Concept
Margin of Safety
Margin of safety is the gap between a cautious estimate of what an asset is worth and the price you pay for it, kept wide enough that mistakes in your analysis and plain bad luck still leave you without a permanent loss.
Benjamin Graham closes The Intelligent Investor by naming this the central idea of sound investing. The logic is humble. Your valuation is an estimate built on forecasts, and forecasts are often wrong. Suppose a business looks worth $100 a share on conservative assumptions. If you pay $95, one small error wipes out your cushion. If you pay $60, you can be wrong about growth, margins or timing and still come out fine. The margin does not predict returns. It absorbs error. Two conditions matter. First, the estimate itself must be cautious, so use normal earnings rather than a record year. Second, the discount must be real and not manufactured by optimistic inputs. When no price offers a wide gap, the disciplined move is to wait. The same principle works outside markets. Budgets, deadlines and the loads a bridge is built to carry all benefit from a buffer you hope never to use.
A real-life example
Maya values a regional grocery chain using its average earnings over the last seven years, not last year's record, and lands at about $40 a share. It trades at $38. That is tempting, but the gap is only 5%. She writes down a buy price of $26 and waits. Eleven months later a weak quarter knocks the stock to $25, and she buys. Growth later disappoints and fair value looks closer to $32, but she is still ahead.
How to use it
- 1Estimate value conservatively. Use average or normal earnings, assume modest growth, and note the two or three assumptions that would hurt most if they turn out wrong.
- 2Set a buy price well below that estimate. Graham-style investors often demand a discount of 30% or more. Write the number down before you look at the daily price.
- 3If the price never reaches your number, pass or hold cash. Never shrink the margin to justify a purchase you already want to make.