the assessment: the new economy - university of … · the assessment: the new economy ... economy...

241© 2002 OXFORD UNIVERSITY PRESS AND THE OXFORD REVIEW OF ECONOMIC POLICY LIMITED

THE ASSESSMENT: THE NEWECONOMY

OXFORD REVIEW OF ECONOMIC POLICY, VOL. 18, NO. 3

JONATHAN TEMPLEUniversity of Bristol1

The remarkable economic success of the United States in the 1990s led many observers to talk about a ‘NewEconomy’. This paper provides an overview of the main issues, including faster productivity growth, thestability of inflation despite very low unemployment, the reduction in output volatility, the role of monetarypolicy, and the boom in the stock market. The paper also considers whether or not the acceleration inproductivity growth can be sustained, and the possible implications for the rest of the world.

I. INTRODUCTION

On 26 November 2001 the National Bureau ofEconomic Research announced that the USeconomy had entered a recession in March 2001.This announcement, made by the Bureau’s busi-ness-cycle-dating committee, provided official con-firmation that the heady days of the 1990s wereover. The collapse in the stock-market valuations ofInternet companies had already taken the shine offthe optimism surrounding the ‘New Economy’. Atthe time of writing, in the wake of the spectacularcollapse of Enron and mounting evidence of dubiousaccounting practices elsewhere, that optimism seemsvery distant.

It is therefore a good idea to restate some of thereasons why the New Economy has been thesubject of so much excited comment and analysis.The US expansion which came to an end in 2001 hadlasted exactly 10 years, one of the longest unbrokenexpansions ever recorded by an industrial country.The rate of inflation stayed low throughout, eventhough the unemployment rate fell to a 30-year low.Faster productivity growth ultimately translated intofaster growth in real wages. The incidence ofpoverty fell, and wage inequality finally stabilized.

The expansion had also been accompanied by amassive boom in the stock market so that, by the late1990s, price–earnings ratios for the aggregate US

1 I am grateful to Andrew Glyn for very helpful comments and discussion. Any errors are my responsibility.

242


market were at the highest levels ever observed inthe twentieth century. The most dramatic and glam-orous aspect of the boom was the rise of the Internetcompanies or dot-coms, many of which createdhuge paper fortunes for their founders almost over-night, and gave plenty of leeway to those whoclaimed that the old economic rules were beingrewritten.

As a result, the remarkable performance of the USeconomy in the 1990s has been much discussed.This article reviews the current state of knowledgeon the macroeconomics of the New Economy in theUSA. It highlights the principal reasons why thename has become so popular, and the sense in whichit probably remains justified. A consensus hasemerged that, away from the clamour of the stockmarket, something genuinely important was hap-pening—an acceleration in the rate of productivitygrowth, with implications for living standards and, inthe short run, for inflation and unemployment.

These events are important beyond the USA. To theextent that new technologies are easily transferredacross national borders, the economic dynamism ofthe USA could rapidly spread outwards, to otherdeveloped countries and perhaps even to the devel-oping world. At the peak of the Internet gold rush,the more enthusiastic dot-com entrepreneurs evenspoke of Silicon Valley as the equivalent of Florenceduring the Renaissance, the birthplace of ideas thatwould affect societies worldwide.

In retrospect, that view looks like one of the wilderclaims for the New Economy, but rapid productivitygrowth in the USA certainly has implications for thegrowth prospects of other countries. In compilingthis special issue, one of our aims has been to extendthe analysis of the New Economy beyond the USA,and to invite some genuinely rigorous analysis ofdevelopments elsewhere. Popular commentatorshave tended to imply that the US experience isunique, and have used it to criticize the apparent lackof progress in other countries, especially those ofEurope. Governments outside the USA are rou-tinely castigated for presiding over sluggish econo-mies that are over-regulated and slow to innovate.

This common view, however, may be something ofa caricature. In this respect, it is worth rememberingthat Europe had its own success stories. At the

height of the boom, European companies in newtechnology sectors, such as Nokia and Vodafone,were among the largest in the world by stock-market capitalization. Moreover, by the late 1990s,some European countries were starting to investheavily in information and communications technol-ogy (ICT). Several contributors to this issue exam-ine precisely this question, using new data to providesome of the first rigorous evidence on ICT adoptionand productivity growth outside the USA.

This is especially important, because the emergenceof the New Economy in the USA has encouragedother countries to examine their own records with amore critical eye. In tacit acknowledgement of this,European Union leaders have ambitiously commit-ted themselves to establishing Europe as the mostdynamic ‘knowledge-based’ region in the world by2010. The prospects for such a change in fortunesare discussed later in the paper, after an overviewof the New Economy in the USA.

II. AMERICA’S NEW ECONOMY

The American economy of the mid-1990s has beena source of envy for the world and of puzzlement formacroeconomists. (Robert Gordon, 1998)

The US Department of Commerce has defined theNew Economy as ‘an economy in which IT andrelated investments drive higher rates of productiv-ity growth’. The term became popular in consider-ing the remarkable economic performance of theUSA in the 1990s, and especially the years 1995–2000. This section describes the key features of theperiod, as background for the subsequent discus-sion.

The first point to make is perhaps a surprising one:over the 1990s as a whole, productivity growth wasunremarkable by the standards of the 1980s. Thecase for the importance of the New Economy restson the second half of the decade, when the boomcontinued for much longer than most observers hadexpected, and growth rates in excess of 4 per centa year were repeatedly recorded.

Over this period, productivity growth appears tohave risen considerably. Using data that incorporatethe August 2002 revisions, Figure 1 plots the annual

243

J. Temple

-2-1012345

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

% c

hang

e on

yea

r ear

lier

Figure 1US Annual Productivity Growth, 1950–2001

Table 1US Growth, 1950–2001

Real GDP growth (%) Productivity growth (%)

1950:2–1972:2 3.9 2.71972:2–1995:4 2.9 1.41995:4–2001:4 3.5 2.4

Notes: Both sets of figures are based on the annualized growth rate between the two quarters listed. Thefigures for real GDP growth are based on the Bureau of Economic Analysis (BEA) chain-weighted seriesfrom the US National Income and Product Accounts. The figures for productivity growth are based on theBureau of Labor Statistics (BLS) series for output per worker hour for the non-farm business sector (seriesPRS85006093).

growth rate of output per worker hour for 1950–2001. This is shown as the thin line. The darker lineis a centred 5-year moving average, which high-lights the slow upwards trend in productivity growthover the course of the 1990s. It also shows thatrecent productivity growth, although better than inthe late 1970s and early 1980s, has not yet achievedthe heights of the early 1960s.

Table 1 compares growth rates for the period fromthe fourth quarter of 1995 (1995:4) until the fourthquarter of 2001, with growth rates for two earlierperiods, essentially the 1950s and 1960s, and theperiod of the productivity slowdown that began inthe early 1970s.2 It can be seen that in the NewEconomy period—the latter half of the 1990s—theannual growth rate of output per worker hour was

nearly one percentage point higher than in thepreceding 20 years. As is discussed later, the pro-ductivity improvement was particularly strong in themanufacturing sector. Hansen (2001) compiles strongstatistical evidence for a structural break in labourproductivity growth in durable manufacturing in themid-1990s.

America’s improving productivity growth was ac-companied by an unprecedented boom in equipmentinvestment. The proportion of GDP devoted to fixedinvestment as a whole rose only slightly, but thisconceals an ongoing shift in the composition ofinvestment, away from buildings and towards equip-ment (including computer software). Over the post-war period, equipment investment as a share ofGDP has steadily risen, and over the 1990s it rose

2 The exact timing of these periods follows Gordon (2002) quite closely. This means that the quarters have been chosen tocorrespond to points in the business cycle of roughly the same rate of unemployment (around 5.5 per cent). Note that oneachievement of the New Economy may have been a reduction in the lowest sustainable rate of unemployment, a possibility discussedin section IV.

244


Table 2Greater Economic Stability in the 1990s

Standard deviations 1950s 1960s 1970s 1980s 1990s

Inflation volatility 1.90 1.18 2.18 2.54 1.05Unemployment volatility 1.28 1.07 1.16 1.48 1.04GDP growth volatility 3.84 2.11 2.77 2.65 1.52Productivity growth volatility 2.09 1.92 2.11 1.62 1.03

Notes: Standard deviations for the 1950s are calculated from the first quarter (or month) of 1950 until thelast quarter (or month) of 1959, and similarly for other decades. Inflation corresponds to the rate of changein the implicit deflator for personal consumption expenditure over the previous four quarters, calculated fromBEA data. The figures for unemployment are based on the BLS series for the monthly rate ofunemployment for the civilian labour force. The standard deviations for GDP growth and productivitygrowth are each based on a quarterly series for growth rates, using the same data as Table 1. The growthrate in a given quarter is the growth rate of GDP (or productivity) from four quarters earlier to the currentquarter. The form of presentation follows Mankiw (2001b).

from 7.5 per cent of GDP to almost 10 per cent. Thisis shown in Figure 2.

Since the relative price of equipment has beendeclining at a fast rate, the rising nominal share ofequipment spending translates into very rapid growthin real investment. Tevlin and Whelan (2002) calcu-late that growth in real equipment investment overthe period 1992–8 averaged 11.2 per cent a year,exceeding all previous 7-year intervals in the post-war era. They find that the primary reason for theequipment boom was soaring investment in comput-ers, and the effects of these investments are consid-ered in the next section.

There are still further reasons for seeing the macro-economic record of the 1990s as special. By April2000, the unemployment rate had fallen to just 3.9per cent, a 30-year low. Even more remarkably,despite this very low rate of unemployment, inflationremained stable. As measured by the deflator forpersonal consumption expenditure, inflation aver-aged 2.5 per cent in the 1990s as a whole, comparedto an average of around 5 per cent in the 1980s, andmore than 6 per cent in the 1970s.

Overall, the economy displayed much lower volatil-ity in growth, unemployment, and inflation than inprevious decades, leading some to hail ‘the death of

Figure 2The Long-term Boom in Equipment Investment

0

0.02

0.04

0.06

0.08

0.1

0.12

1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Shar

e of

GD

P

245

J. Temple

the business cycle’. This is illustrated in Table 2,which presents the standard deviations of monthlyunemployment rates, and of quarterly series forinflation, real GDP growth, and productivity growth.It is clear that the 1990s compare favourably withthe 1980s, and especially the 1970s, in terms ofoverall economic stability.

A sense that the economy was starting to behavedifferently contributed to a dramatic bull run in theAmerican stock market. The ratio of stock-marketcapitalization to GDP rose to an unusually high level,and valuations increasingly departed from the long-run historical relationships between share pricesand corporate earnings. The Federal Reserve chair-man, Alan Greenspan, warned of the dangers of‘irrational exuberance’, and eminent economistsqueued up to reaffirm that the market was beginningto look dangerously overvalued.

By now, it should be clear that the US economy inthe 1990s was beginning to look rather differentfrom that of the 1970s and 1980s. In the press, theevidence for a New Economy was often closelylinked to the increasingly high profile of the Internet.Considerable excitement surrounded the spectacu-lar rise of the Internet companies or ‘dot-coms’.Later in this issue, Eli Ofek and Matthew Richardsonreview some of the features of the boom in Internetstocks. As they point out, the rise in Internet stockvaluations between 1998 and 2000 has to be consid-

ered one of the most remarkable asset pricingphenomena of our time.

Yet it is not straightforward to link the productivitygrowth of the late 1990s to the Internet. Even at itspeak the Internet-related sector accounted for arelatively low share of total US stock-market capi-talization. Most Internet shares were trading onvery large multiples of earnings, and this suggeststhat relatively few Internet companies were makinga noticeable contribution to growth in productivityand output. Later in the paper, I sometimes discussthe role of the Internet, but overall I follow theliterature in emphasizing a much broader definitionof new technology.

As we have seen, the economy was ‘new’ in otherregards. This can be revealed simply by listing someof the questions that surround this period, as in Table3. The choice of questions reflects the general focusof this special issue on the macroeconomics of theNew Economy.3

It is important to recognize that many of thesequestions are interconnected. For example, theunusual length of the expansion can be explained bythe continued strength of productivity growth, theboom in equipment investment, and the stability ofinflation despite falling unemployment. Similarly, agood understanding of the sources of productivitygrowth is essential in deciding whether or not faster

3 Overviews that cover microeconomic issues in more detail can be found in Brynjolfsson and Hitt (2000) and Van Reenen (2001).The Internet has been discussed in the summer 2001 issue of this journal, Vol. 17 No. 2. More general surveys of the New Economyperiod include Baily (2001, 2002), Bosworth and Triplett (2001), Brenner (2002), Gordon (2002), Stiroh (1999) and Wadhwani(2001).

Table 3The Macroeconomics of the New Economy

What were the sources of the rise in productivity growth?

What led to the boom in equipment investment?

Why did inflation remain stable in the face of low and falling unemployment?

What explains the greater overall stability of the 1990s compared to previous decades?

Did the stock-market boom represent a bubble or improving fundamentals?

What were the effects on wage inequality, poverty, and social change?

How long can the faster rate of productivity growth be sustained?

What are the implications for the rest of the world?

246


growth is sustainable, and in assessing whether ornot other countries will be able to benefit from fasterproductivity growth.

In seeking to answer these questions, it is useful tofollow Mankiw (2001b) in observing that the USboom of the 1990s looks much like the 1970s inreverse. The 1970s were notable for substantialadverse supply shocks, the 1990s for their absence.The 1970s saw a steep fall in stock-market capitali-zation relative to GDP, an unexpected decline in therate of productivity growth, rising unemploymentand unstable inflation, which is the reverse of the1990s experience on all counts. In what follows, Isometimes argue that our understanding of the NewEconomy in the USA can be enhanced by contrast-ing this period with the 1970s.

The structure of the paper follows the order of thequestions in Table 3 relatively closely. Section IIIconsiders the evidence on the acceleration in pro-ductivity growth and the accompanying boom inequipment investment. Sections IV and V look atinflation, unemployment, the role of monetary policy,and the greater overall stability of the US economyduring the 1990s. Section VI examines the erraticbehaviour of the stock market, taking a long-termperspective. Section VII discusses recent move-ments in wage inequality and poverty, changes inlabour markets, and the wider social impact. Thenext two sections, VIII and IX, examine futureprospects, and the implications for the rest of theworld. Finally, section X offers some conclusions. Afew of the more technical measurement issues arebriefly discussed in an Appendix.

III. PRODUCTIVITY GROWTH

The productivity gain that will be delivered by IT willbe at least as great as the electricity, transportationand telephone revolutions put together. (John Cham-bers, CEO of Cisco Systems, quoted in Wadhwani(2001))

The case for seeing the New Economy as a sub-stantial and important development rests at leastpartly on the unexpected acceleration of productiv-ity growth in the mid-1990s. This productivity growthsustained the expansion for an unusually long time,raised real wages, fuelled the stock-market boom,

and may have also been responsible for a temporarychange in the relationship between inflation andunemployment.

In this section, I review the available evidence on thesources of the productivity acceleration. Econo-mists have tended to emphasize the productivityrecord in their research on the New Economy,partly because of its significance for living stand-ards, and partly because productivity improvementsare likely to be sustainable, unlike the stock-marketboom or the increasingly rapid growth of householdexpenditure.

One plausible hypothesis is that faster productivitygrowth was driven by the new technology sectorsand, more specifically, a combination of productivitygains in computer production and increased invest-ments in computing. Oliner and Sichel (2000) notethat in real terms, business investment in computersand peripheral equipment in the USA rose morethan fourfold between 1995 and 1999. Stiroh (2002a)points out that, over the decade as a whole, US firmsinvested more than $2.4 trillion in ICT assets.

This sharp rise in computer investment is almostcertainly related to the increase in productivity in themanufacturing of computers. These productivitygains have led to the price of computing powerdeclining at an ever faster rate, and have thereforeencouraged much wider use of computers. Morespeculatively, the recent rapid growth in computerinvestment may also have been driven by the grow-ing importance of the Internet, which has arguablyraised the usefulness of computers in a variety ofapplications.

In seeking to understand the aggregate impact ofsuch changes, researchers have used the standardframework of growth accounting. Since severalcontributors to this special issue adopt this approach,a brief review may be useful. The basic idea is todecompose observed output growth into the contri-butions of various inputs, and a residual term knownvariously as growth in multi-factor productivity(MFP) or total factor productivity (TFP). This latterterm quantifies the extent to which inputs are beingcombined more efficiently. A rise in TFP indicatesthat more output can be achieved from given inputsor, in other words, that the production frontier hasshifted.

247

J. Temple

In the standard textbook exposition of growth ac-counting there are just two inputs, capital and labour.Output growth is then a weighted average of thegrowth rates of these two inputs, plus the residualterm, TFP growth. The weight on the growth rate ofa particular input is equal to the elasticity of outputwith respect to that input. Typically, the researcherdoes not observe this elasticity directly, but underthe assumptions of perfect competition, constantreturns to scale, and marginal productivity factorpricing, the elasticity is equal to the share of the inputin total income, which often is observable. Forexample, labour’s income share is just the totalwage bill divided by national income.

When only two inputs are considered, this approachmay seem more than a little simplistic, especially inits mechanical aggregation of different types ofcapital and labour. What this criticism overlooks isthat exactly the same principles can be extended toconsidering output as a function of many differentinputs. For example, we can easily distinguish be-tween different types of labour, and weight theircontributions to output by their respective shares inincome.

This suggests that one approach to analysing theNew Economy would be to distinguish betweendifferent types of capital and, in particular, betweenICT capital and other forms. The definition of ICTcapital usually includes computer hardware, andsometimes software and telecommunications equip-ment as well. Using growth accounting, we candirectly assess the contribution of ICT capital tooverall productivity growth.

Along the way, this exercise sheds light on a famous1987 saying of Robert Solow’s: ‘we see the compu-ter age everywhere but in the productivity statis-tics’. Solow was drawing attention to an apparentparadox, namely that the increasingly high profile ofcomputers in business did not appear to be reflectedin faster aggregate productivity growth. If anything,during the period when computers were first be-coming widely adopted—the 1970s and 1980s—productivity growth in the USA was rather slowerthan previously. This may be less paradoxical thanit seems at first, given that for much of this period,computers still accounted for only a small fraction ofthe total capital stock (Oliner and Sichel, 1994).

In recent years that has started to change. Aconsensus has now emerged that faster productivitygrowth in the latter half of the 1990s can be tracedback, in large part, to the production and adoption ofICT capital. Part of this gain is direct, since techno-logical advances in the production of computerhardware—particularly in semiconductors—havecontributed an increasing proportion of growth inaggregate TFP. Although ICT production is only asmall part of total output, technical progress in thissector has been so rapid that it makes a noticeablecontribution to TFP growth even at the aggregatelevel.

The size of this effect has been estimated byJorgenson and Stiroh (2000a) and discussed furtherby Jorgenson (2001). The latter paper indicates thatproductivity gains in the production of ICT werealone responsible for 0.5 percentage points of theannual rate of aggregate TFP growth over 1995–9.This is a sizeable amount given that, historically,TFP growth has rarely been over 1 per cent a yearin the USA, even in the decades of strongestperformance.

Clearly we are interested not only in the sectorproducing ICT, but also what happens to productiv-ity when ICT equipment is purchased by firms. Inprinciple at least, this form of capital deepeningshould raise labour productivity, as workers havemore equipment to work with. More controversially,it is possible that adoption of ICT equipment couldaffect TFP, through various externalities or spillovers.The remainder of this section discusses both theseissues. It will also consider the need for cyclicaladjustments in measuring productivity, and whetherthe New Economy productivity improvements werepervasive, or concentrated in a few sectors.

In calculating the growth contribution of capitaldeepening, there are two effects at work. First, realICT investment grew at a faster rate after 1995 thanpreviously, as the productivity gains in ICT produc-tion translated into increasingly rapid price declinesof these goods. Second, the income generated bycomputers has accounted for a rising share of totalincome, as computers supply a larger share of totalcapital services than previously. In terms of growthaccounting, both the rate of growth of the ICTcapital input, and the weight placed on that growth

248


rate, have increased. The overall consequence is asubstantial rise in the contribution of ICT invest-ments to aggregate growth.

Oliner and Sichel (2000) find that the increasingadoption of ICT capital raised the annual rate oflabour productivity growth by half a percentagepoint between the first half of the 1990s and thesecond half. Jorgenson (2001) presents a similarestimate. Overall, it seems that developments in theproduction and adoption of ICT can account formost of the acceleration in labour productivity growthbetween the first and second halves of the 1990s.

Table 4 shows this in more detail, by reporting therate of growth in aggregate labour productivity andits various components, using estimates fromJorgenson (2001). Note that these are aggregatecontributions to growth in labour productivity. Forexample, the aggregate contribution of TFP growthin ICT production reflects the extremely rapid TFPgrowth in this sector, but also its small share of totaloutput.

This analysis raises two interconnected questions.The first question is whether or not one should takeinto account cyclical effects in thinking about therecent acceleration in productivity growth. Thesecond question is why, if the New Economy repre-sents a pervasive change in the way of doingbusiness, productivity growth outside the manufac-turing sector has not been faster.

To see the connection between these two questions,consider the analysis of Gordon (2000). He pointsout that when output grows faster than its sustain-able trend, so does labour productivity, partly be-cause more use is made of previously underutilizedlabour, and partly because the hiring of additionallabour tends to lag behind surges in output. Thissuggests that some of the productivity gains re-corded in the late 1990s may be temporary. Further-more, his adjustments for cyclical effects indicatethat there has been no sustained productivity accel-eration outside the durable manufacturing sector,which casts some doubt on the scope of the NewEconomy.

Nobody doubts that cyclical effects have played arole, especially given the steep fall in unemploy-ment. The magnitude of the required adjustments ismore controversial, partly because trend/cycledecompositions are inherently difficult, and partlybecause the 1990s expansion was one of the longeston record. The length of the expansion suggests thatmuch of the productivity improvement should besustainable. This view is supported by the carefulanalysis of Basu et al. (2001), who find that theproductivity record of the late 1990s cannot beexplained by increased factor utilization. If anything,measured productivity may understate the underly-ing technological improvement. This is because theunusually rapid accumulation of capital is likely tohave imposed substantial adjustment costs, tendingto reduce standard measures of final output.

Table 4Sources of Labour Productivity Growth, 1973–99

1973–90 1990–5 1995–9

Labour productivity growth 1.26 1.19 2.11Aggregate contributions of

ICT capital deepening 0.35 0.43 0.89Non-ICT capital deepening 0.44 0.21 0.35ICT production TFP 0.19 0.25 0.50Non-ICT production TFP 0.06 –0.01 0.25Labour quality 0.22 0.32 0.12

Notes: Average annual percentage rates of growth. Aggregate contributions for 1990–5 do not total exactlyowing to rounding.Source: Jorgenson (2001, Table 8).

249

J. Temple

A second debate concerns the scope of the NewEconomy, and whether or not productivity improve-ment was pervasive. The central puzzle is thatcompanies in the non-manufacturing sector, al-though often intensive users of ICT, did not performas strongly as might have been expected. Over thesecond half of the 1990s, output per hour in non-manufacturing grew at a rate of about 2 per cent ayear, compared to 1 per cent a year over 1972–95.Furthermore, some of this increase can be ac-counted for by the increasing adoption of ICTcapital, which suggests that any improvement inTFP growth in non-manufacturing has been rela-tively modest.

One corollary is that the net social rate of return tocomputer investments must have been little higherthan on other assets, since these higher returnswould be reflected in faster TFP growth understandard growth-accounting assumptions. The rela-tively small acceleration of TFP growth in non-durable manufacturing, and outside the manufactur-ing sector, makes it difficult to argue that newtechnologies have generated widespread spillovers,or otherwise had significant benefits for TFP.4 Thisis not to deny that capital deepening, largely in theform of ICT investments, has raised labour produc-tivity growth throughout the economy.

An alternative and more controversial view is thatnet returns on ICT investments have been lowerthan returns on other assets. The supporting evi-dence is that substantial overinvestment appears tohave taken place in certain areas, notably telecom-munication networks (Brenner, 2002, pp. 256–9). Ifreturns on ICT investment have sometimes beenlow, this tends to imply that the contribution of ICTcapital is overstated in growth-accounting exer-cises, relative to other sources of productivity im-provement.

Another argument often made is that, given theproblems of measuring output in service industries ina meaningful way, some gains in productivity mayhave been undetected in the national accounts.Some support for this argument is that the incomeside of the national accounts has shown fastergrowth than the product side in recent years.Nordhaus (2001) constructs a series for labour

productivity based on the income side, which for1996–8 raises the growth rate by more than 1percentage point compared to the standard esti-mates. Hence Nordhaus concludes that productiv-ity has accelerated throughout the economy, and notsimply in a few New Economy sectors. Baily andLawrence (2001) arrive at a similar conclusion.

If we subject the New Economy to what Young(1995) calls ‘the tyranny of numbers’, the overallverdict is mixed. In some respects, notably therelatively slow rate of TFP growth outside manu-facturing, the New Economy looks much like the oldone. More positively, rigorous analysis suggests thatproductivity gains in the production of ICT havecontributed to faster growth in aggregate TFP.Furthermore, these gains have had an indirect ef-fect, since increasingly rapid price declines for ICThardware have yielded a massive boom in equip-ment investment and ICT adoption. As a result, thegrowth contribution of ICT capital has increasedsubstantially. This is just about consistent with theview that new technology is transforming ways ofdoing business, if not with the view that suchinvestments are yielding abnormally high returns.

It is also important to remember that the case for theNew Economy rests not just on faster productivitygrowth, but also on the remarkable behaviour ofinflation and unemployment, and a level of overalleconomic stability not seen since the 1960s. Thenext two sections of the paper consider thesedimensions of the New Economy in more detail.

IV. INFLATION AND UNEMPLOYMENT

Anyone who follows both mainstream economicsand the newspapers knows that the textbook modelof the wage-price-unemployment process is in trou-ble. (William Nordhaus, in his comments on Ball(1999))

As pointed out above, one of the central puzzlesabout the New Economy is how inflation stayed lowand stable, despite the unemployment rate falling toa level that seemed to make rising inflation inevita-ble. By the year 2000, the US unemployment ratewas hovering around 4 per cent, some 2 percentage

4 For more on this point see Jorgenson and Stiroh (2000a) and Stiroh (2002b).

250


points below the consensus estimates of the naturalrate of unemployment (the lowest rate consistentwith stable inflation in the absence of supply shocks)that had prevailed in the early 1990s.

By this reckoning, monetary policy should havebeen tightened aggressively, yet Alan Greenspan’sFederal Reserve held interest rates steady in theface of a booming economy and falling unemploy-ment. It seems to have worked, and Greenspan hasbeen hailed as something close to a miracle worker.The journalist Bob Woodward (2000) wrote a popu-lar book on the Greenspan era, describing his subjectas ‘the innovative technician who spotted productiv-ity growth in the 1990s and refused to raise interestrates when the traditional economic models andtheories cried out for it’ (as cited in Ball andTchiadze, 2002).

Recognizing the acceleration in productivity growthwas undoubtedly important. Otherwise the 1990sboom might easily have been perceived as unsus-tainable, and policy-makers would have been waryof the economy overheating. Yet productivity growthdoes not in itself explain why inflation remained loweven when the unemployment rate fell below mostestimates of the natural rate. The 1990s pose some-thing of a challenge to adherents of the natural ratehypothesis, although not an insurmountable one.

One potential explanation is that a series of fortu-nate, positive supply shocks lowered inflation, off-setting the inflationary effect of low unemployment.Gordon (1998) pointed to largely exogenous changesin the inflation rates of food and energy, imports,computers, and medical care, and to improvementsin the measurement of inflation rates, as possiblecandidates for the stability of inflation. It is not clearthat the supply shocks explanation can account forthe maintenance of low inflation beyond 1998, how-ever, given a slowdown in the appreciation of theexchange rate, and rising energy prices.

Alternatively, it is possible to find explanations forwhy the natural rate of unemployment may havefallen by the mid-1990s. Shimer (1998) and Katzand Krueger (1999) quantify the possible role ofdemographic shifts. As the baby-boom generation

has matured, young workers with high rates of jobturnover now make up a lower proportion of thelabour force. Furthermore, temporary help agencieshave accounted for an increasing share of employ-ment, albeit from a very low base, a developmentthat may allow workers and vacancies to be matchedmore efficiently. Both these ideas are consistentwith a further stylized fact of the New Economy,namely that there has been a particularly steepdecline in short-term unemployment, so that thecomposition of unemployment durations has shiftedtowards long-term spells.5

Furthermore, Ball and Tchiadze (2002) argue thatmonetary policy under Greenspan was consistentwith awareness of a decline in the natural rate. Theyemphasize that estimates of the natural rate, ex-tracted from standard regression models of unem-ployment and inflation, start to fall from around1994. Although the true extent of this fall probablywent unrecognized, they argue that there was enoughevidence of a decline for the Federal Reserve to becautious about raising interest rates. Ball andTchiadze provide some indirect evidence to supportthis theory, by showing that the path of interest ratescan be predicted quite accurately by estimates of amonetary-policy rule that incorporates inflation anda time-varying natural rate, where the latter is basedon well-known models.

This element of the available evidence is onlyindirect, however. The overall arguments are vul-nerable to a traditional criticism of the natural-ratehypothesis, namely that candidate explanations formovements of the natural rate sometimes look toomuch like ex-post rationalizations. From this point ofview, a potentially more satisfactory explanation ofthe 1990s record is to link the joint behaviour ofunemployment and inflation to the increase in pro-ductivity growth. To the extent that workers wereinitially unaware of the increase, wage demandsmay have lagged behind productivity growth, imply-ing that inflation could remain stable even whileunemployment fell below previous estimates of thenatural rate.

This hypothesis has been much discussed, and hasbeen investigated more formally by Ball and Moffitt

5 There is a paradoxical element to this change in composition. It is sometimes argued that wage and price inflation are moreresponsive to short-term unemployment than long-term unemployment. In this light, given the recent decline in short-termunemployment, the stability of inflation in the 1990s appears all the more remarkable.

251

J. Temple

(2001). Part of its appeal is that it could also explainwhy inflation rose sharply in the early 1970s, be-cause aspirations for real wages took time to adjustto the marked slowdown in productivity growth thatbegan in that decade. Ball and Moffitt show that, ifaspirations for real wage growth are based on therecent behaviour of real wages, the Phillips curveshould include a new variable, the difference be-tween current productivity growth and a weightedaverage of past growth in real wages. As a result,a change in the rate of productivity growth will shiftthe Phillips curve temporarily, until wage aspirationsadjust. Consistent with this story, the new variableperforms well in empirical tests. It raises the ex-planatory power of otherwise standard models ofthe Phillips curve, and can fully explain the stabilityof inflation after 1995, without needing to appeal toany fall in the underlying natural rate of unemploy-ment.

All these explanations are largely within standardtextbook models of the wage–price–unemploymentprocess. It is possible to argue that such modelsneed more radical surgery if they are to explain the1990s experience. Stock, in his comments on Gordon(1998), suggests that the key puzzle is why unem-ployment has ceased to have much predictive powerfor inflation, even while other indices of real activ-ity—such as capacity utilization and housing starts—continue to play a useful forecasting role. Morerecent work, however, suggests that the Phillipscurve is alive and well (Staiger et al., 2001). Overall,the inflation record of the 1990s can be explained bya series of favourable supply shocks, a decline in thenatural rate, unexpected productivity growth, ormore likely some combination of all three. Thetextbooks do not need to be rewritten just yet.

Whatever the reason for the stability of inflation, theGreenspan era appears to have been one of largelysuccessful monetary policy. Views differ on theextent to which this record can be sustained. Someargue that we now know the principles of goodmonetary policy, and point out that the path taken byinterest rates can be predicted quite accurately byregression estimates of simple policy rules thatembody those principles. On this view, future chairs

of the Federal Reserve will be able to implementsuccessful policy in a fairly mechanical fashion, withlittle need for privileged insight into the workings ofthe economy. Others question the extent to whichthese estimated rules are genuinely recovering be-havioural parameters. It might also be pointed outthat the Greenspan era has maintained the longtradition for central bankers, even the most success-ful ones, to remain somewhat opaque in their evalu-ations and judgements.6

Finally, it should not be forgotten that Greenspan’sFed presided over a massive boom on Wall Street.The stock-market boom took equity valuations tounprecedented levels, ones that are increasinglyregarded as having been unrealistic. The bursting ofthis bubble could yet have serious consequences,and lead to a reassessment of the Greenspan era.The headlong euphoria of the financial markets mayhave been sustained by some of his more optimisticpronouncements, and the belief that the Fed wouldintervene whenever the markets looked like falling.

In fairness to Greenspan, it is not clear that theFederal Reserve could have prevented the boom,even if it had sought to do so. Moreover, the keyproblem may have been his own past success,which gave market participants an exaggeratedfaith in his ability to stabilize the economy (Miller etal., 2002). Nevertheless, the experience of Japan inthe 1990s has pointed to the dangers of asset-pricebubbles. The final judgement of recent monetarypolicy may have to wait until the effects of fallingshare prices have fully unwound. At present, as isdiscussed later on, there are some grounds forpessimism.

V. GREATER STABILITY

Most discussions of the New Economy in the USAemphasize the combination of fast productivitygrowth, low unemployment, and low inflation. The1990s were also a period of unusual economicstability, as indicated by announcements in thefinancial press of the ‘death’ or ‘taming’ of thebusiness cycle. Although sometimes ridiculed, and

6 See Mankiw (2001b) for more on this latter point, and Ball and Tchiadze (2002) for a contrasting view on the Greenspan legacyto future policy-makers. Gali (2000) and Gali et al. (2002) provide a more complete analysis of monetary policy, which suggeststhat policy responses to productivity shocks did improve under Volcker and Greenspan. A fuller discussion of the Phillips curvecan be found in Mankiw (2001a).

252


undoubtedly premature, such statements were closerto the truth than one might have expected. As Table2 has already shown, the standard deviations ofinflation, unemployment, and growth rates werelower in the 1990s than in any of the previous fourdecades.

The greater stability of productivity growth is appar-ent from Figure 1 and, to reinforce this point, Figure3 shows a rolling standard deviation of productivitygrowth rates for the period since 1955. The figurefor each quarter is the standard deviation of annualproductivity growth rates, measured each quarter,over the preceding 5-year period. A steep decline involatility, perhaps beginning in the mid-1980s, isclearly evident.

Some recent research has tried to explain thedecline in volatility, with an especial emphasis on thegreater stability of output. Views differ on the timingof the change. McConnell and Perez-Quiros (2000)argue that a reduction in output volatility can betraced back to the mid-1980s. Blanchard and Simon(2001) take a longer view, and argue that there hasbeen a long-term trend towards greater stability,beginning in the 1950s but then interrupted in the1970s and early 1980s.

Certainly the statistical evidence for greater stabilitysince the mid-1980s appears very strong. For theperiod 1953:2 to 1983:4, about a fifth of quarterlygrowth rates were negative. For 1984:1 to 2000:4,the proportion is less than 5 per cent. McConnell andPerez-Quiros find strong evidence for a structuralbreak in the residual variance of a simple

autoregressive model for GDP growth, around thefirst quarter of 1984.

The possibility of an increase in stability has beendiscussed for some time, and Burns (1960) madethis issue the central focus of his presidential ad-dress to the American Economic Association. Hesuggested that changes in the composition of output,in the extent of automatic stabilizers, and in oppor-tunities for consumption smoothing, could all con-tribute to greater stability. Looking at the differentcomponents of GDP, McConnell et al. (1999) andBlanchard and Simon (2001) point to a long-termreduction in the volatility of consumption as one ofthe main reasons for greater stability. The volatilityof non-residential investment has also fallen, andthere appear to have been important changes in thebehaviour of inventories, reflected in a decline in thevolatility of output relative to that of sales.

McConnell and Perez-Quiros argue that one of thekey changes has been a reduction in the volatility ofcapital goods production, which may be sufficient initself to account for the mid-1980s break in outputvolatility. They suggest that inventory managementin the durable goods sector has changed, since thesector’s output is less variable than previously,whereas the same pattern is not present in sales.

Importantly, an increase in output stability does notappear to be confined to the USA. Looking at theexperience of other G7 members, Blanchard andSimon (2001) suggest that stability has increased inCanada, Germany, the UK, and, to a lesser extent,France. This tends to support explanations based on

0

0.5

1

1.5

2

2.5

3

1955 1960 1965 1970 1975 1980 1985 1990 1995 2000

Rol

ling

stan

dard

dev

iatio

nFigure 3

The Declining Volatility of Productivity Growth

253

J. Temple

technological developments, such as changes ininventory management, rather than ones based onpolicy improvements that might be specific to theUSA. Yet as Wadhwani (2001) points out, in-creased stability in the UK cannot be attributed tochanges in the behaviour of inventories. A reductionin the volatility of consumption appears more impor-tant, perhaps driven by financial liberalization.

Overall, it is clear that more research is needed tounderstand the patterns of volatility outside theUSA, and the forces driving any tendency to greaterstability. It should also be remembered that an-nouncements of the ‘taming’ of the business cyclehave been made before, and have rarely lookedconvincing in retrospect.

VI. THE STOCK MARKET

Although unusual economic stability is a striking andimportant feature of the recent macroeconomicdata, no aspect of the New Economy receivedgreater media attention than the booming stockmarket, and the rise of the Internet companies. Ofall the achievements of this period, the stock-marketvaluations of the late 1990s appear the least secure.In this section, I consider the long-term behaviour ofthe stock market, and discuss whether it can bereconciled with notions of market rationality.

The recent bull market should be seen as part of alonger-term pattern of instability. The last 30 yearshave seen dramatic changes in the ratio of stock-market capitalization to GDP in the USA. This ratiofell nearly threefold in 1973–4, and then rose five-fold in the years after 1985 (Hobijn and Jovanovic,2001). From these remarkable swings, many haveconcluded that stock valuations have persistentlydiverged from fundamentals. The even more spec-tacular rise and fall of Internet stocks has furtherencouraged the view that the 1990s valuationsreflected a speculative bubble.

The case is not as clear cut as it may first appear,however. The view that stock valuations reflectsensible forecasts of future pay-outs to sharehold-ers continues to have some prominent adherents. Ina series of papers, Hall (2000, 2001) has argued thatthe recent behaviour of the stock market is consist-

ent with the view that firms have accumulated largeamounts of intangible capital, especially in the last 10years. This could be associated with intellectualproperty rights and brand names for differentiatedproducts, for example. One characteristic of theNew Economy may have been a growing role forintangible capital in generating profits. It is alsopossible that firms are now able to appropriate thereturns to this capital to a greater extent thanpreviously (Nordhaus, 2000).

These arguments are inherently difficult to test,given that intangible capital is hard to measure. Hall(2001) demonstrates how to extract a measure ofthe quantity of total capital from stock-marketvaluations. The basic idea is to use equity capitaliza-tion, plus debt, minus financial assets, to infer thetotal value of non-financial assets. Hence, under themaintained assumption that stock-market valuationsreflect rational valuations of assets, it is possible tomeasure the total capital stock (tangible and intan-gible) based on various specifications for the cost ofadjusting installed capital. This measure of capitalcan then be compared with a more conventionalmeasure of physical capital alone, such as one basedon the perpetual inventory method. The great strengthof Hall’s work is to indicate how intangible capitalmust have evolved over time, if these movementswere genuinely the force behind swings in the stockmarket.

This is important, because the very different eventsof the 1970s and 1990s are likely to be central tofuture discussions of market rationality. TakingHall’s story on its own terms, the first problem it hasto confront is the experience of the mid-1970s, whenthe quantity of total capital inferred by Hall issubstantially below the quantity of physical capitalimplied by the perpetual inventory method.

Taken at face value, this implies that the value ofintangible assets vanished abruptly. The explanationHall seems to lean towards is that of Baily (1981).The first oil shock was associated with a dramaticchange in the relative price of energy, which mayhave rendered a large fraction of the installed capitalstock effectively obsolete almost overnight. Thisimplies that a standard perpetual inventory calcula-tion will tend to overstate the quantity of physicalcapital in use from the mid-1970s onwards.

254


More ambitiously, Hobijn and Jovanovic (2001) linkthe 1970s valuations to the first stages of the NewEconomy, through the growing importance of ICTinnovations. They argue that existing companieswere ill-equipped to exploit these opportunities,which therefore required new companies to emergeand supplant the ‘Old Economy’ firms. Since thefinancial markets may have anticipated this, thevaluations of the older firms declined, explaining aperiod of depressed market capitalization that couldonly be reversed once the new firms were estab-lished and listed on the market.

Some of their evidence is persuasive, but the argu-ment clearly relies on great foresight on the part offinancial markets. In contrast, it has often beenargued that equities were irrationally undervaluedfor much of the 1970s, not least because financialanalysts took time to adjust to their first experienceof high inflation (see Temple, 2000, and the refer-ences therein). And the view that mispricing canpersist for long periods has only been reinforced bythe bull market of the 1990s, and the recent steepdecline in valuations.

The most obvious instance of mispricing occurred inthe new technology sector. This may sound likewisdom after the event, but most economists I spoketo in late 1999 were confidently predicting a crashfor the NASDAQ high-technology index, and thereal surprise was how long it took to arrive. In thisissue, Eli Ofek and Matthew Richardson provideone of the first rigorous analyses of the Internetsector. They show that a wide range of marketevents can be combined to form a very strong caseagainst market rationality, albeit a circumstantialone.

Using a different approach, Bond and Cummins(2000) also compile strong evidence for wider mar-ket irrationality. They use a database of earningsforecasts by analysts to construct an alternative setof company valuations, which can be comparedwith the market valuations. They find persistentdivergences between the two, implying that marketvaluations rarely reflected expert profit forecasts.

This could be explained by analysts underestimatingthe importance of intangible assets, but Bond andCummins provide other evidence for the existenceof a bubble. Importantly, they find that their fore-cast-based valuations have greater explanatorypower for investment behaviour than actual stockprices. This tends to suggest that company manag-ers themselves did not believe the stock-marketvaluations of their assets.7 It has subsequentlyemerged that some analysts had a similar lack ofbelief even in their own forecasts.

A further problem for the intangible capital view isthat profits have not risen accordingly. Althoughcorporate earnings performed quite strongly over1992–2000, they do not appear to have grown at arate anywhere near fast enough to justify the stock-market boom. On some measures, earnings hadessentially stopped increasing as early as 1997.Brenner (2002, p. 209) describes this behaviour ofprofitability as the Achilles’ Heel of the NewEconomy—the ‘one major trend that did not fit’.

Even were this not the case, the outlook for realyields on equities appears unusually weak. Writingin April 2001, Campbell and Shiller (2001) arguedthat valuation ratios ‘imply a stronger case for a poorstock market outlook than has ever been seenbefore’. Given the awesome scale of the 1990sovervaluation, it seems unlikely that recent falls inthe market have been large enough to reverse thisconclusion. As evidence accumulates of creativeaccounting practices, possible overinvestment incertain sectors, and declining profitability, it be-comes ever harder to defend the view that the NewEconomy valuations were anything other than mas-sively over-optimistic.

VII. WINNERS AND LOSERS

In 2002, the stock market has fallen sharply, andmuch of the optimism surrounding the New Economyhas dissipated. A sceptical observer might arguethat the New Economy mainly benefited a fewfortunate insiders, often at the expense of those less

7 This interpretation is a little subtle and is based on Tobin’s q theory of investment. Simplifying a little, new investment shouldtake place when the market valuation of a company’s installed capital exceeds its reproduction cost, or when Tobin’s q is greaterthan one. Bond and Cummins find that a measure of q based on analysts’ implicit valuations has much greater explanatory powerfor company investment behaviour than one based on stock-market valuations. This tends to suggest that those making companyinvestment decisions did not have much faith in the market’s valuation of their capital.

255

J. Temple

well informed. Revelations of dubious accountingpractices at Enron and WorldCom, intentionallymisleading investment advice from financial ana-lysts, and large bonuses for leaders of failing com-panies have all contributed to this darker view ofAmerica’s boom.

This section considers whether the benefits of theNew Economy were widely shared, and examinesthe recent path of income inequality. It also exam-ines whether the New Economy is changing thelabour market, in ways that could have importanteconomic and social consequences. Finally, thesection briefly considers ideas about wider socialchange that are sometimes discussed in the media,but have received relatively little attention fromeconomists.

The first comments to make are positive ones. AsFreeman (2001) points out, the 1990s were a rea-sonably good decade for many, regardless of theirposition within the income distribution. The realwages of the low-skilled increased, and unemploy-ment rates for the less-educated and low-skilled fellto levels not seen since the 1960s. Helped by lowerunemployment, median hourly earnings for workersat the 10th decile increased by as much as 10.2 percent in real terms in just 3 years (1996–9). Theincidence of poverty fell modestly, reversing theincrease of the previous decade. Overall, wageinequality among both men and women was nearlyconstant over the 1990s.

To understand the significance of this last fact, ithelps to stand back and briefly review the longer-term patterns. As is well known, the 1980s saw asteep rise in wage inequality in the USA. Explainingthis rise has been a central goal of recent researchin labour economics. At least until recently, it waspossible to trace an emerging consensus that greaterwage inequality was primarily due to a shift indemand towards skilled labour and away fromunskilled labour. This shift was said to be driven bychanges in technology, and perhaps especially bythe diffusion of ICT.

The hypothesis that recent technological changehas been ‘skill biased’ is widely regarded as themost convincing explanation for the 1980s rise inwage inequality (see, for example, Machin, 2001). Ithas survived investigation rather better than some of

the other competing explanations, such as thosebased on globalization, increased trade, or shifts inthe composition of product demand. It could poten-tially explain the overall rise in wage inequality, therise in the educational wage premium, and the risein inequality that has occurred even among workerswith similar observable skill characteristics. In theircontribution to this issue, Philippe Aghion and PeterHowitt argue that these trends can be explained bythe arrival of a new, pervasive technology such asICT, which puts an increased premium on workeradaptability in addition to conventional skills.

As pointed out by Card and DiNardo (2002), how-ever, the skill-bias hypothesis now faces a severetest: how can it reconcile the very rapid ICT diffu-sion of the 1990s with the barely noticeable in-creases in wage inequality of the same period? Thebehaviour of wage inequality in the New Economyprovides little support for the simpler theories ofskill-biased technical change. It may still be consist-ent with more complex models, such as those ofAghion and Howitt, if the process of adjustment toICT, perhaps with a temporary premium for adapt-ability, was concentrated in the 1980s. More gener-ally, there is likely to be a shift of research attentiontowards institutional explanations, and perhaps es-pecially the erosion of the real value of the minimumwage in the early 1980s.

Economists have studied wage and income inequal-ity in depth, but have given less attention to theimpact of the New Economy on the nature of jobs,working life, and society as a whole, despite muchcoverage of these topics in the media. The popularview is that the 1990s were associated with a rise injob insecurity, more-rapid job creation and destruc-tion, a move away from long-term employmenttowards short-term contracts, and a general in-crease in managerial pressure on workers.

These ideas are among those investigated by Rich-ard Freeman, in his wide-ranging contribution to thisissue. Using a ranking of US states in terms of theirNew Economy characteristics, he points out that therate of job reallocation does not appear to differgreatly between the top ten New Economy statesand the bottom ten. If rates of job reallocation arechanging, it does not appear to be a distinctivelyNew Economy phenomenon. Furthermore, evidencefor the country as a whole casts doubt on the idea

256


that job tenure has noticeably declined, and on theview that high-tech sectors are associated with newkinds of employment relationships.

Freeman gives greater emphasis to rapid growth inthe use of the Internet in recruitment, which is likelyto have implications for the efficiency of job match-ing, and perhaps thereby for the natural rate ofunemployment. He also discusses the prospects fornew technologies to help trade unions in providingservices to their members, in acting as watchdogsfor company behaviour, and in raising union democ-racy and general effectiveness. Such opportunitiescould help to counter the perception that the NewEconomy will necessarily favour those already pow-erful.

Concerns about the labour market are sometimeslinked to a deeper unease about the evolution ofmodern Western economies, and long-term socialchange. Although this may seem beyond the remitof economists, we should perhaps take more seri-ously the standard definition of our subject, namelythe study of the relationship between the allocationof scarce resources and human welfare. That defi-nition implies that we should be keenly interested inthe mutual interaction between the economy andsociety, and especially in those social trends that arepartly shaped by economic decisions and activity.

There has been no shortage of influential commen-tators lining up to register their unease. Among themost important contributions are Robert Putnam’s(2000) vision of Americans ‘bowling alone’ as theyincreasingly disengage from community participa-tion and civic life, and Richard Sennett’s (1998)depiction of the ‘corrosion of character’ in the NewEconomy. Sennett, a distinguished sociologist, ex-plores the personal consequences of working life inthe 1990s, and argues that modern careers toorarely contribute to self-respect or a sense of socialconnection.

In this respect, there are some useful statisticswhich should give us pause. As many observershave pointed out, surveys of Americans since the1970s have revealed long-term declines in job satis-faction, happiness in marriage, satisfaction withfinancial situation, and satisfaction with place of

residence (Lane, 1998). It is too early to say whetherthe boom of the late 1990s will have interruptedthese trends. Some observers, notably Quah (2002),seem to express optimism that the New Economywill change consumption patterns and perhaps life-styles for the better. The Internet has often beenregarded as a positive force, at least overall. Never-theless, a complete account of the New Economyneeds to register the possible dislocation betweeneconomic success and social well-being. It also needsto acknowledge that all forms of economic changehave social consequences that are poorly under-stood, yet are likely to be of profound importance.

VIII. FUTURE PROSPECTS

The previous section drew attention to the new-found stability in wage inequality, and the modest fallin poverty in the 1990s. These achievements wouldbe quickly threatened by a deep recession. In thissection, I briefly review the immediate prospects forthe USA, and consider whether or not the recentacceleration in productivity growth can be sus-tained.

It is clear that, by the year 2000, consumption andinvestment were growing at an unsustainable rate.8

The ratio of the trade deficit to GDP reached arecord high in 2000, the counterpart of a massiveprivate-sector financial deficit. Household borrow-ing, helped along by apparent capital gains from thestock-market boom, has raised personal debt expo-sure to the highest level seen for 70 years. A periodof difficult adjustment seems inevitable, as consum-ers start to save again, and firms cut back oninvestment in the face of declining demand andovercapacity. The difficulty of adjustment will bereinforced by the steep fall in the stock market, andits effects on consumer and investor confidence.

Contrary to this gloomy prognosis, the recession of2001 initially appeared one of the mildest on record(Nordhaus, 2002), but that view may not surviveongoing revisions to the data. A forecasting modelof US GDP developed by Muellbauer and Nunziata(2001) indicates that a sharp correction to USgrowth is likely, as the effects of falling share pricesand high debt exposure unwind. This is consistent

8 Alarm bells were ringing slightly earlier, as in the analysis of Godley (1999).

257

J. Temple

with the current view that the USA could experi-ence another ‘double dip’ recession. The morepessimistic observers, notably Godley (1999) andBrenner (2002), argue that a more serious crisiscould be in the making.

Assuming that the USA can avoid a prolongedslump of the kind seen in Japan, the prospects for thenext decade may still be relatively bright. Somecommentators have drawn attention to recent pro-ductivity growth figures, which appear unusuallystrong for this stage of the cycle (Baily, 2002). Intaking a longer view, the key question formacroeconomists is whether or not the faster rate ofproductivity growth can be sustained.

In forecasting future rates of productivity growth, itis essential to distinguish between the mediumterm—say 5–10 years—and longer horizons. It isalso useful to think in terms of forecasts that specifyan entire distribution over future growth rates. TheNew Economy has encouraged many to revise themean of such forecasts upwards, but also to empha-size that the degree of uncertainty has risen consid-erably.

Considering the medium term, Jorgenson and Stiroh(2000a) make a strong case for revising the sustain-able rate of US productivity growth upwards slightly.They also emphasize that this is dependent oncontinuing rapid technical progress in the productionof semiconductors. If the rate of technologicalchange in this sector slows down again, the economywill be hit by a ‘double whammy’. TFP growth inICT production will fall, and the price of computerswill decline at a slower rate, possibly calling a halt tothe boom in equipment investment.

Over the longer term, the degree of uncertainty iseven greater. Modern theoretical models of ‘endog-enous growth’ relate the long-run growth rate tovariables such as the productivity of researchers,via the generation of new ideas. We have verylimited insight into how such variables will evolveover time. For example, it is possible that theInternet will raise the productivity of researchers,but equally possible that it is continually becomingharder to extend the technological frontier, tendingto restrict the long-term growth rate.

In a careful empirical analysis, Jones (2002) hasemphasized two reasons to predict a long-term fallin the growth rate of the USA. He points out that itwill be hard to sustain current growth rates oncerising educational attainment starts to level off, andrising research intensity starts to approach a steady-state level. The latter transition could take placeover such a long period of time that we should notattach much weight to this consideration. Neverthe-less, the analysis of Jones is an important reminderthat some of the factors behind America’s post-wargrowth—notably, sustained improvements in edu-cational attainment—cannot be maintained forever.

IX. THE REST OF THE WORLD

Most discussions of the New Economy take UStechnological leadership for granted. It is not diffi-cult to identify possible reasons why Americancompanies are consistently at or near the techno-logical frontier, and why they are usually the onespushing it forward. Popular explanations include alarge domestic market, high educational attainment,dynamic research institutions, and advanced capitalmarkets. Gordon (2002) provides a more detaileddiscussion of these and other factors.

A keen awareness of America’s advantages, andsuperior performance, often motivates criticism ofan apparently moribund Europe. The stark contrastbetween the American and European records in the1990s has given an extra impetus to calls for Euro-pean reform, and thus the New Economy stands atthe centre of some crucial policy debates. In March2000, EU leaders at the Lisbon summit expressedtheir ambition to make Europe the world’s mostdynamic region by 2010. With this end in mind, theBarcelona summit of March 2002 reaffirmed acommitment to wide-ranging economic reform.

The first point to make is that Europe has laggedbehind before, and then started to catch up. Theprocess of convergence was particularly clear in the1950s and 1960s, as European countries grew atunprecedented rates, in many cases more rapidlythan the USA. The convergence perspective sug-gests that, given the widening gap of the 1990s, oneshould revise upwards medium-term forecasts for

258


European growth, as European firms identify oppor-tunities to exploit new technologies. Although Eu-rope will benefit less from the direct effect oftechnical progress in semiconductor production,there is no reason why the accelerating price de-clines of computers should not have spurred anequipment investment boom similar to that observedin the USA. This would be the case even in theabsence of reform.

In this respect, it is essential to quantify the extent towhich ICT investment has already contributed togrowth in Europe and elsewhere. This task has beenundertaken by several of the contributors to thisissue. The paper by Nicholas Oulton analyses ICTand productivity growth in the UK. His estimatessuggest that ICT investments have accounted for alarge and increasing share of capital deepening inthe UK, and that they account for perhaps as muchas half of the labour productivity growth observedover 1994–8. Nevertheless, these developmentshave not been sufficient to raise the overall rate ofproductivity growth.9

This suggests that the Solow paradox is alive andwell: high-profile ICT investments are failing toraise the overall growth rate. Francesco Daveri, inhis contribution to this issue, asks whether the Solowparadox has indeed migrated to Europe. Contrary toconventional wisdom, he finds that perhaps two-thirds of the EU population lives in countries wherethe rate of ICT adoption is not dissimilar to that of theUSA. The period of more rapid adoption occurredrelatively late in the 1990s, so it is perhaps too soonto assess the effects on productivity growth. As yet,it is hard to discern a productivity accelerationcomparable to that of the USA.

This view is confirmed by a detailed comparison ofthe growth record of European countries, and otherOECD members, carried out by Andrea Bassaniniand Stefano Scarpetta of the OECD. Their paper inthis issue highlights the diversity of experienceacross countries, but confirms that the larger econo-mies of continental Europe have continued to growrelatively slowly. Their analysis also indicates thatchanges in ICT investments are often found to behigher in countries that also show a rise in aggregate

TFP growth. This could indicate that, as one mightexpect, some OECD countries are adapting to theNew Economy more readily than others, and sodoing well on several fronts. Explaining this varia-tion is a difficult task, but Bassanini and Scarpettadraw attention to the relationship across countriesbetween productivity performance and severalmeasures of the domestic regulatory framework.

Considering this research as a whole, one of themost interesting questions about the New Economyis why the acceleration in productivity growth wasconfined mainly to the USA, even though the boomin ICT investment was more widespread. As brieflydiscussed in the Appendix, measurement issuesmay be at the heart of the puzzle. Alternatively, onecould plausibly argue that, given the adjustmentcosts associated with these investments, and thelater onset of ICT adoption outside the USA, the restof the OECD can look forward to an improvementin productivity performance over the next fewyears.

In this respect De Long (2002) emphasizes that theaggregate growth contribution of ICT depends ontwo elements—the rate of growth of ICT inputs,and the weight on this rate of growth, which dependson the proportion of total income generated by ICTcapital services. In the USA, both input growth andthe weight on that growth have increased substan-tially over time, and this explains why the Solowparadox now appears less relevant. De Long arguesthat exactly the same pattern is likely to hold forEuropean investments in ICT, which suggests thatEurope can look forward to a rise in productivitygrowth.

Most of the world’s population lives outside theOECD member states, and so we should alsoconsider the impact of New Economy technologieson the developing world. In principle at least, theNew Economy would appear to offer considerableopportunities. A large recent literature has docu-mented the way in which developing countriesappear to suffer from an ‘ideas gap’ or ‘knowledgegap’ rather than simply a lack of capital equipmentor skilled labour. From this perspective, some newtechnologies, and most obviously the Internet, have

9 For other discussions of aspects of the New Economy in the UK, see Greenhalgh and Gregory (2001) and Kneller and Young(2001).

259

J. Temple

the potential to make an important contribution todevelopment over the next few decades.

Realizing this potential could be very difficult, asMatti Pohjola emphasizes in his contribution to thisissue. Relatively few developing countries haveinvested heavily in ICT, and even fewer show anoticeable impact. Policies designed to promoteICT investment may run into problems similar tothose of earlier investment projects in less-devel-oped countries, perhaps including misallocation anda lack of complementary inputs.

In principle, the production and consumption ofcertain new technology goods and services, such ascomputer software, can spread across nationalboundaries more easily than traditional goods andservices. Yet some of the familiar barriers to knowl-edge diffusion may also turn out to be barriers to thediffusion of ‘knowledge products’. In this respect,the policy implications for developing countries couldlook much the same in the New Economy as in theold. Investments in ICT are not a good substitute foraddressing concerns of longer standing.

X. CONCLUSIONS

One of the surprising aspects of the academicliterature on the New Economy in the USA is thedegree of consensus on some of the major issues.This seems particularly remarkable for commen-tary on a very recent event, and one that took evenexpert observers by surprise. It may be that greatercontroversy will arise as time progresses, and newdata become available. Before such a debate be-gins, it helps to know where we stand at present,even if that is sometimes on thin ice. We cancertainly provide some preliminary answers to sev-eral of the questions raised earlier.

There is a consensus that much of the recentimprovement in US productivity growth can betraced back to faster technical progress in ICTproduction. This has a direct effect on aggregateproductivity, and has also encouraged greater in-vestments in ICT goods, as the relative price ofcomputing power has fallen at an ever faster rate.

The combined effect of these forces was to sustainthe boom of the 1990s for an unusually long time.This has encouraged some commentators to revisetheir medium-term growth forecasts upwards, al-though it should be emphasized that the degree ofuncertainty has risen as well.

We know less about the reasons for the stability ofinflation, and for the greater overall stability that wassuch a marked feature of the USA in the 1990s. Inthe first case, the problem is to discriminate betweenmany competing explanations, some of which arebased on movements in a variable that is not directlyobserved, namely the natural rate of unemployment.In the second case, although we again have sometentative explanations, there is continued disagree-ment among macroeconomists about the most im-portant sources of short-run fluctuations. This tendsto imply that it will be hard to establish widelyaccepted explanations for the decline in volatility,although changes in the management of inventoriesand in the volatility of consumption appear to beleading candidates.

We know much less about New Economy develop-ments elsewhere. Although few countries havematched the dynamism of the USA, there are somemore general developments that could have far-reaching implications, notably greater stability andwidespread increases in ICT adoption. Some of thecontributors to this issue provide new evidence onICT investments, and on the productivity recordmore generally. As they emphasize, we know sur-prisingly little about the reasons for the recentdivergence in performance, even though such evi-dence is a vital component of any informed debateabout the need for economic reform.

Overall, this dictates caution, and it does seem tooearly to draw conclusions about the long-term ef-fects of America’s New Economy on the otherdeveloped nations, and hence to formulate policyimplications. In the absence of rigorous empiricaland historical research, one should be a littlewary of the traditional caricature of a moribundEurope urgently in need of radical reform. Europehas started to close the gap before, and may do soagain.

260


APPENDIX: MEASUREMENT ISSUES

This appendix provides a brief introduction to someof the measurement issues that arise in studying theNew Economy. It gives greatest emphasis to issuesarising in the measurement of output and productiv-ity. Additional discussion of some of these issuescan be found in van Ark (2002), while Jorgenson(2001) includes an overview of measurement issuesspecific to the ICT sector.

A first distinction to make is between gross nationalproduct (GNP) and net national product (NNP),where the latter takes into account the depreciationof capital assets. NNP is a useful concept becauseit acknowledges that some fraction of current GNPwill be devoted simply to replacing worn-out assets,rather than contributing to consumption or net in-vestment. Hence long-run growth in living stand-ards rests on growth in net output (NNP) not grossoutput (GNP).

More formally, the classic analysis of Weitzman(1976) provided a theoretical foundation for the useof NNP as a measure of welfare. He established aset of circumstances under which NNP will beproportional to the highest present discounted valueof future consumption that can be attained, given theeconomy’s production constraints. This implies that,if the discounted value of future consumption is usedas a measure of welfare, higher NNP correspondsto higher potential welfare.

Historically, GNP and NNP have moved closelytogether in the USA, since estimates of the extent ofdepreciation relative to GNP have been stable. TheGNP and NNP measures are likely to have divergedin recent years because, as discussed by Tevlin andWhelan (2002), the equipment investment boom ischanging the composition of the capital stock to-wards assets with shorter service lives. The out-come will be a faster rate of depreciation, and thisimplies that NNP will grow more slowly than GNPfor a transitional period. Adjustments for this effectwill make the New Economy output record appearless impressive, as pointed out by Baker (2002) in abrief discussion.

As discussed in the main text, the strong productivityrecord of the New Economy may also be partlycyclical in nature. The main opposing argument is

the unusual length of the boom. A brief review of thedebate can be found in Berry and England (2001).

In contrast, many other measurement issues sug-gest that standard figures may understate recentproductivity growth. One of the most familiar pointsis that real output and price changes are very hardto measure in some service sectors, such as financeand insurance, which are assuming greater impor-tance over time. It is often argued that the rate ofproductivity growth is being understated as a result,perhaps because the quality of services is improvingover time in unmeasured ways. Current measuresof output in some service sectors imply negativeproductivity growth even over periods of severaldecades, which should be regarded as implausible(Jorgenson and Stiroh, 2000b).

The critique of standard New Economy argumentsby Gordon (2000) takes a relatively sceptical viewof mismeasurement. One key point is that service-sector firms are often providing services to otherbusinesses, as intermediate inputs. These servicesoften make intensive use of ICT. If computers orother technological advances have raised the qualityof these intermediate industries in unmeasured ways,these effects should feed through into higher meas-ures of TFP for the producers of final goods andservices. Given that productivity gains outside themanufacturing sector appear to be modest, it is notclear that such an effect is present in the data.

There is, however, some alternative evidence forthe view that output mismeasurement is prevalent inthe New Economy. As pointed out in the main text,in recent years there has been an increasing dis-crepancy between the different methods of meas-uring GDP (Moulton, 2000; Nordhaus, 2001). Thefaster rate of growth seen on the income side of thenational accounts is consistent with the view thatsome components of GDP, perhaps including serv-ice-sector outputs, are being mismeasured. Theprecise extent of the problem is hard to gauge; vanArk (2002) provides a more detailed discussion.

Another key issue in New Economy measurementconcerns the rate of relative price declines forICT goods. A good estimate of real output in the ICTsector, and of the quantity of real investment inICT, requires an accurate price deflator, which isdifficult to achieve when computer technology is

261

J. Temple

advancing at current rates. The challenge is todistinguish changes in price that are due to changesin quality, from changes in price that hold qualityconstant. The latter changes are needed to form thebasis of a ‘constant quality’ price index for ICTgoods. In other words, how fast is the price of agiven amount of computing power declining overtime?

To address this challenge, the USA is ahead ofmany other countries in adopting ‘hedonic’ meas-urement of certain price deflators, whereby pricesof differentiated goods are related to their charac-teristics. This means that one can estimate the rateof price decline while holding these characteristicsconstant. For example, the prices of computersmight be related to memory size, hard disk capacity,and so on, together with a set of time dummies,allowing an estimate of the relative decline in thecost of computing power over time. A less sophis-ticated method is to compare prices of identicalproducts across periods, but this ‘matched model’approach is hard to implement when models arechanging rapidly, as in personal computers or mobilephones. Jorgenson (2001) and Stiroh (2002b) pro-vide more detailed introductions to the measure-ment of ICT prices.

The US national accounts now incorporate theseconstant-quality price indices for several ICT prod-ucts, including computer hardware, semiconduc-tors, and prepackaged software. As Jorgenson(2001) points out, a remaining challenge is to intro-duce a similar approach for other forms of softwareinvestment, and for communications equipment.Investment in software is now an important compo-nent of total business investment in the USA, whichreinforces the case for better measurement.

Differences in the construction of price deflatorscould partially account for observed differences ineconomic performance across countries. For ex-ample, because the relatively sophisticated USdeflators for ICT equipment are likely to capturefalling ICT prices more successfully than thoseelsewhere, simply the method of data constructioncould account for differing productivity growthacross countries. Schreyer (2002) examines thispoint in detail and makes the reassuring point that, atthe aggregate level, some of the effects are offset-ting, not least because ICT goods are often interme-

diate inputs. The effects of alternative measure-ment conventions are likely to be rather moreserious for comparing disaggregated data acrosscountries, such as growth in real ICT investment, orthe output and productivity of particular industries.

Further measurement issues relate to capital inputsand their treatment in growth accounting. A growth-accounting exercise requires an estimate of growthin capital services. In Jorgenson and Stiroh (2000a)a measure of capital services is computed whichtakes into account the differing productivity ofdifferent capital assets, and therefore changes in thequality of capital over time. Jorgenson (2001) em-phasizes that such an approach is essential at a timewhen the composition of the capital stock has beenundergoing rapid change, as in the recent boom ofequipment investment.

This approach does raise one issue of interpretation.In principle, changes in the measured quality ofcapital services could partly reflect technologicaladvances that raise the performance of certaintypes of capital goods. Hence, when this approachis used, it should be remembered that the resultingmeasure of TFP growth (the residual term in thegrowth-accounting equation) will not capture all theeffects of technical progress as conventionally un-derstood.

Some authors use a ‘wealth’-based measure of thecapital stock as a proxy for capital services. Thesewealth-based measures are based on the currentvalue of assets, rather than their productive capac-ity. In calculating these stocks, standard deprecia-tion profiles take into account declines in the valueof assets as they wear out, and their useful lifeshortens. For ICT goods, however, these deprecia-tion rates may be misleading, at least when the aimis to measure changes in productive capital servicesand their contribution to output growth. To clarifythe distinction, Whelan (2002) gives the example ofa light bulb, which retains its productive capacityuntil the date it expires, even though its market valueis diminishing as it nears the end of its service life.

It is likely that ICT equipment will usually maintainits productive capacity until it is finally replaced dueto technological obsolescence. Whelan (2002) pro-vides a clear treatment of these issues and theimplications. His measures of the growth of ICT

262


REFERENCES

Baily, M. N. (1981), ‘Productivity and the Services of Capital and Labor’, Brookings Papers on Economic Activity,1, 1–50.

— (2001), ‘Macroeconomic Implications of the New Economy’, Institute for International Economics, WorkingPaper WP01-9, September.

— (2002), ‘The New Economy: Post Mortem or Second Wind?’, Journal of Economic Perspectives, 16(2), 3–22. — Lawrence, R. Z. (2001), ‘Do We Have a New E-conomy?’, American Economic Review, 91(2), 308–12.Baker, D. (2002), ‘Is the New Economy Wearing Out?’, Challenge, 45(1), 117–21.Ball, L. (1999), ‘Aggregate Demand and Long-run Unemployment’, Brookings Papers on Economic Activity, 2, 189–

251. — Moffitt, R. (2001), ‘Productivity Growth and the Phillips Curve’, NBER Working Paper No. 8421. — Tchaidze, R. (2002), ‘The Fed and the New Economy’, American Economic Review, 92(2), 108–14.Basu, S., Fernald, J. G., and Shapiro, M. D. (2001), ‘Productivity Growth in the 1990s: Technology, Utilization, or

Adjustment?’, NBER Working Paper No. 8359.Berry, S., and England, D. (2001), ‘Has There Been a Structural Improvement in US Productivity?’, Bank of England

Quarterly Bulletin, 41(2), 203–9.Blanchard, O., and Simon, J. (2001), ‘The Long and Large Decline in US Output Volatility’, Brookings Papers on

Economic Activity, 1, 135–74.Bond, S. R., and Cummins, J. G. (2000), ‘The Stock Market and Investment in the New Economy: Some Tangible Facts

and Intangible Fictions’, Brookings Papers on Economic Activity, 1, 61–124.Boskin, M. J. (2000), ‘Economic Measurement: Progress and Challenges’, American Economic Review, 90(2), 247–

52.Bosworth, B. P., and Triplett, J. E. (2001), ‘What’s New about the New Economy? IT, Economic Growth and

Productivity’, International Productivity Monitor, Spring, 19–30.Brenner, R. (2002), The Boom and the Bubble: The US in the World Economy, London, Verso.Brynjolfsson, E., and Hitt, L. (2000), ‘Beyond Computation: Information Technology, Organizational Transformation,

and Business Performance’, Journal of Economic Perspectives, 14(4), 23–48.Burns, A. F. (1960), ‘Progress Towards Economic Stability’, American Economic Review, 50(1), 1–19.Campbell, J. Y., and Shiller, R. J. (2001), ‘Valuation Ratios and the Long-run Stock Market Outlook: An Update’, NBER

Working Paper No. 8221.Card, D., and DiNardo, J. E. (2002), ‘Skill Biased Technological Change and Rising Wage Inequality: Some Problems

and Puzzles’, NBER Working Paper No. 8769.De Long, B. (2002), ‘Boom-time in Europe’, Financial Times, 24 July.Freeman, R. (2001), ‘The Rising Tide Lifts . . . ?’, NBER Working Paper No. 8155. — Schettkat, R. (2001), ‘Marketization of Production and the US–Europe Employment Gap’, Oxford Bulletin of

Economics and Statistics, 63, 647–70.

capital services show faster growth than widelyused alternatives, and therefore tend to raise thecontribution of ICT capital deepening to recent USgrowth, and lower the contribution of other forces.

Finally, there are some additional measurementissues which are long-standing and intractable, butwhich may be assuming greater importance in theNew Economy. These include the difficulties fornational income accountants caused by the intro-duction of new goods, and the changing boundarybetween market and non-market production. Boskin(2000), Moulton (2000), and Nordhaus (2000) re-

view some of the relevant problems, and the pros-pects for solving them.

The issue of market versus non-market production,especially, is gaining greater attention. Freeman andSchettkat (2001) argue that differences in themarketization of production can account for differ-ences in labour-force participation between theUSA and Germany. This would also have implicationsfor international comparisons of GDP per capita.More generally, Nordhaus (2000) argues that ac-counting for non-market production could challengethe conventional wisdom in many research areas.

263

J. Temple

Gali, J. (2000), ‘The Conduct of Monetary Policy in the Face of Technological Change: Theory and Postwar USEvidence’, Universitat Pompeu Fabra, manuscript.

— David Lopez-Salido, J., and Valles, J. (2002), ‘Technology Shocks and Monetary Policy: Assessing the Fed’sPerformance’, Universitat Pompeu Fabra, manuscript.

Godley, W. (1999), ‘Seven Unsustainable Processes’, Jerome Levy Economics Institute, Special Report.Gordon, R. J. (1998), ‘Foundations of the Goldilocks Economy: Supply Shocks and the Time-varying NAIRU’,

Brookings Papers on Economic Activity, 2, 297–346. — (2000), ‘Does the “New Economy” Measure Up to the Great Inventions of the Past?’, Journal of Economic

Perspectives, 14(4), 49–74. — (2002), ‘Technology and Economic Performance in the American Economy’, NBER Working Paper No. 8771.Greenhalgh, C., and Gregory, M. (2001), ‘Structural Change and the Emergence of the New Service Economy’, Oxford

Bulletin of Economics and Statistics, 63, 629–46.Hall, R. E. (2000), ‘E-capital: The Link Between the Stock Market and the Labor Market in the 1990s’, Brookings Papers

on Economic Activity, 2, 73–102. — (2001), ‘The Stock Market and Capital Accumulation’, American Economic Review, 91(5), 1185–202.Hansen, B. E. (2001), ‘The New Econometrics of Structural Change: Dating Breaks in US Labour Productivity’, Journal

of Economic Perspectives, 15(4), 117–28.Hobijn, B., and Jovanovic, B. (2001), ‘The Information-technology Revolution and the Stock Market: Evidence’,

American Economic Review, 91(5), 1203–20.Jones, C. I. (2002), ‘Sources of US Economic Growth in a World of Ideas’, American Economic Review, 92(1), 220–

39.Jorgenson, D. W. (2001), ‘Information Technology and the US Economy’, American Economic Review, 91(1), 1–32. — Stiroh, K. J. (2000a), ‘Raising the Speed Limit: US Economic Growth in the Information Age’, Brookings Papers

on Economic Activity, 1, 125–235. — — (2000b), ‘US Economic Growth at the Industry Level’, American Economic Review, 90(2), 161–7.Katz, L. F., and Krueger, A. B. (1999), ‘The High-pressure US Labour Market of the 1990s’, Brookings Papers on

Economic Activity, 1, 1–87.Kneller, R., and Young, G. (2001), ‘The New British Economy’, National Institute Economic Review, 177, 70–84.Lane, R. E. (1998), ‘The Joyless Market Economy’, in A. Ben-Ner and L. Putterman (eds), Economics, Values and

Organization, Cambridge, Cambridge University Press.Machin, S. (2001), ‘The Changing Nature of Labour Demand in the New Economy and Skill-biased Technology

Change’, Oxford Bulletin of Economics and Statistics, 63, 753–76.Mankiw, N. G. (2001a), ‘The Inexorable and Mysterious Tradeoff Between Inflation and Unemployment’, The

Economic Journal, 111, C45–61. — (2001b), ‘US Monetary Policy During the 1990s’, NBER Working Paper No. 8471.McConnell, M. M., and Perez-Quiros, G. (2000), ‘Output Fluctuations in the United States: What has Changed since

the Early 1980s?’, American Economic Review, 90(5), 1464–76. — Mosser, P. C., and Perez-Quiros, G. (1999), ‘A Decomposition of the Increased Stability of GDP Growth’, Current

Issues in Economics and Finance, 5(13), Federal Reserve Bank of New York.Miller, M., Weller, P., and Zhang, L. (2002), ‘Moral Hazard and the US Stock Market: Analysing the “Greenspan Put” ’,

The Economic Journal, 112, C171–86.Moulton, B. R. (2000), ‘Getting the 21st-century GDP Right: What’s Underway?’, American Economic Review, 90(2),

253–8.Muellbauer, J., and Nunziata, L. (2001), ‘Credit, the Stock Market and Oil: Forecasting US GDP’, CEPR Discussion Paper

No. 2906.Nordhaus, W. D. (2000), ‘New Directions in National Income Accounting’, American Economic Review, 90(2), 259–

63. — (2001), ‘Productivity Growth and the New Economy’, NBER Working Paper No. 8096. — (2002), ‘The Mildest Recession: Output, Profits, and Stock Prices as the US Emerges from the 2001 Recession’,

NBER Working Paper No. 8938.Oliner, S. D., and Sichel, D. E. (1994), ‘Computers and Output Growth Revisited: How Big is the Puzzle?’, Brookings

Papers on Economic Activity, 2, 273–317. — — (2000), ‘The Resurgence of Growth in the Late 1990s: Is Information Technology the Story?’, Journal of

Economic Perspectives, 14(4), 3–22.Putnam, R. D. (2000), Bowling Alone: The Collapse and Revival of American Community, New York, Simon & Schuster.

264


Quah, D. T. (2002), ‘Technology Dissemination and Economic Growth: Some Lessons for the New Economy’, CEPRDiscussion Paper No. 3207.

Schreyer, P. (2002), ‘Computer Price Indices and International Growth and Productivity Comparisons’, Review ofIncome and Wealth, 48(1), 15–31.

Sennett, R. (1998), The Corrosion of Character, New York, Norton.Shimer, R. (1998), ‘Why is the US Unemployment Rate So Much Lower?’, NBER Macroeconomics Annual, 13, 11–

61.Staiger, D., Stock, J. H., and Watson, M. W. (2001), ‘Prices, Wages and the US NAIRU in the 1990s’, NBER Working

Paper No. 8320.Stiroh, K. (1999), ‘Is There a New Economy?’, Challenge, 42(4), 82–101. — (2002a), ‘Are ICT Spillovers Driving the New Economy?’, Review of Income and Wealth, 48(1), 33–57. — (2002b), ‘Measuring Information Technology and Productivity in the New Economy’, World Economics, 3(1),

43–58.Temple, J. R. W. (2000), ‘Inflation and Growth: Stories Short and Tall’, Journal of Economic Surveys, 14(4), 395–426.Tevlin, S., and Whelan, K. (2002), ‘Explaining the Investment Boom of the 1990s’, Journal of Money, Credit, and

Banking, forthcoming.van Ark, B. (2002), ‘Measuring the New Economy: An International Comparative Perspective’, Review of Income and

Wealth, 48(1), 1–14.Van Reenen, J. (2001), ‘The New Economy: Reality and Policy’, Fiscal Studies, 22(3), 307–36.Wadhwani, S. B. (2001), ‘The New Economy: Myths and Realities’, Bank of England Quarterly Bulletin, 41(2).Weitzman, M. L. (1976), ‘On the Welfare Significance of National Product in a Dynamic Economy’, Quarterly Journal

of Economics, 90(1), 156–62.Whelan, K. (2002), ‘Computers, Obsolescence, and Productivity’, Review of Economics and Statistics, 84(3), 445–

61.Woodward, B. (2000), Maestro: Greenspan’s Fed and the American Boom, New York, Simon & Schuster.Young, A. (1995), ‘The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth

Experience’, Quarterly Journal of Economics, 110(3), 641–80.

the assessment: the new economy - university of … · the assessment: the new economy ... economy...

Documents