The Foundation of the New Economy: Computing Power Doubles Every Eighteen Months

• The foundation of the new economy is the revolutionary explosion of computer processing power.
  
• Back in 1965 Intel Corporation founder Gordon Moore projected that the density of transistors that could be placed on a silicon chip would double every twelve months. (He was overoptimistic: now people say every eighteen months.)
  
• Information processing power grows along with transistor density--and the cost to make a chip of given size stays roughly constant.
  
• And Moore's Law still has at least another ten years--seven generations--to run.
  
• Our computers today have 66,000 times the processing power at the same cost as the computers of 1975. Our computers in 2010 will have ten million times the processing power of the computers of 1975.
  

Source: "Processor Hall of Fame," Intel Museum; http://www.intel.com/intel/museum/25anniv/hof/moore.htm

The Speed of the Information-Processing Revolution Dwarfs the Speed of the Electric-Power Revolution

• How does the information-processing revolution compare to other, previous waves of innovation? Let's compare the past forty years of progress in information processing with the replacement of the steam engine by the electric motor.
  
• In 1869 America's steam engines delivered 1.2 million horsepower to America's manufacturing firms.
  
• By 1939 America's electric motors delivered 45 million horsepower to America's manufacturing firms.
  
• This was roughly a forty-fold increase in mechanical power in seventy years--a five percent per year increase in muscle power.
  
• At the end of the 1950s, when computers had largely replaced electromechanical calculators, there were roughly 2000 installed computers in the world--machines like Remington Rand UNIVACs, IBM 650s or 702s, or the DEC PDP-1 with processing power that averaged perhaps 10,000 instructions per second.
  
• Today--forty years later--there are approximately 200 million active computers in the world with processing power that averages approximately 100,000,000 instructions per second.
  
• This is a million-fold increase in forty years--a thirty-five percent per year increase in information processing power.

Sources: Warren Devine (1983), "From Shafts to Wires: Historical Perspectives on Electrification," Journal of Economic History (June), pp. 347-372; Martin Campbell-Kelly and William Aspray (1996), Computer: A History of the Information Machine (New York: Basic Books).

Productivity in Information-Technology Sectors Skyrocketed in the 1990s

• The measured productivity performance of America's whole economy--the rate of increase in labor productivity in all of nonfarm business--has not been especially impressive in the 1990s: only some 1.6% per year.
  
• The measured productivity performance of America's information technology industries has been much more impressive: an average annual rate of increase of value added per worker of nearly 11% per year.
  
• And the measured productivity performance of those industries that produce information technology goods--semiconductors, fiberoptics, and so forth--has been astonishing: an average annual rate of increase of value added per worker of nearly 25% per year.

Source: Industry Standard; http://www.thestandard.com/, using BEA and BLS data; updated by the authors.

Total Trade in Information Technology Is Doubling Every Six and a Half Years

• The total volume of American trade--imports and exports--in information technology is now doubling in less than seven years.
  
• Trade in information technology goods is still some twelve times larger than trade in information technology services--but that won't last long: services trade is doubling every four years while goods trade is doubling every seven.
  
• Information technology trade will amount to a quarter of all U.S. trade by the year 2000.
  
• America benefits not just from information technology exports but from information technology imports as well: imports give downstream manufacturers the power to build better, cheaper products, and they give users the power to perform their information processing and data communication tasks.

Source: Department of Commerce, Bureau of Economic Analysis

The Volume of Semiconductor Production Is Doubling Every Eight Years

• Worldwide shipments of semiconductors reached 300 billion this year.
  
• The total volume of worldwide semiconductor production is growing at 8.5% per year--fast enough to double every eight years.
  
• The increase in the unit volume of shipments understates the increase in total production by an order of magnitude. Today's semiconductor products are much more sophisticated than those of fifteen years ago. Back then semiconductor products had only one-thousandth as many transistors per square centimeter as today's semiconductor products.

Source: Robert Atkinson and Ranolph Court (1999), "The New Economy Index" (Washington: PPI); http://www.neweconomyindex.org/

The Internet Economy Is Now at Least $135 Billion a Year

• The Progressive Policy Institute has estimated that $135 billion of value added was provided by the "internet economy" in 1999.
  
• They project that a year and a half hence--in 2001--that the internet economy will generate $360 billion of value added.
  
• Nearly a third of the current internet economy is made up of business-to-business transactions.
  
• The Progressive Policy Institute's projections are for that share to grow: to half of the internet economy over the next eighteen months.

Source: Robert Atkinson and Ranolph Court (1999), "The New Economy Index" (Washington: PPI); http://www.neweconomyindex.org/

Total Federal Spending on Research and Development Has Been Falling as a Share of GDP

• Back in the 1960s and 1970s, federal spending on research and development averaged one and a half percent of GDP.
  
• By the mid-1980s the U.S. government had already scaled back its relative commitment: research and development spending then was only about 1.15% of GDP.
  
• The macroeconomic policy mistakes of the early 1980s left the United States with huge federal budget deficits. As a result federal spending on research and development has come under further pressure.
  
• Today federal spending on research and development is only 0.8% of GDP.
  
• Only health and medicine have avoided stringent cutbacks.
  
• The extremely tight "discretionary spending caps" that will govern the congressional budget debate over the next five years will place enormous further downward pressure on federal research and development spending.
  
• Given the extraordinarily high returns earned by federal investments in research and development during the Cold War generation, is this really a sector of federal spending where we want to economize?

Source: Robert Atkinson and Ranolph Court (1999), "The New Economy Index" (Washington: PPI); http://www.neweconomyindex.org/

In the 1980s and 1990s Measured Household Income Has Stagnated for the Non-Rich

• At the end of the 1990s, the measured household income of those in the bottom fifth of the American income distribution averaged less than the household income of their predecessors at the start of the 1980s.
  
• At the end of the 1990s, the measured household income of those in the middle fifth of the American income distribution averaged only five percent morethan the household income of their predecessors at the start of the 1980s.
  
• Yet the measured household income of the highest fifth of Americans at the end of the 1990s was some 40% higher than that of their predecessors two decades before.
  
• Problems in measurement make it very likely that measured household income growth understates true household income growth: we have probably seen only real wage stagnation, not real wage decline, in the past two decades among groups like white blue-collar workers in America.
  
• Nevertheless the remarkably increased spread of incomes in the U.S. over the past two decades is a potential political fault line. The "new economy" is almost surely not responsible for it; but it may well be blamed for it.

Source: Robert Atkinson and Ranolph Court (1999), "The New Economy Index" (Washington: PPI); http://www.neweconomyindex.org/

The Internet Economy Was $300 Billion a Year in 1998

• The University of Texas's Center for Research in Economic Commerce, in a study funded by Cisco, has estimated the internet economy at $301 billion of value in 1998.
  
• They attribute 1.2 million jobs to the internet economy as well.
  
• Note that the U.T. estimates are fully three to four times the size of the PPI's estimates once one takes account of the different periods covered.
  
• To our minds the major lesson of the U.T. study is that soon it will be impossible to talk about an "internet economy." The internet will be so much a part of daily life that it will be as impossible to separate out the internet economy as it is to separate out the electricity economy.

Source: University of Texas Center for Research in Economic Commerce (1999), "Internet Indicators" (Austin: U. of Texas); http://www.internetindicators.com/

The Price of Computers Has Fallen More than Ten Thousand-Fold in a Single Generation

• The price of computers has fallen more than ten thousand-fold in a single generation; the price of semiconductors has fallen even faster.
  
• Such steep price declines produce enormous problems of measurement in calculating the rate of economic growth.
  
• How to resolve this index number problem? If we value computers at today's (low) price, we grossly underestimate the value and utility of the first expensive machines. If we value computers at yesterday's (high) price, we grossly overestimate the value and utility of machines that today are often used mostly to run screen saver programs and play solitaire.
  
• This is just one of the reasons that national income accountants despair of ever producing a consensus estimate of the impact of the computer revolution on GDP.

Source: Jack Triplett (1999), "Computers and the Digital Economy" (Washington, DC: Brookings Institution); http://www.digitaleconomy.gov/powerpoint/triplett/sld001.htm

More than 60 Million Computers Are Now on the Internet

• Ever since 1987 the Internet Software Consortium (http://www.isc.org/) has run a semiannual survey to count the number of "hosts" on the internet.
  
• By the end of 1999 their count will hit 60 million: 60 million computers, all accessible one to another through the global internet.
  
• In October 1990 there were only 300,000 computers on the internet.
  
• In August of 1981 there were only 213.

The U.S. in 1997 Had One Internet Computer for Every Twenty-Three People

• In 1997 the U.S. had one internet computer for every twenty-thre people.
  
• But even then the U.S. was not the most "wired" nation: Finland had one internet computer for every fifteen people.
  
• The number of internet "hosts" per capita in the United States is growing at more than fifty percent per year.
  
• The number of internet "hosts" per capita in other major industrialized countries--Britain, Germany, Japan, Canada--is growing even faster, but from a lower base.

Source: Robert Atkinson and Ranolph Court (1999), "The New Economy Index" (Washington: PPI); http://www.neweconomyindex.org/

The Direct Contribution of Computers to GDP Has Become Large

Year:	1990	1991	1992	1993	1994	1995	1996	1997	1998	1999*
GDP Growth	0.7%	-1.8%	2.7%	2.3%	3.5%	2.3%	3.4%	3.9%	3.9%	3.5%
Without Computers	0.6%	-1.9%	2.5%	2.2%	3.4%	1.9%	2.9%	3.4%	3.3%	2.8%
Computer Contribution	0.1%	0.1%	0.2%	0.1%	0.1%	0.4%	0.5%	0.5%	0.6%	0.7%

*Projected.

• Back in 1987 economist Robert M. Solow wrote in the New York Review of Books: "How come we see the computer revolution everywhere except in the [aggregate] productivity statistics?"
  
• One answer has been that economic growth is underestimated--that the areas in which computers have greatly multiplied productivity have been areas in the service sector in which measurements of output are the worst.
  
• A second answer is that the U.S. economy is very large, and that until the 1990s investments in information processing were not all that large a share of total investment: hence we would not have expected a large contribution from computers to measured GDP.
  
• But in the 1990s--even with potential underestimates, and even with flaws in the statistical system (like those that count investments in software as a current cost of doing business rather than as investment proper)--the contribution of final sales of computers to GDP growth has become quite large. Without growing sales of computers, economic growth over the past six years would have been half a percent per year lower.

Source: Brent Moulton (1999), "GDP and the Digital Economy" (Washington, DC: Department of Commerce)

Standard Modem Technology Proved Capable of Extraordinary Improvement

• Back in the early days of networking--in the 1960s and 1970s--it was thought that high-speed data communications would require special data-friendly phone lines: ISDN or some similar service.
  
• It was expected that "ordinary" phone lines used for POTS would be capable of carrying data transmissions at the 103 standard 300 bits per second, or perhaps at most the V.22bis standard 2400 bits per second, but no faster.
  
• Yet over the past two decades we have seen a 22-fold increase in the speed of data transmission obtained over ordinary telephone lines.
  
• This extraordinary 18 percent per year improvement in data-transmission-over-POTS is not the pace of Moore's Law, but it is very rapid. It allows everyone with a phone line today the potential to connect at speeds that 20 years ago it was thought would require expensive, dedicated equipment.
  
• This tremendous improvement in standard modems has greatly fueled the growth of the internet. It has allowed the build-out of the world wide data network on top of the already-existing phone network.
  
• It has thus shaved a telephone-equipment generation off of the time it would have otherwise taken to wire the United States for the internet.
  
• And looking forward, at least half the phone lines in the U.S. are suitable for high-speed DSL service. And approximately three-quarters of the households not suitable for DSL service are potential customers for cable modems.

Source: Kim Maxwell (1999), Residential Broadband (New York: Wiley).