Defensive Investor's Manual — Ch. 5: Computing a Margin of Safety
阅读中文版 (with Audio)Turning margin of safety from a concept into three computable forms — earnings yield vs bond yield, reverse DCF, and the portfolio-level margin that matters most to a retiree.
🔊 Listen to Article (Chinese Audio)
The Defensive Investor's Operating Manual — Chapter 5: Computing a Margin of Safety
"We have another word for the secret of sound investment: MARGIN OF SAFETY." — Benjamin Graham
The Problem This Chapter Solves
The canonical Chapter 4 uses the bridge analogy to explain what a margin of safety is and why price determines risk. This chapter computes one.
Margin of safety is rarely used in practice not because people fail to understand it, but because it sounds like a number requiring precise valuation — enterprising work. This chapter shows that to be a misconception: three forms are computable by a defensive investor in ten minutes.
Form One: Earnings Yield vs. Bond Yield
Earnings yield = EPS ÷ price = 1 ÷ P/E
A stock at 20× earns 5%. A stock at 12× earns 8.3%.
Graham's test: a stock's earnings yield should be at least twice the yield on high-grade corporate bonds.
| Environment | Bond yield | Required earnings yield | Implied max P/E |
|---|---|---|---|
| Low rates | 3% | ≥ 6% | ≤ 16.7 |
| Neutral | 5% | ≥ 10% | ≤ 10 |
| High rates | 7% | ≥ 14% | ≤ 7.1 |
This table shows something important and routinely ignored: the threshold moves with interest rates. The same 15× stock is reasonable when bonds yield 2% and expensive when they yield 6% — the company has not changed; what you gave up to own it has.
Which is why "its P/E is 15 and the historical average is 15, so it is fairly valued" is a faulty inference. It omits the risk-free rate as the frame of reference.
Form Two: Reverse DCF
A forward DCF requires forecasting future cash flows — enterprising work, and usually wrong. A reverse DCF inverts the question:
"At today's price, what growth is the market implying — and is that implication absurd?"
You need not forecast the future; you need only judge whether the market's forecast is reasonable, a far easier question.
A rough but usable approximation: implied growth ≈ the premium of (P/E ÷ 15) over the baseline. A company at 45× has a price implying earnings will roughly triple.
The question is not "can it?" but "if it achieves only half of that, what do I lose?" That is what margin of safety actually means: not being right, but surviving being wrong.
Form Three: Portfolio-Level Margin (The One That Matters in Retirement)
The first two forms concern individual stocks. For a defensive investor the most important margin is not at the stock level at all.
Portfolio margin of safety = how large a decline your portfolio absorbs without changing your retirement plan.
| Item | Calculation |
|---|---|
| Required annual spending | e.g. $60,000 |
| Cash + short-bond buffer | e.g. $150,000 |
| Years the buffer covers | 150,000 ÷ 60,000 = 2.5 years |
| Equity portion | e.g. $600,000 |
| If equities fall 40% | $360,000 remains |
| Does the plan still hold? | This is the margin of safety |
This requires valuing no company and forecasting no growth rate, and takes ten minutes. The question it answers — "if the market halves, can I still retire?" — is closer to Graham's intent than any individual P/E.
Graham's bridge: one built for 30,000 pounds carries 10,000-pound trucks. In a retirement portfolio, the truck is your spending requirement and the bridge is what your assets are worth in the worst case. The margin is the gap.
Procedure
- Always read earnings yield beside the current bond yield. Valuation divorced from rates is meaningless.
- Use a reverse DCF to test whether implied expectations are absurd; do not attempt forward forecasts.
- Run the portfolio-level calculation annually. It matters more than any single-stock analysis.
- Make "does the plan survive a 40% equity decline?" the core test. 40% is not pessimistic — the S&P 500 fell over 45% after both 2000 and 2007.
- If the answer is no, change the allocation — not the expected return. Raising your assumed return substitutes hope for margin.
- Never use a margin of safety to justify concentration. "I calculated that it's cheap" is precisely the reasoning Chapter 3's step 6 exists to prevent.
Relevance to a Retirement Portfolio
One misuse is especially dangerous in retirement: treating a margin of safety as permission to take more risk.
The reasoning runs: "I calculate a 40% margin here, so I can size up." The flaw is that a margin of safety protects against valuation error, not against being wrong about the business. If the earnings forecast underlying it is mistaken, so is the margin. A margin of safety cannot protect you from an error in the assumption used to compute it.
For a retiree the correct use runs the other direction: it is a reason to reduce risk, never a licence to increase it.
This is the largest difference in tone between this manual and the canonical volume. Graham wrote for investors capable of security analysis; this manual is written for investors who have admitted they are defensive. For them the most reliable margin is not a discount found in a stock but low costs, broad diversification, an adequate buffer, and a policy statement written before the panic. None of those four requires you to value anything correctly — which is precisely their value.