Everybody knows about Moore's Law, named after Gordon Moore, founder of Intel. This states that computer power doubles every 18 months or so. The law has worked for the past 30 years and looks safe for the a good number more.

By 2010, our computers will have 10 million times the processing power of a 1975 computer. The price of computers has already fallen 10,000-fold within a single generation.

Wordlwide, total computer power has been growing at a rate of 35 per cent a year throughout my lifetime, compared to the 5 per cent a year growth delivered by steam engines and their successor electric engines in any comparable period between 1869 and 1939.

At the same time, the expansion of computer use has been extraordinary.

In 1943, Thomas Watson, the chairman of IBM, predicted a world market of maybe five computers. In 1977, Ken Olsen, chairman and founder of Digital Equipment Corporation, famously said: "There is no reason for any individual to have a computer in their home."

He's got into all the anthologies of disastrous forecasts. Now more than a third of households in the OECD have a PC in the home, and that's not to mention all the computers and embedded chips in the workplace. And use of the internet is growing literally exponentially.

It took radio 37 years to reach a global audience of 50 million, and TV 15 years. It took the world wide web just three years after the development of the Mosaic web browser in 1994. There were only 300,000 computers linked to the precursor net in 1990. Now, there are approaching 100 million.

UUNET, one of the backbone internet providers, estimates that the volume of internet traffic is doubling every hundred days.

I'm starting to get breathless. Never before have we seen anything like this. These are staggering numbers.

And, of course, we have the dot com craze, which even big business has finally cottoned onto. I'm reminded by the huge enthusiasm some of our business leaders have discovered for the Internet of the no doubt apocryphal tale about Robespierre. Seeing a mob forming in the street, the great revolutionary said: 'I must see which way the crowd is headed, for I am their leader.'

It is really hard at one level not to believe that the improvement and extension of technology on this scale has made no improvement to economic growth and living standards. Unfortunately (as Nick has shown) there is little evidence of it in the statistics.

The dawn of the computer age in fact coincided with the sunset of the economic post-war golden age. Specifically, average productivity growth in the advanced industrial economies has slowed between the 1970s and 1990s.

This is Robert Solow's famous productivity paradox: there are signs of computers everywhere except in the statistics.

There are several ways, as Nick said, you could think of resolving this paradox, and I'm going toconcentrate on two of them. One basic tack is to argue that there is measurement error. An alternative is to argue that there is a long lag between innovation and growth.

But let's just start by thinking about exactly what we're trying to explain. The basic question is: is this anything special? After all, economic growth is always led by a particular sector where innovation is progressing at an heroic rate. And often these have been linked to communication.

It was railroads in the 1840s and 1870s, cars and radio in the 1920s, TV in the 1950s, air travel in the 1960s. What makes us think telecommunications and computers are different? After all, every generation thinks its own experience is unique.

But it is hard to argue that the ability to download David Bowie's new album direct from the internet is a step improvement in economic prosperity, no matter how big a fan you are. What we are doing, the debunker says, is producing nicer graphs in the same old reports in free moments between downloading pornography.

In fact, it is only in the US that there is any sign the so-called new paradigm might be delivering faster productivity growth. From an average of 1.2 per cent a year in 1970-90, growth in GDP per worker has picked up to an average of 1.6 per cent in the 1990s, which is a bit more than during previous expansions, and it ihas surged since mid-1995 to an annual rate around 2 per cent a quarter or more rather than diminishing as you might expect at this stage of a business cycle.

But no other country has yet shared this experience, and certainly not the UK.

Every new technology has its sceptics, of course. Harry M Warner, founder of Warner Brothers, said in 1927: "Who the hell wants to hear actors talk?" Some people could not imagine any use for wireless broadcasts - who would pay for messages broadcast over the ether to no-one in particular?

The fact some people were wrong in the past does not mean the self-proclaimed visionaries are right now. Even so, I want to debunk the debunkers.

Impatience

Let's start with the question of long and variable lags.

To help me I have here a visual aid. This is a shuttle rescued from a Lancashire cotton mill that closed in 1978. It was actually used by my auntie. The loom had been bought second hand in the 1950s - the shuttle is probably pre-war. If she hadn't brought it home after her last day at work it might have gone to the local textile museum with a lot of the other machinery that was being scrapped then .

Of course, it is a standard explanation for the decline of the Lancashire cotton industry that it failed to modernise. America's ring-spinning defeated Britain's old spinning mules.

Keynes criticised the industry's backwardness in a Treasury memo of April 1945 reported by the historian Peter Hennessy. Keynes said British industry was world class as long as it was unshackled by history.

But, he wrote, none of the existing industries had a shadow of a hope of competing internationally. It would have been better starting with a clean slate: "If by some sad geographical slip the American airforce (it is too late now to hope for much from the enemy) were to destroy every factory on the north east coast and in Lancashire (at an hour when the directors were sitting there and no-one else) we should have nothing to fear."

The point is simply that vintages of physical capital can be and often have been , and they are in fact a result of social and political as much as technical and economic influences.

Improvements in spinning and weaving technology ought to have boosted productivity substantially in the cotton industry. Yet they did so only to the extent that manufacturers installed new machines, and used them efficiently. The Lancashire cotton industry had done little of either between the late 1930s and late 1970s.

Many of the most important changes in the computer industry had taken place by the mid-1970s; some key telecoms advances date from the early 1980s. You might argue that 20 years is long enough for business to have adopted a brand new technology widely. You'd be wrong.

For although, as we saw earlier, new technologies are now spreading very rapidly, the spreading has all taken place very recently. In particular the network technologies as opposed to basic computer capacity are pretty recent. You'll often hear it said that the Internet is 30 years old, dating it to the first Arpanet link in the US. That's true, sort of, but I think the Internet as we understand it actually does not date back before 1990.

Certainly, for anything that involves building a network, the spread of technology can be very slow. There is a classic S-shaped curve for their diffusion - slow, quick, slow.

It took 50 years between the development of the technical capacity to generate electricity and the building of the first power station in the United States (in 1882) and then another 50 years before electricity powered as much as four-fifths of American businesses and households.

In the well-known paper Nick referred to earlier the economic historian Paul David at Stanford described how it took 40 years for US industry to reorganise in order to exploit efficiently the electric motor. It needed more than the spread of the technology - the list of pre-conditions surely includes limited liability, development of the banking industry, free trade and a reliable framework of contract law and competition policy.

Arguably, it also needed the extension of primary education universally to create the industrial workforce. When it did, however, what the world got was mass production and mass consumption. This is dramatic stuff. The fruits of the earlier technological revolution were ultimately the assembly line and corporatism.

New technology has to be embodied in new investment.

Large-scale investment in the new information and communicatios technologies is actually a recent phenomenon even in the US. Businesses just weren't spending all that much of their investment budgets on new technologies until the current expansion started in 1992.

Since then, of course, this investment has been growing at a double-digit annual rate. The IT share of the US economy, as estimated by the Commerce Department, almost doubled from 4.2 per cent in 1977 to 8.2 per cent in 1998. And outlays on equipment and capitalised software as a proportion of GDP are now at their highest in post war history.

In fact, the contribution of IT to growth of the US economy in this cycle has been dominant. It accounts for over a quarter of the growth in GDP.

In a debunking paper that got some publicity this summer, Robert Gordon of Northwest University found he could attribute all of the productivity rebound at least to mid-1999 to three things: better measurement of prices which raised estimates of real volumes; the normal increase any economic expansion brings; and productivity growth in the computer hardware industry itself.

He wrote, gleefully: 'There has been no productivity growth in the 99 per cent of the economy outside the sector which manufactures computer hardware, beyond that which can be explained by price measurement and a normal (and modest) pro-cyclical response.'

To Professor Gordon this proves the computer revolution is not going to have any broader impact on the economy. To me, on the other hand, it is an encouraging sign that the anecdotal evidence of change is showing up for the first time in conventional economic statistics. You would after all expect the industry at the heart of it all to experience the earliest productivity gains.

As I argued a little earlier, new technologies perhaps only have their full economic effect when they have also achieved much of their ultimate cultural impact. Maybe there is a parallel in the industrial revolution of the 19th century.

Steam and then rail were both the triggers and more importantly symbols of the industrial revolution.

The railroads and steam laid the cultural and intellectual foundations for a shift to new technologies such as telegraphy, photography, optics, chemicals.

The railways in particular fired the Victorian imagination in much the same way as the internet inspires many of us. It was not until rail companies had linked up that the UK had a consistent national time. Many parts of the country, even individual towns, had their own time zones. The rail network transformed the everyday notion of time and punctuality, and made uniformity of time and ultimately factory time-keeping possible.

Railways also changed people's understanding of distance. The death of distance was first proclaimed more than a century ago. It underpinned 19th century globalisation and the creation of empires.

Railways changed the mental geography in a way that some contemporaries immediately understood - think of Middlemarch, where the drama is entirely shaped by the emotional gap between provinces and centre, and by the arrival of an outsider, or many of George Gissing's novels.

Computers have fundamentally changed the mental landscape again. It hardly matters if the physical landscape is slow to catch up. It will.

The railroads allowed humans to master distance. Computers and telcoms have made it irrelevant, which is why we see both globalisation and the renewed importance of local geography.

The industrial revolution also, for the first time, took the work and worker out of the home and left dependant family members at home. It broke up the family unit while at the same time making it more necessary for social reasons.

Computer technology is having an equally fundamental effect on the link between work and workplace, and perhaps on human relationships too. It is not necessarily putting work back in the home. Rather, it's dissolving the boundaries of space and time that confined work to particular offices and factories and to monitored hours of the day. Fundamental technical changes fundamentally changes society.

And I think you can sometimes read that backwards. Big social change can reflect technological forces.

We do have to hope that something fundamental is really going on. As DeAnne Julius pointed out in a recent paper for the OECD's Forum for the Future, any scenario positing a Long Boom for the developed world anything like the post-war golden era must rely to some degree on a technology boost.

There is a dusty and unfashionable old theory of long booms - and long declines. It is the Kondratiev cycle.

This is the hypothesis that there are cycles of around 50 years or two generations in the economy. Comrade Kondratiev himself was exiled to Siberia by Stalin in the late 1920s; after all, Marxist dogma said there was only one possible direction for capitalism.

But his cycles have been extrapolated by followers, and you end up with the following rough timetable.

1787-1814 were up; it was the springtime of the Industrial Revolution.

1815-1849 down.

1850-1873 up, with a railway boom at either end.

1873-1896 down.

1898-1929 up. Clearly in this one you have to let the Edwardian boom and roaring twenties dominate the memory of the first world war. I don't need to tell you what happened in 1929.

1929-1948 down.

1948-1973 up. This is roughly the post-war golden era or what the French call les trentes glorieuses.

1973-1998 down.

I know of no explanation for the Kondratiev cycle that makes economists happy, although curiously many complex systems in nature also display a periodicity of 56 years.

Yet by a happy coincidence you will notice that the hypothesis puts us at the start of a 25 year upswing. I hope it won't unduly depress anybody to know that the followers of Kondratiev noticed that the early years of a long upswing often bring revolution and the later years war. It is as if, in something parallel to Paul Kennedy's theory of imperial overreach whereby the dominant power spends too much on the military, humanity sees economic security as an excuse for political adventurism and over-confidence.

Mismeasurement

I'd like to turn now to the second explanantion of the productivity paradox, and that is the possibility of mismeasurement.

I am consoled in resorting to what can seem like the excuse of inadequate measurement by the fact that Alan Greenspan spends about half of every speech airing it. In his latest, given the day revised figures showed the US economy had been growing even faster with even less inflation than first thought, he noted that the gap between US GDP measured by output and by income had widened dramatically. The latter is far higher.

It is very hard for formal statistics to keep up with the economy. Not so much for any practical reasons as for intellectual ones.

History represents a massive investment, financial and emotional. We get rather attached to it and are, if not blind to change, then rather myopic.

For example, there were almost no factories in England in the 1820s but they were obviously going to become exceedingly important. That was already 20 years after the visionary William Blake had written of dark satanic mills.

It was not until the second half of the 19th century that we at least got around to counting the number of mills. A statistical abstract for the UK for 1871 to 1885, showed there were 7,465 textile mills with 773,704 power looms and employing just over 1m people by the end of that period.

However, nearly three generations on from Blake's Jerusalem, the number of factories and their employees were amongst the only statistics in the volume to track manufacturing in the world's foremost industrial economy. We also counted the number of miles of rail track laid and numbers working in mines.

Some manufactures were classified in exports, but the categorisation reveals how much the collection of statistics did not understand industrialisation the way we do now. For example, woollen manufactures are a sub-category of wool. Metals are lumped together next to musical instruments. There are lots more figures for agricultural production.

It took another half century again to reconceptualise the economy to take account of its shape at the dawn of the last century.

Maybe it helps to think about the sorts of new output being made possible now by new technologies to see how they might not be captured by conventional statistics. We have faster design-times. Faster and more accurate medical diagnosis. Customisation - we perhaps even have personalised cancer drugs on the near horizon.

In one of his papers Brad DeLong, the Berkeley economist, looks at improvements in real living standards in the US since the end of the Civil War. The statistics say GDP per worker today is around $60,000. In the late 1860s it was $7,500 measured at 1998 prices. But, as he points out, the way a typical person would spend that amount of money now is completely different from the way he might have spent it 160 years ago.

After all, in the 1860s, there was no TV, no antibiotics, no convenience food, no central heating. The average 19th century worker would actually have been much worse off than sombody with a paltry $7,500 income in 1998 even though their measured real incomes are equal.

The paper turns to a late 19th century bestseller called ÔLooking Backward' to get an idea of what technological improvements people at the time envisaged. The writer, Edward Bellamy, imagined people being able to dial up live orchestras over the telephone. Tower Records and Blockbuster Videos were beyond his imagination, although he might actually have understood being able to download that Bowie album.

DeLong writes: 'The answer implicit in Looking Backward is that modern economic growth has been so great as to carry us off the scale of measurement that past generations could imagine. ..... Thus in some respects there is no-one in the America of 1867 who had then the same material standard of living as an average American today.'

And he goes on to point out that Nathan Meyer Rotschild, the richest man in the world, died of an abscess that could have been cured today with a $10 antibiotic.

In other words, technology-driven growth delivers economic benefits so large that they are not captured in any conventional price index or measure of real output.

So perhaps we have not fully got our minds round all the new kinds of output the new economy might be delivering. After all, parents notoriously don't think any of their offspring has a proper job - perhaps that means one generation does not think of the output of those jobs as proper output.

The examples I just gave are pretty utilitarian and uncontentious, but what about the example from one of Bruce Sterling's science fiction novels of somebody who is an online importer of designs for lampshades?

The more I think about it, the more I wonder whether we even have the right metric for measuring our economy. Should we not be thinking about an altogether new kind of accounting?

For one thing, it might not be national.

It would certainly not be built around the unit of production as its boundaries are becoming increasingly hazy. More and more companies are no longer isolated entities but operate in fairly fluid networks, contracting out and in as need be.

The traditional distinction between manufacturing and services will have to be discarded. Who can say whether a given company is now a manufacturer or a service provider when all but the most basic goods have such a large element of customer support and after-sales service.

My hunch is that the future unit of economic well-being will involve time.

I'm not referring to the fact that we are all incredibly pressed for time - except for the unemployed, who have far too much. We are, of course, all much too busy. I was very amused to read in a new book called Faster that those 'close door' buttons in lifts are purely decorative. I always press them, but they don't work at all. They are there to make those of us always in a rush feel better.

I'm more interested in the possibility of using time as a metric for the economy.

Already in a service economy the input of employment, essentially time spent working, is in some cases used as a measure of output. It gives us no handle on quality of output but it definitely does capture the essential feature of personal services, which is somebody else's time.

Equally, time use is a pretty good measure of consumption and well-being more broadly in an economy dominated by services. It has the additional big advantage of counting non-marketed services such as childcare and housework in the economy. That would be a big improvement in my view.

In that speech I referred to earlier, Alan Greenspan characterised the impact of technology as reducing the information void in which businesses operate. Companies have lacked timely information about what their customers want. Decisions were based on information days or weeks or even months old. This meant keeping large stocks of inventories, and enough people to cope with the unexpected.

But information technology is allowing businesses to respond more accurately and more swiftly to demand and at lower cost. Design times have fallen dramatically. Delivery lead times are down. Medical diagnoses can be made far faster and more accurately, and treatment speeded up. Fewer worker hours are needed for a given quality and quantity of output of goods and services.

We are moving towards an economy operating in real time. John Naughton's new history of the Internet captures in one passage the essence of how much time matters in any web-based business. He writes:

'Without the need to assemble a huge manufacturing and distribution operation, just about anyone could get into your market - and gain a sizeable share of it in the blink of an eye. All of which meant that from now on there were only two kinds of operation on the Net - the quick and the dead.'

Many people who enthuse about the New Economy claim it opens up a world of abundance. A typical statement goes: 'Software is the only medium in which the limits are exclusively those asset by your imagination.' And there are certainly aspects of the software business that live up to this vision of abundance.

But it is wrong, of course. It omits the limit set by the fact that there are just 24 hours in a day. So perhaps our imaginations can improve the quality of our software without limit, but at least until we are able to program computers to program computers, there comes a point when producing more software means human beings programming longer and faster.

Herbert Simon, the nobel laureate, looked at it another way. He said: 'What information consumes is rather obvious. It consumes the attention of its recipients. Hence a wealth of information creates a poverty of attention, and a need to allocate that attention efficiently among the over-abundance of information sources.'

He suggested the big growth profession would be information management.

Futurology is a dangerous business even for a Nobel Laureate. Technologies are never used the way people expect, and are either over-hyped or under-estimated. Examples of the latter are always very amusing. A Western Union memo of 1876 dismissed the telephone: 'It has too many shortcomings to be seriously considered as a means of communication. In 1895 Lord Kelvin, president of the Royal Society proclaimed: 'Heavier-than-air flying machines are impossible.

The people devising the new technologies are pretty unpredictable too.

Perhaps we will turn out to be completely on the wrong track with todays obsession with the world wide web. Maybe time travel is the next revolutionary technology.

Copyright, Diane Coyle 1999.