China's Economic Growth 1978-2025

China’s Economic Growth 1978-2025:
What We Know Today about China’s Economic Growth Tomorrow




Views of the future China vary widely. While some believe that the collapse of China is inevitable,
others see the emergence of a new superpower that increasingly poses a threat to the U.S. This paper
examines the economic growth prospects of China over the next two decades.

Extrapolating past real GDP growth rates into the future, the size of the Chinese economy surpasses
that of the U.S. in purchasing power terms between 2012 and 2015; by 2025, China is likely to be the
world's largest economic power by almost any measure. The extrapolations are supported by two types of
considerations. First, China’s growth patterns of the past 25 years since the beginning of economic
reforms match well those identified by standard economic development and trade theories (structural
change, catching up, and factor price equalization). Second, decomposing China’s GDP growth into
growth of labor and other variables, the near-certain information available today about the quantity and
quality of Chinese laborers through 2015, if not several years after, allows inferences about future GDP
growth.

Short of some cataclysmic event, demographics alone suggests China’s continued economic rise. If
talent is randomly distributed in the world population and if agglomeration of talent is important, then the
odds are strongly in China’s favor.

JEL codes: O1 (O10, O11), O4 (O40, O47), O53, J11, O3, I21
Keywords: economic growth, growth accounting, growth forecasts, development theories, human capital
formation, education (all: China)





Carsten A. Holz
Social Science Division, Hong Kong University of Science & Technology
Clear Water Bay, Kowloon, Hong Kong
E-mail: socholz@ust.hk; Tel/Fax: +852 2719-8557





I am grateful for comments from seminar/ panel participants at the Center for East Asian Studies,
Stanford University (23 February 2004), the Association for Asian Studies meetings, San Diego (5 March
2004), the Universities Service Center for China Studies, Hong Kong (30 June 2005), the Western
Economic Association Meetings 2005, San Francisco (7 July 2005), and the Center on China’s
Transnational Relations, HKUST (30 September 2005). Thomas P. Lyons has provided suggestions at
various turns, and the paper has benefited from communication with Thomas G. Rawski on Chinese
statistics. A grant from the Research Grants Council of Hong Kong (Project no. HKUST 6122/03H)
provided financial support.


2 November 2005
(minor stylistic revision of 3 July 2005 version)

---

1. Introduction


The rapid economic growth of China since the beginning of the economic reforms in 1978
has captured the imagination of Western commentators and researchers. The responses range
from outright pessimism about China’s future to fear of a strong China. Lester Brown (1995)
wonders who will feed China. Gordon Chang (2000) announces the coming collapse of China.
Callum Henderson (1999) sees China as on the “brink,” while Ross Terrill (2003) writes of the
“illusory nature of the market in most of the Chinese economy” and that “a crash looms because
the Leninist core of the regime is unchanged from Mao’s construction of it in Yan’an six decades
ago” (pp. 329, 313). Nicholas Lardy (1998) stresses the large economic problems and the
unprecedented potential for social unrest due to ever more indebted state-owned enterprises, the
extent of nonperforming loans, and a decline in government revenue.


At the other end of the range are those who project a strong China. Geoffrey Murray (1998)
describes China as the next superpower. A number of authors view an all-powerful China as a
threat (Bill Gertz, 2000, or Edward Timperlake and William Triplett II, 2002).1 News items
imploring the “devastation Chinese competitors are inflicting on U.S. industries, from
kitchenware and car tires to electronic circuit boards” and the “futility of trying to match the
China price” have become common fare.2


What is undeniable is China’s rapid economic growth over the past 25 years since the
beginning of economic reforms in 1978, of, measured in gross domestic product (GDP), on
average 9.37% per year. In economic size China is surpassed today only by the U.S., Japan,
Germany, and France.3 Its share in global growth 1995-2002 has been estimated at 25%,
compared to 20% for the U.S.4


While China’s economic growth has received much attention in the popular literature, China
researchers, when it comes to economic growth, tend to focus on narrow topics, measuring total
factor productivity growth in enterprises in different ownership forms or investigating the growth
prospects of individual industries or firms. For example, Douglas Allen (2002) documents the
transformation in four industries in China and predicts the imminent emergence of world-class
Chinese brands. Ming Zeng and Peter Williamson (2003) examine the international growth
prospects of Chinese firms (positive). Colin Carter and Scott Rozelle (2001) ask if China will
become a major force in world food markets (yes, if structural transformation occurs).


A number of studies try to explain China’s past economic growth. Much of the transition
literature explains past reform successes—usually, if only implicitly, measured as economic
growth—through institutional changes (for example, Qian Yingyi, 2000, 2003, or Wing Thye
Woo, 1994, 1999). Justin Lin, Cai Fang, and Li Zhou (2003) explain China’s reform success
with adoption of a “comparative-advantage following strategy,” allowing full play to China’s
plentiful endowment with labor. Others analyze China’s past economic growth within the
aggregate production function framework (Wang Yan and Yao Yudong, 2003; Wu Yanrui, 2004,
Alwyn Young, 2000). The aggregate production function framework is also occasionally
extended into the future. Gregory Chow and Kui-Wai Li (2002) estimate an economy-wide,
aggregate production function for the years 1952-98 and obtain GDP values for the years through
2010 by extending variations of past factor input growth rates into the future; Gregory Chow
1

---

(2002) further includes an estimate of year 2020 GDP. A 1997 World Bank policy analysis of
current economic issues is supplemented by an estimate of China’s GDP in 2020 based on a
closed-economy model. A systematic study of future economic growth in China that makes full
use of the hard facts already available today, however, is still lacking.


Answers to the question of what we know today about the economic size of China ten or
twenty years from now are relevant not only for the popular discourse on the new superpower
and for China threat theorists and military strategists, but also for economists. For example,
conventional economic wisdom holds that free trade benefits, by the laws of comparative
advantage, all countries in the long run. Thus, a growing China, and growing trade between
China and the U.S., benefits the U.S. But as Paul Samuelson (2004) within the framework of
standard trade theory shows, a productivity gain in one country alone may, in fact, lead to a
permanent loss in per capita real income in the other country; references to China abound.


This paper examines China’s future economic growth in three steps. First, it conducts a
straightforward extrapolation of a stable, past growth trend into the future. A number of
scenarios are played through in a comparison with the U.S. But extrapolation is not a particularly
convincing research tool. Why should the past continue into the future?


One argument why past economic growth may continue into the future is that China’s
economic growth matches standard growth patterns identified by theories of economic
development and trade. These are structural change, catching up, and factor price equalization.
China’s past economic growth fits well with all three. Furthermore, China’s reform period
growth, within these three analytical frameworks, matches those of Japan, Korea, and Taiwan at
an earlier stage of their development. Obviously, these four countries differ, as do the domestic
and international circumstances under which they experienced a particular stage of development.
But the differences need not be systematic with respect to the impact on the dependent variable
economic growth. These growth patterns may well apply to China in the future, just as they
already have to Japan, Korea, and Taiwan in the past.


The final step is to decompose GDP growth. Growth accounting in the production function
framework turns out to not be useful due to the lack of a stable aggregate production function
over time. An alternative is to decompose GDP growth following the income approach to the
calculation of GDP into growth of labor and of other variables. We already know today with
near-certainty the size of China’s working age population through 2015, and can project with
high reliability for several years after. In as far as gains in high school and university enrollment,
most of them very recent, are unlikely to be reversed, a bottom line scenario on the quality of
China’s future labor force can also be developed. Future GDP growth can be re-composed using
solely the hard facts about the future quantity and quality of labor available today.


Demographics in form of the quantity and quality of labor is enough to drive continued
economic growth in China. Demographics means, for example, that for every “genius” in the
U.S., China potentially has four. If talent is randomly distributed in the world population and if
China’s education system is able to identify the brightest students, then China has a larger pool
of talent to draw from than any other country in the world. If growth and innovation depend on
the agglomeration of talent (geographically, nationally, culturally, or linguistically), then China
2

---

is in an excellent position to continue to grow and innovate. The recent explosion in tertiary level
education may yet supply the productivity gains that Paul Samuelson identifies as potential
causal factors for a permanent loss through trade in U.S. per capita real income.


Projections cannot predict the future with certainty. Potential complications range from a
collapse of Chinese Communist Party rule to war in the Taiwan Straits; China researchers
ruminate about such issues as the bad loan problem in China’s banking system, loss-making
state-owned enterprises, pilfered pension funds, local government budget deficits, rural-urban
inequality, and environmental degradation. This paper works within these limitations. It does not
attempt to provide an exhaustive list of complications, to examine the individual issues, to
quantify their potential impact, and to attribute a likelihood to their occurrence. It predicts
China’s future economic growth under the assumption that current and future problems continue
to be resolved as they arise and that events of catastrophic dimensions do not occur.5


2. Extrapolation of Past Growth into the Future


China’s year 2002 nominal GDP of 10.48 trillion RMB at the official exchange rate of
approximately 8.277 RMB per USD translates into USD 1.27 trillion, compared to year 2002
U.S. GDP of 10.49 trillion.6 The U.S. economy in 2002, thus, was eight times larger than
China’s economy; by 2004, following preliminary figures, the difference was seven-fold.7
However, such a comparison neglects to take into account the differing price levels in the two
countries. The Penn World Tables (PWT), which report comparable GDP data in “international
dollars,” assume that the general price level in China in 2000 was equivalent to only 23.13782%
of the U.S. price level. The U.S. economy then, today, in purchasing power terms, is less than
twice as large as China’s economy.


Table 1 in the first row extrapolates PWT (version 6.1) data into the future. The PWT do not
report economy-wide GDP. These values can, however, be obtained, by multiplying per capita
GDP in international dollars (and constant 1996 prices, using the chain method) by the
population numbers. Extrapolation into the future is based on year 2000 economy-wide GDP (in
international dollars, at constant 1996 prices) because 2000 is the most recent year for which
PWT data are available. The annual real GDP growth rate used to extrapolate into the future is
the geometric mean of the period 1978-2000; 1978 is chosen as starting year since it marks the
beginning of economic reforms in China. If China and the U.S. after 2000 each continue to grow
with the same average annual growth rate they enjoyed between 1978 and 2000, then, as Table 1
reports, economy-wide GDP (in international dollars, at constant 1996 prices) of China surpasses
that of the U.S. starting in 2015, a decade from now.


But the manipulations of official Chinese data in the PWT may not be correct.8 The second
scenario in Table 1 therefore uses the PWT year 2000 per capita GDP values (in international
dollars, at constant 1996 prices) times population—accepting for the time being the PWT
adjustments to Chinese data in levels—but then applies the average annual real GDP growth rate
of the years 1978-2000 as implicit in the official data published by the Chinese National Bureau
of Statistics (NBS) and the U.S. Bureau of Economic Analysis to the PWT GDP values.9 Using
the official national data to obtain the average annual real GDP growth rate yields a larger
3

---

growth rate discrepancy in favor of China. Consequently, China’s aggregate purchasing power
parity GDP begins to exceed that of the U.S. less than a decade from now, namely in 2012.


The third scenario in Table 1 abandons both the PWT real growth rates and PWT GDP
values. Instead, it uses national data (for details see notes to table). These are adjusted upward in
the case of China by the PWT’s purchasing power parity factor of 4.321928 (1 divided by
China’s price level of 23.13782% of the U.S. price level). The effect is almost the same as in the
previous case. China’s aggregate purchasing power parity GDP now begins to exceed that of the
U.S. in 2013 instead of 2012.10 The fourth scenario, using national base year data of the year
2002, the latest year for which final (revised) GDP values are available, and national average
annual real growth rates of 1978-2002, yields a cross-over year of 2013. Assuming lower
average annual real growth rates for China of 8% or even 7%, while maintaining that of the U.S.
at 3%, the cross-over year moves further into the future to 2016 and 2020 (fifth and sixth row).


The price adjustment factor in the case of China carries “substantial uncertainty” (Alan
Heston, 2001, p. 1).11 The multiplicative factor of 4.321928 could be significantly off (in either
direction). The seventh row in Table 1 reports the results if a smaller factor of 3 is used; in
comparison to the PWT, this means underestimating China’s purchasing power parity GDP. In
this case, the cross-over year is 2019. If the adjustment factor were to fall gradually over time—
in the PWT it stayed virtually constant throughout the 1990s—assume from a value of 4 in 2002
to a value of 1 (no adjustment) in 30 years’ time, the cross-over year moves far into the future,
namely to 2038 (eighth row). Since 2038 is more than 30 years into the future, this is the same as
making no purchasing power adjustments at all (ninth row). Dropping the purchasing power
concept and the PWT data altogether, but assuming a 3% appreciation in the official exchange
rate every year after 2002 results in a cross-over year of 2026 (tenth row). In this scenario, the
year 2002 exchange rate of 8.2770 yuan RMB/USD falls by approximately half to 4.0717 in
2026. Including, with the appreciation, some kind of price adjustment, which, even if only at a
level of 2, currently makes eminent sense, would bring the cross-over year forward to 2019.


The results reported so far covered aggregate nationwide GDP, with China overtaking the
U.S. as early as 2012 in the scenario most (and not unrealistically) favorable to China and as late
as 2038 in a highly unrealistic scenario.


In terms of per capita GDP, the year when China surpasses the U.S. is far into the future (see
additional columns in Table 1). But what is far into the future for the average Chinese person
moves closer to the present for the richer areas of China. Little more than two decades from now,
coastal China (as defined in the notes to Table 1), accounting for 41.54% of China’s population
in 2000 and exceeding the U.S. population by 92.04%, may well be as rich, per capita, as the
average U.S. citizen. Focusing on the five fastest-growing provinces in coastal China only, plus
Shanghai, with a joint population in 2000 that is 27.14% larger than the U.S. population, moves
the cross-over year in per capita GDP a few years further to the present. In other words, by the
mid-2020s, a share of China’s population that exceeds the size of the U.S. population could
enjoy the same standard of living as the U.S.; in addition, China has a per capita poorer
hinterland with roughly three times the population in the richer areas.12

4

---

In these calculations, all variables were held constant except those explicitly allowed to vary.
In particular, in the extrapolation of aggregate GDP, the use of an average annual past real
growth rate reflects the product of the real GDP growth rate per capita and the population growth
rate. In the short run, such as one decade, changes in the population growth rate are likely to be
minor, but this may not hold in the long run. Similarly, the exchange rate was held constant
except in the last two scenarios, as was the price adjustment factor except in the eighth through
eleventh scenarios. The further into the future, the less valid are the assumptions. A more
sophisticated approach would, furthermore, focus on the employed instead of the population.13
But none of these adjustments is likely to have a significant impact on the determination of
cross-over years which are as little as one decade away.


Extrapolations do not constitute a reliable research tool. Factors that promoted economic
growth in the past may be absent in the future, or factors constraining economic growth may
become newly relevant in the future. The next section examines explanations of economic
growth.


3. Development Theories and Asian Precedents


Explanations of China’s economic growth tend to focus on economic transition. The
“success” of the reform process is explained by transition facts and strategies, where success is
usually taken to imply economic growth (or a rise in living standards). For example, Wing-Thye
Woo (1994) lists as crucial the creation of non-state firms in every sector of the economy, a high
savings rate, good initial conditions (such as a small extent of central planning, unemployment in
the countryside that could be taken up by township and village enterprises, or a limited social
security net), historical conditions, and the Chinese Diaspora. Qian Yingyi (2000, 2003) ascribes
much of China’s reform success to the unorthodox economic policy measures adopted by
China’s leadership; the key to China’s economic growth was the unleashing of incentives and
competition while making reform interest-compatible for those in power.


Yet these explanations of past economic reform successes (and thus economic growth) face
the problem of lack of a counterfactual. Wing-Thye Woo offers cross-country comparisons, but
the number of explanatory variables appears larger than the number of comparison countries.
Qian Yingyi makes an in part historical argument, but offers no explicit time series evaluation
that examines the status before and after a particular reform measure was implemented. These
explanations, thus, are not as strong as one might wish them to be.14


If one claims that economic transition “in total” has caused China’s past economic growth,
then one could argue that key elements of transition have been in place since the early 1990s
(price and domestic trade liberalization, the entry of private enterprises) and that therefore the
gains from transition have already been exhausted. Consequently, economic growth should since
have slowed, which it didn’t, or it could slow any time now. Alternatively, one could argue that
past transition measures impact on economic growth over an extended period of time, or that
transition is as yet incomplete, with further measures to go, in which cases the gains continue.

5

---

Growth patterns identified by economic theories of development and trade perhaps offer
firmer ground. Economic transition could then be viewed as the removal of constraints that
prevented well-known development patterns from unfolding.15 Furthermore, if China’s reform
period growth patterns match those exhibited by other countries at stages in their economic
development similar to China’s development level in the reform period, then China’s future
growth patterns may also match those of the other economies later. The argument has two
foundations: one is a focus on standard growth patterns, the other is a cross-country comparison
with countries with which a comparison is likely to be meaningful.


The growth patterns are structural change, catching up, and factor price equalization.16 These
three patterns are not independent of each other, nor do they hold irrespective of the larger
economic environment. They represent uni-causal explanations of economic growth that have the
advantage that they do not rely on individual transition measures and have been identified by
development economics and trade theory as of relevance. At the same time, in as far as they
reveal China to be at the very early stages of accepted development patterns, they support the
growth projections into the future.


The comparison countries are Japan, Korea, and Taiwan. These three countries are relevant
for China only if the assumption of constant effect among the four countries is met: different
countries do not differ systematically with respect to the impact of the explanatory variable
(specific to each growth pattern) on the dependent variable (economic growth). While these four
countries differ, as do the domestic and international circumstances under which they
experienced a particular stage of development, this does not necessarily invalidate the
assumption of constant effect.


The choice of countries to compare China to is a subjective choice, motivated only by the
desire to make comparisons across a relatively homogeneous group of countries. China may
share some economic growth patterns with Japan, Korea, and Taiwan due to cultural similarities,
geographic location, similar economic development strategies, or, in the case of Japan, relatively
large size of the domestic economy. Limiting the analysis to Japan, Korea, and Taiwan allows
the careful compilation of the necessary data from each individual country’s statistical office,
with a very few holes filled using the World Bank Development Indicators database, the
International Financial Statistics (IFS), and the PWT.17 (The individual data sources are
documented in Appendix 1.) An attempt was made to cover the years 1960 through 2002, with,
in some instances, data available only beginning around 1970. Ideally, data for the 1950s should
also have been included but are not available. The earlier in the development process of Japan,
Korea, and Taiwan, the closer to the case of China the initial conditions might have been. Data
on China are for the years of the reform period (the years since 1978).


Throughout, the variable to be explained is real GDP growth per laborer since GDP is
produced only by the active working population, namely laborers (one PWT figure, Figure 6, by
necessity uses per capita figures). Employment data are midyear data when related to a variable
that measures an annual flow, and end-year data otherwise (and then noted with the figure). Each
data point in a figure is a particular country in a particular year.18


6

---

a. Structural
change

As labor shifts from low-productivity agriculture to higher-productivity industry and services,
economy-wide real GDP per laborer, i.e., (partial) labor productivity, increases, if only because
those laborers who have shifted sectors now produce a multiple of their former output value.
Figure 1 shows just how big, and increasing, the labor productivity differences are between the
three economic sectors in China. One would expect to see relatively high aggregate (economy-
wide) labor productivity growth in those years when a relatively large number of laborers shifts
out of agriculture.


Figure 2 confirms the expectations. In years with a high absolute reduction in the share of
laborers in agriculture, the growth rate of (real, likewise below) labor productivity is high. This
pattern holds equally for all four countries. Among Japan, Korea, and Taiwan, the shift out of
agriculture was, on average, fastest in Korea, followed by Taiwan and then Japan (Figure 3); this
matches the initial shares of laborers in agriculture, with the highest one in Korea and the lowest
one in Japan. One would consequently expect China, with an extremely high share of laborers in
agriculture in 1978, to exhibit rapid reductions in this share, comparable perhaps to Korea, but
the decline is more gradual. However, if in the official employment statistics the primary sector
were obtained as residual, it could include an over time increasingly undercounted migrant
population. The share of the primary sector in employment of, in 2002, approximately 50%,
would then be an overestimate.


At China’s 1978-2002 rate of decline, with an annual reduction in the share of laborers in
agriculture by approximately an absolute value of 0.01 every year (thus, for example, from 0.7 to
0.69 in one year), China has another forty years to go before its agricultural labor share reaches
the level of just below 10% at which Japan, Korea, and Taiwan appear to bottom out. Even if
China’s primary sector employment of 2002 were somewhat overestimated, China still faces
another two to three decades of labor transfers from the primary sector to the secondary and
tertiary sector. But this implies that structural change as a source of economic growth has up to
twenty to forty more years to contribute to labor productivity growth in China. Furthermore, in
China, a given shift out of agriculture comes with higher labor productivity growth than in any of
the other three countries. In Figure 2, the observations for China tend to lie above those of the
other three countries.


b. Catching
up


Catching up means that production techniques and technologies that have already been
invented and implemented can be copied rather than need to be re-invented; technology transfer
can also happen through the import of foreign equipment, possibly as part of foreign direct
investment. Taking the U.S. as the leader in research and development, and proxying the level of
technological development by labor productivity, the distance between a particular’s country
labor productivity (in USD) and U.S. labor productivity serves as a measure of the potential
scope for catching up. One would expect to see relatively high real GDP growth per laborer in
those cases where the distance to the leading country (the U.S.) is relatively high, i.e., where the
potential for catching up is large.19
7

---

Figure 4 confirms the hypothesis for Japan. Japan experienced its highest growth rates when
its labor productivity was only ten to fifty percent of U.S. labor productivity. As the gap closed,
growth rates in Japan fell (the trend curve shows a slight upswing at the highest labor
productivity levels that appears to be an artifact of the imposed second-degree polynomial).
Korea and Taiwan exhibit constantly high growth rates of labor productivity at around 4% to 6%
throughout all years (Figure 4). But these two countries’ level of economic development is in the
range of 5% to 50% of the U.S. level, which is a range in which Japan also exhibited high and
constant labor productivity growth rates. Korea and Taiwan may yet have to experience the
slow-down that comes when the potential gains from catching up are exhausted. China also
exhibits a downward pattern, but at a much lower level of economic development. Figure 5
shows that China’s labor productivity between 1978 and 2002 was only 1.2 to 2.4 % of that of
the U.S. (using the official exchange rate to translate Chinese yuan RMB values into USD
values). This seems almost too narrow a band to determine long-term trends. As/if more
observations become available at higher labor productivity levels relative to the U.S., the
negative slope could well disappear or turn into a positive one. Whichever direction future
observations are heading, at China’s highest level of catching up in the past, the growth rate of
labor productivity was still at a relatively high 6-7%.


Figure 6 replicates Figure 4 using PWT data. PWT data adjust for differing price levels
across countries and may also, as in the case of China, undertake further adjustments to official
data. Data are available for the 1950s through 1990s and are per capita (rather than per laborer);
the observations for China are as always limited to the years since 1978. The pattern now is one
of first rising labor productivity at low development levels before gradually falling off. This
might reflect initial opening up effects as barriers to foreign direct investment and imports are
removed and access to foreign knowledge increases. The Chinese observations are again
concentrated in a very narrow band, with a negative slope.


Independent of the choice of data, China is at a very low development level compared to the
U.S. It appears to be at a stage of economic development (labor productivity) where other Asian
countries started out more than 30 years ago. While labor productivity growth appears to fall as
China catches up with the technology leader (and still is at a very high level), the scale of
catching up is so small that such factors as exchange rate effects or uncertainties about the scale
of the purchasing power adjustment could well render the slope coefficient insignificant. In the
long run, as (if) China catches up with the U.S. and observations at higher development levels
become available, the negative slope might yet turn into the positive slope that the other three
countries exhibited at their earliest stage of economic development.20


c. Factor price equalization


The factor price equalization theorem (or Heckscher-Ohlin-Samuelson theorem) states that
factor prices, such as skill-specific wages, should equalize between two countries as long as a
range of assumptions are met. I.e., the skill-specific wage rate of one country divided by that of
the other country should equal unity.21 The slightly less restrictive, relative version of the factor
price equalization theorem focuses on the relative prices of factors of production. Thus, for
8

---

example, a country with labor that is cheap relative to capital should see demand for its labor rise.
As underemployed laborers become fully employed, labor productivity rises. An increase in the
demand for labor may also be accompanied by wage rises, which in turn are likely to be
accompanied by labor productivity growth.22 In the absence of reliable prices of capital, the price
of capital is approximated here by the price of investment goods. The price of investment goods
of any particular country relative to the U.S. is available in the PWT.



Figure 7 reveals an inverse U-shaped pattern for all four countries. In other words, labor
productivity growth first rises along with relative wage increases and then gradually falls as the
relative return to factors of production in each country, vs. the U.S., approaches unity. In the case
of China, also reproduced at a larger scale in Figure 8, the inverse U-shaped pattern occurs at an
extremely low level, with wage rates in China, relative to investment prices, at only 2-5% of the
corresponding U.S. ratio. This is well below the early levels of the other three Asian countries.
Labor productivity growth in China is still high in the declining part of the inverse U-shaped
pattern at 6-7%. Again, as more observations become available over time, the inverse U-shaped
pattern may yet change. If this were just the beginning of a long-run trajectory, then the patterns
of the other three Asian countries suggests that China’s potential for economic growth from
relatively low labor costs will continue to exist for another thirty years.23


This section focused on three explanations of economic growth from development economics
and trade theory and applied them to the case of China. China’s past economic growth matches
these standard development patterns, as does economic growth in Japan, Korea, and Taiwan. In
as far as China is located at the early stages of each pattern, in comparison to Japan, Korea, and
Taiwan, there remains much scope for future gains in labor productivity and therefore growth.


4. Growth
Accounting


Growth accounting decomposes GDP growth into growth of variables other than GDP. I.e.,
growth of GDP is “explained” by the weighted growth rates of other variables. One need not
view these other variables as constituting a satisfactory final explanation of GDP growth; for the
purpose of forecasting future economic growth, regularity in the relationship with output
suffices.24 If GDP growth and the growth of the variables into which it is decomposed exhibited
a stable relationship in the past, and if this stable relationship is likely to continue into the future,
then information about the future values of the variables into which GDP has been decomposed
allows derivation of future GDP growth. One growth decomposition is the traditional
decomposition of GDP growth into growth of different factor inputs. A second growth
decomposition follows the income approach to the calculation of GDP.


a. Decomposition of GDP growth into growth of factor inputs


The traditional approach to growth accounting decomposes economic growth into growth of
the factor inputs labor, capital, and “everything else” (also “technological progress,” or growth in
total factor productivity, TFP). The traditional growth accounting equation is

9

---