amounted to about 0.16 percent of GDP and that one year’s improvement is realized for the rest of calendar time. That is, it lowers costs forever, so that next year’s improvement is added to it. If this were compounded for 20 years, the technological improvement would amount to about 5 percent of GDP; if it were compounded over 30 years (which would require a lot of brave assumptions), the improvement would equal about 20 percent of GDP—an enormous figure.
Dr. Flamm then compared the economic role of the semiconductor with that of the nineteenth-century railroads, which have been studied extensively by economic historians as examples of technological change. Those studies show that there were two waves of railroad building in the United States: one from 1830-1860 and a second from 1860-1890. At the end of the first period the total benefit to consumers in lowered transportation costs was about 4 percent of GDP. At the end of the second period, after 60 years of railroad building, the benefit to consumers exceeded 10 percent of GDP (GNP then).
In contrasting the benefit of railroads with that of semiconductors it is important to recognize the role of the government, which has a very price-inelastic demand for semiconductors. Depending on how government purchases are treated, the estimated benefit to consumers from semiconductor improvement is somewhere between two and four times the benefit from railroads.
He concluded by saying that something very significant for the economy has happened, and that another insight from the case of the railroads may also apply to the case of information technology. That is, economic historians have looked at railroads not only in the United States but also in other countries, and they have concluded that the impact of railroad building in Brazil or Mexico or Serbia as a percentage of GNP was considerably higher than in the United States. The reason is that before the railroads were built in other countries there was virtually no competing transportation infrastructure. In the United States railroads had to compete with a well-developed system of canals and macadam roads and turnpikes. Therefore, in Mexico or Brazil the benefit over 60 years might have been 20 to 40 percent instead of 10 percent of GDP.
The present digital economy is an obvious analog to the transportation economy of the nineteenth century. Those countries with well-developed infrastructures of information technology and computing experienced a large benefit from the recent decline in technology costs. Countries lacking a basic information technology infrastructure are likely to gain even larger relative benefits from cost declines in information technologies. If the analogy with railroads is sound, he concluded, other countries may look forward to very large benefits in the near future, especially in the developing world.