Information about http://econ.lse.ac.uk/~dquah/p/9702iwe.pdf
Increasingly weightless economies By Danny T Quah(1) (Centre for…
Tags: academics, aggregate output, bank of england, commodities, consumption, economic activity, economic growth, economic performance, evolution, fraction, gdp, information technology, large scale, london school of economics, macroeconomic policy, occasional series, physical manifestation, proceeds, wealth creation, weightlessness,Pages: 8
Language: english
Created: Mon Jun 2 11:54:48 1997
Display cached document
Page 1
Page 2
Page 3
Page 4
Page 5
Page 6
Page 7
Page 8
Increasingly weightless economies
By Danny T Quah(1) (Centre for Economic Performance, the London School of Economics).
This article is one of an occasional series provided by academics working outside the Bank of England.
The views expressed reflect those of the author rather than those of the Bank of England.
In this article Danny T Quah examines how, when an economy grows, its patterns of production and
consumption systematically change. He describes one such large-scale evolution, namely, the increasing
weightlessness of aggregate output across advanced economies. In all fast-growing successful countries,
growth in information technology has contributed positively both to increasing weightlessness and to
economic growth. In the sample of countries studied here, the richer the country the higher the
contribution to growth of information technology and services; in no country has manufacturing, as
traditionally construed, continued to be as important.
Introduction and macroeconomic policy must adjust as dematerialisation
proceeds.
What happens to the mix of economic activity as developed
economies continue to grow? This article presents empirical In assessing the evidence on sectoral change and aggregate
evidence that shows such economies become increasingly growth, this article examines whether dematerialisation and
weightless over time. By this I mean that greater value, IT will matter increasingly for economic performance and
as a fraction of GDP, resides in economic commodities wealth creation.
that have little or no physical manifestation. Another
description of such structural change is progressive Such an exercise can refine our understanding of the nature
dematerialisation. of economic growth. And the analysis has an immediate
policy implication. Statistical agencies divide their research
Examples of weightlessness and dematerialisation are resources across a range of activities to build up a composite
diverse: they range from economic activities such as picture of the economy. Knowing that some sectors are
stocking supermarket shelves and styling haircuts, all the likely to become more important than others can help
way through financial services and up to improve the allocation of those resources. For instance, as
telecommunications and providing software products on the manufacturing sector shrinks as a percentage of GDP
CD-ROM. Within the diversity of dematerialisation and shows behaviour different from those sectors that grow
different activities might also have widely differing faster (ie, services), the performance of manufacturing will
implications for the ability to sustain economic growth. For reveal progressively less about the macroeconomy.(2)
example, dematerialisation includes high technology but
also low technology activities such as making hotel beds or
The economic issues
providing gardening services. The latter might well carry a
society through hard times but are unlikely to forge nations Two aspects suggest that dematerialisation might be
into enduring economic powerhouses. macroeconomically important. The first is simply
increasing weightlessness deriving from the growth of
One prominent form of dematerialisation is concentration of services--as opposed to, say, manufacturing in particular, or
activity in information technology (IT). Economists are not industry in general. The second is dematerialisation
all agreed that this matters, or that such change differs from deriving from the increasing importance of IT.
any other kind of macroeconomic transformation. Those
who suspect it might be important and different have just These two aspects differ in their economic implications.(3)
begun to study its implications for a range of issues: But they are both controversial. Some economists doubt if
whether economic growth can be sustained in advanced the basis of a strong, growing economy can be provided by
economies; what might happen to the distribution of services in general or IT in particular. Some have argued
income across and within societies; and how trade patterns that weightlessness might be regarded as being without
(1) Director of the National Economic Performance Programme, CEP and Professor of Economics, LSE. The author would like to thank
Louise C Keely for help, and the British Academy for financial support. Anonymous referees provided extremely useful comments and
suggestions.
(2) For studying aggregate business cycles, this point has been emphasised in Lee (1996).
(3) Quah (May and October 1996) discusses some of those in greater detail.
49
Bank of England Quarterly Bulletin: February 1997
substance and, consequently, without value. Two versions q Agriculture includes agricultural and livestock production
of this view can be distinguished. To the extent that and services; fishing; hunting; and logging and forestry.
increasing weightlessness means further development of
multimedia in entertainment and video games, some q Industry includes mining and quarrying; manufacturing;
consider it to be frivolous at best but actually harmful construction; and electricity, gas, and water.
at worst. Others view increasing weightlessness as
irrelevant. q Finally, services includes transport, storage, and
communications; wholesale and retail trade; banking,
It is useful to say explicitly what I mean by insurance, and real estate; ownership of dwellings;
dematerialisation in economic growth, and to consider public administration and defence; other services; and
services and IT separately. I ask below, where in advanced statistical discrepancies.
economies does growth in GDP occur?
The natural step is to identify services as the sector where
Nominal GDP--like any other measure of aggregate output increasing weightlessness matters, and to take agriculture
that macroeconomists study--represents both value created and industry as being the opposite.
and that willingly paid for by consumers. If something is
valued and marketed then it shows up as part of GDP. But, for studying the issues of interest here, this partitioning
Among economists at least, there should be no controversy is not ideal. IT does not show up neatly on only one side of
over whether IT is what people want or whether multimedia either manufacturing or services. It comprises
entertainment adds economic value. These should simply be semiconductors, computers, software, telecommunications,
evaluated at market prices and sized up relative to
and IT services (see, eg, OECD). Thus, IT straddles, among
everything else that enters GDP. Weight is irrelevant for
other things, manufacturing under industry, and transport,
economic value; what matters is how large that economic
storage, and communications under services.
value is.
Production of semiconductors is manufacturing. But,
Examining output through GDP gives a measure of domestic
semiconductors are also prime examples of
value added. But detail on--for instance--UK spending on
dematerialisation. The location of their manufacture is
US imports has to be obtained from the expenditure measure
unimportant because transportation costs are trivial
of GDP. If the typical UK resident had come to depend on
(see, eg, the anecdotal evidence in Quah, May and October
IT products--for example, through banking by computer or
1996). Semiconductor value derives from a logic
shopping over the internet--but all those were imported IT
configuration that sits on top of an ultra-thin wafer of
products from the United States, then little of what really
silicon; the physical material, to all intents and purposes, is
mattered in economic life would show up in UK GDP.
worthless. Yet, semiconductor production is recorded as
Analysing the GDP output measure alone would then
manufacturing--just as are making steel cranes and railway
underestimate IT's true importance in the UK economy. It
sleepers.
is useful, therefore, to look also at the spending side; I will
do so below.
For analysing dematerialisation, difficulties remain even at
more micro levels of disaggregation. The International
Decomposing the United Kingdom's GDP Standard Industrial Classification (ISIC) scheme partitions all
growth: dematerialisation and IT economic activity into categories--the greater the number of
According to the IMF Balance of Payments Statistics digits in the ISIC code, the finer the disaggregation. At the
Yearbook, in all OECD economies exports of services four-digit level, the category ISIC 3825--office, computing,
accounted for more than 10% of total gross exports of goods and accounting equipment manufacture--includes not only
and services over the period 199094. The OECD economies computers and peripherals, but also typewriters, cash
can be split into two groups: first, exporters primarily of registers, and simple accounting machines. Electronic
goods, for example Canada, Germany and Japan; and components--a key IT ingredient--is placed not with
second, more diversified exporters, including the United computers, but elsewhere in ISIC 3832 (radio, TV, and
Kingdom, where services amounted to close to 25% or more communication equipment manufacture) instead, which then
of total gross exports of goods and services. And for the includes many other things unrelated to IT.
United Kingdom, of the value added from exports, services
accounted for 43% in 1990 (the latest data available). Thus, the agriculture, industry and services split provides
little sharp insight on GDP dematerialisation and IT.
To begin, consider how GDP in the United Kingdom has Nevertheless, such a division is the only one available for a
evolved: what has contributed how much to the growth of wide range of countries at differing levels of development.
GDP here? Therefore, it is the one I use. Such a division, while crude,
shows general tendencies reasonably well. And, the
Traditional macroeconomic accounting divides GDP into methodology I describe will readily apply as more apposite
three principal categories: agriculture, industry, and data become available: I use this fact below when I
services. combine IT and GDP data from different sources.
50
Increasingly weightless economies
Let Y(t) denote the time t flow of GDP. Index by j an Table A shows the nominal GDP decomposition for
arbitrary partitioning of the economy--say into the three five-year intervals from 1972 for the
categories agriculture, industry and services--so that: agriculture/industry/services split, but then also peels out the
manufacturing component in industry.
Y = Yj (1)
j Table A
Decompositions of UK GDP growth over five-year
where Yj denotes GDP in category j. Differentiate both intervals
sides of equation (1) with respect to time t, and use dots to Per cent per year
denote time derivatives. Then, normalising by Y, equation Time span /g
g Agr. Ind. (Mnf.) Svc. IT (a)
(1) becomes:
197277 9.8 2.2 39.7 (25.5) 58.1
.
(Y j / Y )
. . 197782 13.1 2.0 42.1 (20.7) 55.8
Y / Y = Yj / Y = Yj / Yj x (2) 198287 7.2 1.0 26.9 (22.6) 72.1
j j 198792 9.0 1.4 21.3 (15.6) 77.3 1.1
(a) The IT figure is for 198794.
In each row, the figures under Agr., Ind. and Svc. sum to 100, subject to rounding error.
Define g to be the proportional growth rate of Y and gj that
of Yj. Let sj be sector share Yj/Y. Then equation (2) can be
rewritten as Next, I used data from the OECD 1995 Information
Technology Outlook to calculate the contribution of IT to
g = gj sj = j with j = gj x sj (3) GDP growth, although only from 1987. Unlike the
j j value-added calculation given in the
Equation (3) decomposes GDP growth into contributions agriculture/industry/services split, IT contribution means
due to the different sectors. Provided that g is not zero, spending on IT, not production. Therefore, division (1) is
equation (3) then says j j/g = 1, so that j/g is sector j's used now with Y as total spending, rather than total
relative contribution to total GDP growth.(1) production. Nevertheless, the same principle applies.
Table A shows that services' contribution to UK growth rose
Each j is the product of the sector's share in total GDP with
from twice manufacturing's in 197277 to more than five
that sector's proportional growth rate. Thus, in general,
times the latter by 198792. Over the same period,
sector j will show a high contribution to growth only when
services' growth contribution rose from one and a half times
both its share sj and its growth rate gj are high. Growth
industry's to almost four times the latter. This increase
rates and sector shares g, gj, and sj will typically all change
seems dramatic, but even by the 1970s, UK growth was
through time; but, at each moment in time, equation (3) will
already more than half due to services alone. Regardless of
hold.
whether historical overall growth is considered strong or
weak, it is unambiguous that the services sector has
Since the split of the economy into categories in (1) is
contributed substantially and increasingly to UK wealth
arbitrary, a sector's contribution to GDP growth can always
creation.
be estimated by gj x sj, regardless of whether complete data
exist on all remaining sectors. When the data are The IT figure of only 1.1% for 198792 is, by comparison,
exhaustive, then the sector contributions sum to 1, but that is tiny--smaller even than agriculture.
the only additional insight from having the entirety of
sectors. Can services' /g--contribution to GDP growth--continue
to be so much larger than all the other sectors'? If the
Using the World Bank's 1996 World Tables, I calculated the economy undergoes balanced growth--when all sectors
decomposition (3) for UK GDP measured in nominal grow at the same rate and thus sector shares are constant--
US dollars at prevailing exchange rates. Use of nominal then j/g ratios simply reflect those different (constant)
GDP data does mean that the results are subject to a couple shares.(2)
of caveats. When considering the contribution of each
sector it might be more informative to look at the real Thus, in balanced growth, our observations on the relative
contribution--excluding the effect of different inflation rates contributions in Table A apply not just to growth dynamics,
(or even industry specific purchasing power parity exchange but to level shares as well. Using Table A then as a
rates) between sectors. But this breakdown is difficult to prediction on long-run, steady-state growth, the
achieve accurately if price measures do not adequately overwhelming importance of services is obvious.
account for changes in quality: this may be important for
computers where prices are measured per computer rather Of course, most economies need not be undergoing balanced
than per unit of computing power. The importance of this is growth just yet. Some sectors might grow much faster than
uncertain and it is left to further research to consider these others and will thus be increasing their share of total GDP.
issues. What then can we learn from the calculations underlying
(1) Even if g is negative with some j positive, so that j/g < 0, the interpretation still goes through: sector j restrained the economy from wherever it
would have otherwise gone.
(2) To see this, notice that equation (3) with gj = g for all j gives j = gj sj = gsj j/g = sj.
51
Bank of England Quarterly Bulletin: February 1997
Table A? By definition, sector j's share is United States below, this possibility needs to be investigated
further. More finely disaggregated and timely data would
sj = Yj/Y help here.
Taking proportional growth rates on both sides gives I now turn to the growth experiences of other economies,
but it is useful to summarise the lessons thus far. Tables A
. and B have provided a picture of the UK economy where
s/sj = gj - g
= (j/sj) - g the performance of the weightless services sector has been
= [j/g - sj] x (g/sj) (4) the outstanding characteristic in aggregate economic growth.
The sector is not only large, but continues to outpace all
others. If the current trend were to continue, within a
Equation (4) says that the sector share's proportional growth
decade manufacturing would contribute no more than one
rate depends on how large that sector share already is
tenth of the total value generated in the economy.
compared to the overall growth rate. Of course, since sector
shares have to lie between 0 and 1, this growth cannot
continue indefinitely, but away from those boundaries, Decomposing economic growth across
equation (4) gives a rough guide as to how sector shares will countries
evolve. Tables C.1, C.2, C.3, C.4 and C.5 provide results for a range
of countries with differing growth experiences--the United
In Table B I present growth dynamics for the services, States, Singapore, Korea, Pakistan, and the Philippines,
manufacturing, and IT sector shares. Within each bloc, respectively: they present the same growth decompositions
column s shows percentage share; column /g shows as given for the United Kingdom in Table A. Similarly,
.
contribution to total growth; and column s/s shows how fast Tables D.1, D.2, D.3, D.4 and D.5 present sector share
the sector share is growing. I emphasised above that these dynamics for the different countries, analogous to Table B.
(For Pakistan and the Philippines, I have been unable to
Table B obtain IT numbers.)
Changes in UK sector shares: services, manufacturing,
and IT On Table C.1 we see that the United States is an economy
Per cent per year where the services sector growth contribution has risen,
Svc.
.
Mnf.
.
IT (a)
. continually, from under 60% in 197277 to over 80% in
s /g s/s s /g s/s s /g s/s
198792. These contributions exceed their UK
197277 54.7 58.1 0.6 30.4 25.5 -1.6
197782 54.9 55.8 0.2 27.8 20.7 -3.4 counterparts. At the estimated 2% annual rate of decrease
198287 57.5 72.0 1.8 25.0 22.7 -0.7
198792 62.8 77.3 2.1 23.6 15.6 -3.1 2.5 1.1 -5.2 (Table D.1, 198792) manufacturing's share of 20% would
A sector is expanding when its growth contribution /g exceeds its share s. The sector growth rate
. decline to 15% in ten years. In the United States the IT
s/s = [/g - s](g/s) is given per cent per year; ratios (s, /g) are in percentage points.
share was only 2.4% of GDP by the end of the 1980s,
(a) The IT figures are for 198794.
marginally lower than in the United Kingdom. But, unlike
the United Kingdom, the IT sector share is estimated to be
figures can show no more than rough tendencies; growing at over 2% per year. This figure though seems
nevertheless, it is useful to provide an interpretation for quite small: if maintained, it implies only a doubling in
them. Thus, take the row for 198792: services' sector 30 years.
share is growing at 2% per year. If this continues, then
services' current 60% share will become 90% in 20 years.(1) Singapore is widely regarded as a successful, fast-growing
By contrast, the manufacturing share continues to decline: economy. The services sector has, throughout the sample,
with a rate of change of -3% the 20% share will decline to accounted for over 60% of GDP growth. However, that
15% in ten years. contribution has remained roughly constant, unlike the
United States and the United Kingdom where it has risen
Interestingly, although in absolute figures IT spending is sharply. In Singapore, manufacturing's growth contribution
increasing, its share of the total is declining: the growth has consistently remained more than one quarter; and its
contribution /g is less than s. Moreover, this low growth share in GDP began to decline only towards the end of the
contribution comes from IT's relatively low growth rate, not sample. Singapore's recent massive IT push (eg, Slavin
a low sector share. The 2.5% share in the United Kingdom 1996) has not yet manifested in these data: the IT share
is high compared with many other countries. The actually declined over 198794.
United States has about the same share, but one that is rising
rather than falling. Two interpretations are possible: first, Korea resembles Singapore in that manufacturing remains
IT is just not an important part of the burgeoning important for growth, but the share declines towards the end
dematerialised economic activity in the United Kingdom. of the sample. On the other hand, Korea has increased its
I think this implausible. Second, the United Kingdom IT share, but the figure of only 0.7% in 1987 is surprisingly
is not yet saturated with IT, and much more room remains small--as is IT's growth contribution of only 0.8% over
for continued expansion. Given the results for the 198792. Although services' growth contribution in Korea
(1) The ratio of 90 to 60 is 1.5, whose natural log is 0.4; dividing this by the growth rate gives time in years needed to make the transition.
52
Increasingly weightless economies
Table C.1 Table D.1
Decompositions of US GDP growth over five-year Changes in US sector shares: services, manufacturing,
intervals and IT
Per cent per year Per cent per year
/g Svc. Mnf. IT (a)
. . .
Time span g Agr. Ind. (Mnf.) Svc. IT (a) s /g s/s s /g s/s s /g s/s
197277 10.3 -5.0 48.3 (29.4) 56.8 197277 62.8 56.8 -1.0 24.3 29.4 2.2
197782 9.8 2.1 30.3 (15.6) 67.6 197782 63.7 67.6 0.6 22.5 15.6 -3.0
198287 7.4 0.8 22.1 (17.2) 77.1 198287 66.8 77.1 1.1 20.2 17.2 -1.1
198792 5.9 1.8 17.3 (13.0) 81.0 3.3 198792 69.5 81.0 1.0 19.2 13.0 -1.9 2.4 3.3 2.2
In each row, the figures under Agr., Ind., and Svc. sum to 100 (subject to rounding error). A sector is expanding when its growth contribution /g exceeds its share s. The sector growth rate
.
s/s = [/g - s](g/s) is given per cent per year; ratios (s, /g) are in percentage points.
(a) The IT figure is for 198794.
(a) The IT figures are for 198794.
Table C.2
Table D.2
Decompositions of Singapore GDP growth over five-year
Changes in Singapore sector shares: services,
intervals
manufacturing and IT
Per cent per year
Per cent per year
/g
Time span g Agr. Ind. (Mnf.) Svc. IT (a) Svc. Mnf. IT (a)
. . .
s /g s/s s /g s/s s /g s/s
197277 17.7 1.6 35.0 (27.9) 63.4
197782 18.4 0.5 39.4 (27.0) 60.1 197277 64.2 63.4 -0.2 24.0 27.0 2.2
198287 5.8 -1.1 40.0 (37.4) 61.1 197782 62.4 60.1 -0.7 26.8 27.0 0.1
198792 19.6 -0.0 35.0 (26.2) 65.0 1.7 198287 61.6 61.1 -0.0 25.4 37.4 2.7
198792 61.0 65.0 1.3 28.0 26.2 -1.3 1.9 1.7 -2.1
In each row, the figures under Agr., Ind., and Svc. sum to 100 (subject to rounding error).
A sector is expanding when its growth contribution /g exceeds its share s. The sector growth rate
.
(a) The IT figure is for 198794. s/s = [/g - s](g/s) is given per cent per year; ratios (s, /g) are in percentage points.
(a) The IT figures are for 198794.
Table C.3
Table D.3
Decomposition of Korea GDP growth over five-year
Changes in Korea sector shares: services,
intervals
manufacturing, and IT
Per cent per year
Per cent per year
/g
Time span g Agr. Ind. (Mnf.) Svc. IT (a) Svc. Mnf. IT (a)
. . .
s /g s/s s /g s/s s /g s/s
197277 28.1 20.3 39.4 (30.8) 40.4
197782 15.3 6.7 43.7 (29.1) 49.6 197277 43.3 40.4 -1.9 25.7 30.8 5.6
198287 12.5 5.0 47.3 (35.6) 47.8 197782 43.7 49.6 2.1 29.9 35.6 2.4
198792 17.7 5.3 45.4 (25.2) 49.2 0.8 198287 46.3 47.8 0.4 29.9 35.6 2.4
198792 47.5 49.2 0.7 30.0 25.2 -2.8 0.74 0.75 0.4
In each row, the figures under Agr., Ind., and Svc. sum to 100 (subject to rounding error).
A sector is expanding when its growth contribution /g exceeds its share s. The sector growth rate
.
(a) The IT figure is for 198794. s/s = [/g - s](g/s) is given per cent per year; ratios (s, /g) are in percentage points.
(a) The IT figures are for 198794.
Table C.4 Table D.4
Decompositions of Pakistan GDP growth over five-year Changes in Pakistan sector shares: services,
intervals manufacturing and IT
Per cent per year
Per cent per year
/g
Time span g Agr. Ind. (Mnf.) Svc. IT (a) Svc. Mnf. IT (a)
. . .
s /g s/s s /g s/s s /g s/s
197277 10.2 25.0 25.4 (14.1) 49.6
197782 15.0 29.9 22.2 (15.5) 47.9 197277 43.4 49.6 1.4 15.3 15.5 0.2
198287 1.6 -35.3 43.8 (34.4) 91.5 197782 45.8 47.6 0.7 15.2 15.5 0.2
198792 7.7 26.0 28.8 (18.9) 45.2 n.a. 198287 48.5 91.5 1.4 15.9 34.4 1.9
198792 49.0 45.2 -0.6 17.0 19.0 0.9 n.a.
In each row, the figures under Agr., Ind., and Svc. sum to 100 (subject to rounding error).
A sector is expanding when its growth contribution /g exceeds its share s. The sector growth rate
.
(a) No IT data were available. s/s = [/g - s](g/s) is given per cent per year ratios (s, /g) are in percentage points.
(a) No IT data were available.
Table D.5
Table C.5
Changes in Philippines sector shares: services,
Decompositions of Philippines GDP growth over
manufacturing and IT
five-year intervals
Per cent per year
Per cent per year
Svc. Mnf. IT (a)
/g s /g
.
s/s s /g
.
s/s s /g
.
s/s
Time span g Agr. Ind. (Mnf.) Svc. IT (a)
197277 35.4 32.6 -1.6 25.9 24.7 -0.9
197277 19.6 29.0 38.4 (24.7) 32.6 197782 36.0 40.7 1.8 25.5 24.9 -0.3
197782 13.6 17.4 41.9 (24.9) 40.7 198287 39.5 6.4 1.9 24.8 27.1 -0.2
198287 -2.2 18.0 75.5 (27.1) 6.4 198792 43.3 51.7 1.9 25.0 23.4 -0.6 n.a.
198792 9.8 17.6 30.6 (23.4) 51.7 n.a.
A sector is expanding when its growth contribution /g exceeds its share s. The sector growth rate
.
In each row, the figures under Agr., Ind., and Svc. sum to 100 (subject to rounding error). s/s = [/g - s](g/s) is given per cent per year ratios (s, /g) are in percentage points.
(a) No IT data were available. (a) No IT data were available.
53
Bank of England Quarterly Bulletin: February 1997
increased sharply from 1977 to 1982 (40% to 50%), that Chart 2
figure has since remained roughly constant. Contribution of IT to GDP growth plotted against
1992 per capita GDP
Finally, turn to Pakistan and the Philippines.(1) The patterns OECD
Dynamic Asian Economies
of change here are less clear-cut. In Pakistan, services have Others Per cent
5
always contributed more than 40% to GDP growth, but
manufacturing continues to increase its share of GDP. 4
United States
Growth patterns show little stability in the Philippines, but
3
manufacturing has consistently declined.
2
United Kingdom
Chart 1
1
Contribution of services to GDP growth plotted +
against 1992 per capita GDP 0
OECD
_
Dynamic Asian Economics 1
Others Per cent
100 2
3
United Kingdom 0 5,000 10,000 15,000 20,000
80
Per capita GDP (1985 US $)
United States Chart 3
60
Contribution of manufacturing to GDP growth
plotted against 1992 per capita GDP
40 OECD
Dynamic Asian Economiess
Others Per cent
100
20
80
0
0 5,000 10,000 15,000 20,000
Per capita GDP (1985 US $)
60
I now expand the sample to include all the other OECD
economies, all Dynamic Asian Economies (DAEs), and a 40
United Kingdom
selection of others. For these countries, Charts 1 and 2 United States
plot the growth contributions of services and IT against
20
per capita incomes. For completeness Chart 3 also gives
the analogous plot for manufacturing's growth
contribution.(2) 0 5,000 10,000 15,000 20,000
0
Per capita GDP (1985 US $)
The figures yield a number of interesting conclusions. First,
for all the relations depicted here, the distinction between IT; only the relatively poorer economies continue to see
OECD and non-OECD membership does not seem to matter. significant contributions from manufacturing.
Once one allows for per capita GDP, the respective scatters
of OECD and non-OECD points are not out of line with each The third conclusion is that for growth, the services sector is
other. Although in 1994, 93% of the IT market of the most important in all advanced economies. In
US $431 billion was concentrated in the OECD area economies with per capita GDP of at least US $5,000, the
(with 80% in just the United States, Japan, Germany, services sector accounted for more than 40% of that
France, and the United Kingdom), that concentration might economy's growth performance. In 80% of economies
reflect only the pattern of income distribution across having per capita GDP of at least US $10,000, the
countries.(3) manufacturing sector contributes less than 20% of that
economy's growth performance.
Second, the slope of the scatter of points in Chart 1 and
Chart 2 is positive, while that in Chart 3 is negative. Across Finally, the United States is distinguished in having
the sample, richer countries are those that have higher experienced the highest contribution from IT to aggregate
contributions to economic growth from services and from growth.
(1) Fairly or unfairly, these are frequently singled out (by, among others, Lucas (1993) and the World Bank) for comparison with Korea and other
successful fast-growing economies.
(2) Per capita incomes are taken as the 1992 values of per capita GDP, in 1985 US dollars, calculated using a chain index (Summers and Heston (1991)
and updates).
(3) In Chart 2 the two highest points are the United States and New Zealand. The two points that show negative IT contribution are India (-2.4%) and
Finland (-0.5%). From 1987 to 1994, India actually experienced negative GDP growth, while IT growth was 11%. In this case, how to interpret my
measure of IT's contribution to growth is subtle, but I have decided to maintain the convention earlier described. Over this period, Finland showed
a slight negative decline in IT spending measured in current exchange rates, as used throughout this paper. Using purchasing power parity
corrections, this would have showed a slight increase instead. Either way, however, the magnitude of its contribution to growth is small.
54
Increasingly weightless economies
What do these estimates tell us about the United Kingdom's Third, increasing weightlessness and dematerialisation in
specialisation in services? Advanced economies all have economic growth take many different forms. To see
the services sector contributing the most to growth. The whether IT has become more important in overall economic
richer the economy, the more it relies on the services activity, it is far from ideal to look simply at the national
sector, and the less on industry, manufacturing, or income accounting distinction between manufacturing and
agriculture. services. IT involves elements of both, and looking at just
the standard classification categories can mislead. While
Conclusion already-developed economies like the United Kingdom and
the United States almost uniformly show continuing
This article has investigated growth facts on GDP's decline in manufacturing, the shift to the services account
increasing weightlessness--through dematerialisation in does not reveal whether IT is becoming more or less
general and IT in particular. important as a fraction of GDP: revisions to the standard
industrial classification categories might be called for,
Although the picture varies across countries, several eventually. New data are critical for further investigation.
generalisations are apparent. First, the services sector is the For the United Kingdom, considerable additional insight
most important in all advanced economies. In richer might result if finer statistical details on this split were
economies (those with per capita GDP of at least available.
US $5,000), the services contribution to growth is always at
least 40%. In almost all advanced economies it is services One overarching conclusion from this analysis is that the
which figure most prominently in growth. term `industrialised countries' no longer carries any
resonance: now, no advanced and growing country is
Second, while increasing dematerialisation matters, it is dependent on production industries. But, whether it is IT
much less clear that a great deal of that has, thus far, arisen and only IT that will subsequently be the main engine of
from IT. Successful economies like the United States growth is not yet apparent from the numbers. The
and Korea do show a rising emphasis on IT, but other United States leads the way, but even there IT has made
successful ones, like Singapore, do not--at least not only a 3% contribution to GDP growth, while the increase
dramatically. in IT share in GDP is, for the time being, no more than
2% per year.
Circumspection is called for in drawing implications from
this last observation. The United States and Korea might be One goal of this article was to stimulate discussion on the
leading the way for all other economies, while Singapore issue of changing industrial structures in economic growth.
might simply show the potential for greater future growth The financial sector is a large part of the weightless
and IT concentration. economy, and this article has said little about it.
Implications for financial and monetary policy, appropriate
The United Kingdom is one economy that has had its emphasis on the manufacturing sector, the importance of the
services sector both contributing strongly to GDP growth exchange rate for the geographical location of economic
and continuing to increase rapidly in share. But, here, the activity--all might follow from better understanding and
transition to an IT emphasis remains far from obvious. more precise measurement of the effects described above.
Again, this might just mean that the scope for high IT My calculations above made simplifying assumptions, and
growth in the United Kingdom remains correspondingly left open a number of other empirical issues that will lead to
large. a programme of further work.
55
Bank of England Quarterly Bulletin: February 1997
References
Lee, J (1996), `Do services temper business cycles? Implications of the rising service sector', Discussion Paper No 95963,
UC Irvine Economics Department, July.
Lucas Jr, R E (1993), `Making a miracle', Econometrica, 61 (2), pages 25171, March.
OECD (1996), Information Technology Outlook 1995, OECD Publications.
Quah, D T (1996), `The invisible hand and the weightless economy', Centre for Economic Performance,
Occasional Paper No 12, May.
Quah, D T (1996), `Growth and dematerialisation: why non-stick frying pans have lost the edge', CentrePiece:
The Magazine of Economic Performance, (3), pages 205, October.
Slavin, T (1996), `Virtual port of call', New Scientist, (2,012), pages 403, 15 June.
Summers, R and Heston, A (1991), `The Penn World Table (Mark 5): An expanded set of international comparisons,
195088', Quarterly Journal of Economics, 106 (2), pages 32768, May.
World Bank (1995), Workers in an integrating world, Oxford University Press.
56