For the developed world, the great economic shift of the 1990s was the apparent productivity miracle in the US. On the basis of published figures, US labour productivity grew at an annual rate of 2.5%, nearly twice the 1973-95 rate of 1.4%. And although there is now dispute about whether productivity really did improve quite so fast, there is little doubt that some improvement took place.
This underpinned a decade of unbroken economic growth in the US and the non-bubble part of the 1990s bull run in share prices. And it had political ramifications around the globe. If governments could find the US recipe for success, they could copy it at home.
No government was more impressed than the UK's New Labour administration. Brown and Blair were persuaded that the US's staunch promotion of competition and its big investment in information and communications technology were the vital ingredients. They have introduced a much more aggressive competition policy in the UK, including giving greater powers to the Office of Fair Trading to crack down on anti-competitive behaviour, and introduced tax incentives for investment.
As yet, there has been no noticeable impact on UK productivity. One explanation is that the Government's greater priority in its first term was employment creation through the New Deal. This would tend to have a depressing effect on productivity levels, since new employees tend to have lower skills.
However, a new study by McKinsey carried out over the past year into the causes of US productivity growth suggests all may not be lost for the UK. Its results are both surprising and mildly reassuring, in that it implies that a crucial cause of US success was competition. Since the UK Government's crackdown on market abuses will take some years to change the business culture, we may yet see an improvement in productivity in Britain.
But probably the most striking McKinsey conclusion is quite how unimportant was US technology investment. Just six sectors, representing 30% of the economy, accounted for all of the US productivity increase. They were retail, wholesale, securities, telecoms, semiconductors and computer manufacturing.
And IT investment was just one factor that led to the improved performance of these sectors.
It becomes clear that IT investment is no guarantee of productivity increases.
The other 53 sectors of the economy contributed 62% of the increase in use of IT in the US. Taken as a whole, they experienced almost zero productivity growth, and many saw a deterioration.
In banking and hotels, for example, there was huge IT investment - and almost no productivity improvement. Banks spent a fortune providing new online services, but these turned out to be an expensive addition to their distribution channels, not a cheap substitute for bricks and mortar.
Other factors are at least as important as investment in IT. The existence of competitive markets is relevant. And so is the quality of management.
Take the retail sector. Here the dominant player, Walmart, greatly increased its efficiency through a series of innovations, such as introducing bigger stores, economies of scale in warehousing and purchasing, wireless barcode scanning and electronic data interchange. Its reward was that between 1987 and the mid-1990s, it pushed up its market share from 9% to 27%.
Because its market is so competitive, many of Walmart's rivals copied its innovations, and between 1995 and ^'99, increased their productivity by 28%, according to McKinsey. The result was a big increase in the whole sector's productivity.
Among wholesalers, there was an almost identical process of copycat reforms to working practices, some helped by technology. Yet the information systems that were exploited to improve stock control and distribution had been around for years. It was the way that management applied technology and the fact that it felt competitive pressure to do so that seems to matter.
Competition helps to explain growth in productivity in the semiconductor industry too: pressure from Advanced Micro Devices forced Intel to shorten the gap between product launches and encouraged exponential improvement in chip performance.
Only the securities industry seems to have seen productivity increases linked primarily to technology changes. Here, the proliferation of low-cost online services and the automation of back offices did wonders.
What does this all mean for the global economy in 2002? Well, if there was nothing special about the causes of the mid-1990s productivity growth, there's no reason to assume that US productivity cannot continue to grow at a reasonable rate.
If the explosion of investment in internet services, data transmission networks or PCs was unrelated to productivity growth, then the collapse of this investment should not lead to a slump in productivity. There are grounds for believing that when the US economy bounces, which it will this year, the recovery may not be too insipid.