Margin of Safety
Margin of Safety
A margin of safety is the discount to intrinsic value at which a position is bought. It is the protection that survives your mistakes. Buy a dollar of value for sixty cents, and even a flawed analysis can still make money. The wider the gap between price and worth, the more room you have to be wrong and still come out ahead. It is the single idea that turns investing from a gamble into a discipline.
Start Learning →
Margin of Safety
LEARNING OBJECTIVES
IMPORTANCE
The margin of safety is the most important idea in value investing, because it is the only protection an investor has against being wrong. Every valuation rests on assumptions about an future. The is what absorbs that uncertainty.
Most investors focus on the forecast. The better question is how much room the price leaves if the forecast is wrong.
Once this is understood, the purchase decision changes:
The margin of safety is the second filter in the Compoundex process. The discount is the protection.
INTELLECTUAL ORIGINS
The margin of safety is Graham's central idea. He gave it the final chapter of The Intelligent Investor and reduced sound investment to three words.
Confronted with a like challenge to distill the secret of sound investment into three words, we venture the motto: margin of safety.Benjamin Graham, The Intelligent Investor, Chapter 20, 1949
His reasoning was structural. The investor cannot know the future, so the price must hold a discount large enough that an error in the estimate still leaves the principal intact. Buffett called it the cornerstone of investment success. Klarman named his book after it.
CORE FRAMEWORK
The margin of safety is the discount between and price. A position is bought only when that discount is wide enough to absorb the error in the valuation.
How large the discount must be depends on the uncertainty of the business:
The discount is set before the purchase and measured against the , never the bull case. A margin of safety measured against an optimistic valuation is not a margin of safety. It is the forecast restated as a cushion.
We insist on a margin of safety in our purchase price.Warren Buffett, Berkshire Hathaway letter, 1992
THE THEORY IN DEPTH
The margin of safety is the single idea that turns careful analysis into disciplined investing. It accepts an uncomfortable truth: you will sometimes be wrong, and the future will sometimes surprise you. Rather than pretend otherwise, it builds protection into every decision from the start.
What is a margin of safety?
A margin of safety is the gap between the price you pay and the value you believe you are getting. It means buying a business meaningfully below your estimate of its intrinsic value, not close to it.
That discount is not greed. It is protection. It is the room you leave yourself to be wrong and still not lose money.
Why is it necessary?
No valuation is exact, and no one can predict the future with certainty. Every estimate rests on assumptions, and assumptions can be wrong.
The margin of safety absorbs those errors. If you buy at a large enough discount, then even if your analysis is somewhat off, or the business hits a difficult year, the price already accounts for it. Without that discount, everything has to go right simply to avoid a loss.
How large should the margin be?
The right size depends on how confident you can honestly be.
The less sure you are of the value, the wider the discount you should insist on. The margin of safety scales with your uncertainty, not with your optimism.
COMPARISON
| Margin of safety | No margin of safety | |
|---|---|---|
| Price paid | Below the base-case value | At or near the value |
| If the valuation is too high | The cushion absorbs the error | The error becomes a loss |
| Source of protection | The discount | The hope the forecast is right |
| Required discount | Scales with uncertainty | Not considered |
| If you are wrong | Principal preserved | Permanent loss |
REAL COMPANY APPLICATION: META
Meta shows the margin of safety on a high-quality business, and shows that the required discount depends on how predictable that business is.
The figures below are a snapshot from the Compoundex valuation as of 28 June 2026 (Q1 FY26). Intrinsic value is an estimate, not an observable fact. The full model and its assumptions are set out in Equity Research, under the Meta analysis.
| Date | Price | Estimated value | Margin of safety |
|---|---|---|---|
| 28 June 2026 | $550 | $792 | 31% |
What this changes for you tomorrow
The discount is the protection. A correct valuation at full price protects nothing.
The prices will date. The principle will not. The margin of safety turns a good analysis into a safe position, and its size is set by the uncertainty of the business, not by conviction in it.
LIMITATIONS
COMPETING VIEWS
The margin of safety assumes price can be measured against an independent estimate of value. The efficient market hypothesis denies that a reliable gap exists, since price already reflects what is known. Modern portfolio theory manages risk through diversification rather than the price paid for any single asset.
It remains useful for two reasons. The gaps it relies on appear repeatedly, most of all under fear and forced selling. And no model is precise enough to justify paying full value for an uncertain future.
IN SUMMARY
A margin of safety is the discount to base-case intrinsic value at which a position is bought. It improves neither the business nor the forecast; it protects the investor from the error in both. Its size is set by the uncertainty of the business. Buy with enough of it, and being wrong costs less than being right pays.
RELATED TOPICS
Topic 1, Investment vs Speculation. The margin of safety is one of its three conditions.
Topic 3, Intrinsic Value vs Market Price. Supplies the value the discount is measured against.
Topic 4, Quality vs Value. Sets how predictable the business is, and therefore how deep the discount must be.
Topic 6, Risk and Position Sizing. Combines the margin of safety with how much to hold.
FURTHER READING
REFLECTION QUESTIONS
The margin of safety is the cushion that protects an investor from the errors built into any valuation. Because value is estimated, not measured, some assumptions will prove wrong, and paying meaningfully less than the estimate is what absorbs that error. It is purchased, not assumed: it exists only when the price is genuinely below value, never because a business is admired. The same gap that guards against error is also where the return comes from, so the discipline is to buy with a real discount and let it do both jobs.
Salesforce, in our analysis, is worth about $278 and trades near $158, a margin of safety of (278 - 158) / 278, about 43 percent. If the true value proved to be $240, the purchase still holds, because the cushion absorbed the error.
Intrinsic value is the worth of a business in its own right, set by the cash it will generate over its life and discounted to today, independent of the price the market happens to quote. It is the figure the margin of safety is measured against, which is why the whole discipline depends on estimating it honestly. Because it rests on the future it is always an estimate and a range, never a fact, and overstating it quietly destroys the very cushion it is meant to support.
If Meta's projected cash flows discount to about $792 a share, that estimate is the anchor. The price near $550 then sits about 31 percent below it, the margin of safety on the position.
The required margin of safety is not a single number. It scales with how predictable a business is, because the less foreseeable the cash flows, the wider the likely error and the larger the cushion needed to absorb it. The conventional bands are a starting discipline rather than a rule: roughly 30 percent for stable, predictable businesses and closer to 50 percent for cyclical or uncertain ones.
About 30 percent: quality, predictable holdings such as Meta and Salesforce, whose earnings hold a steadier line.
About 50 percent: cyclical holdings such as AMD, whose earnings swing with the chip cycle. The same 30 percent discount is not equally safe on a recurring-revenue business and on a chipmaker.
Upside describes the same gap between value and price, but seen from the other side. It answers what the return would be if the price rose to meet value, so it is divided by the price rather than by value. Because the price is the smaller number, the upside is always larger than the margin of safety, and confusing the two flatters a position by making it look safer than it is. The margin of safety governs how much room there is for error; the upside describes the reward if the thesis is right.
Nike, worth about $63 and trading near $41, offers about a 35 percent margin of safety against value but about 55 percent of upside against price. One gap, two denominators: divide by value for safety, by price for reward.
Permanent loss is the kind that does not come back, and it is the thing the margin of safety exists to prevent. It is distinct from a temporary decline in the quote, which reverses for a sound business held with patience. Permanent loss comes instead from overpaying so the price never recovers, from owning a business whose value genuinely erodes, from leverage that forces a sale, or from panicking at the bottom. Protecting against it matters more than chasing the last increment of gain.
AMD bought near $220, about 55 percent below value, can fall in a panic and still recover, because there is a cushion beneath the price. AMD bought near $522 with no margin can turn the same fall into a permanent loss, with nothing to absorb it.
A quality business that falls 30 percent in a panic and later recovers caused volatility, not loss. Capital impaired by overpaying, by a broken business, or by forced selling is a permanent loss.
A valuation is a range, not a point. The base case is the central, most probable estimate, sitting between an optimistic bull case and a pessimistic bear case. The margin of safety must be measured from the base case, because measuring it from the bull case manufactures a cushion that does not exist. Anchoring to the most hopeful scenario is one of the most common ways investors quietly talk themselves into overpaying.
In our Nike analysis the central estimate of value is about $63, with a more pessimistic and a more optimistic case around it. The margin of safety is measured from that central $63 figure, never from a hopeful one, so a price near $41 gives roughly a 35 percent cushion.
A value trap looks like a margin of safety but is not, because the intrinsic value itself is falling. The price is low for a reason: the business is eroding, and the discount is measured against a value estimate that is stale or too optimistic. The guard against it is to ask, before trusting any gap, whether the value is stable or in decline. A cheap price only helps when the value behind it is intact.
The Honest Company screens cheap against an estimate of value, but it is a micro-cap with a thin moat and slim profits. If its value is genuinely eroding, the low price is a trap rather than a margin of safety, which is why it is held only as a very small position, if at all.
Intrinsic value comes from the fundamentals and moves slowly. The margin of safety is simply the gap between it and the price you pay.
The discount is your protection. Buy at a large enough one and you can be somewhat wrong, or hit a bad year, and still not lose money.
Uncertainty, not optimism, sets the size of a margin of safety. A predictable business needs a smaller cushion; a cyclical one needs a much larger one.