THE NEW BRITISH ECONOMY.

Kneller, Richard ; Young, Garry

Richard Kneller [*]

Garry Young [**]

The British economy has performed well since the recession of the early 1990s. How much of this is due to the revolution in information and communications technology (ICT)? We find that the stock of computing equipment has grown at a similar rate to that seen in the US, but there appears to have been no similar ICT-induced pick-up in productivity growth. We suggest that any underlying improvement in productivity growth has been obscured by a slowdown in manufacturing and by the need for the unemployed to be absorbed into employment. We find no evidence yet of any clear effect of ICT on pricing and suggest that any benefit of greater competition is likely to come about by encouraging productivity growth rather than by reducing margins. We argue that the main cause of Britain's improved performance lies in the labour market. Looking forward, we are optimistic that the benefits of ICT will become more apparent as the factors that have obscured it become less important.

Introduction

Since the recession of the early 1990s, Britain has enjoyed a period of rare economic success. Economic growth has averaged about 3 per cent per annum and for the most part exceeded the rate of price inflation, which is now the lowest in Europe. Economic growth has been accompanied by a substantial rise in employment of over 2 million and it is now possible to talk of the economy being close to 'full employment'. Nor is this favourable economic performance expected to change markedly over the near future, even though the recent volatility in equity prices, the foot-and-mouth crisis and the slowdown in the US economy are expected to reduce the rate of GDP growth this year.

The ability of the British economy to generate reasonably strong growth in output and employment over a lengthy period without causing inflation has been a surprise to many. The apparent improvement in the relationship between changes in economic activity and inflation is one of the factors that has encouraged some people to believe that the economy is behaving differently from the way it did in the past. Until recently this view has been supported by the experience of the US, where even stronger growth has not yet led to a pick-up in inflation.

In the second half of the 1990s, an improvement in the inflation-activity relationship has been accompanied by a sharp surge in private non-residential capital spending both in Britain and the US (chart 1). For both households and firms, a substantial part of the increase in capital spending has been in new information and communications technology (ICT). Not only is the equipment new, but it has been applied to new areas such as the internet. The coincidence of an improvement in overall economic behaviour with the prevalence of ICT has led to the suggestion that technological innovations in ICT have caused the economy to function better.

Although the slowdown in the US economy and the collapse of new-economy share prices has tempered the level of enthusiasm for this new view of the economy, it is, however, also true that the economic behaviour of British households and firms has changed in other respects over the last few years. Magazines and newspapers are full of examples; shopping has become many people's favourite leisure activity and in terms of the labour market the 'working week' means different things to different people. This has been associated with a proliferation of products, which has widened consumer choice confusingly, and the development of an 'open all hours' culture.

The primary purpose of this paper is to assess whether there has been or is about to be a fundamental change in the British economy as a result of investment in and application of new ICT technology. Throughout, we concentrate on conventional measures of economic activity and ignore other possible aspects of the new economy. It is clear that mobile phones and the internet are important tools for those trying to keep track of events in parts of this new economy, but it is doubtful whether they have caused it. These and other changes to our lifestyles probably have important implications for economic welfare and management, but we do not consider them further here.

The paper is organised as follows. In the first section we look at the extent of investment in ICT in the UK, comparing developments here with those in the US. We focus next, in the second section, on recent trends in productivity growth in the UK. So far, there has been no UK productivity miracle; on the contrary, productivity growth actually slowed down in the late-1990s. We therefore investigate whether there are any grounds for believing that the beneficial effect of ICT on UK productivity might have been delayed. In the third section we discuss the possible effect of ICT on price setting by firms. In broad terms, our findings are that there is little to link the improvements in the British economy to ICT developments. Finally, we discuss the more likely reasons for Britain's economic improvement since the recession and look forward, arguing that there is a strong case for expecting a more robust pick-up in productivity growth in the near term as the effects of ICT investment finally emerge.

ICT investment in the UK

One of the more important difficulties in discussing the impact of ICT investment and its applications on the UK economy is in measuring its size. [1] There are a number of aspects to this, not least the familiar difficulty in measuring how much company and household spending should be counted as investment. It is clear, for example, that the output of programmers employed within a firm is less likely to be counted as capital investment than the value of software purchases. But even within the traditional definition of investment there is no official data series measuring gross fixed capital formation in computer or other ICT equipment. Instead, figures have to be inferred from input-output tables. A further difficulty is in deflating current spending to a sensible measure of the volume of capital spending. Our purpose in this section is to describe a series measuring investment in computers and to outline estimates of the capital stock in this area. [2]

As a preliminary, chart 2 shows business investment as a share of GDP at both current and constant prices since 1965. This indicates clearly that the volume of business investment is very strong at present, but that capital spending in current prices as a share of national income is similar to its historical average. The difference reflects the historically low relative price of capital goods in the UK and is clearly affected by falling computer prices. [3]

Nominal investment spending on computers

In order to calculate how much of this investment is in computers it is necessary to use input-output tables which break down investment spending into its component products. [4] Computers are included within the office machinery category, which also includes other office equipment. Chart 3 shows estimates of investment in office machinery as a percentage of machinery and equipment investment excluding transport equipment. Also shown is investment in electronic components (available only from 1991).

The share of computer and office equipment investment in total plant investment has been rising fairly steadily over time. The latest figure for 1998 is 18.8 per cent, slightly below the peak level of 19.2 per cent in 1996. In 1998, this amounted to spending of around [pound]9 billion. This is sufficient to supply each employed person with a new [pound]1000 computer once every three years. As we have noted, this estimate of capital spending on computers is unlikely to include spending on developing inhouse software; nor does it include other ICT spending such as that to lay cables and build mobile telephone networks.

Chart 4 shows how total spending on computers in 1997 was spread across five main industrial sectors in comparison with their share of national output. It is clear from this that computer investment was disproportionately high in the business services sector and relatively low in the public sector. [5]

Deflation

It is well known that the price of a standard computer has fallen dramatically in relative terms over the past 25 years at the same time as its processing power has increased. This means that any measure of the volume of investment in computers derived simply by deflating nominal investment by an average investment price deflator will underestimate the growth in computing power. One approach to deflation is to use producer price indices relevant to the computer industry. This shows the price of computing equipment as halving between January 1994 and April 2000. [6]

However, this price fall makes only a small allowance for the improvement in the quality of computers over this period. [7] In the US, the official NIPA computer price index is a constant-quality series constructed from so-called 'hedonic' price regressions. [8] This shows much larger falls in the prices of computers on a constant-quality basis. Since US computers are unlikely to be very different from those used in the UK, we are unlikely to introduce a large error by using such indices to deflate UK capital spending. According to Jorgenson and Stiroh (2000) the price of computers for investment declined at an average rate of 18 per cent per annum between 1960 and 1995 but this has further accelerated to 27.6 per cent since then. Following Broadbent and Walton (2000), we convert this series to UK prices using the average sterling-dollar exchange rate in each period. On this basis, the volume of investment in computers grew at an average annual rate of 45 per cent in the UK between 1975 and 1997.

Despite this very high growth rate over a number of years, the actual stock of computers remains relatively small. This is because the economic lives of computers are very short relative to most other types of capital. For example, Lansbury et al. (1997) calculated the stock of computer capital on the assumption of economic lives shown in table 1. These were informed by discussion with industrial experts and the Office for National Statistics.

On this basis, and at 1990 prices, Lansbury et al. estimated the gross stock of computers in 1994 to be [pound]23.8 billion. This is about 5 per cent of the gross stock of plant and machinery, but less than 1 per cent of the overall gross capital stock when dwellings and other buildings and works that have substantially larger asset lives are included.

However, Lansbury et al. chose not to use hedonic pricing in their calculations. We can update capital stock estimates to 1995 hedonic prices using a simple version of the perpetual inventory method (PIM) which assumes an economic life of four years for computers. Here the gross stock version assumes that investment remains in the stock for four years and then drops out of the calculation, whereas the net stock assumes that decay occurs in an approximately geometric fashion throughout the life of the asset. [9] The former can be thought of as a useful upper limit on the size of the capital stock if the rate of economic decay is higher than the rate of productive decay (Whelan, 2000). Chart 5 shows the estimated stock of computers in 1995 hedonic prices as a proportion of the business sector capital stock (NIESR estimates of the net capital stock covering the business sector) under both assumptions about the rate of decay.

It is clear from these calculations that while the share of computing equipment in the overall industrial capital stock remains small, it is nevertheless increasing very rapidly. By 1998 (the last available observation) it made up around 4.5 per cent of total capital when measured using geometric depreciation rates and just over 6.3 per cent assuming no depreciation. In part the rapidly increasing share reflects the hedonic price assumption, but the chart also indicates that this would be estimated to be growing quickly using conventional deflators.

It is possible to make some comparisons with the US experience by comparing the increase in the stock of computers in the UK with those estimated by Oliner and Sichel (2000) for the US, shown here in table 2. The increase in the capital stock appears to have been faster in the UK than in the US over the periods 1978-90 and 1991-5. It is unclear how much to make of these differences given the uncertainties inherent in constructing estimates of capital stocks in the two cases. Over the period from 1996-8, the rate of growth of the capital stock in the US, as estimated by Oliner and Sichel, is slightly faster than that for the UK economy.

These estimates underlie impressive growth in the share of computers in the capital stock; however, unless the stock is itself significant, this is unlikely to have much effect on the rate of economic growth. According to standard growth accounting, the contribution of growth in the stock of computers will be given by the growth rate weighted by the income share of computers. The income share of computers is itself a function of the nominal stock of computers multiplied by the gross rate of return divided by total national income in nominal terms. We follow Oliner and Sichel in assuming a fixed net real rate of return of 4 per cent, a depreciation rate of 30 per cent and a rate of capital loss equal to -34 per cent. The outcome from combining the estimated change in the capital stock and the income share of computers is shown in table 3. This also shows the estimated revenue shares for the different time periods and the results for the US economy from the Oliner and Sichel (2000), Whelan (2000) and Jorgenson and Stiroh (2000) studies.

The US and UK evidence suggests a similar increasing contribution to growth from the sharp pick-up in industrial investment in computers. Whereas in the 1980s and first half of the 1990s this was worth only around a quarter of one percentage point, it picked up to around half of one percentage point in the second half of the decade.

Other evidence on computer investment

The preceding figures record capital investment in computers by industry. Household purchases are treated as consumption in the national accounts and included with other items in an 'audio-visual equipment' category. [l0] Expenditure in this area has grown by around 2.5 per cent per annum in volume terms since 1996 after very little growth in the preceding years. This probably reflects strong household investment in computers. Certainly survey evidence points to widespread computer ownership. According to the 1998-9 Family Expenditure Survey, 33 per cent of all households have a home computer and 10 per cent have an internet connection (Richardson and Bowman, 2000). More recent information from AC Nielson eRatings.com (released 8 September, 2000) reports that 19 million people access the internet from home in the UK.

Surveys of business confirm the impression formed earlier that computer ownership is now very widespread. The Department of Trade and Industry (DTI) published a study benchmaking aspects of the digital economy in December 1998 with the Competitiveness White Paper and has since updated some of the figures (DTI, 2000). This showed that, in 1999, 80 per cent of UK businesses had a PC with Modem, 60 per cent had access to the internet, 50 per cent had a website and 30 per cent took part in e-commerce. It also showed significant growth in business use of ICT with use of the internet and websites up by 80 per cent and 90 per cent respectively since 1997.

Productivity growth in the UK

One of the key differences between the economic performance of the UK and the US in recent years has been in productivity. Labour productivity growth has tended to be faster in the UK than in the US. This is partly because UK firms can imitate and import the technological advancements made in countries that have a higher level of technology, such as the US. Since the mid-1990s this position has reversed, as labour productivity growth has accelerated in the US but slowed down in the UK. This is illustrated in chart 6. Whereas productivity growth in the US picked up to average 2.4 per cent per annum in the second half of the 1990s after averaging 1.2 per cent per annum from 1986 to 1995, in the UK it fell from 2 per cent to 1.5 per cent per annum over the same period. The US new-economy labour productivity growth 'miracle' is simply not present in the aggregate UK data.

Table 4 breaks down the recent behaviour of aggregate productivity growth in the UK into that of manufactures and services. [11]

The deterioration in UK productivity growth in the second half of the 1990s follows a number of years where productivity growth had been improving. On an output per hour basis, productivity growth picked up from about 2 per cent per annum between the cyclical peaks of 1973 and 1979 to 2.2 per cent per annum between the peaks of 1979 and 1990 and to 2.5 per cent from the 1990 cyclical peak to the first quarter of 2000.

This steady improvement in aggregate productivity growth hides some stark differences in the performance of manufacturing and non-manufacturing industries. While manufacturing productivity growth was fastest in the 1980s, reflecting widespread restructuring, it was more rapid in non-manufacturing in the 1990s, when it was higher by a full percentage point than it had been in the 1980s.

Splitting the period from the 1990 cyclical peak into two indicates that there has been a slowdown in productivity growth in both manufacturing and non-manufacturing industries. This has been most obvious in manufacturing, where productivity growth in the first half of the decade was almost identical to that in the 1980s but in the second half was very similar to that of the 1970s. Chart 7 shows that performance in US manufacturing has been much better.

Thus one of the major differences between the US and UK economies in the late 1990s is in the behaviour of manufacturing productivity. The sharp slowdown in the UK reflects the stagnant output in this sector, due mainly to the impact of the strong pound. It should also be noted that according to Gordon (1999) and others, the pick-up in US manufacturing partly reflects the explosive growth in firms manufacturing computers.

There has also been a slight slowdown in productivity growth outside the manufacturing sector. This can be broken down according to different industries as shown in table 5. Here we concentrate on four industrial groupings: business services including financial intermediaries and real estate companies, wholesale and retail distribution, public sector industries and a composite category including agriculture, mining, construction, transport and communication. This decomposition ignores output from the oil industry and from the letting of dwellings (including imputed rent); neither have we allocated the adjustment for financial services included in the national accounts. Productivity growth in the industries on which we focus is virtually identical in both halves of the 1990s. The slowdown outside manufacturing is mainly accounted for by a reduction in the rate of growth of oil output from about 10 per cent per annum to virtual stagnation (not shown in the table).

These figures indicate the sources of the pick-up in productivity growth in non-manufacturing in the 1990s compared with the 1980s. Much of this is in the production of public sector services where productivity growth from the 1990 peak averaged 1.65 per cent per annum compared with 0.12 per cent per annum in the 1980s.

All the other non-manufacturing sectors showed similar productivity growth in both periods.

Among these industries, there was no slowdown in productivity growth in the second half of the 1990s. However, among the individual sectors, the rise in productivity growth in the business services sector almost compensates for declines elsewhere. It is worth noting that the business services sector can be identified as an area where new ICT technology is being used intensively.

Accounting for productivity growth

It is useful to estimate how much of the productivity change in any industry is due to increases in capital per worker and how much is due to other factors. Exercises of this type need to be painstaking to cover the many possible influences on productivity growth, including labour quality (see O'Mahony, 1999, for recent comparative cross-country evidence). But for current purposes where we are looking for evidence of change in the determinants of growth over a recent period, a more broad-brush approach is acceptable.

Suppose output is produced using a simple Cobb-Douglas production function of the following form:

Y = [AK.sup.[alpha]] [([Ne.sup.[lambda]t]).sup.1-[alpha]]

where Y is output, A is a constant, K is the net capital stock, N is hours of labour input, [lambda] is the rate at which labour is augmented by technical progress and [alpha] is a technological coefficient. Then, by taking logarithms and differencing, it is possible to split productivity growth into one part determined by capital deepening (increasing the capital stock per hour of labour input) and another determined by technological progress:

[delta]ln(Y/N) = [alpha][delta] ln(K/N) + (1-[alpha])[lambda]

Although it is not possible to observe the rate of technical progress, it is possible to use the equation to infer it, on the basis that the equation is an accurate description of how output is determined. While there are many good reasons for expecting the relationship determining output to be considerably more complicated, this expression is useful as a means of classifying how much productivity growth is due to capital deepening and how much to other factors represented here by technology. Table 6 splits productivity growth into the contributions of capital deepening and technical progress. In all cases, [alpha] is set in line with the share of capital in each industry in 1995.

Clearly, the factor labelled 'technical progress' is a catch-all term reflecting the many factors other than capital deepening which influence productivity growth. It appears to have been fastest in the composite industrial sector - made up of agriculture, mining, construction, transport and communication - where productivity growth has been rapid. By contrast, it has been relatively slow in the private sector distribution and business services sectors where productivity growth appears to be due to capital deepening.

On the basis of these figures, we can draw some tentative conclusions about the proximate causes of the slowdown in productivity growth in the UK in the second half of the 1990s. First, there appears to have been a slowdown in the rate of capital-deepening in all of the collections of industries included in these figures. While the capital stock has grown quickly, so has employment. Second, there has been a pick-up in the contribution of technical progress to productivity growth in the business services and public sectors. Since the former uses computers intensively, this could be an indication that computers are now having an impact in these industries over and above their contribution to capital deepening. Note that this contrasts with the finding of Jorgenson and Stiroh (2000) that the finance sector has continued to lag in productivity growth in the US. Third, there has been a substantial slowdown in the rate of 'technical progress in manufacturing. We seriously doubt that this is related to technological factors and see it more as a reflection of the general difficulties faced by manufacturing as a consequence of the high value of the pound.

New-economy industries

One of the features of the recent US experience has been the pick-up in growth in the industries producing goods and services for the new economy. Indeed, it has been claimed by Gordon (1999) that the majority of the pickup in productivity in the US in the late 1990s has been in the computer-producing rather than computer-using industries. Here we examine growth and productivity developments in several similar UK industries in both manufacturing and non-manufacturing. We focus on SIC 30 (manufacture of computers and office equipment), SIC 32.2 (manufacture of television and radio transmitters and apparatus for line telephony), SIC 32.3 (manufacture of television and radio receivers sound or video recording), SIC 72 (computer and related activities) and 64.2 (telecommunications). [12]

As in the US, the new-economy sectors in the UK have grown strongly relative to the other industrial sectors, although there is little evidence of any recent acceleration. SIC 30, which includes the manufacture of computers, grew by an average annual rate of 13.8 per cent between 1979 and 1996 and 11.2 per cent between 1996 and 2000. A similar slight slowdown is evident in SIC 32.30, manufacture of television and radio receivers sound or video recording, which grew at an annual rate of 7.1 per cent between 1979 and 1996 and 4.4 per cent between 1996 and 2000. However, growth has accelerated in SIC 30.20 (manufacture of television and radio transmitters and apparatus for line telephony) from 4 per cent per annum in the first period to 16.7 per cent per annum in the second. Combined, these three sectors make up only 5.8 per cent of total manufacturing output, or around 1 per cent of GDP.

Growth has also been fast compared to the average for the economy as a whole in the new-economy parts of the service sector. SIC 64.2, telecommunications, grew by an annual average rate of 5.9 per cent between 1979 and 1999, with 11.2 per cent per annum after 1996. Growth has also been faster in SIC 72, computer and related activities, the annual growth rate of which is displayed in the table below. This sector accounts for 1.3 per cent of total service sector output, or about 0.85 of value added in the whole economy.

Fast output growth means that the new-economy industries have increased their share of total output over time (chart 8). The manufacturing new-economy sectors accounted for 0.83 per cent of the total economy in 1979, 0.96 per cent in 1995 and 1.26 per cent in 1999. SIC 72 accounted for 0.47 per cent of total output in 1986, 0.85 per cent in 1995 and 1.31 per cent in 1999. SIC 64.2 accounted for 0.77 per cent of total output in 1979 and 1.59 per cent in 2000. As an aggregate, the neweconomy sectors account for about 4.15 per cent of the whole economy in 1999.

The increasing size of the new-economy sectors, combined with their fast rate of growth, means that in the period since 1995 they have contributed on average 0.47 percentage points of growth. As table 7 shows, the greatest relative contribution has been from the service sector rather than manufacturing.

Labour productivity growth in the new-economy sectors has been on average much faster than that in the rest of the UK economy. Data prior to 1996 is readily available only for SIC 30 and SIC 72. The average rate of labour productivity growth in SIC 30 between 1979 and 1999 was 13.7 per cent per annum, whereas in SIC 72 it averaged 2.6 per cent per annum (data for 1981-99), not much faster than that for the economy as a whole.

Since the end of 1995, the rate of productivity growth in US computer manufacturing has increased markedly. Computer output per hour worked grew by 41.7 per cent per annum from the end of 1995. The equivalent figure for the UK is 12.3 per cent per annum, slower than before 1995. Similarly, in SIC 72 the average rate of labour productivity growth has been only 0.6 per cent per annum, now slower than that for the economy as a whole. An increasing share of IT production in the UK has been taken up by services -- software and computer maintenance.

The use of hedonic price data in the US but not the UK may explain part of this difference. In ONS (2000) it is calculated that manufacturing output in the UK in the first quarter of 2000 would be over 6 per cent higher if US methods of deflation were used. This is a substantial change and indicates the degree to which differences between the US and the UK in recent years reflect differences in measurement.

Price setting

One of the key benefits of new ICT technology for the British economy is expected to be in relation to price setting. It is widely believed that the internet will reduce search costs, lower barriers to entry, shorten the supply chain and generally increase product market competition (Wadhwani, 2000).

This process could have three main effects. First, businesses could be forced to cut margins as they lose the economic rents arising from lack of awareness of alternatives on the part of their customers. Second, some businesses could seek to restore their profitability by cutting their costs and so raising productivity. Third, the search for profits could encourage innovation, new investment and faster technical progress.

To illustrate the potential benefits to the British economy of the first of these effects, we have carried out a simulation on the National Institute model of a reduction in margins on consumer goods. The magnitude of the change would be sufficient to reduce consumer prices in the long run by 1 per cent if there were no response in wages or other economic variables. In fact, because the government has an inflation target the full impact of the reduction in margins will not be felt on inflation at all. Instead it will show up mainly in a reduction in interest rates and a rise in the sustainable level of employment and output. This is outlined in chart 9.

The chart shows that there would be sustained real effects on the economy from lower desired profit margins caused by greater domestic competition. In the short run, for this size of shock, interest rates are sometimes about 1/2 per cent lower than they would otherwise have been as policy responds to lower inflationary pressure in the domestic economy. This allows output, employment and real wages to increase by between 1/2 and 1 per cent in the long run. The rise in employment reflects a fall in the NAIRU as inflationary pressure can be controlled with a higher level of employment due to the willingness of firms to concede a lower mark-up. This is reflected in a reduction in profitability in the short to medium term.

More generally there is relatively little evidence that the benign inflationary conditions of the past few years have been caused by a reduction in profit margins. While goods price inflation has been negligible in the UK in the past year, this partly reflects the impact of the high level of sterling on import prices and the effect of technical change on the prices of computers, mobile telephones and other ICT equipment. Certainly, there is nothing in the National Institute's forecasting equations to suggest that prices have been unusually depressed relative to their determinants.

The question then arises as to whether the internet has caused a similar shock to that described above and, if so, what is its likely scale?

The phenomenal recent growth of the use and awareness of the internet is clear from experience and is documented above and in Wadhwani (2000). Wadhwani reported that 18 per cent of the UK population were users of the internet in December 1998. He noted (p.186), "anecdotal evidence suggests a significant increase in internet penetration since then". More recent survey evidence by Nielsen (2000) shows that, in the second quarter of 2000, 42 per cent of UK adults had access to the internet at home and 24 per cent had access at work. Moreover, the same survey recorded that 14 per cent of UK adults had browsed for products online in the previous six months while 9 per cent had made a purchase.

However, it is also clear that it is still too early to see the implications of any increased competition on price setting. While there is survey evidence suggesting that for a range of goods prices are significantly lower on the internet than in the typical High Street, the full impact of the internet must wait until High Street prices are themselves reduced. For example, a Barclays Capital survey carried out in December 1999 reported that books, one of the items most heavily traded on the internet, were 22 per cent cheaper on the internet than in the High Street. But book prices generally were 3.0 per cent higher in April 2001 than they had been a year earlier, as compared to the 0.1 per cent fall in the goods component of the RPI over the same period.

This is borne out by evidence on the profitability of British industry, summarised in chart 10, which shows no decline in the net rate of return on capital in the service sector in recent years. There has been a sharp decline in profitability in manufacturing, reflecting the industry's response to the strong pound. In fact, the response of manufacturing industry to the high level of sterling may have some important lessons for the way in which industry as a whole might respond to greater product market competition. In particular, manufacturers have made it plain that there is a limit to how far margins can be reduced before they stop producing in the UK entirely. That is, margins are not infinitely flexible in a downward direction, since firms are in business to make a profit. Manufacturers have seen profit margins on their export business reduced to a level which many regard as unsustainable. Anecdotal evidence suggests that while the current level of margins is sufficient for firms to cover their variable c osts and remain in business, future investment in the UK is not viable with this rate of profit. Similarly with the internet, margins cannot sustainably be reduced below the level at which marginal firms make normal profits. It would appear that investors in the stocks of these firms have taken a similar view.

Furthermore, there is relatively little evidence that manufacturers have succeeded in using more competitive pressures in the product markets to reduce their costs by raising productivity. As we saw earlier, manufacturing productivity has been adversely affected by the strength of sterling. It is possible that manufacturers will intensify their efforts to improve productivity over the short term since there appears to be no other way of raising their profitability. However, it can reasonably be argued that profit-seeking firms should be cost minimising at all times, irrespective of product market conditions, and as such there is no reason why productivity should improve now. Indeed, to the extent that productivity relies on capital investment which is itself determined by prospective profitability, there is a good chance that productivity will worsen.

Thus, in summary to this section, we see lower profit margins as a possible effect of the internet which would have a generally beneficial impact on the overall economy. However, at this stage there is no obvious evidence of such an effect. Indeed, there may not be much scope for cutting margins significantly given that firms must expect to deliver a sufficient rate of return on capital to their shareholders. In this case, any beneficial effect of the internet on pricing will need to come from the incentives and opportunities it offers to firms to cut their costs. This supports the view of Brookes and Wahhaj (2001) that its most important impact will be to encourage cost reduction through business-to-business trade.

This effect may take some time to have an observable impact and could work differently from a reduction in margins. Chart 11 shows an illustrative simulation on the National Institute model of an increase in technical progress of 10 per cent spread over a ten-year period. In contrast to the situation where profit margins are reduced, there is no long-run impact on employment or the distribution of income. Output rises because of a higher level of productivity which is reflected in higher real wages. It is also worth noting that real and nominal interest rates rise in the short term. This possibility is discussed theoretically in Buiter (2000) and arises because the benefits of higher productivity raise both aggregate demand and supply. In the National Institute model, demand is increased by more than supply in the short term as forward-looking households and firms increase spending in light of higher future productivity levels. This leads to higher interest rates in the short term so as to prevent inflation i ncreasing. An implication of this is that equity prices actually fall initially when future productivity is expected to rise because higher interest rates offset the beneficial effect of higher future profits.

Conclusion

It is clear from casual observation that we are in the midst of a revolution in information technology. However, in contrast to the US experience, there has been little obvious impact on the British economy at the aggregate level. In the second half of the 1990s, productivity growth was about 1 percentage point lower than in the first half of that decade when measured on an output per hour basis. While goods price inflation has remained very subdued as unemployment has fallen, this does not appear to be due to a new-economy effect on pricing. What then is responsible?

A key factor is the improvement in the behaviour of the British labour market in the 1990s, which is evident in low unemployment. That this has been brought about by a willingness of workers to accept a lower share of national income is shown in chart 12. This shows employee compensation as a share of gross domestic product at basic prices, that is, excluding indirect taxes and subsidies on products. We also show the share when self-employment income, measured in the national accounts until 1998, is added to employee compensation. [13] Both series indicate a generally falling share of labour in national income through most of the 1990s. The recent increase reflects the reduction in profits in manufacturing due to the high pound.

It would appear that much of the improvement in Britain's economic performance in the 1990s was due to the fact that the recession at the beginning of the decade caused a large gap to open between the unemployment rate and its equilibrium. This put downward pressure on real wages and the labour share and at the same time allowed output to expand significantly without sparking inflationary pressure while encouraging firms to substitute labour for capital.

The decline in equilibrium unemployment and the consequent increase in employment from the early 1990s recession partly accounts for the fact that new-economy factors have not been evident so far in the UK: productivity growth tends to be slow when employment is increasing rapidly. As the economy reaches a high level of employment, there are some indications that its benefits are just about to reveal themselves:

1. As far as we are able to tell from official figures, there has been extensive investment in ICT equipment and infrastructure in recent years and it is only now becoming an important component of the aggregate capital stock. The growth in computer investment could not be expected to account for a pick-up in productivity growth in the late 1990s of more than a half of a percentage point. Similarly, the growth in the widespread use of the internet is very recent and it is unrealistic to expect it to have had an important impact so far. In this aspect the UK experience has been reasonably similar to the US. Finally, an important difference between the US and UK experience in the second half of the 1990s is that the UK new-economy producing sectors are smaller than in the US and their productivity growth has not been as fast, partly because of the way output is measured. Gordon (1999) and others have claimed that the fast pace of technological change in this sector accounts for much of the increase in productiv ity growth in the US.

2. The slowdown in productivity growth in the second half of the 1990s is relatively easy to explain in terms of the old economy. Its main causes are the slowdown in the manufacturing and oil industries, alongside a reduction in the rate of capital deepening as the economy has generated another 1.5 million jobs. Contrary to the US experience, there is evidence of a strong pick-up in productivity growth in the key ICT-using business services sector. With little slack left in the labour market, the factors that have obscured an underlying improvement in the economy are now expected to evaporate, although this depends on how the manufacturing sector recovers from its recent lacklustre performance. Therefore, there is now a good chance of a strong revival in productivity growth even without allowing for any special new-economy effects. Adding the effects of ICT investment strengthens the likelihood of this.

3. The internet seems certain to increase competition in the product market, but we have argued that the need for companies to make profits leaves relatively little scope for them to reduce profit margins. Instead, the main impact is expected to be in encouraging further cost reduction and raising productivity, thereby adding to the likelihood of a productivity revival. This remains the case despite recent events.

(*.) University of Nottingham.

(**.) Bank of England. This paper was completed while both authors were at the National Institute of Economic and Social Research. The views expressed do not necessarily reflect those of the Bank of England. We are grateful to participants at seminars hosted by NIESR, the University of Kent and the House of Commons Treasury Committee for comments and suggestions. We are responsible for any remaining errors.

NOTES

(1.) Richardson and Bowman (2000) discuss how the Office for National Statistics proposes to measure the use of e-commerce.

(2.) See also Oulton, N. (2001).

(3.) See also Bloom and Bond (2001).

(4.) Because input-output tables are not available for every year, we also used information supplied by the Finance Leasing Association and the Office for National Statistics.

(5.) Output to say how industries are defined.

(6.) See ONS (2000), p.88.

(7.) This is based on the 'manufacturers' cost method' which measures quality improvements on the basis of estimates of the cost of changes in product specification (see ONS, 2000).

(8.) This is discussed in the context of investment equations in Tevlin and Whelan (2000).

(9.) It is assumed that 0.85 of investment lasts into its second year, 0.35 into the third year and 0.15 into the fourth year.

(10.) Consumer Trends, Table 2.4.

(11.) The estimates of productivity growth are based on data prior to the data revisions by the ONS arising from the incorporation of information from the Annual Business Enquiry. The effect of these changes is primarily on the level of productivity, although there is some very mild effect on the calculations of productivity growth over the most recent time periods in the table. These changes have the effect of slightly raising productivity growth in the manufacturing sector (employment falls more quickly than under the old data) and lowering productivity growth in the non-manufacturing sectors and the total. As these changes have no impact on the conclusions reached from this section of the paper it was decided to leave a full assessment of their impact to future research.

(12.) Additional data for this section were provided by the ONS.

(13.) Batini, Jackson and Nickell (2000) show some measures of the labour share which do not display a downward trend over this sample period. However, this is based on an estimate of self-employment incomes which is inferred from relative movements in employment and self-employment and contains no actual information on the incomes of the self-employed.

REFERENCES

Batini, N., Jackson, B. and Nickell, S. (2000), 'Inflation dynamics and the labour share in the UK', External MPC Unit, Discussion paper no. 2, Bank of England.

Bloom, N. and Bond, S. (2001), 'UK investment: high, low, rising, falling?' Institute for Fiscal Studies, Briefing Note no. 18.

Broadbent, B. and Walton, D. (2000), 'How "new" is the UK economy', Goldman Sachs Global Economics Paper no. 43.

Brookes, M. and Wahhaj, Z. (2001)' 'The economic effects of business-to-business internet activity', National Institute Economic Review, 175, January, pp. 95-108.

Buiter, W. (2000), 'The new economics and the old monetary economics', Bank of England Quarterly Bulletin, May, 40(2).

Department of Trade and Industry (DTI) (2000), UK Competitiveness Indicators 1999.

Gordon, R.J. (1999), 'Has the "New Economy" rendered the productivity slowdown obsolete?', mimeo dated June 14, 1999.

Jorgenson, D.W. and Stiroh, K.J. (2000), 'Raising the speed limit US economic growth in the information age', mimeo.

Lansbury, M., Soteri, S. and Young, G. (1997), 'Improving estimates of the UK capital stock', report for the Office for National Statistics prepared at NIESR.

Nielson, A.L (2000), 'Nielsen//NetRatings Global Internet Trends', Q2 2000. http://eratings.com.

Office for National Statistics (ONS) (2000), 'Review of short term output indicators', National Statistics Quarterly Review Series, I.

Oliner, S.D. and Sichel, D.E. (2000), 'The resurgence of growth in the late 1990s: is information technology the story?' Federal Reserve Board mimeo.

O'Mahony, M. (1999), Britain's Productivity Performance 1950-1996, An International Perspective, London, National Institute of Economic and Social Research.

Oulton, N. (2001). 'ICT and productivity growth in the UK', Bank of England Working Paper (forthcoming).

Richardson, C. and Bowman, J. (2000) 'E-Commerce@the.ONS.UK', Economic Trends, April, pp. 89-99.

Tevlin, S. and Whelan, K. (2000), 'Explaining the investment boom of the 1990s', Federal Reserve Board mimeo.

Wadhwani, S. (2000), 'The impact of the internet on UK inflation', Bank of England Quarterly Bulletin, May, 40(2), pp. 184-98.

Whelan, K. (2000), 'Computers, obsolescence, and productivity', Federal Reserve Board mimeo.

Table 1. Economic life assumptions on Computer-related investment
(years)
Year investment was undertaken Average life
1975-1979 10
1980-1981 8
1982-1983 7
1984 6
1985-1989 5
1990 onwards 4
Source: Lansbury et al. (1997).
Table 2. Growth in stock of computers
Increase in computer 1979-90 1991-5 1996-8
capital stock (per cent)
UK - gross stock 40.6 14.6 31.3
 - net stock 34.1 17.9 33.3
Oliner and Sichel (2000) 31.3 17.5 35.9
Table 3. Contribution of computer investment to GDP growth
 1979-90 1991-5 1996-8
Income shares (per cent of GDP)
 UK 1.0 1.4 1.5
 Oliner and Sichel estimate of US 1.0 1.4 1.8
Contribution to growth
 (percentage points)
 UK 0.33 0.26 0.49
 Oliner and Sichel (2000) 0.27 0.25 0.63
 (1996-9)
Whelan (2000) - 0.33 0.82
 (1990-95) (1996-98)
Jorgenson and Stiroh (2000) - 0.15 0.36
 (1996-8)
Table 4. UK productivity growth (peak to peak, annual averages, per
cent)
 Output per head
 Whole economy Manufacturing Non-manufacturing
1973Q2-1979Q3 0.91 0.53 1.05
1979Q3-1990Q1 1.77 3.91 1.20
1990Q1-2000Q1 2.07 2.49 2.09
Note:
1973Q2-2000Q1 1.68 2.58 1.50
1990Q1-1995Q1 2.64 3.97 2.45
1995Q1-2000Q1 1.51 1.02 1.74
 Output per hour
 Whole Economy Manufacturing Non-manufacturing
1973Q2-1979Q3 2.01 1.43 2.17
1979Q3-1990Q1 2.19 4.09 1.57
1990Q1-2000Q1 2.49 2.72 2.49
Note:
1973Q2-2000Q1 2.26 2.95 2.05
1990Q1-1995Q1 2.95 4.05 2.73
1995Q1-2000Q1 2.03 1.41 2.24
Source: ONS, UK National Accounts; ONS, Labour Market Trends.
Table 5. UK non-manufacturing productivity growth (peak to peak,
annual averages, per cent)
 Agriculture, Business Public Distribution
 mining, services
 utilities,
 construction,
 transport and
 communication
Output per hour
1973Q2-1979Q3 2.69 2.39 1.44 0.36
l979Q3-1990Q1 4.21 2.21 0.12 1.94
1990Q1-2000Q1 4.47 2.02 1.65 1.99
1973Q2-2000Q1 3.95 2.18 0.99 1.58
Note:
1990Q1-1995Q1 5.10 1.20 2.12 2.17
1995Q1-2000Q1 3.85 2.86 1.17 1.80
 Total of these Non-
 manufacturing
 industries
Output per hour
1973Q2-1979Q3 1.82 2.17
l979Q3-1990Q1 2.05 1.57
1990Q1-2000Q1 2.49 2.49
1973Q2-2000Q1 2.16 2.05
Note:
1990Q1-1995Q1 2.52 2.73
1995Q1-2000Q1 2.47 2.24
Source: ONS, UK National Accounts; ONS, Labour Market Trends.
Note: Oil industry, output of dwellings (mainly imputed rent) are
excluded. Financial intermediation services indirectly measured are
included.
Table 6. Contribution of capital deepening (I) and technical Progress
(II) to productivity growth (peak to peak, annual averages, per cent)
 Agriculture, Business Services
 mining, utilities,
 construction,
 transport and
 communications
 I II I II
1973Q2-1979Q3 1.74 0.95 2.01 0.38
1979Q3-1990Q1 0.73 3.48 1.80 0.41
1990Q1-2000Q1 3.17 1.30 1.68 0.34
Note
1973Q2-2000Q1 1.87 2.08 1.81 0.38
1990Q1-1995Q1 3.63 1.48 2.24 -1.05
1995Q1-2000Q1 2.72 1.13 1.14 1.72
 Public Distribution Manufacturing
 I II I II I
1973Q2-1979Q3 0.55 0.89 1.90 -1.54 1.46
1979Q3-1990Q1 0.26 -0.15 1.59 0.35 0.92
1990Q1-2000Q1 0.20 1.45 1.47 0.52 1.04
Note
1973Q2-2000Q1 0.31 0.69 1.62 -0.03 1.09
1990Q1-1995Q1 0.72 1.40 1.51 0.66 1.17
1995Q1-2000Q1 -0.32 1.49 1.43 0.37 0.92
 II
1973Q2-1979Q3 -0.03
1979Q3-1990Q1 3.17
1990Q1-2000Q1 1.68
Note
1973Q2-2000Q1 1.86
1990Q1-1995Q1 2.88
1995Q1-2000Q1 0.50
Source: ONS, UK National Accounts; ONS,Labour Market Trends.
Note: Column I shows the contribution to productivity growth of
capital-deepening while column II shows the implied contribution of
technical progress.
Table 7 The contribution of high technology industries to
economic growth (per cent per annum)
Industry code 1980-95 1996-99
30 0.029 0.082
3220 0.009 0.074
3230 0.009 0.007
Manufacturing 0.047 0.163
72 0.051 0.151
64.2 0.041 0.157
Services 0.092 0.308
Total 0.139 0.471
Note: table shows the growth rate of the sector weighted by its
share in GDP.

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

art 8. The [Graph omitted] lustries

4.5

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]

[Graph omitted]