It was not supposed to be like this. Information technology (IT) was going to give the economic aggregates a new precision and the collectors of national statistics a new power to collate and categorise economic activities.
But the opposite seems to be happening. As modern economies develop they are becoming what some of the more prophetic economists describe as increasingly "weightless". This is the conclusion, for instance, of the director of the Centre of Economic Performance at the London School of Economics, Danny T Quah, writing in the February 1997 Bank of England Bulletin.
He points out that an ever-larger proportion of the value of "production" has no physical form. A process of dematerialisation has set in which makes nonsense of, for example, the old distinctions between manufacturing and services. This clearly also makes it much harder for the central authorities, be they tax collectors or economic planners, to know what is really going on.
Weightless economic activity can be low-tech or hi-tech. One could be talking about contract gardening and the resurgence of the domestic service sector, or making hotel beds, or at the hi-tech end, almost any form of endeavour connected with IT, from servicing answering machines or surfing the Internet to the most advanced telecommunications, logic configurations and software programming and usage.
Weightless economic activity raises another problem for the statistics collectors and collators of economic aggregates. Quality changes are becoming much harder to measure than in the old safe world of physical goods. This is in good part because IT now enters into everything. IT is not a sector, not a service, not an industry and not a manufacture. It is all these things. There is literally no product or service, from a farm vegetable to a table to fancy piece of electronics, and virtually no business, which is without an IT element in its production or activities.
Not only do we lack the methods for isolating and categorising this giant and influential chunk of the national product. Price measures themselves become less and less reliable in reflecting the change content of almost every product and service as a result of the IT upgrade effect.
For example, how are the benefits of ever-cheaper mobile phones supposed to be factored into the cost of living? If more powerful laptops cost much less now than less powerful ones two years ago, how does that show up in either the price or the output aggregates? What about better television and video machines, easier banking, smarter corkscrews, more comfortable buses, improved school education (where it applies), the car that does not have to be serviced so often, or ever at all?
These are precisely the issues which seem to have torn a gaping hole in US consumer price inflation figures. The price data have failed to keep up with quality improvements and changes in lifestyle. If US inflation rates have in consequence been persistently and increasingly overstated, then the implications for policy are enormous.
Interest rates have been wrongly set, wage rates have been wrongly calculated, pensions and index-linked salaries have been overpaid, family incomes incorrectly computed and the whole arithmetic of federal expenditure thrown out of gear.
But the headaches of the weightless economy do not stop even there. Even on the old definitions, services are now much the most important sector and make the most powerful contribution to growth. Not only is the quality of service factor becoming progressively harder to pick up through price and output data (what is the difference between a good haircut nowadays and a bad one five years ago?) but the whole sector gives rise to a myriad range of small transactions and small units of enterprise, and to a pattern of outputs which changes so fast with innovation that no statistics can now keep up.
It is not just a question of information technology itself being untraceable and impossible to measure in national income statistics. The effects of IT on the whole traditional production structure make it infinitely harder to track.
This explosion in new services is the link with another failure of centrally computed statistics - the figures cannot keep track of all the services being provided for cash. Partly this is because the black and grey markets flourish as services multiply in variety and number and therefore untraceability.
Partly - and this is particularly a UK phenomenon - it is because the authorities have set a minimum turnover level below which an enterprise does not have to register for VAT. It should be no surprise at all that service enterprises miraculously fragment so as to be just below the threshold.
As invisibility spreads, the familiar aggregates in any single economy - like output or demand, or industrial investment, or unemployment - become more divorced from real levels of activity.
But is all this anything new? The answer is no and yes. Before the rise of modern statistics and Keynesian policies, business life went on quite well, no one lost sleep over the trade figures or consumer price inflation figures, for the simple reason that there were none. Prices in the various sectors and markets, including the price of labour and the price of money, told the business class what they wanted to know. So a return to that kind of world, without the benefit of aggregate government statistics, or the more dubious benefits of economic policy based on those aggregates, would be nothing new.
But the feature that is entirely new is that IT-related activity defies classification. It blurs the traditional statistical classifications of production - agriculture, industry and services. We have therefore entered a world in which the manufacturing sector tells us less and less about what is going on in an economy in which services dominate and drive economic growth, in which almost every manufacture is itself part of a sequential chain of services, largely knowledge-based.
This then leads on to the paradox that the "more industrialised" economies are those with a shrinking manufacturing sector, as classically defined and many more services of higher quality, greater variety and bigger IT content.
Manufacturing as a measurable sector unravels and melts away as more and more component parts of the production process are out-sourced, contracted out, delegated and diversified. The faster this process goes, concludes Professor Quah, the more successful the economy. So one of the favourite moans of the British left for years past - that British manufacturing is in decline and the economy too service-dominated - has to be stood on its head. The bigger the service sector the better - although harder to measure and fine-tune.
David Howell is advisory director at SBC-Warburg.