Imports are subtracted from gross domestic product because GDP measures production within a country. Spending totals can already include goods made abroad. Subtracting imports removes that foreign production. It does not mean that buying an imported item, by itself, destroys an equal amount of domestic output.
Follow both entries for the same purchase
The expenditure identity is GDP equals consumption plus investment plus government purchases plus exports minus imports. These components are often written as C, I, G, X and M. Business investment includes equipment and changes in inventories, not purchases of financial shares.
The US Bureau of Economic Analysis explains that spending data do not always separate domestic goods from imported goods. The import adjustment makes the expenditure total consistent with the domestic production boundary.
Consider a fictional business that buys a newly imported machine for USD 1,000 during one quarter. Assume no delivery margin, installation service, tax or other domestic production is included. Record USD 1,000 in investment and USD 1,000 in imports. The purchase adds zero to domestic GDP: 1,000 minus 1,000 equals zero. Subtracting the import while ignoring the investment entry would incorrectly show a USD 1,000 reduction.
Domestic work still counts
Now suppose the business also buys USD 200 of installation work produced domestically in that same quarter. Assume this service uses no intermediate inputs, so its entire value is domestic value added. The installed asset costs USD 1,200, including the imported machine.
| Entry | Machine only | Machine and installation |
|---|---|---|
| Business investment | 1,000 | 1,200 |
| Imports to subtract | 1,000 | 1,000 |
| Domestic GDP included | 0 | 200 |
The USD 200 is included because of the installation work. It is not evidence that importing the machine caused total GDP to rise by USD 200 relative to every alternative. That causal question requires knowing what the business and installer would otherwise have done. The Federal Reserve Bank of St. Louis explains the same accounting principle using imported final goods and inputs.
What if the imported goods remain in a warehouse?
Take a separate example. A retailer imports USD 800 of goods in quarter one and leaves them unsold. Assume unchanged prices, no retail margin and no other associated services. Inventory investment rises by USD 800 and imports rise by USD 800. Their combined contribution to the level of GDP is zero.
In quarter two, a household buys those goods for USD 800. Consumption rises by USD 800 while inventory investment records a USD 800 withdrawal. The import was already recorded in quarter one, so it is not subtracted again. With the stated simplifications, the sale adds zero to quarter two GDP as well.
This timing distinction is why the BEA advises reading imports alongside inventories and other expenditure components. A negative import contribution in a release can have an offset elsewhere.
Read the identity before making a growth claim
These examples use nominal dollars and fixed quarterly periods. They are entries in an accounting identity, not growth rates or percentage point contributions. Do not apply their simple addition to published chained dollar components. Use the statistical agency's contribution tables when decomposing real GDP growth.
Trade can change domestic production through substitution, access to inputs and other behavioral responses. The minus sign alone establishes none of those effects. When reading a release, check what was imported, whether it entered consumption, equipment or inventories, and what happened to domestic production. Then distinguish the reported arithmetic from the evidence needed for a causal explanation.
For the wider framework, read GDP, growth and productivity. The practical rule is to trace the matching spending entry before interpreting an import subtraction.
Sources and review
The linked BEA explanations and St. Louis Fed teaching article support the accounting principles. All examples here are hypothetical. Educational explanation checked on 13 November 2026 by the FOLITIC content agent with a generative language model, without an independent human review. The original generated cartoon is an illustrative metaphor, not a record of actual people or transactions. Next content review is due on 12 May 2027.



