History shows that new technologies do not move instantly from the inventor's laboratory to everyday usage. It can take a long time for them to increase productivity. The absence of immediate productivity improvement with the advent of new information processing technology was not unlike earlier experiences with general-purpose technologies. Similar delays in the impact of technological progress on aggregate productivity occurred during past industrial revolutions.

Part of the delay, according to Northwestern University economist Joel Mokyr, occurs because, important as they are, fundamental technological breakthroughs often require further inventions to make them broadly applicable: "Such gap-filling inventions are often the result of on-the-job learning or of a development by a firm's engineers realizing ad hoc opportunities to produce a good cheaper or better. Over time, a long sequence of such microinventions may lead to major gains in productivity, impressive advances in quality, fuel and material savings, durability and so on."

For example:

• Although Thomas Newcomen built the first successful steam engine in 1712, it was not until about 1765 that major improvements in the engine by James Watt made it suitable for factory use. Additional improvements, which included the addition of a governor and rotary movement, made the steam engine a huge economic success in the 19th century. Recent estimates suggest that at the height of the British Industrial Revolution (1760 to 1830) output per capita in the United Kingdom grew at less than 0.5 percent per year on average, about the same rate as during the period between 1700 and 1760. By comparison, per capita output increased at an average rate of nearly 2 percent per year from 1830 to 1870. Mokyr argues that despite slow growth during the era of high invention, rapid growth in Britain after 1830 could not have occurred without the technological breakthroughs of the previous 70 years.

• Although Michael Faraday invented the first electric motor in 1821 and the dynamo in 1831, it took nearly a century of additional, substantial breakthroughs to make electricity the dominant source of power in manufacturing. Despite major technological breakthroughs in electricity, chemicals, steel production and other major sectors, American manufacturing productivity slowed in the late 19th century. Whereas output per hour increased at 1.7 percent per year from 1869 to 1889, output per hour increased at just 1.4 percent per year from 1889 to 1909. U.S. manufacturing productivity growth remained modest until after World War I, but grew during the 1920s at an astounding rate of 5.6 percent per year. Productivity growth remained high for another 40 years.

• As with the steam engine and electric motor, the computer chip did not affect productivity in many industries until additional inventions came along to apply the new technology. In banking, for example, microinventions like the ATM, the debit card and credit-scoring software were required to generate the productivity gains promised by the computer.

Stanford University economist Paul David explores the dynamics of technological diffusion by comparing the electric dynamo, a key technological advance of the 19th century, with the modern computer. The dynamo, like the computer and steam engine, is a general-purpose technology, having profound effects on nearly all sectors of the economy. Decades elapsed, however, between the introduction of reliable electric motors and their widespread use in industry. Some of the delay was accounted for by lags in the development of efficient means of electric power generation and by competition between direct and alternating current. Electric power generation was reasonably efficient and commercially viable by 1880, however, and the superiority of alternating current for most applications was clear by 1893. Yet, as the chart to the left illustrates, electricity accounted for just 5 percent of mechanical power in U.S. manufacturing in 1900 and did not exceed 50 percent until 1920.

"Part of the delay in the exploitation of the potential industrial productivity gains offered by the [electric] dynamo," according to David, "was due simply to the durability of old manufacturing plants embodying technology adapted to the regime of mechanical power derived from water and steam." A slow rate of decline in the cost of adopting electric power also contributed to the delay. Between 1907 and 1917, the price of electricity to industrial users dropped sharply, however, and the technology began to spread rapidly.

Once electricity accounted for some 50 percent of the power sources used in American manufacturing, U.S. productivity began to accelerate. Electrification enhanced productivity by affording greater flexibility and more efficient use of labor and capital in manufacturing. For example, electrification enabled more use of continuous-process techniques, such as the factory assembly line, which often reduced production times and waste. Efficiency was improved also by the wide adoption of "unit drive," that is, the use of dedicated electric motors to power individual machines and tools, rather than a system of shafts and belts powered by a single engine. Unit drive brought savings through reduced energy usage, less wear and tear, and more flexible and efficient factory design. Electrification also enhanced productivity by improving factory lighting and safety.

The histories of the steam engine and the electric dynamo show us that delays of years or even decades from the initial invention of a general-purpose technology and its impact on aggregate productivity and standard of living should not be surprising. Follow-up inventions and adaptations of existing workplaces and products to the new technology are required before large productivity gains arise. Is there any way to ensure that such microinventions do occur--that fundamental technological breakthroughs lead to growth in productivity and standard of living?

Many observers believe that a country's economic performance is related to its political and economic institutions. Countries with stable, democratic political systems, limited government involvement in economic decision-making, but strong protection of property rights, are thought to have institutions that are conducive for technological progress and economic growth. We turn next to how public policy might affect growth and what the histories of past industrial revolutions might teach us.