Output per worker can rise in two ways. Workers can get more capital to work with: machines, roads, software. Or the economy can get more output from the same capital and labor, a gain in what is called total factor productivity (TFP). The Solow model shows why the first route runs out. Each extra machine adds less than the one before, so saving more raises the level of income but cannot keep it growing. Lasting growth has to come from the second route, better ideas. An idea, unlike a machine, can be used by everyone at once, which is why ideas escape diminishing returns.
Growth accounting splits a year's growth into these sources. Suppose output grows 4%, capital 6% and hours worked 1.5%, and capital earns one third of national income. Capital contributes one third of 6%, or 2 points. Labor contributes two thirds of 1.5%, or 1 point. The remaining 1 point (4 minus 2 minus 1) is TFP growth.
The US Bureau of Labor Statistics publishes this calculation every year. For 2025 it reported that output in the private nonfarm business sector grew 2.6%, the combined inputs of capital and labor grew 1.7%, and TFP grew 0.8% (release of March 19, 2026). From 2019 to 2025 TFP growth averaged 1.0% a year, up from 0.6% between 2007 and 2019.
Diminishing returns also predict catch-up: a poor economy with little capital should grow faster than a rich one, but only if it has similar saving, schooling and rules. Growth is easy to confuse with three other things. It is not recovery from a recession, which is the business cycle returning output to its trend. It is not a rise in total GDP, which can happen just because the population grew. And it is not the same as development: why some countries never start catching up is the subject of Economic Development, and the rules that make investing worthwhile are covered in Institutions and Governance.