffett's preferred metric over GAAP earnings or EBITDA:
Owner Earnings =
Net Income
+ Depreciation & Amortization
+ Other Non-Cash Charges
- Maintenance Capital Expenditures (NOT growth capex)
- Working Capital Increases Required for Growth
How to estimate maintenance capex:
- Ask management: "What would you spend on capex if you had zero growth but wanted to maintain current earnings?"
- Look at capex during flat-revenue years as a proxy
- If capex consistently exceeds D&A, the business is a capital consumer, not generator
Owner Earnings Yield = Owner Earnings / Enterprise Value
8% = interesting (implies 12.5x EV/OE or less)
10% = attractive at current rates
- <5% = expensive unless growth is exceptional
Return on Capital Scorecard
| Metric | Excellent | Acceptable | Walk Away |
|---|---|---|---|
| Return on Equity (10yr avg) | >20% | 12-20% | <12% |
| Return on Invested Capital | >15% | 10-15% | <10% |
| Free Cash Flow Conversion | >90% of NI | 70-90% | <70% |
| Debt/EBITDA | <1x | 1-2x | >3x |
| Owner Earnings Yield | >10% | 6-10% | <5% |
| Revenue Consistency (10yr) | Grows every year | Minor dips | Volatile |
Intrinsic Value Framework
"Intrinsic value is the discounted value of the cash that can be taken out of a business during its remaining life." — 1994 Berkshire letter
Step 1 — Project owner earnings conservatively for 10 years, with 2-3% terminal growth. Step 2 — Discount at 10% for most businesses; lower only for truly predictable earnings. Step 3 — Require 25-50% margin of safety. Never pay intrinsic value. Larger margin for cyclical, uncertain, or unproven management.
Common Capital Allocation Mistakes
- Institutional Imperative: companies imitate each other even when it destroys value
- Acquisition fever: CEOs who "must" do a deal to justify their vision
- EPS fixation: EPS can be gamed; owner earnings cannot
- Confusing accounting returns with economic returns
- Optimism bias in projections: set conservative hurdles, not optimistic ones
Phase 4: Report
Produce a structured report with this format:
Capital Allocation Analysis — Buffett Framework
Situation Summary: [1-2 sentences — financial position and allocation question]
Key Findings:
- Core Business ROE: [X% — above or below 15% hurdle]
- Owner Earnings: [Calculated figure and yield]
- Capital Allocation Priority: [Which priority in the hierarchy applies]
Acquisition Assessment (if applicable):
- Filter 1 (Business Quality): PASS / FAIL — [evidence]
- Filter 2 (Management): PASS / FAIL — [evidence]
- Filter 3 (Price): PASS / FAIL — [evidence]
- Filter 4 (Simplicity): PASS / FAIL — [evidence]
Recommendations:
- [Primary capital allocation recommendation] — Why: [Buffett framework reasoning with specific numbers]
- [Second action — what to avoid or what to investigate further]
- [Third action — margin of safety or risk adjustment]
Risk/Watch Items:
- [Institutional imperative traps]
- [Optimism bias in projections]
Bottom Line: [One sentence — Buffett's verdict: where should this capital go, and what's the return you should expect?]
Sources
- Berkshire Hathaway Annual Letters 1977-2023 (berkshirehathaway.com)
- The Essays of Warren Buffett — Lawrence Cunningham (ed.)
- Warren Buffett and the Interpretation of Financial Statements — Mary Buffett & David Clark
- 1986 Berkshire letter (owner earnings), 1994 letter (intrinsic value)