units-of-production depreciation

Think about a machine that stamps out metal parts. It does not really wear out by the calendar — it wears out by use. If it sits idle all year it barely ages; if it runs flat out making millions of parts, it ages fast. So instead of charging the same depreciation every year regardless, it feels fairer to charge depreciation in proportion to how much work the machine actually did. That usage-based idea is units-of-production depreciation.

Units-of-production depreciation allocates an asset's depreciable cost (cost minus salvage value) based on output or usage rather than the passage of time. You first work out a rate per unit: (cost − salvage value) / total estimated units over the asset's life. Then each period's expense is that rate times the units actually produced (or hours run, miles driven, etc.) that period. For example, a 90,000 machine with a 10,000 salvage value expected to produce 400,000 units has a rate of (90,000 − 10,000) / 400,000 = 0.20 per unit; in a year it makes 50,000 units, so depreciation is 50,000 × 0.20 = 10,000. Make more, depreciate more; make less, depreciate less.

This method matters where an asset's wear truly tracks its use, like vehicles (by mileage), aircraft (by flight hours), or factory machines (by units). Its great virtue is that it links the expense tightly to the matching principle — depreciation rises and falls with the revenue-producing activity. The trade-off is that it needs reliable usage records and a good estimate of total lifetime output, and the yearly expense becomes lumpy and hard to forecast. Like all methods, it never depreciates the asset below its salvage value, no matter how much it is used.