At the start of the 1990s, a group of economists became unhappy with the mathematical models used to assess the impact of technology on economic growth.
Their key insight was simple. Only very rarely does a truly transforming technology appear which has an impact on virtually all aspects of economic and social life. But when they do they are referred to as general purpose technologies.
Initially, a new GPT may not be as productive as the older technology used by many existing applications. Bugs have not been worked out and working practices have yet to be adapted. For example, the efficient use of electricity in industry required steam powered plants to rearrange completely their factory floors.
In the early stages of adoption, both productivity and output stagnates while the technology is refined and society adjusts. But once the infrastructure has been laid and a critical mass of users emerges, growth rates shoot up and stay up. Remember electricity did not have its biggest impact in the States until the 1920s - more than 40 years after Thomas Edison's first commercial power plant.
The full effects of the technology will not be felt until it has been widely adopted and the economic and social structures have adapted. Key technologies are surrounded by social institutions dependent on them. When technologies change these institutions are threatened and either change or are replaced.
As the technology becomes more sophisticated, more uses are developed and more sectors adopt it, which, in turn, spurs the development of complementary technologies, which spurs faster adoption and so the chain reaction starts.
Eventually, a critical mass of users is reached and economic growth and social changes take place at a much faster rate.
To demonstrate the internet is a GPT it is worth looking at some of its unique
-It is owned and controlled by no one.
-It exists, but it is not a physical thing. It is an idea implemented via physical things -computers, switches, phone lines - but which is none of those things, nor inherently imbedded in any of them.
The internet is one of the purest examples of network economics. A network is useless if only one person uses it. To have value a network must have lots of users. But networks have a unique and peculiar property - as it grows, the cost of adding someone to the network is close to zero.
Thinking of the internet as a GPT goes a long way towards explaining many recent economic controversies and helps us project its likely future economic impact.
Is the world enjoying the benefits of a new economy? Not yet. There are as many versions of the new economy theory as there are analysts, but the most common version holds that during the mid-1990s, the US had finally adjusted to the last big GPT - the computer - which caused a tremendous burst of productivity and higher levels of sustainable economic growth.
In other versions of the theory, it was the adoption of the internet itself that caused the burst in productivity. In this vision of the new economy, the US will enjoy higher growth rates than any other developed economy, without inflation or recession.
Parts of this theory are consistent with the GPT theory. It is plausible that the adjustments needed to fully take advantage of the computer would have pulled down productivity for an extended period - the 1970s through until the mid-1990s - and once this adjustment occurred a permanent spike in output and productivity would take place. The problem with this theory is that its underlying premise
is wrong - there has been no economy-wide jump in productivity over the last five years.
All recent productivity growth that has stemmed from technical change has taken place in the durable goods manufacturing sector. The biggest gains from computers have come in the industries tied to the production of computers and related equipment. In the 88% of the US economy not connected to durable manufacturing there has been a slowdown of productivity growth.
This does not mean the entire new economy theory is bogus and that the internet is not a GPT. The world has just not yet fully assimilated and adjusted to the changes triggered by the internet.
Estimates vary as to the critical percentage of users needed to realise the benefits of a network. Most economists would agree that if over 50% of a country is networked - the US has almost reached this level - the impact would be significant.
But the exact percentage depends on the development level of each economy and its degree of concentration. In many economies, 40% usage would have a sizeable impact. But in an emerging economy where the bulk of gross domestic product is generated by only 10% of the population then just 10% usage would have a major impact.
All of this is totally consistent with the concept of the internet as a GPT.
What it suggests is that there are further productivity and growth gains to come in the future and, given adjustment lag times, possibly far into the future.
-Phil Fortuna, San Francisco Office, Zurich Scudder Investments.