When tony blair spoke at the CBI conference in November 1999, he used the phrase "It's the e-conomy, stupid." It was delivered somewhat apologetically, but it captured New Labour's fascination with the "new" economy. The Competitiveness White Paper of 1988 was entitled "Building the knowledge driven economy," and identified the US as the model to emulate because of its success in information and communications technology (ICT). Britain, it argued, must lead Europe in the e-commerce revolution and to that end it was equipped in summer 1999 with an "e-tsar." Education and training policy was to be designed to prevent the emergence of a "digital divide." The government reflected the fashionable belief of 1999 and early 2000 that a new paradigm of economic performance and competition had emerged, carrying profound implications for all facets of government policy. Charles Leadbeater's paean to that new paradigm, Living on thin air, was hailed by Peter Mandelson as setting out "the agenda for the next Blair revolution."
Evidently, the shine has gone off the new economy over the last year. Boring things like profit and positive cash flow have come back into fashion, Nasdaq stock prices have halved, telecom companies are struggling with huge debt burdens and a host of whimsically named e-retailers-boo.com, clickmango, foodoo and many others-have shut up their virtual shops. More fundamentally still, confidence in the US economic miracle of perpetually rapid growth has faltered. And in Britain the problems of the last six months-fuel protests and floods, train crashes and cattle diseases-have a physical rather than thin air feel. So with the hype of a year ago gone, we are better placed to assess the substance-is there something fundamentally new about the new economy, does ICT have a peculiarly profound economic effect?
The answer is a qualified yes. There is (probably) something distinctive and significant about the economic impact of information and communications technology. But the key effects are quite different from those often ascribed and many of the policy implications frequently drawn are confused. For while ICT is distinctive in technological terms, its paradoxical effect is to increase the importance of many non-ICT-intensive activities, and our greatest policy challenges-domestically and internationally -lie precisely in all the many problems which ICT does not solve and largely does not affect. Those challenges moreover are deeply political, requiring choice between conflicting objectives and interests. Their resolution will therefore often be incompatible with a third way rhetoric designed to deny the existence of political conflict within an inclusive big tent.
Ict is distinctive in several ways. First because it is ubiquitous. Unlike many technologies relevant only to specific sectors of the economy, ICT gets just about everywhere, increasing potential productivity in almost all sectors of the economy, making possible the automation of all the information and paper processing, record-keeping and reconciliation, and administrative activities which on some estimates account for as much as 50 per cent of all hours worked. That ubiquity does not in itself make ICT unique-electricity was similarly pervasive when introduced in the first decades of this century. But it becomes distinctive when combined with a second feature electricity did not share, which is a pace of technological advance an order of magnitude faster than in any previous wave of technological change. The rule of thumb (originally expressed in more technical form by Gordon Moore, a founder of Intel), is that computing power purchasable per pound spent doubles every 18 months. That produces a 10,000 times increase in 20 years, and means that the personal computer on your desk now incorporates computing power affordable only to the Pentagon and Nasa in 1970. Similar trends are at work in communications-the wholesale price of data communication per kilobit falls over 50 per cent each year.
Finally, ICT software is technologically distinctive because of almost zero marginal costs-the simple fact that once one copy of Microsoft Windows has been produced it costs close to nothing to make the next 100m copies. Other sectors and technologies-pharmaceuticals for instance-have high fixed costs and low marginal costs, but no sector as important as ICT has marginal costs so close to zero.
Ubiquitous relevance, super-rapid productivity growth and zero marginal cost: these three factors together make it highly likely that ICT is having an economic impact more profound than previous waves of technological change. For individual companies the latter two factors explain both huge private wealth creation and huge losses as well. Rapid volume growth combined with collapsing prices create inherent uncertainty and vol-atile valuations: volume up five times with prices only halving is corporate nirvana, volumes doubling with prices down 80 per cent can spell disaster. Zero marginal costs can mean massive profits for any company which enjoys defensible intellectual property rights and a natural monopoly, and can mean nil profits for any which doesn't, as prices are competed down towards zero.
At the macro level too the impact should be profound and distinctive. The three factors above make it likely that ICT should stimulate a period of rapid productivity growth. Indeed the oddity is not the recent spurt of US productivity, but the previous 15-20 years of poor performance when, in Robert Solow's quip, ICT was visible everywhere except in the productivity statistics. But the explanation of this delayed impact seems clear-it simply takes time for a new technology to be absorbed and used effectively by business, and it takes time especially for business to progress from islands of automation (the secretary's typewriter replaced by a word processor but all other processes unchanged) to complete business redesign (entire processes performed by computers and managers communicating directly by e-mail). Just as electricity produced its greatest productivity impact several decades after its initial introduction, so probably with ICT. The best judgement is that the US's productivity spurt of the last five years-productivity growth of 2.5-3 per cent per annum versus 1.0-1.5 per cent growth over the previous 15 years-is a "new economy" phenomenon. It will not be a permanent effect even if it survives the present cyclical downturn; it would be surprising if it lasts more than 10 to 15 years. And it is not a unique event in economic history-the 1950s and 1960s, for similar technological reasons, saw faster rates of productivity growth than either the 1930s or the 1970s. But a significant increase in the sustainable rate of productivity growth for a decade or more is still a big development, changing dramatically, for instance, governments' ability to balance demands for increased spending with constrained or reduced taxation. The proposition that there is an ICT-based "new economy" effect should therefore survive the recent demise of dotcoms and the slowdown of the US economy.
But even if it does, that will not mean that ICT is the unique key to national or corporate competitive advantage, nor that all workers will become information technology workers, nor that ICT is the or even an important public policy priority. For the impact of a technology depends not just on its technological characteristics but on its consequences for consumer demand, and even the purely technological characteristics can have surprising and paradoxical effects.
The impact of a technological advance on the shape of an economy depends on the relationship between two factors-first, the pace of productivity growth (both in the technology's own specific goods and services, and that which it enables in other sectors), and second, the extent to which it creates new consumer demand. If-like agricultural technology between 1800 and today-productivity growth hugely outpaces new demand created (people just didn't need much more food), the benefit derives primarily from fewer people employed in the existing industry, freeing resources to do other things. If-as with the car industry-the new technology creates a new category of demand, it can lead to the emergence of a big new sector of employment. Cars and lorries have created growing global employment for most of the 20th century.
The impact of ICT depends on its position along these two dimensions. Several factors will make it more similar to agricultural technology than to the car industry in its economic effect. The first is the super-rapid productivity growth achieved in its own sector-which drives falling prices, thus limiting the value of consumer demand even while volumes soar. In the US, where penetration rates for ICT are highest, consumer expenditure on ICT products-PCs and internet connections, mobile phones and mobile phone calls and even plain old voice telephony-accounts for just 2.4 per cent of consumer expenditure, up only marginally from 1.9 per cent in the late 1980s. As a consumer product, ICT is cheap and keeps getting cheaper relative to goods and services where productivity growth is more difficult to achieve or where supply is inherently limited-such as restaurant meals, healthcare, recreational facilities or houses in desirable locations. Over time consumption (and employment too) shifts to the goods and services whose production cannot be automated. In 20 years time, given the operation of the 18 months rule, it is quite possible that consumers will purchase and use 10,000 times as much computing power and communications capacity as they do today. But it is also quite possible that the economic value of all that consumption will have fallen relative to the economic value of the world restaurant and hotel business. It is precisely because ICT is so technologically powerful that it will be self-limiting as a percentage of GDP.
This self-limiting nature is reinforced yet further by the phenomenon of consumer value accruing for free. The fact that my daughters can gain information for homework projects via internet access to freely- provided academic web sites represents a very welcome increase in consumer value, but not in economic activity. The ability to find out flight availability or bank balances over the internet gives consumers increased leisure time previously spent on phone calls or branch visits, but that increased leisure appears nowhere in GDP, nor are more than a trivial number of people gaining employment or profit from providing that benefit. As one of the world's most successful investors, Warren Buffet, puts it, the internet could be simultaneously a huge benefit for consumers and a relatively small opportunity for capitalists.
Finally ICT will be more like agricultural technology than the car industry because it enables big productivity improvement in sectors of activity where the potential for volume growth is limited. Automation of back office processing in utilities, public administration and financial services, is not going to stimulate growing demand for energy, government activities, or bank accounts-it's simply going to mean fewer people doing these activities, freeing them to perform other jobs, satisfying different demands. That effect-releasing people from existing processes to find jobs elsewhere-is indeed the really big story of the ICT revolution and the concrete reality behind the US's accelerated productivity growth. It results from the massive application of ICT as a tool of business process redesign. ICT products and services may account for only 2.4 per cent of US consumer expenditure: but they account for about 44 per cent of all business investment.
This pattern of impact means that we will not all be ICT workers, living on thin air. Of course the ICT industries themselves create some jobs and often highly paid ones, but not that many. In the US, ICT- producing companies employed just 3.7 per cent of workers in 1998 and while that percentage may grow slightly, it will not grow much and within five to ten years it may fall. For it just doesn't take many people to manufacture computers or to run telecom companies, and zero marginal costs mean that a software company can have a huge economic impact while employing very few people. Microsoft employs 31,000 people worldwide; Walmart, the world's largest retailer, employs over 1m. Combine this limited employment creation with the large numbers released through automation, and it is clear that ICT has been and will continue to be a job-destroying rather than job-creating technology.
But that does not matter. For a technology does not have to be job creating or even job neutral in its direct effects in order to be economically beneficial. It can also be beneficial, like agricultural technology, because it eliminates jobs in some sectors, making possible new employment growth elsewhere. And the evidence is that provided labour markets are reasonably flexible and provided demand growth is maintained, that benign shift does indeed occur. The US productivity spurt, resulting from the intense application of ICT, has been accompanied by rapid employment growth. That growth has occurred primarily in personal face-to-face service activities which are difficult to automate. US Bureau of Labour Statistics forecasts for 1998 to 2008, extrapolating existing trends, predict 1.9m more employees in the ICT producing industries. But the really big increases are forecast to be 4.4m new jobs in hotels, restaurants, retailing and amusement parks, auto- cleaning and parking and a host of other personal face-to-face services, and 5.5m new jobs in health services, education, the voluntary sector and other social services. The most rapidly growing jobs in the US and Europe include cooks, nurses, waiters, leisure centre employees, sports club professionals and gardeners. The paradox of the new economy is that ICT's profound effect on productivity growth increases the importance of categories of consumer expenditure and of employment which we might more validly call "high-touch" rather than high-tech.
Public policy priorities need to reflect the reality of the high-touch as well as high-tech economy. The fashion has been to define policy priorities in "new economy" terms-information technology skills as a key to success, the "digital divide" as a social challenge, and closing Europe's competitiveness gap in ICT a priority. But once the true nature of ICT's impact is understood different priorities emerge.
The US's lead in the development of consumer internet applications is not a valid matter for public policy concern in Europe. The important US advantage is accelerated growth through higher business investment in ICT, and that carries some implications for European policy-namely regulatory frameworks to encourage competition in high speed internet access, and reasonably flexible labour markets to enable rapid industrial restructuring. But the take-off of consumer internet services has no specific importance, and statements that Britain is ahead of continental Europe because it has more internet start-ups reflect a dangerous tendency to congratulate ourselves for irrelevant achievements. Successful e-retailing businesses, such as lastminute.com, are expressions of praiseworthy individual entrepreneurship, providing useful new forms of customer service, but they are no more or less fundamental to the economy than were Tie Rack or Sock Shop. And how rapidly consumers care to sign up to high-speed internet access is no more a fundamental measure of economic competitiveness than was the pace of penetration of colour television or CD players.
The striking feature of internet penetration rates between the US and Europe, moreover, is not how long they are but how short. European internet penetration lags the US by 18 to 24 months. When cars or television sets were first introduced, the lag was more like 15 years. That reflects two factors: the first is how expensive cars and televisions were and how cheap is internet access; the second is how effective are the free market processes for transferring any new internet concept from the US to other markets. When the history of the present period is written, the fact that European consumer internet penetration lagged the US by two years will hardly feature.
The shortness of the lag also suggests that social concern about a "digital divide," whether within or between nations, is largely misplaced. When cars were first introduced, there really was a huge social divide-the middle class enjoyed journeys to the countryside, the working class did not, and that divide lasted for many decades. Time lags between different income groups in the penetration of personal computers, internet connections or mobile phones are much shorter, once again because all these products are cheap. Getting internet access and computers in all schools, meanwhile, is a sensible policy, but also a fairly easy and cheap one-far cheaper than reducing pupil-staff ratios in the face-to-face unautomatable world of primary or secondary school teaching, far cheaper than making teaching an adequately paid and high-status profession, and far easier than creating the culture of discipline essential to attract good teachers to more difficult schools. Precisely because ICT is cheap and virtual rather than human, ICT is unlikely to be either the cause of, or the solution to, educational inequalities.
At the global level the same scepticism about a digital divide should prevail. Africa may lag 15 years or so behind US levels of PC and internet penetration, but it lags more like a century behind in basic literacy and health care. Anti-malaria programmes, good schools, and the attainment of clean government are far higher priorities for the world's poor countries than avoiding a digital divide. Africa could be an economic disaster zone even with mobile phone and internet access as widely spread as in Europe today.
One reason why ICT is cheap and getting cheaper lies in the zero marginal cost of software. That zero cost makes software an extreme example of what economists call an "uncontested" good-a good where my enjoyment of it in no way limits another's ability simultaneously to purchase and enjoy its benefits. The most contested goods are those whose supply is inherently scarce, or whose value derives from specific physical location or attribute. Social conflict, requiring political resolution, is far more likely to arise from disputes over contested and positional goods than from the uncontested world of computer software. Hardware and software prices will keep falling: prices of desirable houses are likely, especially in crowded countries, to rise not only in real terms but faster than real earnings, as increasingly prosperous people compete for a limited supply. Looser planning rules could help offset that effect, to the benefit of new purchasers but at the expense of existing owners and of environmental harm. Debates over appropriate planning rules will be far more contentious than any digital divide.
Within ten years, meanwhile, we will all be able to enjoy video telephone calls for the price we now pay for voice telephony-any temporary problems of congestion will be soluble with more optical fibre. But congestion of roads, railways and airports is likely to increase, and disputes over road pricing, road building and levels of public transport subsidy will be continuous and contentious. The management of transport assets will be a key challenge. Britain's New Labour government has lauded the entrepreneurship of the internet start-ups, and been receptive to pleas for favourable tax treatment to tempt talented executives from staid "old economy" firms. But the quality of management attracted to Railtrack and London Underground may be more important to the quality of life in Britain in ten years' time than the number of e-retail entrepreneurs. And the decisions now being made about the ownership and management structure for London Underground will leave an imprint far longer lasting than policies to stimulate e-commerce.
Internationally too, it is the physical which creates the greatest potential for conflict. Disputes over scarce water supplies could cause wars, not disputes over internet access. And the scarce resource of the world's atmosphere poses a challenge far more difficult than the limitless possibilities of information processing and exchange. Nothing is certain, but the best judgement is that CO2 emissions are generating harmful climate change. If that is true, then the right to emit global-warming gases is one of the world's most inherently contested goods and the negotiation of those rights one of the most difficult global issues. The internet will flourish without global rules or global political leadership: global warming and other environmental issues make an element of global governance essential. Persuading electorates to accept policies-such as higher fuel taxes-required to meet global responsibilities will demand real political leadership. Governments could fall on this issue or succeed at the expense of long-term harm. None will fall because their e-commerce legislation is badly designed.
Britain's Labour party, concerned to banish forever the belief that it was bad for business, took from the New Democrats the philosophy that "it's the economy, stupid," and the policies which followed have on the whole been sound. Macro-stability, fiscal prudence, free markets and a commitment to improve education are all sensible priorities. New Labour also responded to the "new economy" fashion with enthusiasm, and some of the resulting initiatives have also been sensible-competition in local telecom access, encouragement of venture capital. But as a definition of the government's priorities, of the truly difficult issues it faces, the "e-conomy" is certainly a diversion, and even the economy (without the e-) may not be as hard to manage as is often assumed.
For the most difficult issues facing modern governments are not those of economic growth, or of international competition, and certainly not those posed by the uncontested world of software, but issues of social preference, distributional equity and physical constraints. Healthcare demand will grow relentlessly because of the insatiable customer preference for a service where ICT-based productivity improvements (though present) are insufficient to deliver falling prices. Either we accept higher taxes or we increase the role of private provision-constraining demand by rationing is not acceptable. Income inequalities are growing because of technology and globalisation: we cannot therefore avoid a debate on how far the state should redistribute. Rising demand for physical mobility is creating intractable problems of congestion and environmental damage. Resolving these problems will be truly difficult-requiring political choice and courage.
On one of these issues the government has been more radical than its own rhetoric. Gordon Brown's taxation policies have been significantly redistributive. In healthcare, meanwhile, the unnecessary expenditure squeeze of 1997-99 has been followed by big planned increases. But there remains no answer to the long-term challenge of limitlessly rising demand, nor a willingness to debate the radical options which may need to be considered. And on transport and the environment, we have seen four years of contradictory messages and minimal political courage. Such vacillation is rooted in the implicit philosophy of the third way-the belief that all interests and objectives can be embraced, old conflicts of right and left, green and developmentalist, reconciled by technocratic efficiency. It is this hankering for a non-political politics which made the e-conomy such a beguiling theme-an opportunity to escape difficult physical and distributional issues, focusing instead on the delusion of a conflict-free modernity.
The "new economy" of ICT application is important, but at the public policy level it is rather simple-establish a framework, provide incentives and leave the free market to do the rest. But free markets work least well where goods are positional and contested, where environmental externalities and distributional issues abound, and where lead times are beyond the time horizon of private business. The job of government is to tackle the difficult issues which the market alone cannot solve. As the Blair government prepares for a second term, it's not the e-conomy, stupid.
