Joachim Andersen

Accelerated Learning

Margin of Safety: Why I Leave Room for Being Wrong

A tightrope walker in a black costume balancing with a long pole on a golden wire inside a dark circus tent at night, with a short safety leash from a cuff on his ankle to a steel ring around the wire
The walker trusts the wire, but the leash is what lets him survive a mistake.


A Margin of Safety is the gap between what something can handle and what I expect it to face. Engineers build bridges stronger than the heaviest load they expect, pilots carry more fuel than the trip needs, and careful investors pay well below what they think a business is worth. The gap is there to absorb what I can’t foresee: errors in my estimates, bad luck and surprises nobody has thought of.

The essence of a Margin of Safety is to ask: “Am I leaving enough room to be wrong, and what happens to me if I am?”


Origin: From Engineering to the Stock Market

Engineers have long designed structures to be stronger than they need to be. They express this as a factor of safety, the ratio between what a structure can carry and the load it is expected to carry. A beam that can hold three tonnes, but will only ever bear one has a factor of safety of three. Its margin of safety, the extra capacity measured against the load, is two.

In 1934, Benjamin Graham and David Dodd, who both taught at Columbia University, brought the term into investing in Security Analysis. Graham returned to it in The Intelligent Investor (1949), and in later editions he wrote: “Confronted with a like challenge to distill the secret of sound investment into three words, we venture the motto, MARGIN OF SAFETY.” For Graham, the margin was the gap between price and value, and its job was “that of rendering unnecessary an accurate estimate of the future”.

Warren Buffett, Graham’s student at Columbia, made the link to engineering explicit in his 1984 talk The Superinvestors of Graham-and-Doddsville: “When you build a bridge, you insist it can carry 30,000 pounds, but you only drive 10,000-pound trucks across it.” In 1991, Seth Klarman, who runs the Boston investment firm Baupost Group, gave his only book the title Margin of Safety. His argument was that because valuation is imprecise, the future is unpredictable and investors make mistakes, they should only buy at a significant discount to underlying value.


Key Concepts

  • Capacity vs. Load
    A margin is always relative: what a system can bear compared with what it is asked to bear. Either side can move. The load can grow, and the capacity can shrink through wear, mistakes or fraud.

  • Protection Against the Unknown
    The margin isn’t mainly for the risks I have calculated, but for those I haven’t. As Graham saw, a wide enough margin means I don’t need to predict the future precisely to come out fine.

  • The Price Decides the Margin
    Graham wrote that “the margin of safety is always dependent on the price paid”. A wonderful asset bought at too high a price has no margin at all. The same goes for time: a plan with no slack is a bet that nothing will go wrong.

  • Margins Are Quietly Spent
    In good times, a buffer looks like waste, so it gets trimmed, until the day it was needed.


Some Examples

  • The Bridge That Absorbed a Fraud
    John Roebling designed the Brooklyn Bridge to be 6 – 8 times as strong as he thought it needed to be. During construction, his son Washington’s assistant engineers discovered that the wire contractor, J. Lloyd Haigh, had delivered inferior wire: of 80 rings tested, only five met the standard. It was too late to replace it, but the bridge would still be about four times as strong as necessary. The bad wire stayed, 150 extra wires were added to each cable, and the bridge has carried traffic since 1883.

  • The Walkways with No Margin
    On July 17, 1981, two suspended walkways in the atrium of the Hyatt Regency hotel in Kansas City collapsed during a crowded tea dance, killing 114 people. Investigators from the National Bureau of Standards found that even the original hanger design would have reached only about 60% of the capacity the Kansas City Building Code required. A change during construction then essentially doubled the load on the fourth-floor connections. From the day they were built, the walkways “had only minimal capacity to resist their own weight”.

  • Fuel for the Unexpected
    On January 25, 1990, Avianca Flight 052 from Colombia to New York was held three times by air traffic control in bad weather, for about 1 hour and 17 min in total. After a missed approach at Kennedy Airport, all four engines lost power, and the Boeing 707 crashed on Long Island, about 16 miles from the airport. Of the 158 people aboard, 73 died. The National Transportation Safety Board found that the crew had failed to manage the fuel and to declare a fuel emergency before the reserve was gone.

  • Genius Without a Cushion
    Long-Term Capital Management, founded in 1994 by John Meriwether with two Nobel laureates among its partners, earned 20 – 43% a year in its first three years. It borrowed heavily, and by the end of 1997 it held about $30 of debt for every $1 of capital. When Russia devalued its currency and stopped paying its debt in August 1998, the fund lost 44% of its value in a single month. On September 23, fourteen banks and brokerage firms put $3.6 billion into it to prevent a disorderly collapse.


Applying Margin of Safety

When I want to use this mental model, this is the simple process I follow:

  1. Name the Load and the Capacity
    I write down the worst I realistically expect, and what I can actually withstand.

  2. Ask What I Might Be Wrong About
    I list the estimates, assumptions and people the plan depends on, and how far each could be off.

  3. Size the Margin to the Stakes
    The more uncertain the estimate, and the harder a failure would be to undo, the bigger the margin I want.

  4. Keep a Cash Cushion
    In the Federal Reserve’s survey of 2025, 63% of American adults said they would cover a $400 emergency expense with cash or its equivalent, and 55% had set aside three months of expenses. In this example, I go well beyond the second group.

  5. Buy at a Discount, Not at Fair Value
    When I invest, I want a price well below my own estimate of value, because my estimate is the part most likely to be wrong.

  6. Guard the Reserve
    Like the 45 min of extra fuel that U.S. rules require for flights in instrument conditions, I decide in advance what is reserve, and I treat dipping into it as an alarm, not a routine.


Pitfalls and Limitations

  • Too much margin is also a cost. Over-engineered products are expensive, and cash hoarded forever loses value to inflation and misses the growth I discuss under Compounding.

  • A margin only covers the risks it was sized for. A cushion calculated inside a model is only as good as the model, as I discuss in The Map Is Not the Territory.

  • Margins erode quietly. After a calm period, every small cut looks harmless, and stability invites more risk, as I describe in the Minsky Moment.

  • Stacked margins can backfire. When everyone in a chain adds their own buffer, plans swell, so I prefer and strive for one honest buffer where it matters.


Related Mental Models

A Margin of Safety reminds me that the future doesn’t have to match my forecast for me to be fine. I don’t need to be right about everything, as long as I leave enough room for being wrong. By building more strength than I think I need, and protecting it, I can turn disasters into inconveniences.



To learn more about mental models and philosophy,
check out: Wise: How to Attain a Deep Understanding of the World, History, Our Place in It, and How to Live Well



📚 Sources and References