27
1 Pattern and Sustainability of China’s Economic Growth towards 2020 * Xiaolu Wang, Gang Fan, and Peng Liu I. Introduction In this paper, we attempt to examine evidence and influential factors of changing economic growth pattern in China, and examine future growth sustainability towards 2020. China has maintained a high economic growth rate for nearly three decades, since the beginning of its economic reform in 1978. The average GDP growth rate during the period of 1979-2000 was 9.2%, and further accelerated to 10.1% from 2001 to 2006. GDP in real term by the end of 2006 has expanded to 13 times of that in 1978, reached US$ 2.62 trillion. However, economic growth in China has been described as “unsustainable” by economists (e.g., Paul Krugman, 1994) or “extensive pattern of growth” by Chinese government leaders (Premier Wen Jiabao, 2006), both meaning growth driven by massive input with limited technological progress or total factor productivity (TFP) growth. This pattern of economic growth in China can be characterized by very high rates of savings and investment, massive transfer of unskilled labor from the agricultural to urban non-agricultural sectors (which was the main source of TFP growth at the economy level), cheap labor cost, low level of labor education, low level of technical innovation, large income inequality, heavy dependency on external demand, inefficient energy consumption, heavy environment pollution, and so on. * This paper is an outcome of the research project “Pattern and Sustainability of China’s Economic Growth”. The authors thank Mastcard for its financial support.

Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

  • Upload
    hadang

  • View
    216

  • Download
    1

Embed Size (px)

Citation preview

Page 1: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

1

Pattern and Sustainability of China’s Economic Growth

towards 2020∗∗∗∗

Xiaolu Wang, Gang Fan, and Peng Liu

I. Introduction

In this paper, we attempt to examine evidence and influential factors of changing

economic growth pattern in China, and examine future growth sustainability towards

2020.

China has maintained a high economic growth rate for nearly three decades, since the

beginning of its economic reform in 1978. The average GDP growth rate during the

period of 1979-2000 was 9.2%, and further accelerated to 10.1% from 2001 to 2006.

GDP in real term by the end of 2006 has expanded to 13 times of that in 1978, reached

US$ 2.62 trillion.

However, economic growth in China has been described as “unsustainable” by

economists (e.g., Paul Krugman, 1994) or “extensive pattern of growth” by Chinese

government leaders (Premier Wen Jiabao, 2006), both meaning growth driven by massive

input with limited technological progress or total factor productivity (TFP) growth.

This pattern of economic growth in China can be characterized by very high rates of

savings and investment, massive transfer of unskilled labor from the agricultural to urban

non-agricultural sectors (which was the main source of TFP growth at the economy level),

cheap labor cost, low level of labor education, low level of technical innovation, large

income inequality, heavy dependency on external demand, inefficient energy

consumption, heavy environment pollution, and so on.

∗ This paper is an outcome of the research project “Pattern and Sustainability of China’s Economic Growth”. The authors thank Mastcard for its financial support.

Page 2: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

2

Nevertheless, some sighs of changing growth pattern was observed in recent years.

Demand for skilled labor and professional workers is increasing, and supply of unskilled

rural labor in some source regions is exhausted, both causing increases in labor costs.

Workers’ education level is getting higher. There are shifts towards more capital

intensive production. Some authors suggest that China’s economy development is now at

the turning point (see e.g., Garnaut, 2006).

Investment in R&D is still low, but its ratio to GDP doubled in the past decade, and is 1.3

percent in 2005. Both values of annual technical transactions in markets (in real term) and

granted patents nearly quintupled during the same period. Some studies have identified

higher growth of TFP or capital productivity at either the national or firm levels in recent

years (see, e.g., Jefferson, Rawski and Zhang, 2007).

Evidence shows that large, and increasing, income inequality is a main reason for relative

weak domestic consumption growth and high savings rate. Some new policies has been

adopted since 2004 to encourage rural income and consumption growth and to reduce

poverty, such as abolition of agricultural tax, exemption of school fees in rural nine-year

education, and improvement in rural and urban social security systems.

It will be interesting to find out whether or how growth pattern in China is changing, and

what will be the impact on growth in the future. To do so, we carry out empirical tests in

this paper to examine possible effect of a number of potentially influential factors on

growth, and use these results to forecast future growth in China towards 2020.

In 2006, some modified historical GDP data, based on the new economic census, were

published by the National Bureau of Statistics (NBS, 2006). GDP in 2004 were upward

modified by 16.8%. GDP and GDP growth rate in earlier period were also adjusted

accordingly. This enables us to get a more accurate growth accounting result.

In section II, we review the roles of a number of influential factors in economic growth

and their recent changes. Section III specifies a growth model for empirical tests and

reports the results. Section IV carries out growth accounting to calculate contribution of

factors to growth, and to forecast future economic growth towards 2020. Finally, Section

V is the conclusion.

Page 3: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

3

II. What Contributed to Growth and Do These Show Any Chang in Growth Pattern?

Capital stock

There has been an increasing rate of capital formation in the past, resulting from high

savings and massive capital inflow. The average rate of capital formation rose from

35.2% of GDP in the 1980s to 40.2% in 2001-2005. Our calculated growth rate of fixed

capital stock was 9.2% in average for the pre-reform period of 1952-1978, 10.6% from

1979 to 1998, and 12.8% from 1999 to 2005. In 2005-06 it was 15.6%. Rapid capital

formation has played a major role in China’s economic growth. Because the savings rate

remains high, capital growth rate is unlikely to slow down in the near future. We assume

an annual rate of 15% in 2006-2010, and 14% in 2011-2020. Improvement in social

security systems and reduction in income inequality can help to resume consumption and

to moderate the rate of capital formation, but this depends on the government policies.

Details of our calculation of capital stock, human capital stock and other variables can be

seen in the notes of Table A1 in the Appendix.

Labor and Human capital

The increasing supply of unskilled cheap labor to industrial and service sector in the past

two to three decades is an important source of China’s rapid economic growth. However,

there has been some evidence showing that the trend is slowing down, and the labor cost

is increasing. Meanwhile, the roles of human capital in economic growth are getting

more important. While workers’ education level is increasing, skilled labor and

professional workers are in short supply.

Following Lucas (1988), human capital in this study is defined as effective labor that

enhanced by their year of schooling. According to our calculation, average year of

schooling of labor force is increasing. It was 6.1 years in 1980, 7.1 years in 1990, 8.6 in

2000, and 9.2 in 2005 (calculated from NBS data, various years; same below for un-

sourced data). However, average growth rate of human capital stock was diminishing,

7.3% in 1979-1988, 3.2% in 1989-1998, and 2.5% in 1999-2005. This was mainly caused

by diminishes of labor growth rate, as the result of one-child policy. We expect fast

increases in workers’ year of schooling in coming years due to expansion of tertiary and

secondary vocational education, and improvement in rural nine-year education. Therefore

Page 4: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

4

the current growth rate of human capital (around 2.6%) in 2006-2010 can be continued,

and it may be slightly lowered to around 2% in 2011-2020.

Marketzation

Many evidences indicate that high economic growth during the reform period was

benefited from the market oriented institutional changes, especially development of the

non-state enterprise sector (private sector). Due to lack of overall historical data, shares of

the non-state enterprise sector in industrial output are used in this study as a proxy of

marketization. Data between 1996 and 1998 are found overstated, and data between 1999

and 2003 are not completed (small enterprises with annual sales below five million Yuan

were missing). They are therefore modified according to data from the two industrial

censuses in 1995 and 2004. According to the modified data, the shares of non-state

enterprises were 22.4% in 1978, 43.2% in 1988, 63.8% in 1998, and 69.2% in 2005.

Because many SOEs have already been privatized, further increase in the non-state share

will be minor. We assume a three percentage point increase by year 2010, and another

four percentage point increase by year 2020.

Urbanization

As one of the most important result of market oriented reform, urbanization was

accelerated during the reform period, especially since the 1990s. The urbanization ratio,

that is, urban share in total population, increased from 18% in 1978 to 43% in 2005,

roughly meaning 300 million rural residents have transferred to urban areas. This has

sustained increasing labor supply to the rapid growing industrial and services sectors.

Meanwhile, reallocation of labor from sectors with low productivity to those with higher

productivity became a major source of China’s productivity growth. In the past five years,

the urbanization ratio increased at 1.4 percentage point per year. We expect same speed

of increasing urbanization ratio from 2006 to 2010 to reach 50%, and possibly another ten

percentage point or higher increase from 2011 to 2020 to reach at least 60%.

Trade

One of the most significant characteristics of China’s past growth pattern was export

orientation. The trade dependency ratio, i.e., total value of exports and imports as a

proportion of GDP, increased dramatically from 9.7% in 1978 to 63.9% in 2005.

Economic theories state that trade contribute to productivity growth via exploitation of an

Page 5: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

5

economy’s comparative advantages, may lead technological transfers among countries,

and lead to efficiency increases via international competition. However, given the high

achievement, we expect a slower increase in this ratio by annually two percentage points

in 2006-2010, and then remain unchanged in 2011-2020.

Foreign direct investment

FDI has been another source of capital formation in China. It remains at the level of 50-

60 billion US dollars annually in recent years, making China one of the largest FDI

recipient countries in the world. However, the growth rate of foreign capital diminished

from above 30% in the 1990s to only 6% in 2005. In addition, compared with the huge

domestic investment, FDI only accounts for a small part in China’s capital formation. Our

calculation shows a nine percent or lower foreign share in total fixed capital stock in

20051. In this study, we assume an average 0.5 percentage point decrease of foreign share

in total fixed capital stock between 2006 and 2010, and then a 0.3 percentage point

decrease between 2011 and 2020.

Foreign capital may contribute to TFP growth if it had a higher productivity than

domestic capital, or transferred new technology to the economy, although an earlier study

using foreign shares in total investment in fixed assets did not find significant

contribution of FDI to TFP (Wang, 2006). To further test these effects, we calculate

foreign capital stock and use the foreign share in total capital stock in the model. A

positive and significant estimate indicates a higher productivity of foreign capital than

total capital, thus a contribution to TFP growth.

Infrastructure

Better infrastructure makes more efficient use of other factors. In the past decade, there

was a rapid improvement in infrastructure conditions, especially the highway system. The

length of highway increased from 1157 thousand (1995) to 1930 thousand km (2005).

The quality of the road system also much improved. Of the total length of highway,

freeway increased from 2 to 41 thousand km. To make the data comparable, we

converted different grade of road length into a grade II equivalent highway length

according to the road capacity of transport volumes. In this study, it is called standard

1 There should be a deduction of the part invested for working capital from total FDI. However, data are unavailable. Therefore the share of nine percent is an upper limit.

Page 6: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

6

highway length. The standard highway length increased dramatically in the past decade,

from 275 thousand (1995) to 830 thousand km (2005).

Infrastructure investment forms part of total capital stock, and contribute to economic

growth via growth of capital. Meanwhile, improvement in infrastructure may also

generate externalities to the economy and therefore contribute to TFP growth. To test this

effect, we calculated a ratio of standard highway to population. This ratio was 23 km per

thousand persons in 1995, 44 km in 2000, and 64 km in 2005. In the future, growth of

natural length of highway will slow down, but quality improvement in the road system

will continue. We assume another 15 km increase in this ratio in 2006-2010, and 20 km

further increase in 2011-2020.

Research, development and technical innovation

R&D in China was government-leading and generally at a low level in the earlier stage

even during the reform period. No significant contribution of R&D on growth was found.

However, the recent trend shows increases in fund raising for science and technology

activities, mainly led by enterprises. The funds raised by enterprise accounted for 44% of

the total in 1995, increased to 66% in 2005. During the same period, R&D expenses as a

proportion to GDP increased from 0.60% to 1.34%, and total patents granted increased

from 45 to 214 thousand items per year (data are from NBS, 2006 and 2005b). The

increase in enterprise spending is a clear evidence for positive market returns to R&D.

One can expect that the ratio of R&D expenses to GDP will continue to increase in the

future until achieve a reasonable level, possibly around 3%. This implies an accelerated

growth of R&D expenses until the middle of 2020s.

Structural bias

Although high saving and investment have contributed largely to economic growth, the

continued decreases in the share of final consumption in GDP have drawn much concern

from economists and the government. From 1980 to 2005, the share of final consumption

in GDP dropped from 65.5% to 51.9%; and the share of private consumption dropped

from 50.8% to 38.0%. A major drop of final consumption, by ten percentage points,

occurred in recent years from 2001 to 2005.

Due to relative weak consumption growth, the rapid increasing production capacity relies

more and more on investment demand and net export to be utilized, and leading more

Page 7: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

7

trade disputes with other countries. Net export is increasing, accounting for 6.7% of GDP

in 2006. Foreign exchange reserves, resulted from both trade surplus and capital inflow,

mounting up to 1.066 trillion US dollars by the end of 2006; a large part of which was

used to buy US treasure bonds. Balance of bank deposit is huger, reached 33.54 trillion

Chinese Yuan, or 1.6 times of GDP by the end of 2006. The ratio of total bank loans to

bank deposit dropped from 93.8% in 1995 to 67.8% in 2005. All these indicates

inefficient uses of resources.

A number of factors are responsible for weaker consumption growth. Firstly, income

inequality is getting larger. The Gini coefficient was around 0.30 in early 1980s, then

0.45 in 2001. Larger income gap caused more savings and less consumption because the

saving rates are uneven between rich and poor. Secondly, public services on education,

healthcare and housing were largely withdrawn during the reform period, and social

security systems are incomplete, causing heavy burden and future uncertainties to, and

forced savings of, middle and low income people. Finally, due to incompletion of the

taxation system, enterprise savings, from undistributed profits, were built up to a huge

amount. In short, the continued decreases in final consumption ratio are results of

institutional defects, calling for public sector reforms.

Recently, the government adopted a number of policies to restore consumption growth,

including abolition of agricultural tax, exemption of school fees for rural nine year

education, more government expenses on poverty reduction, and improvement in social

security systems. These will certainly have a positive impact on final consumption.

However, whether these are enough to turn back the decreasing trend of consumption

ratio is still uncertain. Improvement in public sector governance is more basic.

In this study, we define the ratio of final consumption in GDP as a structure variable, and

hypothesize that it had a negative effect on economic growth when dropped below a

certain level. This hypothesis will be empirically tested.

Government administration cost

As a result of inefficient use of public resources and possible corruption, the government

administration expenses, as a part of fiscal expenditure, is increasing. It accounted for

1.35% of GDP in 1978, 1.80% in 2000, then increased to 2.63% in 2005. This figure

exclude the operating cost of government departments and their subordinates in the areas

Page 8: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

8

such as agriculture, industry, transport, communication, commerce, culture activities,

publication, education, healthcare, public medication, sports, sciences, social sciences,

government and party training, and superannuation of retired administrative persons.

With these included, it mounts to 1.35 trillion Yuan in 2005, accounting for 7.4% of GDP.

Inefficient and inadequate use of public resources may form some deduction from

economic growth and TFP changes. In this study, fiscal expenses on government

administration (the narrow sensed) as a ratio to GDP is used to test this possible effect.

The increases in administration cost may be treated as a reflection of a wider range of

government activities with low efficiency, including government investment and

government distribution of other public resources.

In this study, the same trend of increases in the share of administrative cost in GDP is

assumed to continue in the future till 2010, and slightly slower increases in 2011-2020, as

a result of possible improvement in government administration, is assumed. Different

scenarios are also simulated.

III. Empirical Tests

In an earlier study, Wang (2006) estimated contribution of a number of factors to

economic growth. A similar method is used in this study with modifications of the model

to examine the roles of more factors played in China’s economic growth.

A Lucas typed growth model is employed for the study. We use time series data at the

national level from 1952 to 2005. Most of the data, except those otherwise referenced, are

calculated from China Statistical Yearbook (NBS, various years) and China Compendium

of Statistics 1949-2004 (NBS, 2005b). Original input and output data used in this study

are shown in Table A1 in Appendix at the end of the paper.

The basic empirical model is specified as follows:

lnY(t)=C+a1lnK(t)+a2lnH(t-3)+a3Ha(t)+R(t) (1)

where Y(t) is GDP in 1978 constant price at year t; K(t) is fixed capital stock in 1978

constant price at year t; H is human capital stock or effective labor that is defined as total

labor force enhanced by their year of schooling; Ha is works’ average year of schooling

for possible spillover effect of human capital on economic growth, C is the intercept term,

and R(t) is the residual term, which contains both unexplained TFP changes and random

Page 9: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

9

errors (thus, without inclusion of other related variables, there should be an

autocorrelation problem).

This is much the same as Lucas (1988) model except a few modifications. First, lnHa is

replaced by Ha so that the effect of workers’ average year of schooling can be explicitly

measured. Second, the residual R(t) is not simply defined as a random error term, since,

obviously, Ha can not catch the entire TFP change. Third, human capital H takes a three-

year lag in this model. This is because our preliminary study using a distributed lag

model found a largest and most significant coefficient of lnH with three-year lag. This is

reasonable since H is only education-enhanced human capital without considering the

learning-by-doing effect. Experiences tell us that school graduates usually become more

productive a few years after they being employed. It is therefore reasonable to believe

that the three-year lag of lnH is a best representative of human capital stock in our case.

To test the possible effects of technical innovation, marketization, urbanization, foreign

investment, foreign trade, and government administration cost on TFP growth, above

model is expanded into the following form:

lnY(t)=C+a1lnK(t)+a2lnH(-3t)+a3Ha(t)+a4DlnRK(t)+a5m(t)+a6u(t)+a7fk(t)+a8td(t)

+a9ga(t)+a10hw(t)+ε1(t) (2)

where RK is a research capital stock that is accumulated by R&D investment. Due to

insignificance of lnRK in a preliminary estimation, we take its first difference, i.e., DlnRK,

in the model. According to this specification, a significant estimate of it meaning that

only an accelerating growing RK contributes to economic growth. m is the share of non-

state sector in industrial output as a proxy of marketization; u is the urbanization ratio,

that is, the share of urban population in total; fk is the share of foreign capital in total

capital stock; td is the trade dependency ratio; ga is the government administration cost as

a proportion to GDP; hw is the highway-population ratio; and ε1 is a random error.

In an alternatively model, we hypothesize that economic growth was restricted by a

structural bias, i.e., too low final consumption and too high saving and investment results

in overcapacity of production. This may work as a deduction of TFP. To test this

hypothesis, a structure variable fc, i.e., the ratio of final consumption in GDP, and its

quadratic term, is included in Model 3:

lnY(t)=C+a1lnK(t)+a2lnH(-3t)+a3Ha(t)+a4DlnRK(t)+a5m(t)+a6u(t)+a7fk(t)+a8td(t)

Page 10: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

10

+a9ga(t)+a10hw(t)+a11fc(t)+a12fc2(t)+ε2(t) (3)

To impose a restriction for constant-returns-to-scale hypothesis (a2=1-a1), both Y(t) and

K(t) are divided by H(-3t), so Model 3 can be transformed into the following version:

lny(t)=C+a1lnk(t)+a3Ha(t)+a4DlnRK(t)+a5m(t)+a6u(t)+a7fk(t)+a8td(t)

+a9ga(t)+a10hw(t)+a11fc(t)+a12fc2(t)+ε3(t) (3’)

where lny(t)=lnY(t)-lnH(-3t), and lnk(t)=lnK(t)-lnH(-3t).

We also want to investigate possible changes in productivity of capital and human capital.

For this purpose, both the capital and human capital variables are multiplied respectively

by three dummy variables for the first and second decades of the reform periods and the

most recent period (1979-1988, 1989-1998, and 1999-2005). Below is the modified

Model:

lnY(t)=C+a1lnK(t)+a1’lnK1979-88(t)+a1’’lnK1989-98(t)+a1’’’lnK1999-2005(t)+a2lnH(-3t)

+a2’lnH1979-88(-3t)+a2’’lnH1989-98(-3t)+a2’’’lnH1999-2005(-3t)+a3Ha(t)+a4DlnRK(t)

+a5m(t)+a6u(t)+a7fk(t)+a8td(t)+a9ga(t)+a10hw(t)+a11fc(t)+a12fc2(t)+ε4(t) (4)

where, for instance, lnK1979-88(t)=lnK(t) for years from 1979 to 1988, and lnK1979-88(t)=0 for

other years. The coefficient a1’, for instance, indicates changes of a1 in the period 1979-

1988. Therefore for this period, the elasticity of capital should be a1+a1’.

In Table 1, empirical results of the models defined above are reported, which are obtained

from Prais-Winsten AR(1) Regressions.

Page 11: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

11

Table 1. Estimation result: Prais-Winsten Regression AR(1)

Model 1 Model 2 Model 3 Model 3’ Model 4

lnK(t) 0.6045 0.3950 0.2815 0.2722 0.3229

(4.89**) (4.00**) (4.61**) (3.88**) (4.95**)

lnK1979-88(t) 0.0111

(2.99**)

lnK1989-98(t) 0.0121

(2.93**)

lnK1999-05(t) 0.0167

(3.05**)

lnH(t-3) 0.8323 0.6615 0.5893 0.5897

(3.28**) (4.65**) (6.75**) (6.18**)

lnH1979-88(t-3) 0.0025

(0.98)

lnH1989-98(t-3) 0.0072

(2.08*)

lnH1999-05(t-3) 0.0131

(2.67*)

Ha(t) -0.1015 0.0282 0.0619 0.0251 -0.0015

(-1.38) (0.72) (3.10**) (1.73’) (-0.05)

DlnRK(t) 0.7419 0.3584 0.3973 0.3100

(9.25**) (4.43**) (5.08**) (3.02**)

m(t) 0.3331 0.3249 0.2821 0.3849

(2.37*) (3.48**) (2.98**) (4.28**)

u(t) 1.1125 0.9824 0.9460 0.4832

(1.32) (2.06*) (1.77’) (0.94)

fk(t) 0.2214 1.1780 1.3280 0.8634

(0.27) (2.33*) (2.43*) (1.52)

td(t) 0.1773 0.2906 0.2043 0.1896

(0.88) (1.99’) (1.35) (1.32)

ga(t) -22.237 -11.491 -10.292 -15.193

(-4.71**) (3.26**) (-3.03**) (3.75**)

hw(t) 0.0291 0.0294 0.0322 0.0340

(1.27) (2.34*) (2.23*) (1.75’)

fc(t) 4.1182 2.6788 4.6895

(2.54*) (1.71’) (2.93**)

fc2

(t) -3.8605 -2.7047 -4.2725

(3.25**) (-2.37*) (-3.68**)

C -6.5602 -3.5624 -3.0078 -3.9580 -3.1308

(-2.95**) (3.24**) (-4.00**) (-7.37**) (-3.87**)

Adj. R2 0.964 0.996 0.999 0.995 0.999

DW (original) 0.323 1.643 1.792 1.563 2.075

DW(transformed)

1.059 1.734 1.847 1.694 2.067

n 51 51 51 51 51

Note: figures in parentheses are t-ratios. Those with ’ are significant at the 10% level, with * are at the 5% level, and with ** are at the 1% level.

Page 12: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

12

In Model 1, both capital and human capital show significant contribution to economic

growth, and the estimated spillover effect of human capital is negative and insignificant.

As expected, the Durbin-Watson statistic indicates existence of autocorrelation, therefore

unreliable estimates.

Model 2 produces better results after including a number of relevant variables. Estimates

of capital, human capital, marketization, difference of research capital, and government

administration cost (negative) are significant. Others are insignificant; they are spillover

effect of human capital, the urbanization ratio, foreign capital share, trade dependency

ratio, and highway-population ratio. Durbin-Watson statistic is much improved from

Model 1. It is in an inconclusive interval but close to dU.

Model 3 rejects the null hypotheses and confirms the negative impact of the structural

bias. The estimate of final consumption ratio, fc, is positive and significant; and the

estimate of its quadratic term is negative and significant. This indicates an invert U-shape

curve for effect of final consumption ratio on economic growth: positive before a certain

critical point and negative after that point. Figure 1 illustrates this effect using the

estimated coefficients of fc and fc2, for an interval of fc between 90% to 10%. Shown by

the simulated curve, the critical point of final consumption in GDP is around 55%.

According to this result, the current final consumption ratio, 51.9% (year 2005), already

has a negative effect on economic growth; and the current trend of changing fc indicates a

greater negative effect on growth in the future.

Page 13: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

13

Figure 1. Simulating the negative effect of structural bias on growth

0.2

0.4

0.6

0.8

1.0

1.2

10%20%30%40%50%60%70%80%90%

Final consumption ratio

Dem

and

eff

ect

Source:Data are from estimates of the model.

Model 3 also shows further improvement in Durbin-Watson statistic. As a result, the

spillover effect of human capital, urbanization, share of foreign capital, trade dependency,

and per capital standard highway all become significant at least at the 10% level (only for

the trade dependency ratio). All the other estimates are significant at either 1% or 5%

levels.

The restricted version of the model for constant returns to scale (Model 3’) and the model

with periodical dummies on capital and human capital (Model 4) obtained very similar

estimates to Model 3, except that the spillover effect of human capital is smaller or

becomes negative and insignificant (indifference from zero in Model 4), due to that more

contribution is attributed to capital and the own effect of human capital, and the Durbin-

Watson statistic for Model 3’ is lower.

Results of Model 4 show significant increases in returns to both capital and human capital

during the reform period. Returns to capital increased by 0.0167 in the 1999-2005 period

compared with the average of pre-reform period, and returns to human capital increased

by 0.0131 in the 1999-2005 period. These equal to a five-percent increase in marginal

productivity of capital and two-percent increase in marginal productivity of human

capital.

Using statistical data and estimates of Models 2, 3, 3’ and 4 to simulate economic growth

for the past periods, all the three models provide generally good simulations, but Model 3

Page 14: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

14

provides the closest ones compared with the actual growth rates. Especially, the

simulated average GDP growth rate for the recent two periods, i.e., 1989-1998, and 1999-

2005, are only -0.05 and -0.2 percentage points away from the actual growth rates. The

estimated parameters from Model 3 are therefore used in growth accounting in the next

section. Table 2 compares different simulation results. Statistics for growth of inputs and

contributing factors in different periods can be seen in Table A2 in Appendix.

Table 2. Simulating GDP growth rates for the past periods (average growth rate, %)

1953-1978 1979-1988 1989-1998 1999-2005

Actual rates 6.15 10.06 9.59 9.11

Simulated rate

Model 2 7.29 10.19 9.33 8.72

Model 3 6.57 9.56 9.54 8.90

Model 3’ 6.72 9.64 9.49 8.67

Model 4 6.48 8.56 8.62 7.32

Error

Model 2 1.15 0.13 -0.26 -0.39

Model 3 0.43 -0.49 -0.05 -0.21

Model 3’ 0.58 -0.41 -0.10 -0.43

Model 4 0.34 -1.50 -0.96 -1.79

Source:Data are from the author’s simulation based on estimates of the models and statistical data, NBS

(various years).

IV. Growth Accounting

Based on the estimates of above and statistical data, the author carries out growth

accounting to decompose economic growth rate into different sources for different period.

Table 3 shows the result. Contribution of inputs and productivity changes to economic

growth are calculated.

The result indicates that capital and human capital growth (or, input-driven growth) have

played a crucial role in driven economic growth in the past, both pre-reform and reform

periods. It contributed 5-6 percentage points in most periods in the past.

TFP growth contributed a neglected 0.3 percentage point to growth in the pre-reform

period, increased significantly to 3.3 percentage points in the early stage of the reform

period, and then between 3.7 - 4.4 percentage points in the later stage of the reform

period until 2005.

Page 15: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

15

Most important sources of TFP growth during the whole reform period was marketization

and urbanization, they together made average 1.5 – 1.7 percentage point contribution to

TFP growth. The market-oriented institutional changes lead to increasing efficiencies via

factor reallocation and better incentive system, and urbanization also leads to reallocation

of labor and other resources from the low-productivity agricultural sector to higher-

productivity urban non-agricultural sectors. The results show that, while the effect of

marketization diminishing, the urbanization effect increased to 1.3 percentage point in

recent years.

The effects of foreign investment and foreign trade, via spillover of technology and

management skills, together contributed only 0.6 percentage point to TFP growth in the

earlier reform period, then increased to 1.0 - 1.3 in the 1990s and recent years. While the

foreign share in total capital stock is decreasing in recent years, the effect of foreign trade

is playing a more important role in TFP growth, made 1.3 percentage point contribution

to TFP. Nevertheless, one can hardly separate the demand effect that driven by export

growth from the effect of trade-sourced productivity growth.

Spillover effect of human capital was also found to make important contribution to TFP

growth, at 0.8-1.0 percentage point.

Infrastructural improvement, reflected by growth of per capital highway, made little, but

increasing, contribution to TFP growth in the 1980s and 90s. It became important source

of TFP growth in recent years, i.e., 1.3 percentage point.

While research capital (i.e., accumulated R&D investment that calculated by Perpetual

Inventory Method) was not found to have significant impact on growth, its annual

difference is significant. This suggests that the faster accumulation of research capital has

lifted up economic growth rate to some extent. The effect is positive but very small in the

1990s, and then increased to 0.5 percentage point in recent years. Considering the fact

that either the research capital stock or the annual R&D expenses is still fairly small, this

is likely an indicator for growth effect of technical innovation in the future.

The high administration cost made an increasing negative impact on economic growth. It

reduced TFP growth by 0.1 percentage points in the 1980s and 1990s, but this effect

jumped to -1.7 percentage points in recent years, heavily reduced TFP growth.

Page 16: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

16

One should note that, the negative effect of the structural bias on growth and TFP

changes only exists when the final consumption ratio in GDP falls below the estimated

critical point of 55%, because in this situation it reduce economic efficiencies. However,

when it varies in the range above 55%, its impact on growth is actually via the effect of

changing savings and investment, and therefore should be reclassified as a part of

contribution by growth of capital. In Table 3, this was added back to input-driven growth

as a part of capital contribution.

The final consumption ratio dropped dramatically from 62.3% in 2000 to 54.3% in 2004,

then to 51.8% in 2005. This made negative contribution to economic growth in 2005, but

not explicitly reflected in the average contribution in the period of 1999-2005. A

continuation of the current trend will lead to further drop of the ratio in the coming a few

years, although will be moderated by the government policy reinforcing income

redistribution and poverty reduction.

There is a small part of unexplained TFP growth for different periods, i.e., differences

between the actual and simulated growth rates. It was negative in the pre-reform period,

indicating a technical deterioration, half percent in the earlier reform period (which was

likely a unexplained allocative effect leading by rapid rural industrialization in that

period), diminished to a neglectable level in the 1989-1998 period, and then increased to

0.21% in recent years. The recent change may be considered as a result of technical

progress that was not indicated by the effect of research capital growth.

Table 3. Growth accounting: input-driven and TFP growth (annual growth rate, %)

1953-78 1979-88 1989-98 1999-05

Economic growth rate 6.15 10.06 9.59 9.11

Input-driven growth 5.83 6.70 5.16 5.36

By capital 2.59 2.58 2.70 3.59

By human capital 2.39 4.26 2.19 1.56

TFP growth – explained 0.74 2.86 4.37 3.53

Spillover effect of human capital 0.40 1.02 0.84 0.79

Increasing R&D expenses 0.11 -0.18 0.16 0.47

Marketization -0.45 0.68 0.92 0.32

Urbanization 0.21 0.78 0.74 1.35

Foreign capital effect 0.00 0.16 1.15 -0.35

Page 17: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

17

Foreign trade effect 0.00 0.46 0.19 1.33

Government administration cost 0.35 -0.14 -0.12 -1.73

Infrastructure effect 0.11 0.10 0.49 1.35

Effect of structural bias 0.86 -0.14 0.28 0.21

TFP growth - unexplained -0.43 0.49 0.05 0.21

Source:Same as Table 2.

The above result rejects the judgment on China’s economic growth as “input-driven

growth without productivity changes” (see, e.g., A. Yang, 2000; P. Krugman, 1994). It

clearly indicates a 3-4 percentage point TFP growth during the reform period. Meanwhile,

it also shows that one-third or half of the TFP growth was from “allocative effect”, i.e.,

productivity growth induced by improvement in factor allocation. This effect is reflected

by contribution of marketization and urbanization. Economic theories have proved that

reallocation of economic resources is a short run effect which does not sustain a high

growth rate in the long run, although urbanization effect in China is unlikely to diminish

in the next 10-15 years. Similar role may be attributed to contribution of infrastructural

improvement. It generates positive externalities to the economy, may continue to push up

economic growth in the coming decades or so, but the effect on growth rate may not be

sustainable in the ‘long run’ that defined in growth theories.

Another important part of TFP growth, at least one percentage point, was identified as

spillovers of technology and management skills from foreign investment and foreign

trade. This is partially a sustainable source of growth since it injects new technology into

the economy. However, sustainability of these effects on growth is discounted to some

extent because the source of TFP is not endogenously generated.

A new finding in this study is that the increasing R&D expenses and the spillover effect

of human capital made important contribution to TFP growth, accounting for 1.3

percentage points in 1999-2005. With inclusion of the unidentified TFP growth, which

should be considered as normal technical innovations that did not reflected from the

effect of increasing research capital, their total contribution to TFP was 1.5 percentage

points. This is a clear signal to indicate high possibility of changing growth pattern

towards a more sustainable way in the future.

Page 18: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

18

While optimistic results were obtained, the outcome of growth accounting also delivers

some pessimistic massages on economic growth. The first one is the significant negative

effect of government administration cost on economic growth, which made a 1.7

percentage point deduction to growth and productivity changes during the 1999-2005

period. This indicates that government inefficiency and corruption is becoming a serious

threat to sustainability of economic growth. The second one is insufficient domestic

consumption, which already made deductions to economic growth in 2005, and likely to

further restrict economic growth in the near future. These two effects make uncertainties

to future economic growth.

Based on these results, economic growth in 2006-2010 and 2011-2020 is projected in the

following in two scenarios. The basic scenario is obtained generally based on the current

trend of changing contributing factors. The alternative scenario makes two optimistic

assumptions on changing government efficiency and increasing final consumption, as

results of government reform, with other conditions the same as in first scenario. For the

2011-2020 period, the effect of increasing R&D is assumed to be zero, and a same effect

is attributed to a common technical progress and added to the item: TFP from

unidentified factors. The two scenarios are compared in Table 4. Predictions and

assumptions on changing contributing factors can be seen in Table A2 in the appendix.

Table 4. Growth forecasts: different scenarios (annual growth rate, %)

Basic Scenario (I) Alternative Scenario (II)

2006-10 2011-20 2006-10 2011-20

Input-driven growth 5.84 5.12 5.84 5.12

By capital 4.31 3.94 4.31 3.94

By human capital 1.53 1.18 1.53 1.18

TFP from identified factors 1.58 0.10 3.64 3.34

Spillover effect of human capital 0.74 0.74 0.74 0.74

Increasing R&D expenses 0.47 0.00 0.47 0.00

Marketization 0.19 0.13 0.19 0.13

Urbanization 1.35 1.08 1.35 1.08

Foreign capital effect -0.59 -0.35 -0.59 -0.35

Foreign trade effect 0.58 0.00 0.58 0.00

Government administration cost -1.73 -1.73 0.00 1.15

Infrastructure effect 0.88 0.59 0.88 0.59

Page 19: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

19

Effect of structural bias -0.31 -0.35 0.00 -

TFP from unidentified factors 0.40 0.80 0.40 0.80

Projected growth rate 7.82 5.92 9.88 9.26

Source:Same as Table 2.

There are two differences in assumptions used for the two scenarios:

First, in the basic scenario, the author assumes that government administration cost as a

share in GDP will continue to expand in 2006-2010 and 2011-2020 at the same rate as in

1999-2005, i.e., increase by 0.15 percentage point per year, whereas in the alternative

scenario, it is assumed to stop expansion in 2006-2010, and decrease by 0.1 percentage

point per year in 2011-2020 as a result of possible government reform and therefore

increasing government efficiency.

Second, the author assumes that, in the basic scenario, the share of final consumption in

GDP will continue to decrease, resulting from further expansion of income inequality in

initial income distribution, but at a slower rate, due to increases in government transfer

payment. In 2000-2005, this ratio actually dropped by 10.5 percentage points. It is

assumed to further decrease by 5.0 percentage points in 2006-2010, and another 5.0

percentage points in 2011-2020. In the alternative scenario, the consumption ratio is

assumed to be stable in 2006-2010, and increase to 55% in 2011-2020. This is considered

as a result of a series policy adjustment toward a more healthy income distribution and

improvement in social security systems and public services.

The results of the two scenarios are very different. In the first scenario, economic growth

rate will drop from average 9.5% (2000-2005) to 7.8% (2006-2010), then to average

5.9% (2011-2020). The rapid growth period since 1978 will end in the 2020s. In the

second scenario, economic growth rate will be sustained at 9.9% in 2006-2010, and 9.3%

in 2011-2020. A continued rapid growth in longer term can be expected.

V. Conclusion

In this study, the author examines China’s economic growth pattern in the past and future

and growth sustainability towards 2020. Empirical study using a Lucas typed growth

model and data after statistic revisions identifies a 3-4 percentage point TFP growth

Page 20: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

20

during the past period of economic reform from 1978 to 2005. Of which, 1.5-1.7 was

contributed by marketization and urbanization during the whole reform period, mainly

via improvement in factor allocation; 1.0-1.3 was contributed by spillover effect of

international trade and foreign direct investment since the 1990s; 1.3 was contributed by

externalities from improvement in infrastructure in most recent years. Increases in R&D

expenses and spillover effect of human capital together contributed 0.8 percentage point

in the 1980s, 1.0 in the 1990s, and 1.3 in most recent years, indicating an increasing trend

of technological progress.

Two negative impacts on TFP growth are identified. One negative impact is the

increasing government administration cost, representing effect of government

inefficiencies and corruptions, which was found to make a rapid increasing deduction to

TFP growth, accounts for -1.7 percentage points in recent years. Another is a structural

bias, i.e., continued drop of the share of final consumption in GDP, which started to

generate negative impact on GDP growth in the immediate past years.

The effect of the structural variable is found to be non-linear. The effect of diminishing

consumption ratio is positive above a critical value around 55%, and turns into negative

after dropping below this point.

Based the result of growth accounting and analysis on changes in various contributing

factors, economic growth rate is projected for the periods of 2006-2010 and 2011-2020. It

suggests a continued trend of growth driven by inputs in the two coming periods with

minor decreases, a diminishing contribution of marketization but a continued strong

contribution of urbanization and infrastructure improvement, and a continued, if not

further increasing, contribution of TFP growth by R&D, human capital spillover, and

other sources of technical innovation.

Changes in the government administration cost and final consumption are found to be

crucial determinants for future growth. With the current trends of increasing government

administration cost and decreasing final consumption, economic growth rate will be

around 7.8% in average of the 2006-2010 period, and then drop to an average level of

5.9% in 2011-2020. The rapid growth period since 1978 will end in the 2020s. However,

with possible increases in government efficiency via government reforms, and recovery

of domestic consumption that can be induced by improvement in public services, social

Page 21: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

21

security systems, and more equity of income distribution, economic growth can be

sustained at above 9% level in both the 2006-2010 and 2011-2020 periods. A more

sustainable and continued rapid growth in longer period can be expected.

Page 22: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

22

References:

Chow, G. C. 1993, “Capital Formation and Economic Growth in China”, The Quarterly

Journal of Economics, August, pp. 809—842.

Garnaut, R. 2006, “The Turning Point in China’s Economic Development”, in Garnau and Song (eds), The Turning Point in China’s Economic Development, Asia Pacific Press at the Australian National University, Canberra.

Jefferson, G, T. Rawski, and Y. Zhang, 2007, “Productivity Growth and Convergence Across China’s Industrial Economy”, paper presented in International Workshop on Chinese Productivity 2007, Tsinghua University.

Krugman, P. 1994, “The Myth of Asia's Miracle”, Foreign Affairs, November/ December 1994, Vol 73, Number 6.

Lucas, R. E., 1988, “On the Mechanics of Economic Development”, Journal of Monetary

Economics, 22, 3-42.

National Bureau of Statistics (NBS), various years, China Statistical Yearbook, China Statistics Press, Beijing.

_______, 2005b, China Compendium Statistics 1949-2004, China Statistics Press, Beijing.

Wen Jiaobao, 2006, “Government Work Report”, at the Fourth Session of the Tenth National People’s Congress, Xinhuanet, http://news.xinhuanet.com.

Wang Xiaolu, 2006, “Growth Accounting after Statistical Revisions”, in Garnaut and Song (eds), The Turning Point in China’s Economic Development, 35-52, ANU E Press, Australian National University, Canberra.

Young, A., 2000, “Gold into Base Metals: Productivity Growth in the Peoples Republic of China during the Reform Period”, NBER Working Paper W7856, National Bureau of Economic Research, Cambridge.

Zhang Jun, et al., 2007, “Estimation of Capital Stock for Chinese Provinces”, paper presented in International Workshop on Chinese Productivity 2007, Tsinghua University.

Page 23: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

23

Appendix

Table A1. Input-output data for China’s economic growth

Year GDP Total

employment Capital stock

Foreign capital

Research Capital

Human capital

Year of schooling

RMB 100 mil,

1978 price 10000

persons RMB 100 mil,

1978 price RMB 100 mil,

1978 price RMB 100 mil,

1978 price 10000

person year Year per laborer

1952 773 20729 700 0.0 3.0 74727 3.605

1953 894 21364 852 0.0 3.3 75047 3.513

1954 931 21832 1020 0.0 4.3 75905 3.477

1955 995 22328 1158 0.0 6.2 77949 3.491

1956 1144 23018 1388 0.0 11.4 79341 3.447

1957 1202 23771 1602 0.0 16.0 81834 3.443

1958 1458 26600 1939 0.0 26.6 83362 3.134

1959 1587 25173 2350 0.0 44.2 86470 3.435

1960 1582 25880 2809 0.0 75.1 88827 3.432

1961 1150 25590 2888 0.0 87.2 93728 3.663

1962 1086 25910 2858 0.0 92.5 100338 3.873

1963 1196 26640 2862 0.0 102.1 106922 4.014

1964 1415 27736 2931 0.0 116.5 112839 4.068

1965 1656 28670 3071 0.0 133.5 117526 4.099

1966 1833 29805 3235 0.0 147.4 128377 4.307

1967 1729 30814 3307 0.0 150.8 136632 4.434

1968 1658 31915 3337 0.0 153.5 146365 4.586

1969 1938 33225 3496 0.0 165.9 152306 4.584

1970 2314 34432 3806 0.0 183.5 155017 4.502

1971 2476 35620 4160 0.0 207.7 158376 4.446

1972 2570 35854 4483 0.0 228.2 165296 4.610

1973 2773 36652 4822 0.0 245.5 172877 4.717

1974 2837 37369 5176 0.0 261.4 179460 4.802

1975 3084 38168 5609 0.0 281.7 186501 4.886

1976 3034 38834 5989 0.0 299.1 191801 4.939

1977 3265 39377 6371 0.0 317.1 198498 5.041

1978 3645 40152 6878 0.0 344.6 212410 5.290

1979 3922 41850 7358 1.7 377.0 236568 5.653

1980 4228 43850 7811 6.4 406.6 267700 6.105

1981 4451 45950 8234 13.6 429.8 293978 6.398

1982 4852 48150 8829 26.7 453.7 318083 6.606

1983 5380 50100 9509 40.1 486.8 338705 6.761

1984 6197 52450 10408 64.1 528.7 357099 6.808

1985 7032 54800 11670 103.3 567.0 374509 6.834

1986 7655 57050 13114 150.5 605.3 392615 6.882

1987 8541 59300 14766 196.5 635.4 410318 6.919

Page 24: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

24

1988 9503 61650 16526 251.2 655.7 427752 6.938

1989 9889 63450 17669 299.9 668.2 445292 7.018

1990 10269 64749 18590 361.3 676.5 462533 7.143

1991 11213 65491 19697 441.3 692.6 478822 7.311

1992 12809 66152 21343 658.7 712.5 494272 7.472

1993 14595 66808 23714 1131.6 727.3 509303 7.623

1994 16505 67455 26657 1914.2 735.8 524132 7.770

1995 18310 68065 29887 2636.6 768.4 539186 7.922

1996 20143 68950 33390 3348.2 807.5 553964 8.034

1997 22013 69820 37004 4052.6 868.5 569558 8.158

1998 23738 70637 41234 4694.7 938.4 585766 8.293

1999 25549 71394 45488 5162.2 1040.0 601840 8.430

2000 27700 72085 50077 5583.8 1188.3 617415 8.565

2001 30000 73025 55355 6099.3 1361.6 632901 8.667

2002 32727 73740 61805 6694.4 1587.3 647730 8.784

2003 36007 74432 70583 7224.2 1852.4 663166 8.910

2004 39638 75200 81484 7789.4 2177.7 679444 9.035

2005 43695 75825 95576 8261.5 2578.9 696327 9.183

Note: 1. Modifications were made to employment data, because there was originally a large gap between the 1989 and 1990 statistics on total employment, equals to 17% of the 1989 employment. This was a result of data inaccuracy between the two national population censuses in 1982 and 1990. Data for the period of 1983-1989 are modified accordingly. 2. Fixed capital stock is calculated using historical data of total investment in fixed assets and price index for fixed investment, using a Perpetual Inventory Method (see Wang, 2006). To avoid shortage of data on various categories of capital stock, an overall depreciation rate of 5% is used for the pre-reform period. Some studies found faster capital depreciation during the reform period. In this study, we adopt a final 9.6% rate that recommended by Zhang Jun (2007), but assume it was gradually achieved during the 1979-1992 period with a 0.3 percentage point change per year. For the initial fixed capital stock in 1952, we take RMB 70 billion (in 1978 price), mainly based on Chow (1993). He calculated that capital stock in the non-agricultural sectors was 58.3 billion Yuan in 1952 (1952 price), of which fixed capital was 31.6 billion, and capital stock in the agricultural sector was 45 billion including non-fixed capital. We assume 70% of total agricultural capital being fixed capital, and upward modify the total fixed capital by 10% with consideration that Chow’s capital stock maybe more or less underestimated due to data incompletion, we get a total 69.4 billion fixed capital stock in 1952 at 1952 price (see Wang, 2006). 3. Foreign direct investment data are used for calculation of foreign capital stock. They are converted to Chinese Yuan at the official exchange rates and deflated using Fixed Asset Investment Price Index of the NBS. A depreciation rate of 9.6% is assumed. 4. The research capital is defined as an accumulation of knowledge and technology, and approximately measured by the accumulated expenses on R&D at constant prices. Because of data incompletion, and considering that enterprise expenses on R&D were rare in earlier period, we use fiscal expenses on science and technology for the period before 1990. Data are deflated into 1978 prices using a GDP deflator. The invisible depreciation rate is assumed to be 8%. 5. Annual formation of human capital from 1952 to 2005 is calculated from graduation and enrollment data (with one education-period lag) of all kinds of schools, from primary to postgraduate education. The calculation referenced the information from the four national population censuses in 1964, 1982, 1990 and 2000. Unfinished school education (i.e., the differences between graduation and period lagged enrollment) is assumed to have an average 50% length of the corresponding education period. Vocational education, adult education, special education, overseas study, and informal training programs are also included. Human capital depreciation is calculated based on the death rate of the population and the calculated average year of schooling of population (not of labor force) with time lag. The initial human capital stock in 1952 is projected as average 1.3 year of schooling of the total population. This is based on 1964 national census data on education and detailed education data between 1952 and 1964. 6. Workers’ average year of schooling is calculated using human capital and employment data.

Page 25: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

25

Table A1. continued

Year Standard highway

Urbaniza-tion ratio

Non-state share

Non-state share (adj)

Trade ratio Administra-

tion cost Final

consumption

km Urban/total population

In industrial output

In industrial output

To GDP To GDP To GDP

1952 9031 0.1246 0.5850 0.5850 0.0951 0.0214 0.7892

1953 10049 0.1331 0.5700 0.5700 0.0982 0.0213 0.7723

1954 11004 0.1369 0.5290 0.5290 0.0986 0.0213 0.7447

1955 12930 0.1348 0.4870 0.4870 0.1207 0.0206 0.7726

1956 17946 0.1462 0.4550 0.4550 0.1057 0.0235 0.7471

1957 20700 0.1539 0.4620 0.4620 0.0978 0.0203 0.7409

1958 35134 0.1625 0.1080 0.1080 0.0985 0.0165 0.6603

1959 43319 0.1841 0.1140 0.1140 0.1038 0.0185 0.5660

1960 45347 0.1975 0.0940 0.0940 0.0881 0.0192 0.6184

1961 42626 0.1929 0.1150 0.1150 0.0743 0.0219 0.7803

1962 42316 0.1733 0.1220 0.1220 0.0704 0.0189 0.8379

1963 44323 0.1684 0.1070 0.1070 0.0695 0.0191 0.7844

1964 45662 0.1837 0.1050 0.1050 0.0671 0.0173 0.7481

1965 50054 0.1798 0.0990 0.0990 0.0632 0.0148 0.7111

1966 53977 0.1786 0.1000 0.1000 0.0680 0.0139 0.6848

1967 56471 0.1774 0.1150 0.1150 0.0633 0.0129 0.7470

1968 59054 0.1762 0.1160 0.1160 0.0630 0.0133 0.7429

1969 63243 0.1750 0.1130 0.1130 0.0552 0.0128 0.7318

1970 68311 0.1738 0.1240 0.1240 0.0501 0.0112 0.6614

1971 73820 0.1726 0.1410 0.1410 0.0498 0.0127 0.6512

1972 77895 0.1713 0.1510 0.1510 0.0583 0.0138 0.6701

1973 81078 0.1720 0.1600 0.1600 0.0810 0.0131 0.6560

1974 85082 0.1716 0.1760 0.1760 0.1047 0.0132 0.6608

1975 91919 0.1734 0.1890 0.1890 0.0969 0.0130 0.6397

1976 98232 0.1744 0.2170 0.2170 0.0897 0.0139 0.6635

1977 103777 0.1755 0.2300 0.2300 0.0851 0.0135 0.6500

1978 109763 0.1792 0.2240 0.2240 0.0974 0.0135 0.6210

1979 109734 0.1896 0.2150 0.2150 0.1119 0.0141 0.6435

1980 112613 0.1939 0.2402 0.2402 0.1254 0.0147 0.6549

1981 116457 0.2016 0.2520 0.2520 0.1503 0.0145 0.6711

1982 119813 0.2113 0.2560 0.2560 0.1449 0.0153 0.6645

1983 123364 0.2162 0.2660 0.2660 0.1442 0.0171 0.6638

1984 127735 0.2301 0.3090 0.3090 0.1666 0.0174 0.6582

1985 133571 0.2371 0.3514 0.3514 0.2292 0.0145 0.6595

1986 141596 0.2452 0.3770 0.3770 0.2511 0.0164 0.6492

1987 152505 0.2532 0.4030 0.4030 0.2558 0.0149 0.6357

1988 164196 0.2581 0.4320 0.4320 0.2541 0.0147 0.6394

1989 173701 0.2621 0.4390 0.4390 0.2446 0.0154 0.6449

Page 26: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

26

1990 184691 0.2641 0.4539 0.4539 0.2978 0.0162 0.6249

1991 193751 0.2694 0.4383 0.4383 0.3317 0.0158 0.6242

1992 206101 0.2746 0.4848 0.4848 0.3387 0.0158 0.6241

1993 224837 0.2799 0.5305 0.5305 0.3190 0.0152 0.5929

1994 245657 0.2851 0.6266 0.6266 0.4229 0.0151 0.5823

1995 274513 0.2904 0.6603 0.6603 0.3866 0.0144 0.5813

1996 304256 0.3048 0.6368 0.6217 0.3391 0.0146 0.5922

1997 444028 0.3191 0.6838 0.6487 0.3415 0.0144 0.5895

1998 391918 0.3335 0.7176 0.6379 0.3181 0.0157 0.5962

1999 439258 0.3478 0.5108 0.6376 0.3334 0.0170 0.6116

2000 554138 0.3622 0.5267 0.6359 0.3958 0.0180 0.6230

2001 579256 0.3766 0.5557 0.6446 0.3847 0.0200 0.6137

2002 639305 0.3909 0.5922 0.6602 0.4270 0.0248 0.5957

2003 693469 0.4053 0.6246 0.6736 0.5189 0.0253 0.5678

2004 756825 0.4176 0.6519 0.6841 0.5976 0.0254 0.5430

2005 831022 0.4299 0.6672 0.6919 0.6386 0.0263 0.5186

Note (continued): 7. Standard highway is calculated from natural highway length to grade II highway equivalent according to .the road capacity on transport volumes. 8. There are data inconsistencies in unadjusted non-state share in industry after late 1990s. Data are adjusted according to two national censuses on industry in 1995 and 2004. Source: Calculated from NBS (various year); NBS (2005b).

Page 27: Pattern and Sustainability of China’s Economic Growth …cerdi.org/uploads/sfCmsContent/html/203/FanGang_alii.pdf · 1 Pattern and Sustainability of China’s Economic Growth towards

27

Table A2. Contributing factors in different period:

Average growth rate and changing percentage point (%)

Scenario I Scenario II

1953-1978

1979-1988

1989-1998

1999-2005 2006-

2010 2011-2020

2006-2010

2011-2020

Capital 9.19 9.16 9.57 12.76 15.30 14.00 15.30 14.00

Human capital 4.06 7.22 3.71 2.65 2.60 2.00 2.60 2.00

Spillover effect of human capital

6.48 16.48 13.54 12.72 12.00 12.00 12.00 12.00

Increasing R&D expenses

0.32 -0.52 0.46 1.32 1.32 0.00 1.32 0.00

Marketization -1.39 2.08 2.83 0.99 0.60 0.40 0.60 0.40

Urbanization 0.21 0.79 0.75 1.38 1.38 1.10 1.38 1.10

Foreign capital effect

0.00 0.14 0.98 -0.29 -0.50 -0.30 -0.50 -0.30

Foreign trade effect 0.01 1.57 0.64 4.58 2.00 0.00 2.00 0.00

Government administration cost

-0.03 0.01 0.01 0.15 0.00 -0.10

Infrastructure effect 3.78 3.39 16.62 45.92 30.00 20.00

Effect of structural bias

70.89 65.22 60.96 58.37 51.90 55.00

TFP from unidentified factors

Source: Calculated from NBS (various year); NBS (2005b).