Technological Progress

Technological progress, also called technical progress or technical change, is the improvement in the knowledge and methods used to turn inputs into output, so that the same amount of labor, capital and materials produces more, or produces new and better goods. Economists regard it as the main source of sustained growth in output and income per person, because simply adding more machines or workers runs into diminishing returns.

In growth accounting, technological progress is not observed directly. It is measured as the part of output growth that cannot be explained by growth in measured inputs, known as total factor productivity or the "Solow residual". It is closely related to innovation, which is the process of bringing new technology into use, and to research and development, one of its main inputs. Whether it falls from outside the economy or is created by it is the question that separates the Solow model from endogenous growth theory.

Imagine a bakery. Hiring more bakers or buying more ovens lets it bake more bread, but each extra oven helps a little less once the kitchen is crowded. Now suppose the owner finds a better recipe and rearranges the kitchen so nobody walks back and forth. The same staff and ovens now turn out more loaves every morning.

That extra bread, which comes from knowing how to work better rather than from having more of anything, is what economists mean by technological progress. Across a whole country it is the reason people today are far richer than their great-grandparents.

Robert Solow's 1956 growth model and his 1957 study of the US economy put technical change at the center of growth economics; he received the Nobel memorial prize in 1987. A common version writes output as a function of capital, labor and a technology term:

Taking growth rates gives the growth-accounting equation:

• $Y$ is output, such as real GDP, and $g_Y$ its growth rate.
• $K$ is the stock of capital goods and $L$ is labor input, two of the factors of production.
• $\alpha$ is capital's share of income, often around 0.3 to 0.4.
• $A$ is the level of technology, and $g_A$ its growth rate: the Solow residual.

Worked example: suppose output grows 3% a year, capital 4% and labor 1%, with $\alpha = 0.3$. Then $g_A = 3 - 0.3 \times 4 - 0.7 \times 1 = 3 - 1.2 - 0.7 = 1.1$. About 1.1 percentage points of the 3% growth come from technological progress. In his 1957 paper Solow estimated that roughly seven-eighths of the rise in US output per worker-hour between 1909 and 1949 came from technical change rather than more capital. Because of the law of diminishing returns, the model implies that without continuing technical progress, growth in income per person eventually stops.

• Neutral and biased change. Hicks-neutral progress raises the output of all inputs equally; labor-augmenting (Harrod-neutral) progress makes each worker effectively worth more; capital-augmenting progress does the same for machines.
• Embodied and disembodied change. Some progress arrives only through new equipment, while some, such as better management, needs no new machines.
• Skill-biased change. Skill-biased technological change raises demand for educated workers more than for others.
• General-purpose technologies. Timothy Bresnahan and Manuel Trajtenberg (1995) described technologies such as steam power, electricity and computing that spread across the whole economy and trigger further innovation. Many economists ask whether artificial intelligence belongs in this group.

In 1987 Solow remarked that the computer age could be seen everywhere except in the productivity statistics. US productivity growth did pick up from the mid-1990s to the mid-2000s, then slowed again in most advanced economies. Explanations include slower diffusion of new methods from leading to lagging firms, a shift of spending toward services covered by Baumol's cost disease, and mismeasurement: many services in the internet economy are free to users, and statisticians use hedonic adjustment to capture quality improvements in prices. The US Bureau of Labor Statistics publishes official total factor productivity estimates, which it called multifactor productivity until 2021.

New technology destroys some tasks and creates others. The belief that it must reduce the total number of jobs is usually described as the lump of labor fallacy, yet the transition can still cause structural unemployment for displaced workers. Research by Daron Acemoglu and Pascual Restrepo frames this as a race between automation of existing tasks and the creation of new ones. Effects on wages and economic inequality depend on who owns the technology and whose skills it complements.

For developing countries, progress usually comes through adopting technology developed elsewhere, which allows fast catch-up growth. Slowing adoption and weak institutions can leave economies stuck in what is called the middle-income trap.

• US Bureau of Labor Statistics: Productivity
• Nobel Prize: Robert M. Solow, facts (1987)