An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009 - IA de Investigación especializada en el MIT

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009
An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

An Empirical Analysis of China’s
Dualistic Economic Development:
1965–2009*

Marco G. Ercolani
Departamento de Economía
Birmingham Business School
University of Birmingham
Edgbaston, B15 2TT
Reino Unido
M.G.Ercolani@bham.ac.uk

Zheng Wei
Nottingham University
Business School
University of Nottingham
Ningbo China
Ningbo, 315100, Porcelana
Zheng.Wei@nottingham.edu.cn

Abstracto
We analyze China’s rapid economic development in the context
of the dualistic development theory. Over the period 1965–2009,
we find that China’s economic growth is mainly attributable to the
development of the non-agricultural (industrial and service) segundo-
colina, driven by rapid labor migration and capital accumulation. Nosotros
find that the sectoral reallocation of labor plays a significant role in
promoting China’s economic growth. Más, we find that the
marginal productivity of agricultural labor stopped stagnating in
1978, which indicates that China entered quickly into phase two
of economic development with the initiation of market reforms.
Además, por 2009, the marginal productivity of labor has likely
exceeded the institutional wage, as defined by the initially low av-
erage labor productivity, indicating that China may be now in the
process of entering phase three of economic development.

1. Introducción

China has a long history of dualistic economic develop-
mento. Prior to the 1978 economic reforms led by Deng
Xiaoping, China was a very poor and largely peasant
agrarian economy. Eighty percent of the labor force was
engaged in the rural agricultural sector. Agricultural labor
productivity was extremely low due to the presence of
surplus labor relative to other scarce resources. El

* The authors are grateful to the participants of the Asian Eco-

nomic Panel meeting in New York on 24–25 March 2011, partícipe-
ularly Professors Wing Thye Woo, Jeffery D. Sachs, Mary-
Françoise Renard, Fang Cai, Yunwing Sung, and Prema-
chandra Athukorala, for their helpful comments. Los autores
also thank the Earth Institute at Columbia University and the
Asian Economic Papers panel for their kind sponsorships.

Asian Economic Papers 10:3

Institute of Technology

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

economic reforms initiated in 1978 propelled the Chinese economy into a new phase
of development. Rural liberalization led to development in the agricultural sector
and to the establishment of township and village enterprises (TVEs). It also helped
release surplus labor from agricultural production. The steady marketization of eco-
nomic activities and the integration of China into the international economy gener-
ated a large demand for labor. Como resultado, there was a great increase in labor migra-
ción, albeit on a temporary basis due to the restriction on labor mobility by the
household registration system (the Hukou system). En 1978, 4.38 million laborers mi-
grated to the non-agricultural sector (Zhang and Song 2003); por 2009, the number
of rural–urban migrants rose to over 145 millón (Yao 2011). Labor in the non-
agricultural sector increased from 29.5 percent of total employment in 1978 a
61.9 por ciento en 2009. Over the same period, the contribution of non-agricultural pro-
duction to the value-added GDP increased from approximately 62 por ciento a 89 por-
centavo. Rapid development of the non-agricultural sector, fuelled by great inºows of
rural labor, has driven the Chinese economy to grow at an average rate of 8 por ciento
per annum for more than three decades since the 1978 economic reforms.

China’s remarkable economic growth has been interpreted in various ways. Desde
the 1990s, two schools of thought, institutional experimentalism and institutional
convergence, have emerged in the literature (Sachs and Woo 2000). Experimentalists
attribute China’s rapid economic growth to the discovery of new non-capitalist
mechanisms (such as collectively owned TVEs) that were very efªcient. A diferencia de,
proponents of institutional convergence regard China’s growth as the natural conse-
quence of an initially poor peasant economy converging toward a developed market
economía. Both schools, though disagreeing on the degree of institutional innova-
ción, strongly emphasize the role of marketization and internationalization in spur-
ring economic growth in China. It is generally accepted, sin embargo, that economic re-
forms are only able to lift an economy out of poverty if they are compatible with
that economy’s characteristics. China has been characterized as a dualistic economy
with so-called “backward advantages,” namely, the presence of surplus agricultural
labor. The rapid development of China’s economy is in fact largely attributable to its
dualistic nature and the ºows of surplus labor between the two sectors. Though
China’s economic reforms are undoubtedly important, they facilitated rather than
caused the process of economic transition. Por eso, it is more appropriate to interpret
China’s remarkable economic growth as a consequence of the development of a
dualistic economy with surplus labor.

The most well-known theory of dualistic development was proposed by Lewis
(1954) and formalized by Ranis and Fei (1961). The Lewis-Ranis-Fei theory of
dualistic economic development describes the development of an over-populated
and underdeveloped economy with vast amounts of surplus agricultural

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

labor.1 Economic growth in such an economy can be achieved by rapid development
of the non-agricultural (industrial and service) sector, facilitated by capital accumu-
lation and the drawing of surplus labor from the agricultural sector. Sectoral reallo-
cation of surplus labor changes the productivity of labor and thereby drives the
economy to transit from a labor-surplus stage to a labor-scarce stage of develop-
mento. The Lewis-Ranis-Fei theory of dualistic economic development therefore pro-
vides a suitable theoretical framework for studying the growth path of labor-surplus
developing economies such as China.

en este documento, we analyze the dualistic development of the Chinese economy using
the Lewis-Ranis-Fei theory. En particular, we focus on addressing the pattern of
China’s economic development by highlighting the role of sectoral reallocation of
surplus labor. Además, we identify the three phases of China’s economic develop-
ment and test if China has yet reached the Lewis turning point.2 To answer these
questions we ªrst estimate Cobb–Douglas production functions for China’s agricul-
tural and non-agricultural sectors using national-level data over the period 1965–
2009. Our results show that China’s economic growth is driven by the rapid devel-
opment of the non-agricultural sector, which results from the fast accumulation of
non-agricultural capital as well as expansion of non-agricultural employment. Nosotros
then evaluate the contribution of labor reallocation away from agriculture to non-
agriculture by applying the Labor Reallocation Effects equation speciªed by the
Banco mundial (1996). Our results suggest that labor reallocation has a positive impact
on China’s economic growth, accounting for 1.37 percent of per annum GDP
growth. This result coincides with the ªndings of Kuijs and Wang (2005), Cortejar (1998)
and the World Bank (1996). It conªrms the signiªcant contribution of sectoral labor
reallocation on the development of China’s dualistic economy. This ªnding also un-
dermines other explanations of China’s growth miracle, such as the gradualist ap-
proach by Lin, Cai, and Lee (1994) or the experimentalist strategy proposed
Naughton (1995), both of which make no reference to the important role of surplus
labor in enabling the fast growth.

The phases of China’s economic development, and particularly the Lewis turning
punto, have been studied using various methods but no deªnitive conclusions have

1 Throughout the paper we refer to the two sectors as agricultural and non-agricultural. Vari-
ous authors have used different terms interchangeably for these two sectors. Luis (1954)
originally named the two sectors as the subsistence and the capitalistic sectors and later in
Luis (1979) referred to them as the traditional and modern sectors. Jorgenson (1967, 291)
elaborates further on the distinction between the two sectors and narrows this down to the
stylized fact that the two sectors do not share the same production technology, particularly
when it comes to capital accumulation.

2 The Lewis turning point refers to the point beyond which a developing country has de-

pleted its surplus labor and its industry has to face rising wage rates (Yao 2011).

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

been reached. en este documento, we identify the phases of development and correspond-
ing turning points by examining the evolution of labor productivities over time as
indicated in the Lewis-Ranis-Fei model. We ªnd that the Chinese economy has fully
absorbed the redundant agricultural labor and entered phase two of development,
as shown by the rising marginal productivity of labor since the 1978 economic re-
formas. We also ªnd that labor’s marginal productivity has continued to rise and by
2009 may even have exceeded the institutional wage as deªned by the initial low av-
erage productivity of labor. This indicates that by 2009 China may near, or have ac-
tually reached, the Lewis turning point, which signals entry into phase three eco-
nomic growth as deªned in the Lewis-Ranis-Fei model.

The paper proceeds as follows. Sección 2 reviews the Lewis theory, the Ranis-Fei
modelo, and the related literature. Sección 3 discusses China’s dual-sector economic
development and rural–urban labor migration. Sección 4 presents the model spe-
ciªcations for estimating the production functions, decomposing dual-sectoral eco-
nomic growth rates and evaluating the effect of labor reallocation away from agri-
culture toward non-agriculture. Sección 5 explains the data in relation to China’s
employment, capital stock, labor migration, and technological progress. Sección 6
presents our estimation results. Sección 7 provides detailed numerical analysis re-
garding China’s sources of dual-sector economic growth, the role of labor realloca-
ción, and phases of development and turning points that China has passed through.
A ªnal section concludes and provides tentative policy recommendations.

2. Literature survey

2.1 The Lewis-Ranis-Fei model
The Lewis (1954) theory of dualistic economic development provides the seminal
contribution to theories of economic development particularly for labor-surplus and
resource-poor developing countries. In the Lewis theory, the economy is assumed to
comprise the agricultural and non-agricultural sectors. The agricultural sector is as-
sumed to have vast amounts of surplus labor that results in an extremely low, close
to zero, marginal productivity of labor. The agricultural wage rate is presumed to
follow the sharing rule and be equal to average productivity, which is also known as
the institutional wage. The non-agricultural sector has abundant capital and re-
sources relative to labor. It pursues proªt and employs labor at a wage rate higher
than the agricultural institutional wage by approximately 30 por ciento (Luis 1954,
150). The non-agricultural sector accumulates capital by drawing surplus labor out
of the agricultural sector. When the surplus labor is exhausted, the Lewis turning
point is reached and the economy completes the transition from a labor-surplus
stage to a labor-scarce stage of development.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 1. Lewis-Ranis-Fei phases of economic development in relation to agricultural labor
aporte (LA) and output (control de calidad)

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

Fuente: Ranis and Fei (1961, diagrama 1.3).

Más tarde, Ranis and Fei (1961) formalized the Lewis theory by combining it with
Rostow’s (1956) three “linear-stages-of-growth” theory. They subdivided Lewis’s
ªrst stage into two phases and make Lewis’s stage two correspond to their phase
tres. These three phases, illustrated in Figure 1, are deªned by the marginal pro-
ductivity of agricultural labor. The entry into each phase is formalized by three turn-
ing points:

• The breakout point leads to phase one growth with redundant agricultural labor.
• The shortage point leads to phase two growth with disguised agricultural

unemployment.

• The commercialization point, or Lewis turning point, leads to phase three self-

sustaining economic growth with the commercialization of the agricultural sector.

2.2 Relevant empirical studies on China’s growth
China’s remarkable economic performance has been closely studied and intensively
debated, with the focus being the identiªcation of the sources of China’s growth

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

miracle. The debate over China’s growth has been dominated by two schools of
thought and has evolved over time. Initially the debate was over the speed of the re-
formas, a saber, whether China’s rapid growth was attributable to its gradual and
incremental reforms or to its fast and comprehensive reforms. The accumulated evi-
dence from Eastern Europe started to show that slow-reforming economies gener-
ally performed worse than fast-reforming economies. Además, Vietnam provided
another case-study for an economy where rapid reforms led to rates of economic
growth similar to China’s. As a result of this evidence, the debate moved on to one
of interpreting the nature of China’s economic reforms. Most economists who had
favored the big-bang approach now view China’s high growth as a consequence
of the steady institutional convergence of China’s socialist institutions towards capi-
talist ones. Por otro lado, most economists who had favored the gradualist
explanation now regard China’s high growth rate a consequence of institutional
innovation. They argue that the gradual nature of the Chinese reforms was due to
economic experimentation, and the experiments succeeded in producing new non-
capitalist institutions that are at least as efªcient as capitalist ones. These gradual
innovations include 15-year land leases instead of land privatization, collec-
tively owned enterprises instead of private enterprises, and incentive-compatible
proªt-sharing contracts between state and state-owned enterprises (SOEs) en cambio
of privatization.

El (initially gradualist) experimentalist school is led by Lin, Cai, and Lee (1994),
Naughton (1995), and Nolan and Ash (1995). The convergence school is represented
by Fan (1990, 1994), Bruno (1994), and Sachs and Woo (1994, 2000). En esencia, el
gradualist school claims that China’s economic experimentation and innovation
have helped China evolve toward a unique Chinese economic model. En cambio,
the convergence school argues that China’s growth is the consequence of rapid con-
vergence from an initial labor-intensive peasant agrarian economy to an East Asian
market economy. Both schools stress the common point that market reforms have
spurred China’s economic growth.

We note that China’s economic reforms were ªrst initiated in the backward rural ag-
ricultural sector, which featured surplus labor, and then extended to the urban in-
Industrias. In our judgment, this reform sequence made the most of China’s dualistic
economy and thereby helped it to undertake its economic takeoff. A diferencia de,
the economic transition that occurred in the already industrialized and urbanized
Soviet economies led to economic downturns. Por lo tanto, it was the dualistic econ-
omy that allowed economies like China, Vietnam, and Laos to grow, regardless of
the speed of reform, during their transition from central planning. China’s growth
path is more relevant to the dualistic explanation, such as the Lewis-Ranis-Fei the-
ory, than other debates.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

The Lewis theory has been applied to several economies such as Japan, Korea, y
Taiwan with various degrees of success (p.ej., Ohkawa 1965; A 1972; Fei and Ranis
1973). In view of China’s remarkable economic growth, the Lewis theory has re-
cently been applied to its analysis, with particular focus on identifying the Lewis
turning point, but no deªnitive consensus has been reached. Por ejemplo, Cai (2007)
argues that China had already reached the Lewis turning point. This was illustrated
by increases in the wages of the rural migrant wages and the phenomenon of “rural
labor-scarcity,” which occurred in the Zhujiang triangle coastal area in 2003. Este
was also accelerated by the substantial decline in China’s population growth during
the demographic transition.

Cai’s view is supported by scholars like Park, Cai, and Du (2010) and Wang (2008),
while challenged by other such as Meng and Bai (2007), Caballero (2007), Yao (2011),
and Knight, Deng, and Li (2010). Knight Deng, and Li (2010) argue that evidence on
the rising wages of migrants is subject to sample changes. Yao (2011) points out that
recent agricultural wage increases are largely due to reductions in agricultural taxes
en 2006, direct subsidies to grain production, and high price inºation for food. Vari-
ous studies even indicate that China is a long way from the Lewis turning point. Para
instancia, Caballero y otros. (2010) reports that in the 2002 y 2007 national household
surveys, no signiªcant rise was observed in migrants’ wages and that a substantial
pool of potential migrant workers, aproximadamente 80 millón, is still available in ru-
ral areas. Yao and Zhang (2010) conªrm Knight’s ªnding and argue that it may not
be until 2018 that all of China’s potential migrants are fully absorbed.

Disagreement between current studies is largely attributable to the diverse
methods3 used to identify China’s Lewis turning point. Many studies (see Cai 2007;
Caballero 2007; Meng and Bai 2007) examine changes in rural household’s net incomes
and migrants’ wages, whereas others (see Kwan 2009; Caballero, Deng, and Li 2010;
Minami and Ma 2010; Yao and Zhang 2010) estimate the amount of surplus labor in
rural China. There are also studies (Cai and Wang 2008; Yao 2011) that estimate ur-
ban labor demand or the supply and demand functions for migrant workers.
Studies based on estimates of labor supply and demand functions rely on rural
wages, and therefore, they share the same problems as studies that focus directly on
rural wages. These problems were ªrst discussed by Minami (1968, 384):

when there is a rising trend in the real wage, we cannot ascertain straightfor-
wardly whether that increase comes from a change in the marginal productivity
of labor or from an increase in the subsistence level itself.

3 Similar discussion is also provided by Knight, Deng, and Li (2010, 7–9).

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Identifying the development turning point by measuring the amount of rural sur-
plus labor seems close to the basic premise of the Lewis theory. Sin embargo, el
theory of surplus labor refers to the situation where the marginal productivity of la-
bor is lower than the institutional wage. Por lo tanto, surplus labor need not be limited
to rural labor, it could also exist in urban areas too, but this fact is difªcult to verify
en la práctica.

Sin embargo, one technique that applies the Lewis-Ranis-Fei framework most directly
to identifying the phases of economic development has, hasta la fecha, not been applied to
the Chinese case. This technique examines the change in the marginal productivity
of agricultural labor over time and was ªrst applied by Minami (1968) to study Ja-
pan’s agricultural commercialization. It not only provides intuitive implications on
the phases of economic development but also identiªes the time when the Lewis
turning point is achieved. en este documento, we apply this technique to identify whether
China has reached the Lewis turning point.

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

3. The Chinese experience

3.1 China’s dualistic economic development
China has had a long history of dualistic economic development. According to
Putterman (1992), prior to the 1978 economic reforms, the rural agricultural sector
was run using collective farms and wages were set by the government. The pursuit
of proªt was only allowed in the urban industrial sector. El 1978 economic reforms
did not bring this dualistic structure to an end. Instead it allowed the urban sector to
develop further by creating a dynamic service sector and a new class of TVEs. De este modo,
China’s dualistic structure involves the agricultural sector in rural areas and the
non-agricultural sector4 in mainly urban areas.

As shown in Figure 2, the contribution of agricultural production to GDP falls from
aproximadamente 39 percent in the 1960s to 10 por ciento en 2009, whereas the contribu-
tion of non-agricultural production rises from 61 por ciento a 90 percent during the
same period. Además, the non-agricultural sector grows at a faster pace than the
agricultural sector. This indicates that economic growth in China is driven more by
the non-agricultural sector than the agricultural sector.

4 Más precisamente, the agricultural sector includes farming, animal husbandry, forestry, y
ªshery. The non-agricultural sector includes construction, industria (es decir., manufacturing,
mining and quarrying, electricidad, gas and water supply), transport, post and telecommuni-
cation services, wholesale and retail trade, and catering services. The output of TVEs is in-
cluded in the non-agricultural sector, though they are in semi-urban locations.

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 2. Development of the Chinese dual-sector economy (1960–2009)

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

Fuente: World Development Indictors (Banco mundial 2010).

3.2 China’s sectoral labor reallocation
China is a labor-surplus economy and most of this surplus is engaged in the agricul-
tural sector. Before the 1978 economic reforms, labor mobility was controlled by the
government through the “Hukou system.” According to Zhao (2000), the average
annual rural–urban migration rate was only 0.24 percent in 1949–85, much lower
than the world average rate of 1.84 percent in 1950–90. Since the early 1980s, the re-
strictions on labor mobility have been relaxed to accommodate labor demand in the
non-agricultural sector. Sin embargo, the one-child policy introduced in the 1970s has
been imposed more stringently, particularly in urban areas. This has slowed the
growth of the urban-born labor force and aggravated the labor shortage in the non-
agricultural sector (Caballero 2007). Recently the restrictions on labor mobility have
been relaxed and increasing numbers of rural laborers have migrated to towns and
cities. Como resultado, relative employment in the agricultural sector (illustrated in Fig-
ura 3) dropped from 70 por ciento en 1978 a 38 por ciento en 2009. Correspondingly, em-
ployment in the non-agricultural sector rose rapidly and reached 62 percent of total
employment. Note that even with the relaxation of restrictions on labor mobility,
most of the migrants are only allowed into cities on a temporary basis.

There are no consistent and reliable data on China’s rural–urban labor migration
con el tiempo. Intermittent migration data are collected in population censuses at 8- a

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 3. Employment in the two sectors

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

Fuente: Data before1965 are from China Statistical Yearbook (NBS 2004), and data after 1965 are from the World Development Indicators

(Banco mundial 2010).

10-year intervals, or in surveys covering a few provinces. Many studies (p.ej., Wu
1994; Zhang and Song 2003) apply the residual method suggested by the United
Nations (1970) to derive a consistent time series for China’s rural–urban labor mi-
gration. This method assumes that without international labor migration, el en-
crease in urban population is attributable to the natural growth of the urban popula-
tion and net rural-to-urban migration. De este modo, net labor migration can be derived by
subtracting the natural population growth from the aggregate population increase
in urban areas. Zhang and Song5 (2003, Mesa 1) apply this method and compute the
series for rural–urban labor migration in 1978–99. This is illustrated in Figure 4. El
abrupt drop in labor migration during 1989–91 may be due to events following the
Tiananmen Square incident. Yao (2011, Cifra 6) presents a time series for rural–ur-
ban migration in recent decades. As shown in Figure 4, the series reveals a large dis-

5 Zhang and Song (2003) compute the natural growth of urban population as the product of
the total urban population and the natural urban population growth rate, which is proxied
by the ofªcial “natural city growth rates.” The data for the natural city growth rates in 1978–
82 and 1988–99 are sourced from the NBS Statistical Yearbook (2000). For the missing data in
1982–88, they use a combination of correlations with city growth and projections from the
available years. Similar patterns of the rural–urban labor migration are observed in the data
generated by Wu (1994, Cifra 4).

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 4. China’s rural–urban labor migration (1978–2009)

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

Fuente: Data for the period 1978–92 are from Zhang and Song (2003, Mesa 1). Data for 1993–2009 are from Yao (2010, Cifra 5).

crepancy, compared to that computed by Zhang and Song (2003), in the overlapping
period 1993–99.

4. Model speciªcation

In this section we introduce the production function speciªcations of the two sec-
tores, the equations used in the growth decomposition to identify phases of develop-
ment and the equations used to compute the contribution to growth of labor reallo-
cation away from agriculture.

We assume a dualistic economic framework with the agricultural and non-agricul-
tural sectors representing the traditional and modern sectors of the Lewis theory.
Respectivamente, agricultural output (control de calidad) is a function of agricultural hectares (HA), la-
bor input (LA), and agricultural capital (EL). Output in the non-agricultural sector
(QN) depends on employed labor (LN) and capital stock (KN). Both production func-
tions feature Hicks neutral technological progress (fA(t), fN (t)) where T denotes
tiempo; the exact functional form of these contain trends that reºect socio-economic
events and possibly dummies for structural shifts. The resulting Cobb–Douglas pro-
duction functions for the agricultural and non-agricultural sectors are

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

f L K H T
(
A
A

)

(

)

L H K

a
l
A

a
A

norte

g L K T
,
(
norte

norte

)

norte

(

)

b
L K
l
norte

b
norte

By taking logarithms, we derive the log-linear forms in equations (3) y (4). El
parameters with a hat “ˆ ” are those to be estimated:

lnQA

(cid:2) ln (cid:2)a0

(cid:3) ƒA(t) (cid:3) (cid:2)aL lnLA

(cid:3) (cid:2)aH lnHA

(cid:3) (cid:2)aK lnKA

(cid:3) eA ,

lnQN

(cid:2) ln

(cid:2)
b0

(cid:3) ƒB(t) (cid:3) (cid:2)

bL lnLN

(cid:3) (cid:2)

bK lnKN

(cid:3) eN .

(1)

(2)

(3)

(4)

(cid:3)

We test for, but do not impose, constant returns to scale in each sector by the condi-
(cid:2) 1. We differentiate functions (3) y (4) con
ciones
respect to time and obtain the following equations for decomposing sectoral eco-
nomic growth rates:

(cid:2) 1 y yo

(cid:3)

gramo

(

)

A
¶

(cid:2)
a

gramo

(cid:2)
a

gramo
H H

(cid:2)
a

gramo
K K

gramo

norte

(

)

norte
¶

(cid:2)
b

gramo

norte

(cid:2)
b

gramo
K K

norte

(5)

(6)

where the exponential growth rates for each factor X is calculated by either the in-
stantaneous percentage growth rate in continuous time,

gramo

registro
dt

⋅

100

X
(registro
2009
2009
(

−
−

X
registro
1965
)

1965

)

⋅

100

or the true annual compounded percentage growth rate in discrete time,

AGR (cid:2) (expgx

(cid:4) 1) 100.

In empirical studies, the annual compounded growth rate (AGR) is often used to
represent the exponential growth rate. Sin embargo, in theoretical models growth rates
are usually expressed in continuous time by gX. Unless the actual rate of growth is
extremely high, the discrete-time measures of growth are very close to the continu-
ous-time measures of growth. Finalmente, note that the time derivatives with respect to
the Hicks-neutral technological changes (fA(t), fN(t)) are captured by the appropri-
ate time-trends and time-dummies in the estimated models.

We apply the labor reallocation effect (LRE) approach suggested by the World Bank
(1996) to measure the contribution made to growth by the migration of laborers

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

from the agricultural to the non-agricultural sector.6 According to the World Bank,
the agricultural labor reallocation effect is deªned as:

LRE

l
Y

(

MPL MPL g l
NN

)

norte

−

where l

norte

l
N=
l

(7)

in which MPLN and MPLA refer to the marginal productivities of labor in the agricul-
tural and non-agricultural sectors, respectivamente. They are calculated by

MPL

norte

dQ
dL
A
dQ
dL

norte

= (cid:2)
a

= (cid:2)
b

q
l
A
q
l

norte

(cid:2)
a

APL

(cid:2)
b

APL

norte

(8)

(9)

(cid:2)
bL are the estimated parameters in equations (3) y (4), and APLA and

dónde (cid:2)aL and
APLN refer to the respective average productivities of labor in the agricultural and
non-agricultural sectors respectively. Ecuación (7) shows that a reallocation of labor
away from agriculture will have a positive net effect on growth so long as the value
of the marginal productivity of labor in the non-agricultural sector exceeds that in
the agricultural sector. The size of this effect depends on how much more produc-
tive the non-agricultural sector is and on how large the share of non-agricultural la-
bor (lN) in the total labor force is (Banco mundial 1996, 67–68). En esencia, the approach
uses the difference in the labor productivities of the two sectors to quantify the con-
tribution of sectoral labor reallocation to economic growth.

5. Los datos

Our data span the period 1965–2009 and are mainly from the World Bank’s 2010
World Development Indicators. Data on China’s total employment and percentages
of sectoral employment are from China Statistical Yearbooks (2009) by China’s Na-
tional Bureau of Statistics (NBS) y el 2009 annual report by China’s Ministry of
Human Resources and Social Security (MOHRSS). Data on agricultural capital and
on arable land and cropland are from the United Nation’s Food and Agricultural
Organization (FAO) online statistics database.7 Output and capital stock values are
in real RMB deºated to 2000 prices. Mesa 1 provides summary statistics and vari-
able descriptions.

6 This approach has also been used to measure the effect of labor reallocation from the state to

the non-state sector.

7 See www.faostat.fao.org.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Mesa 1. Summary statistics

Variables

control de calidad

LA
EL
HA

T1979–84

T1985–92

T1993–on

LN
KN

T1984–on

D1967

D1968

Significar
9.6(cid:5)1011

3.32(cid:5)108
756,452.4
1.18(cid:5)107

3.80

3.82

3.40

mín.
2.25(cid:5)1011

2.55(cid:5)108
73,021
1.00(cid:5)107

máx.
2.52(cid:5)1012

3.80(cid:5)108
3,010,658
1.34(cid:5)107

4.84(cid:5)1012

3.43(cid:5)1011

2.18(cid:5)1013

2.33(cid:5)108
9.72(cid:5)1012

5.76(cid:5)107
3.79(cid:5)1011

4.83(cid:5)108
4.56(cid:5)1013

7.8

0.02

Fuente: Authors’ calculations.

Descripción

Agricultural output: in yuan at constant 2000
prices.

Agricultural labor: total workers.

Agricultural capital: total number of tractors.

Agricultural hectares of arable land and
cropland.

Agricultural technological trend: trend starts
en 1979 and stops increasing in 1984, es igual
zero before 1979.

Agricultural technological trend that starts in
1985 and ends in 1992. It equals zero before
1985 and equals 8 de 1992 en.

Agricultural technological trend that starts in
1993 and equals zero before 1993.

Non-agricultural output: in yuan at constant
2000 prices.

Non-agricultural labor: total workers.

Non-agricultural capital: calculated by the
perpetual inventory method, 2000 prices.

Non-agricultural technological trend that
starts in 1984 and equals zero before 1984.

Cultural Revolution dummy: equals one in
1967, zero otherwise.

Cultural Revolution dummy: equals one in
1968, zero otherwise.

Agricultural and non-agricultural outputs are derived from the multiplication of the
relative sectoral shares value-added in GDP by the real values of GDP. The data for
China’s total employment by the NBS show a spurious jump in 1990 due to ac-
knowledged statistical adjustments. Such a jump is not revealed in the employment
series by MOHRSS, but this is only available up to 2002. To generate a complete and
consistent series for employment, we rescale the data from NBS for the period 1965–
95 using the average ratio of the two employment series. De este modo, sectoral employment
series are derived by multiplying the total employment data by the sectoral employ-
ment shares. Agricultural capital is represented by the number of tractors, cual
is consistently available over the entire sample period from the FAO. Capital stock
in the non-agricultural sector is obtained by applying the conventional Perpetual
Inventory Method with an annual depreciation rate of 10 por ciento. The initial value
of the capital stock is set at 378,847,553,230 yuan at 2000 constant prices based on
the calculation by King and Levine (1994) that non-agricultural capital was equal
a 63.9 percent of GDP in 1965. The two series of capital stock are illustrated in
Cifra 5.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 5. Capital stock (1965–2009)

Fuente: Agricultural capital represented by the number of tractors is from World Development Indicators (Banco mundial 2010). No-

agricultural capital is authors’ calculation, based on the World Development Indicators (Banco mundial 2010) and Penn World Tables

(Heston, Summers, and Aten 2002).

We model technological progress in the agricultural sector by three segmented de-
terministic time trends. Following the work of Sachs and Woo (1994), the ªrst trend
covers 1979–84 to capture the decentralization of farming and the second trend cov-
ers 1985–92 to capture the introduction of the market system in the rural economy.
The third trend, starting in 1993, captures the slight slowdown in agricultural tech-
nological progress. No technological trend is included before 1979, it being well es-
tablished that agricultural technological progress had been negligible due to
destabilizing socio-economic events (see Chow 1993).

Technological progress in the non-agricultural sector is modeled by a time trend
para 1984 onward. This trend is justiªed by the fact that experimental reforms
on SOEs began in August 1980 and this translated into widespread reforms by
1984. Political events surrounding the Cultural Revolution justify the inclusion
of year dummies for 1967 y 1968. Other events may have justiªed the in-
clusion of year dummies for 1976 and 1990–91 but this would have removed al-
most all dynamics from the model and would necessitate a substantial number
of dummies.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

6. Estimates of the production functions

6.1 Stationarity tests
Before estimating the production functions, we test the stationarity of variables us-
ing Augmented Dickey-Fuller (ADF) (Dickey and Fuller 1979, 1981) and KPSS
(Kwiatkowski et al. 1992) pruebas. The ADF tests are for the null hypothesis that the se-
ries are non-stationary, the KPSS tests are for the null hypothesis that the series are
stationary. The results of these tests are reported in Table 2 and they suggest, en el
5 percent signiªcance level, that all the variables are non-stationary and integrated
of order one I(1). The one exception is the log of agricultural capital that is border-
(cid:6) I(1/2), but it seems that this am-
line integrated of order one I(1) or two I(2), lnKA
biguity may be due more to the long cycle in the data rather than it being I(2).
Aware of the non-stationarity in the data we take steps to address it in the estima-
tion of the models.

6.2 Result estimates of the production functions
We run regressions on the data described in the previous section to estimate the
log-linear production functions in equations (3) y (4). We estimate these pro-
duction functions8 by Ordinary Least Squares (OLS), Generalized Least Squares
(GLS), and Maximum Likelihood (ML) with robust t-tests based on White’s (1984)
heteroscedasticity-consistent standard errors. Regression results are reported in
Tables 2 y 3.

The OLS production function estimates are reported in columns (1) in Tables 3 y
4, and they represent our initial base cases. The estimated elasticity parameters seem
reasonable as do the technological trend parameters. F-tests suggest that neither sec-
tor exhibits constant returns to scale. The diagnostics on the residuals highlight two
problems not uncommon to time-series regressions. The ªrst is the large degree of
residual serial correlation in the agricultural sector and the second is the heterosce-
dasticity in the non-agricultural production function. The heteroscedasticity has al-
ready been accounted for by using the White (1984) heteroscedasticity-consistent
standard errors for the t-tests and F-tests. The autocorrelation is accounted for in the
GLS and ML estimates that follow.

8 We estimated the production function in the agricultural sector by involving fertilizer con-

sumption and irrigation but the results suffered from severe multi-collinearity problems. Nosotros
therefore settled on the parsimonious parameterization reported in Table 3. Note also that
although it has been suggested that the panel estimates could have been carried out using
provincial-level data, the data for some variables, Por ejemplo, agricultural machinery, son
not available before 1978 across provinces. In that case, the sample period would not be long
enough to test the Lewis theory, nor would it be long enough to identify the stages of eco-
nomic development in China.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Mesa 2. Stationarity tests on variables

ADF on
nivel
(cid:4)0.042
(cid:4)1.888
(cid:4)0.779
(cid:4)1.070
(cid:4)1.083
(cid:4)1.847
(cid:4)0.341

ADF on
diferencia
(cid:4)5.803
(cid:4)2.266
(cid:4)0.908
(cid:4)6.166
(cid:4)5.016
(cid:4)3.962
(cid:4)4.295

ADF
resultado

I(1)
I(1)/I(2)
I(2)
I(1)
I(1)
I(1)
I(1)

KPSS on
nivel

1.78
0.97
1.37
1.38
1.78
1.75
1.78

KPSS on
diferencia

0.0645
0.805
0.409
0.274
0.237
0.489
0.229

lag

2
2
2
2
2
2
2

lag

2
2
2
2
2
2
2

KPSS
resultado

I(1)
I(1) /I(2)
I(1)
I(1)
I(1)
I(1)/I(2)
I(1)

Vars

ln QA
ln LA
ln KA
ln HA
ln QN
ln LN
ln KN

Fuente: Authors’ calculations.

Nota: ADF(norte): Augmented Dickey-Fuller test with n autoregressive lags. Reported value is t-statistic on lagged levels variable. Null
hypothesis is that the variable contains a unit root (is non-stationary). Critical values are: (cid:4)2.608 en el 10 percent signiªcance level,
(cid:4)2.950 en el 5 percent signiªcance level, y (cid:4)3.628 en el 1 percent signiªcance level.
KPSS: Kwiatkowski et al. (1992) prueba. Null hypothesis is that the variable does not contain a unit root (is stationary). Optimal lag-

length is chosen by the Newey–West (1994) automatic bandwidth selector applied by Hobijn, Philip, and Ooms (1998). Critical values

son 0.347 en el 10 percent signiªcance level, 0.463 en el 5 percent signiªcance level, y 0.739 en el 1 percent signiªcance level.

Mesa 3. Agricultural production function estimates

Dependent variable: lnQA

Estimation method

Variables

lnLA

lnKA

lnHA

T1979–84

T1985–92

T1993–on

Constant

AR(1)

Observaciones
R2

Constant returns to scale test

Non-structural residual (et) diagnostics:
Breusch-Godfrey autocorrelation LM (cid:7)2 prueba

White heteroscedasticity (cid:7)2 prueba

Jarque-Bera normality (cid:7)2 prueba

Ramsey Reset F test

Fuente: Authors’ calculations.

(1)

OLS

0.369*
(2.21)

0.178**
(9.72)
(cid:4)0.131
((cid:4)0.65)

0.088**
(9.90)

0.060**
(6.53)

0.038**
(8.30)

19.475**
(3.86)

0.9952
F (cid:2) 4.736
[0.036]

16.188
[0.000]

4.271
[0.118]

2.184
[0.336]

18.61
[0.000]

(2)

GLS

AR(1)

0.341
(1.46)

0.181**
(5.82)

0.027
(0.10)

0.089**
(6.77)

0.053**
(4.63)

0.040**
(6.12)

17.077*
(2.48)

0.626**
(5.37)
45

0.9993
F (cid:2) 1.515
[0.226]

2.9342
[0.087]

2.230
[0.328]

1.914
[0.384]

18.27
[0.000]

(3)

AR(1)

0.340
(1.53)

0.181**
(6.23)

0.028
(0.10)

0.089**
(6.57)

0.053**
(4.37)

0.040**
(6.38)

17.076**
(2.61)

0.628**
(4.05)
45

(cid:7)2 (cid:2) 1.680
[0.195]

2.041
[0.153]

1.268
[0.530]

2.075
[0.354]

10.12
[0.000]

Nota: * y ** denote signiªcance at the 5 percent and the 1 nivel porcentual, respectivamente.

Parentheses surround t statistics, square brackets represent densities in the tail of each distribution for rejection of the respective null

hypotheses. OLS, GLS, and ML estimates of t statistics are based on the White (1984) robust covariance estimator.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

The GLS and ML estimates reported in columns (2) y (3), respectivamente, of Tables 3
y 4 are for models that accommodate ªrst order autoregression, AR(1), en el
structural residuals. Ecuaciones (10) y (11) illustrate how AR(1) in the structural re-
siduals is accommodated by adding a second equation to the production function:

lnQt

(cid:2) (cid:2)

fL lnLt

(cid:3) (cid:2)

fK lnKt

(cid:3) . . . (cid:3) ut,

(cid:2) (cid:2)rut(cid:4)1

(cid:3) et,

(10)

(11)

where ut are the structural residuals and et are the non-structural residuals. Estos
equations are valid for both the agricultural and non-agricultural production func-
tions in equations (3) y (4). The GLS estimator is based on the Cochraine–Orcutt
(1949) iterative procedure with the Prais–Winsten (1954) transformation to retain the
ªrst observation. The ML estimator is based on a uniªed log-likelihood equation
that subsumes equations (10) y (11) into one. The parameter estimates in the GLS
and ML estimates are similar to one another. This indicates that the estimates are
robust to the estimation method. We expect that the GLS and ML parameter esti-
mates are slightly better deªned than the OLS ones because the error autocorrela-
tion has been corrected for. The structural residuals have signiªcant autoregressive
parameters of magnitude 0.626 y 0.628 in agriculture, y 0.315 y 0.309 in non-
agricultura. The diagnostics now pass the Breusch–Godfrey (Breusch 1978; Godfrey
1978) AR(1) test suggesting the non-structural residuals are, apart for the heterosce-
dasticity, white noise. There is evidence of non-normality in the residuals of the non-
agricultural production function but this is due to large negative socio-economic
shocks associated with 1968, 1976, y 1990.

Model estimates on non-stationary series may lead to spurious regressions. Allá-
delantero, we carry out Engle–Granger (1987) co-integration tests on the two estimated
least squares production functions to verify that these are not spurious regressions.
The Engle–Granger test results indicate that for both sectors the null hypothesis
of non-stationarity on the residuals can be rejected at the 10 percent signiªcance
nivel. The computed tau statistics of the ADF tests on the residuals are –3.335 and
–4.995 for the agricultural and non-agricultural sectors, respectivamente. Ellos son
lower than the Engle–Granger co-integration critical values at the 10 percent and
5 percent signiªcance levels.9 This implies that the series, despite being non-
stationary, are co-integrated and the estimates on the non-stationary variables are
reliable and superconsistent.

9 The estimated error correction parameters on the lagged error correction terms are –0.404

and –0.740 respectively, indicating a more rapid speed of adjustment in the non-agricultural
production than agricultural production.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Mesa 4. Non-agricultural production function estimates

Dependent variable: lnQN

Estimation method:

Variables:

lnLN

lnKN

T1984

D1967

D1968

Constant

AR(1)

Observaciones
R2

Constant Returns to Scale test

Non-structural residual (et) diagnostics:
Breusch-Godfrey autocorrelation LM (cid:7)2 prueba

White heteroscedasticity (cid:7)2 prueba

Jarque-Bera normality (cid:7)2 prueba

Ramsey Reset F test

Nota: Same conventions as Table 3.

(1)

OLS

0.732**

(12.82)

0.208**
(5.59)

0.058**

(25.41)
(cid:4)0.148**
((cid:4)7.94)
(cid:4)0.268**

((cid:4)15.06)

8.015**

(19.80)

0.9993
F (cid:2) 5.015
[0.031]

3.049
[0.081]

4.942
[0.085]

2.635
[0.268]

1.91
[0.146]

(2)

GLS

AR(1)

0.734**
(9.61)

0.212**
(4.40)

0.057**

(20.13)
(cid:4)0.140**
((cid:4)6.32)
(cid:4)0.238**
((cid:4)5.47)

7.892**

(15.59)

0.315*
(2.18)

0.9993
F (cid:2) 2.332
[0.135]

1.222
[0.269]

5.644
[0.059]

2.609
[0.271]

1.91
[0.145]

(3)

AR(1)

0.734**

(10.31)

0.212**
(4.74)

0.057**

(20.67)
(cid:4)0.140**
((cid:4)5.19)
(cid:4)0.238**
((cid:4)4.37)

7.894**

(14.60)

0.309
(1.26)

(cid:7)2 (cid:2) 2.395
[0.122]

0.921
[0.337]

5.079
[0.079]

2.473
[0.276]

1.94
[0.140]

All the results are consistent with each other within each sector. All estimates seem
reasonable with most diagnostic tests being passed. Given that all the structural pa-
rameter estimates are so similar, within each sector the growth decomposition anal-
ysis and other analyses could be carried out equally with any set of parameter esti-
compañeros. Por lo tanto, we opt to use the ML estimates for the analyses that follow, como
these estimates represent the most parsimonious model estimates that satisfy all the
diagnostic tests.

7. Empirical analysis on sectoral growth

7.1 Sources of China’s dual-sector economic growth
Ecuaciones (5) y (6)10 are used to decompose China’s sectoral economic growth
and display the results in Table 5. We ªnd that the 4.84 percent exponential annual

10 In the literature, growth accounting is often applied to decompose economic growth.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

growth rate of non-agricultural labor is much higher than the 0.34 percent rate of ag-
ricultural labor. Capital inputs in both sectors rise rapidly, 10.64 percent in the non-
agricultural sector and 8.28 percent in the agricultural sector. Agricultural land,
sin embargo, remains relatively constant, changing by an annual mean of just 0.30 por-
cent during 1965–2009. Además, el 5.99 o 5.78 percent annual economic
growth in the non-agricultural sector is over threefold that of the agricultural sector
en 1.80 percent when measured by instantaneous growth rates, o 1.87 percent when
measured by annually compounded growth rates. This implies that economic
growth is mainly driven by the expansion of the non-agricultural sector, as sug-
gested by the Lewis theory. Además, we ªnd that growth in the agricultural sector
is mainly driven by capital accumulation (83.06 por ciento), whereas growth in the
non-agricultural sector is attributable to both labor accumulation (59.99 por ciento) y
capital accumulation (39.02 por ciento). In both sectors, technological progress, a pesar de
being statistically signiªcant in the estimation, only accounts for a relatively small
share of economic growth. These results show that China’s economic growth is due
largely to factor accumulation. This ªnding is consistent with most studies in the lit-
erature (p.ej., Cortejar 1998). Además, we note that capital accumulates at a faster
rate per annum than labor, whereas it accounts for a relatively smaller share of
growth in the non-agricultural sector. This indicates that the expansion of labor
employment plays a prominent role in promoting the development of the non-
agricultural sector. En resumen, consistent with the Lewis-Ranis-Fei theory, China’s
economic growth is driven by the rapid expansion of the non-agricultural sector,
which is affected by capital accumulation and employment growth fuelled by sec-
toral labor reallocation.

7.2 The contribution of sectoral labor reallocation
We apply equation (7) to account for the sectoral labor reallocation effect (LREWB)
suggested by the World Bank (1996). The merit of this approach is that it is inde-
pendent of the number of migrant laborers and therefore one does not need to know
the number of migrant workers to measure the contribution of sectoral labor reallo-
cation to growth. As shown in Figure 6, the reallocation of labor away from agricul-
ture has had a positive effect on China’s economic growth for most of 1966–2009 ex-
cept for a few years like 1967–68, 1989–90, and 1999–2000, associated with speciªc
socio-economic disturbances. The computations suggest that sectoral labor realloca-

Sin embargo, it is well established that growth accounting has many drawbacks. Por ejemplo, él
treats the contribution other than that by factor input as the total factor productivity. Él
therefore cannot distinguish the pure effect of technological progress on growth. Además,
the result is subject to the assigned input shares. en este documento, we carefully estimate the in-
put elasticity and decompose economic growth by factor contributions. Chow and Li (2002)
and Ho (1972) have used this approach to decompose economic growth in their studies.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Mesa 5. Dual-sector growth decomposition (1965–2009)

Parameter
estimados

Instantaneous
annual growth
rate gX*
[or AGR**]
%

Product of
parameter
and growth
[AGR product]

(1)

0.340

0.181

0.028

0.089

0.053

0.040

0.734

0.212

0.057

Agricultural sector:

Labor

Capital

Land

T1979–84

T1985–92

T1992_on

Total

Non-agricultural sector:

Labor

Capital

T1984_on

Total

(2)

0.34
[0.34]

8.28
[8.63]

0.30
[0.30]

4.73
[4.84]

10.64
[11.23]

(3)

0.115
[0.115]

1.498
[1.562]

0.008
[0.008]

0.089

0.053

0.040

1.803
[1.867]

3.469
[3.552]

2.257
[2.381]

0.057

5.783
[5.991]

Contribution
to sectoral
growth
[AGR version]
%

(4)

6.39
[6.18]

83.06
[83.63]

0.45
[0.44]

4.94
[4.77]

2.94
[2.84]

2.22
[2.14]

100
[100]

59.99
[59.30]

39.02
[39.75]

0.99
[0.95]

100
[100]

Fuente: Authors’ calculations.

Nota: Columna (1): Coefªcients are from the ML estimates for both sectors and are taken from column (3) in Tables 3 y 4.

Columna (2): *Instantaneous annual growth rates in column (2) are derived by gX
represents each of the inputs in turn: Labor, Capital, or Land.
**AGR is the annually compounded growth rate given by AGR(cid:2)(exp gX
Columna (3): Each value is simply the product of the value in columns (1) y (2) and represents the components in equations (5) y

(cid:4)1)(cid:5)100, values reported in square brackets.

(cid:4)lnX1965)/(2009(cid:4)1965)](cid:5)100, where X

(cid:2) [(lnX2009

(6).

Columna (4): The contribution to sectoral growth calculated as the corresponding value in column (3) divided by the respective total for

columna (3) in each sector.

tion contributes on average 1.37 percent to annual economic growth over the period
1966–2009. This amounts to approximately 81.34 billion yuan per annum to China’s
real GDP. Our ªnding is close to the 1.3 percent annual contribution to growth in
1985–93 reported by Woo (1998) and is consistent with other studies. Por ejemplo,
the World Bank (1996) found that the effect of labor reallocation away from agricul-
ture accounted for 1 percent of China’s rapid economic growth during 1985–94. Cai
and Wang (1999) report a 1.62 percent average contribution in 1982–97.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 6. The labor reallocation effect by the World Bank (1996) acercarse

Fuente: Authors’ calculations.

Además, En figura 6, we ªnd that the contribution of sectoral labor reallocation
achieved its highest values, en 7.09 y 3.93 por ciento, in 1978–84. This is closely asso-
ciated with the boom of the TVEs, which has effectively absorbed large amounts of
rural surplus laborers (Caballero 2007). Sin embargo, since the mid-1990s, the contribution
to the growth of labor reallocation has declined, indicating that labor absorption in
the non-agricultural sector has slowed down. This ªnding is supported by Kuijs and
Wang (2005), who also detect the slower pace of expansion in urban employment
desde 1993. The demise of TVEs as well as the stable share of industry employment
in the 1990s contributes to the decline in labor reallocation effect. Además, el
low values of the reallocation effect in 1998–2002 may reºect the adverse impact of
the 1997–98 Asian ªnancial crisis on China’s exports. Para concluir, we ªnd that the
reallocation of labor away from agriculture has made a substantial contribution
to China’s economic growth. This ªnding agrees with the core of the Lewis-Ranis-
Fei theory.

7.3 Phases in China’s economic development and the turning points
As suggested by Minami (1968), we identify the phases in China’s economic devel-
opment, as well as the turning points, by comparing the agricultural marginal pro-
ductivity of labor (MPL) to the agricultural institutional wage. Recall from Section
2.1 that in the Lewis-Ranis-Fei model the initial agricultural institutional wage is
presumed to follow the sharing rule and be equal to the initially low average agri-

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 7. APLs and MPLs

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

Fuente: Authors’ calculations based on the estimates in Table 3.

cultural labor productivity. In the model, an economy is deªned as entering phase
two of economic development if the agricultural MPL starts to increase above its
near-zero value but is still lower than the low agricultural institutional wage. El
advantage of this method is that it does not require knowledge of sometimes un-
available measures such as rural wages and surplus labor. Admittedly, the applica-
tion of this technique is sensitive to the speciªcation of the agricultural institutional
wage. We address this drawback by comparing the agricultural MPL to various
proxies for the institutional wage. We represent three alternative institutional wages
by using the agricultural APL averaged over three time periods: 1965–78, justo 1978,
and 1965–2009. These three proxies allow us to verify the sensitivity of the results to
the choice of institutional wage. entre ellos, the mean value of the agricultural
APL over the entire sample period of 1965–2009 is regarded as the best proxy in that
it accommodates potential increases in the subsistence level itself, as suggested by
Minami (1968).

We compute the MPL in the agricultural and non-agricultural sectors using the esti-
mated elasticities of output to labor using the estimates for equations (8) y (9), re-
spectively. The series for MPL, APL, and the institutional wage are illustrated in Fig-
ures 7 y 8. En figura 7 it is obvious that in both sectors, both the marginal and

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Cifra 8. Agricultural APL and MPL

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

Fuente: Authors’ calculations based on the estimates in Table 3.

average productivities of labor are stagnant before the 1978 economic reforms and
then rise rapidly, particularly in the non-agricultural sector. The agricultural MPL is
low before 1978 but begins to rise rapidly after the 1978 economic reforms. The ris-
ing agricultural MPL indicates that the redundant labor has been reallocated away
from agriculture and the shortage point (ver figura 1) has been passed after 1978.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

Cifra 8 allows for the comparison of the agricultural MPL to various proxies for the
agricultural institutional wage. It appears that by 2009, at the latest, the agricultural
MPL surpassed the assumed institutional wage, as measured by various criteria,
thus suggesting that China has reached the Lewis turning point of economic devel-
opment. Por ejemplo, the MPL reaches the institutional wage assumed as the mean
value of the 1965–78 agricultural APL in 1994; whereas it reaches the 1978 APL in
1995. When comparing this with our favored proxy for the institutional wage,
a saber, mean agricultural APL over the entire sample period 1965–2009, we ªnd
that the rising MPL crosses it in 2009.

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

Notwithstanding our results, we also note that China is a large and diverse econ-
omy. De este modo, different Chinese regions may pass through the Lewis turning point in

Asian Economic Papers

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

different years. Por eso, entry into phase three economic growth may span several
years rather than just one point in time. This way of looking at the Lewis turning
point as occurring over a number of years is also adopted by Minami (1968),
Garnaut and Song (2006), and Knight, Deng, and Li (2010). En breve, Concluimos
that by 2009 China entered a transition period for the Lewis turning point and may
take several years to fully complete it.

8. Conclusion and policy recommendations

Having tested the Lewis-Ranis-Fei theory for China’s economy over 1965–2009,
we found that China’s economic growth is mainly attributable to the development
of the non-agricultural sector. This is driven by rapid capital accumulation and
employment growth. The reallocation of labor away from agriculture to non-
agriculture has made a positive net contribution to China’s rapid economic growth
by around 1.37 percent per annum. By generating measures of the marginal produc-
tivity of labor, we found that the marginal productivity of agricultural labor started
to increase noticeably after the 1978 economic reforms. This indicates that China en-
tered phase two, as deªned by the Lewis-Ranis-Fei theory, of economic develop-
ment after 1978. The marginal productivity of agricultural labor has continued to
rise thereafter and by 2009 it seems to have exceeded the agricultural institutional
wage. Considering the huge diversity within an economy as large as China’s we feel
comfortable with the conclusion that China is now in a transition period of entry
into phase three of economic growth.

Our evidence leads us to make two tentative policy recommendations. First and
foremost, more resources should be dedicated to improving labor productivity par-
ticularly through rudimentary job training. As indicated by the dualism theory, el
third phase of economic development relies on increasing labor productivity. Pro-
viding on-the-job training may improve the working skills of rural migrants and
thereby raise their productivity.

Segundo, labor migration in China, though having made a signiªcant contribution to
economic growth, is largely temporary. This special characteristic of Chinese labor
migration highlights the discrepancy between the rate of urbanization and the pace
of industrialization. The Hukou registration system is the reason for this discrep-
ancy; y, hence, it is important to remove the Hukou restrictions on migration. Es
also important to provide more facilities and social services to accommodate rural
migrants in urban areas. These policy changes would facilitate China’s full transi-
tion into phase three of economic development.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Referencias

Breusch, Trevor. 1978. Testing for Autocorrelation in Dynamic Linear Models. Australian Eco-
nomic Papers 17:334–355.

Bruno, Miguel. 1994. Our Assistance Includes Ideas as Well as Money. Transition (Banco mundial)
5(1):1–4.

Cai, Fang. 2007. The Lewisian Turning Point of China’s Economic Development. En: The Coming
Lewisian Turning Point and Its Policy Implications, Reports on China’s Population and Labor, edited
by Fang Cai and Yang Du, páginas. 147–169. Beijing: Social Sciences Academic Press.

Cai, Fang, and Dewen Wang. 1999. The Sustainability of China’s Economic Growth and Labor
Contribution [Chino]. Journal of Economic Research 10:62–68.

Cai, Fang, and Meiyun Wang. 2008. A Counterfactual of Unlimited Surplus Labor in Rural
Porcelana. China and the World Economy 16(1):51–65.

Perro chino, Gregory. 1993. Capital Formation and Economic Growth in China. Journal of Economics
108(3):809–842.

Perro chino, Gregory, and Kui-Wai Li. 2002. China’s Economic Growth: 1952–2010. Economic Develop-
ment and Cultural Change 51(1):247–256.

Cochrane, Donald, and Guy Orcutt. 1949. Application of Least Squares Regression to Relation-
ships Containing Autocorrelated Error Terms. Journal of the American Statistical Association
44(245):32–61.

Dickey, David, and Wayne Fuller. 1979. Distribution of the Estimators for Autoregressive Time
Series with a Unit Root. Journal of the American Statistical Association 74(366):427–431.

Dickey, David, and Wayne Fuller. 1981. Likelihood Ratio Statistics for Autoregressive Time Se-
ries with a Unit Root. Econometrica 49(4):1057–1072.

Engle, Roberto, and Clive Granger. 1987. Co-Integration and Error Correction: Representation,
Estimation, and Testing. Econometrica 55(2):251–276.

Admirador, Gang. 1990. Macroeconomics Analysis of Public Ownership [Chino]. Shanghai: Shanghai
United Press.

Admirador, Gang. 1994. Incremental Changes and Dual Track Transition: Understanding the Case of
Porcelana. Economic Policy 9(19 Supplement):99–122.

fei, John, and Gustav Ranis. 1973. A Model of Growth and Employment in the Open Dualistic Econ-
omy: The Cases of Korea and Taiwan. Nueva York: Asia Society.

Garnaut, ross, and Ligang Song, editores. 2006. The Turning Point in China’s Economic Develop-
mento. Canberra: Asia Paciªc Press.

Godfrey, Leslie. 1978. Testing Against General Autoregressive and Moving Average Error
Models when the Regressor Includes Lagged Dependent Variables. Econometrica 46(6):1293–
1302.

Heston, alan, Robert Summers, and Bettina Aten. 2002. Penn World Table Version 6.1, Center
for International Comparisons at the University of Pennsylvania (CICUP). Available at http://
pwt.econ.upenn.edu/.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

A, Yhi-Min. 1972. Development with Surplus Population. The Case of Taiwan: A Critique of
the Classical Two-Sector Model, à la Lewis. Economic Development and Cultural Change
20(2):210–234.

Hobijn, Bart, Philip Franses, and Marius Ooms. 1998. Generalizations of the KPSS-test for
stationarity. Erasmus University Econometric Institute Report 9802/A. Rotterdam: Economet-
ric Institute.

Rey, Roberto, and Ross Levine. 1994. Capital Fundamentalism, Economic Development, y
Economic Growth. Policy Research Working Paper No. 1285. Washington, corriente continua: Banco mundial.

Caballero, John. 2007. Porcelana, South Africa and the Lewis model. World Institute for Development
Economic Research Research Paper No 2007/82. Helsinki: UNU-WIDER.

Caballero, John, Quheng Deng, and Shi Li. 2010. The Puzzle of Migrant Labor Shortage and Rural
Labor Surplus in China. Department of Economics Discussion Paper Series No. 494. Oxford:
Universidad de Oxford.

Kuijs, luis, and Tao Wang. 2005. China’s Pattern of Growth: Moving to Sustainability and Re-
ducing Inequality. World Bank Policy Research Working Paper No. 3767. Washington DC:
Banco mundial.

Kwan, Fung. 2009. Agricultural Labour and the Incidence of Surplus Labour: Experience from
China during Reform. Journal of Chinese Economic and Business Studies 7(3):79–94.

Kwiatkowski, Denis, Peter Phillips, Peter Schmidt, and Yongcheol Shin. 1992. Testing the
Null Hypothesis of Stationarity against the Alternative of a Unit Root. Journal of Econometrics
54(1–3):159–178.

Luis, W.. Arthur. 1954. Economic Development with Unlimited Supplies of Labor. The Man-
chester School 22(2):139–191.

Luis, W.. Arthur. 1979. The Dual Economy Revisited. The Manchester School 47(3):211–229.

lin, Justin, Fang Cai, and Zhou Li. 1994. China’s Economic Reforms: Pointers for Other Econ-
omies in Transition. Policy Research Working Paper No. 1310. Washington, corriente continua: Banco mundial.

Meng, Xin, and Nansheng Bai.2007. How Much Have the Wages of Unskilled Workers in
China Increased? Data from Seven Factories in Guangdong. En: Porcelana: Linking Markets for
Growth, edited by Ligang Song and Ross Garnaut, páginas. 151–175. Canberra: Asian Paciªc Press.

Minami, Ryoshin. 1968. The Turning Point in the Japanese Economy. Quarterly Journal of Eco-
nomics 82 (3):380–402.

Minami, Ryoshin, and Xinxin Ma. 2010. The Lewis Turning Point of Chinese Economy: Com-
parison with Japanese Experience. China Economic Journal 3(2):163–179.

Ministry of Labor and Social Security of China. 2009. 2009 Human Resources and Social
Security Statistical Report. Available at http://w1.mohrss.gov.cn/gb/zwxx/2010-05/21/
content_382330.htm.

National Statistical Bureau of China. 2001, 2003, 2004, 2009. The Chinese Statistical Yearbook.
Beijing: PRC.

Naughton, Barry. 1995. Growing Out of the Plan: Chinese Economic Reform, 1978–1993. Leva-
puente: Prensa de la Universidad de Cambridge.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

Newey, Whitney, and Kenneth West. 1994. Automatic lag selection in covariance matrix esti-
formación. Review of Economic Studies 61(4):631–653.

Nolan, Peter, and Robert Ash. 1995. China’s Economy on the Eve of Reform. The China Quar-
terly 144(special issue):980–998.

Ohkawa, Kazushi. 1965. Agriculture and Turning Points in Economic Growth. The Developing
Economies 3(4):471–486.

Parque, Alberto, Fang Cai, and Yang Du. 2010. Can China Meet Her Employment Challenges? En
Growing Pains: Tensions and Opportunities in China’s Transformation, edited by Jean Oi, Scott
Rozelle, and Xuegang Zhou, páginas. 27–55. stanford, California: Stanford Asia-Paciªc Research Center.

Prais, Sigbert, and Christopher Winsten. 1954. Trend Estimators and Serial Correlation. Cowles
Commission Discussion Paper No. 383. Available at http://cowles.econ.yale.edu/P/ccdp/st/
s-0383.pdf.

Putterman, luis. 1992. Dualism and Reform in China. Economic Development and Cultural
Cambiar 40(3):467–493.

Ranis, Gustav, and John Fei. 1961. A Theory of Economic Development. American Economic Re-
vista 51(4):533–565.

Rostow, Walt. 1956. The Take-off into Self-Sustaining Growth. Economic Journal 66(261):25–48.

Sachs, Jeffrey D., and Wing Thye Woo. 1994. Structural Factors in the Economic Reforms of
Porcelana, East Europe and the Former Soviet Union. Economic Policy 18(1):102–145.

Sachs, Jeffrey D., and Wing Thye Woo. 2000. Understanding China’s Economic Performance.
Journal of Policy Reform 4(1):1–50.

United Nations. 1970. Methods of Measuring Internal Migration. Nueva York: United Nations.

Wang, Dewen.W. 2008. Lewisian Turning Point: Chinese Experience [Chino]. In Reports on
Chinese Population and Growth No. 9: Linking up Lewis and Kuznets Turning Points, edited by Cai
Fang, páginas. 88–103. Beijing: Social Sciences Academic Press.

Blanco, Halbert. 1984. A Heteroscedasticity Consistent Covariance Matrix Estimator and a Di-
rect Test of Heteroscedasticity. Econometrica 48(4):817–838.

Cortejar, Wing Thye. 1998. Chinese Economic Growth: Sources and Prospects. In The Chinese Econ-
omy, edited by Michel Fouquin and Francoise Lemoine, páginas. 17–48. Londres: Economica.

Banco mundial. 1996. The Chinese Economy: Controlling Inºation, Deepening Reform. Washington,
corriente continua: Banco mundial.

Banco mundial. 2010. World Development Indicators. Washington, corriente continua: Banco mundial.

Wu, Xiaoying. 1994. Rural to Urban Migration in the People’s Republic of China. The China
Quarterly 139:669–698.

Yao, Cual. 2011. The Relationship between China’s Export-led Growth and Its Double Transi-
tion of Demographic Change and Industrialization. Asian Economic Papers 10 (2):52–76.

Yao, Cual, and Ke Zhang. 2010. Has China Passed the Lewis Turning Point? A Structural Esti-
mation Based on Provincial Data. China Economic Journal 3(2):155–162.

Asian Economic Papers

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d

F
/

1
0
3
1
1
6
8
2
8
1
5
a
s
mi
pag
_
a
_
0
0
1
0
8
pag
d

b
y
gramo
tu
mi
s
t

oh
norte
0
7
S
mi
pag
mi
metro
b
mi
r
2
0
2
3

An Empirical Analysis of China’s Dualistic Economic Development: 1965–2009

zhang, Kevin, and Shunfeng Song. 2003. Rural-Urban Migration and Urbanization in China:
Evidence from Time-Series and Cross-Section Analyses. China Economic Review 14(4):386–400.

zhao, Yaohui. 2000. Rural to Urban Labor Migration in China: the Past and the Present. In Ru-
ral Labor Flows in China, edited by Loraine West and Yaohui Zhao, páginas. 15–33. berkeley, California: En-
stitute of East Asian Studies, Universidad de California.

D
oh
w
norte
oh
a
d
mi
d

F
r
oh
metro
h

t
t

pag

:
/
/

d
i
r
mi
C
t
.

metro

i
t
.

mi
d
tu
a
s
mi
pag
a
r
t
i
C
mi
–
pag
d