Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China
Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

International and Intranational
Technological Spillovers and
Productivity Growth in China*

Xiaolan Fu
Director of Sanjaya Lall
Programme for Technology
and Management for
Desarrollo
Department of International
Desarrollo
Universidad de Oxford
3 Mansfield Road
Oxford, OX1 3TB, Reino Unido
xiaolan.fu@qeh.ox.ac.uk

Yundan Gong
Aston Business School
Aston University
Aston Triangle
Birmingham, B4 7ET, Reino Unido
y.gong1@aston.ac.uk

Abstracto
Technological spillovers from foreign direct investment (FDI) tener
been regarded as a major source of technical progress and pro-
ductivity growth. This paper explores the role of international and
intranational technological spillovers from FDI in technical change,
efficiency improvement, and total factor productivity growth in
Chinese manufacturing firms using a recent Chinese manufactur-
ing firm-level panel data set over the 2001–05 period. interna-
tional industry-specific research and development (R&D) stock is
linked to the Chinese firm-level data, international R&D spillovers
from FDI and intranational technological spillovers of R&D activi-
ties by foreign invested firms in China are examined as well. Política
implications are discussed.

1. Introducción

Three decades of the Chinese economy’s fast economic
growth has attracted substantial research interest. This im-
pressive growth performance is not only due to factor ac-
cumulation, but to productivity growth as well. As the
largest foreign direct investment (FDI) recipient in the de-
veloping world, how much has China beneªted from the
huge inºows of FDI? Technology transfer through FDI has
been regarded as a major engine of technological upgrad-
ing in developing countries for a long period. What are the

* The authors would like to thank Wing Thye Woo, V. norte.

Balasubramanyam, Shigeyuki Abe, Chul Chung, Bhanupong
Nidhiprabha, and conference participants at the Asian Eco-
nomic Panel Annual Meeting (Seoul, Abril 2008), North Amer-
ica Productivity Workshop (Nueva York, Junio 2008) and the Chi-
nese Economic Association (Reino Unido) Annual Conference
(Cambridge, Marzo 2008) for helpful comments.

Asian Economic Papers 8:2

© 2009 The Earth Institute at Columbia University and the Massachusetts

Institute of Technology

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

roles of international and intranational research and development (R&D) spillovers
in the technical progress and productivity growth in China? Can developing coun-
tries rely on foreign technology to catch up with the industrialized countries?

Empirical evidence on the impact of FDI on the productivity growth of indigenous
ªrms is mixed (Blomstrom and Kokko 1998; Aitken and Harrison 1999; Görg and
Greenaway 2001; Javorcik 2004). It is found that intranational knowledge spillovers
are a more important source of technological progress than the international spill-
overs for the United States and Japan (Branstetter 2001). In the context of China, Hu
and Jefferson (2002) ªnd signiªcant productivity depression rather than positive
spillover effects of FDI on domestic ªrms. Using cross-section data for 1995, Buckley,
Clegg, and Wang (2002) ªnd that non-Chinese multinational enterprises (MNEs)
generate technological and international market access spillover beneªts for Chinese
ªrms, whereas overseas Chinese investors confer only market access beneªts. Spa-
inicialmente, Chen, li, and Shapiro (2008) ªnd that in locations with a strong clustering of
innovative foreign ªrms, local ªrms beneªt from knowledge spillovers, but not in
locations where foreign concentration is measured by employment or capital. Estos
studies provide useful insights. Sin embargo, foreign knowledge and the spillovers
from FDI are often tested using an output/employment/asset share of foreign in-
vested ªrms or productivity or R&D activities of foreign ªrms in the same industry
or region. Although there are some studies that have tested the spillovers of interna-
tional knowledge stock through international trade and FDI at the country and in-
dustry levels directly (p.ej., Coe and Helpman 1995), no study at ªrm level has tested
the international knowledge on spillovers directly.

This paper explores the role of inter- and intranational technological spillovers from
FDI in technical change, efªciency improvement, and total factor productivity (TFP)
growth in Chinese manufacturing ªrms using a recent Chinese manufacturing ªrm-
level panel data set of 56,125 ªrms over the 2001–05 period. International industry-
speciªc R&D stock is linked to the Chinese ªrm-level data by corresponding indus-
try and adjusted by industry- and ªrm-level degrees of openness. Por lo tanto, nosotros
employ two sources of R&D spillovers from FDI: R&D spillovers from innovation
activities by foreign invested ªrms in the same industry and international R&D
spillovers through FDI. We use the non-parametric frontier technique to decompose
the TFP growth of ªrms into technical change and efªciency improvement. A diferencia de
most of the existing studies that estimate TFP using a single unchanging production
function across industries, this study allows for the differences in production tech-
nology across industries, and TFP is estimated for each industry separately.

This paper is organized as follows: Sección 2 discusses the theoretical framework on
international and intranational R&D spillovers. Sección 3 discusses the data, modelo,

2

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

and methodology. Sección 4 presents empirical evidence. Sección 5 provides conclu-
siones.

2. Theoretical framework and innovation in China

The literature presents two alternative perspectives for the choice of technology de-
velopment paths for developing countries. One perspective proposes that FDI tech-
nology transferred from developed countries has positive effects on developing
countries (Eden, Lecitas, and Martinez 1997; Kokko, Tansini, and Zejan 1997), y
por lo tanto, the technology spillover effects of FDI may be more important than the ef-
fects of domestic investments (Borensztein, Gregorio, and Lee 1995). The degree of
technology diffusion from FDI grows with the increase in technology distance be-
tween the hosts and the foreign countries (Findlay 1978). The greater the technology
distancia, the more difªcult it becomes for developing countries to boost independ-
ent innovation.

Another outlook is that the introduction of FDI will make the competing domestic
ªrms worse off (Aitken and Harrison 1999), and will reduce the R&D efforts of local
ªrms (OECD 2002). Además, the beneªts of FDI technology spillovers are lim-
ited because most techniques transferred from foreign investment ªrms are usually
mature techniques, not core techniques; and as the working conditions and rewards
of overseas-funded ªrms are better than that of native ªrms, knowledge diffusion
caused by turnover of native talented personnel is usually one-way from the native
ªrms to overseas-funded ªrms. Además, technologies created in the industrialized
countries are argued to be biased to the factor endowment of the country where the
technology is developed, and therefore are capital and skilled-labor augmenting
(Basu and Weil 1998; Acemoglu 2002). The advanced foreign technology, por lo tanto,
may not be appropriate for developing countries given their different factor, eco-
nomic, and social conditions (Atkinson and Stiglitz 1969; Stewart 1983). Finalmente, estafa-
sidering that technology progress has the characteristic of path dependence, a coun-
try that is dependent on the technology spillover of FDI for a long period will later
limit its independent innovation. Por lo tanto, strengthening R&D and enhancing the
independent creative abilities should be the main path for developing countries’
technological advancement. Taking into account the pros and cons of the foreign
and indigenous innovation, Lall (2003) argues that neither autonomous innovations
nor FDI-reliant strategies can be used independently.

Teóricamente, FDI contributes to technological upgrading in the host economy in
varias maneras. Primero, advanced technology embedded in imported machinery and
equipment can lift the level production technology of the host economy. Segundo,
R&D and other forms of innovation generated by foreign ªrms and R&D labs of

3

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

MNEs increase the innovation outputs in the country directly (Athreye and Cant-
Bueno 2007). Tercero, FDI may contribute to the local innovation system by bringing in
advanced management practices and thus improving the innovation efªciency of
the local innovation system (Fu 2008a). Finalmente, technological spillovers from foreign
innovation activities may inºuence technical change and the catch-up of indigenous
ªrms. Knowledge spillovers from foreign to local ªrms may take place through
knowledge transfer within the supply chain, skilled labor turnovers, demonstration
effects when local ªrms are learning by imitation, and competition effects when the
competitive pressure caused by foreign presence forces the local ªrms to improve
their production technology and management.

Sin embargo, foreign R&D activities could also generate negative externalities to the
domestic innovation activities. These negative effects could occur if foreign ªrms
exploit their superior technology and marketing power to force local competitors to
reduce their outputs or if they attract the most talented researchers and compete in
the markets of innovation products that threaten local ªrms, or SMEs in particular
(Aitken and Harrison 1999; Fu 2004, 2007; Aghion et al. 2005; UNCTAD 2005). Más-
encima, there are several reasons that local ªrms might not be able to enjoy the FDI
spillover effects efªciently. Primero, knowledge transfer via supply chain requires effec-
tive linkages between foreign ªrms and local suppliers and customers (Balasubram-
anyam, Salisu, and Sapsford 1996; Fu 2004). Segundo, signiªcant spillovers from FDI
on local ªrms are also subject to sufªcient absorptive capacity of the local ªrms and
organizaciones (Cohen and Levinthal 1989; Girma 2005; Fu 2008a). Tercero, the appro-
priateness of the technology embedded in FDI affects the sign and signiªcance of
the productivity effects of FDI spillovers. Technologies created in industrialized
countries are argued to be biased to the factor endowment of the country where the
technology is developed (Basu and Weil 1998; Acemoglu 2002). Finalmente, diferente
types of FDI have markedly different productivity spillover effects (Drifªeld and
Amar 2003).

Since it launched the economic reforms and invited foreign capital participation in
its economy in 1979, China has received a large volume of international direct in-
vestment ºows and stands as the second largest FDI recipient in the world. En 2004,
FDI inºows into China reached a historical peak of US$ 60.63 billion (Cifra 1). The sources of inward FDI in China have also evolved over time. While investment from overseas Chinese ªrms in Hong Kong, Macao, and Taiwan were the major sources of inward FDI in the 1980s, the 1990s saw increasing inward FDI from the major in- dustrialized countries and other OECD countries. Innovation efforts in China have grown rapidly during the past two decades. The total R&D expenditure in China has grown from 7.4 billion Yuan in 1987 a 4 Asian Economic Papers l D o w n o a d e d f r o m h t t p : / / directo . mi t . / e d u a s e p a r t i c e – pd / l f / / / / / 8 2 1 1 6 8 2 5 6 2 a s e p 2 0 0 9 8 2 1 pd . . . . . f por invitado 0 7 septiembre 2 0 2 3 International and Intranational Technological Spillovers and Productivity Growth in China Figure 1. Trade and FDI in China, 1985–2004 l D o w n o a d e d f r o m h t t p : / / directo . mi t . / e d u a s e p a r t i c e – pd / l f / / / / / 8 2 1 1 6 8 2 5 6 2 a s e p 2 0 0 9 8 2 1 pd . . . . . f por invitado 0 7 septiembre 2 0 2 3 300.3 billion Yuan in 2006, with an average annual growth rate of 15 por ciento (Cifra 2). Since the late 1990s, with the increasing globalization in innovation, R&D activi- ties of foreign ªrms in China have been increasing, at a faster pace than that of the domestic ªrms. The average annual growth of R&D expenditure over the 1998–2004 period was 38 y 33 percent in foreign invested enterprises and Ethnic Chinese in- vested ªrms,1 respectivamente. This is much higher than that of indigenous ªrms at 25 percent over the same time period. Given the fact that foreign investors in China are mostly market- or resource- or cheap labor–seeking processing types, R&D spillovers from foreign invested ªrms are likely to be limited. Además, motivations, technology levels, endowments, and access to advanced technological and managerial knowledge all are different be- tween foreign and ethnic investments. The productivity effects of foreign and ethnic investments are likely to be different. 3. Data and methodology 3.1 Data The empirical work is based on the Chinese manufacturing ªrm-level data set and the international industry-speciªc R&D stock data set. The Chinese ªrm-level data 1 Ethnic Chinese–invested ªrms refer to ªrms that have more than a 25-percent share of capital invested by Hong Kong, Macao, and Taiwan investors. 5 Asian Economic Papers International and Intranational Technological Spillovers and Productivity Growth in China Figure 2. R&D expenditure, 1987–2006 (billion yuan) l D o w n o a d e desde h t t p : / / directo . mi t . / e d u a s e p a r t i c e – pd / l f / / / / / 8 2 1 1 6 8 2 5 6 2 a s e p 2 0 0 9 8 2 1 pd . . . . . f por invitado 0 7 septiembre 2 0 2 3 set is from the Annual Report of Industrial Enterprise Statistics compiled by the State Statistical Bureau of China, covering all state-owned ªrms and other types of ªrms with an annual turnover of over 5 million Renminbi (about US$ 0.6 millón).
The data set includes all the variables we are interested in, such as ªrm ownership
estructura, industry afªliation, establishment year, employment, gross output, ex-
ports, R&D, and employee training expenditures.2 The data cover the period 2001 a
2005. They are broadly classiªed under ªve ownership categories: (i) state-owned,
(ii) collectively owned, (iii) privately owned, (iv) foreign-owned, y (v) otros. Para-
eign-owned ªrms are further divided into ªrms with investments from Hong Kong,
Taiwán, and Macao investors (so-called ethnic ªrms) and from other foreign sources
(foreign invested enterprises; FIEs). “Other” ªrms are mainly shareholding enter-
tomado.

As we are interested in the technology spillover effects from foreign ªrms on do-
mestic ªrms, the econometric work is conªned to domestic-owned enterprises. Nosotros,
sin embargo, use the full sample to construct myriad variables of interest, como el
share of foreign ªrms in an industry-region or the Herªndahl index of market con-
centration. The ªnal data set consists of 269,905 observations from 53,981 ªrms. Nosotros
include only those ªrms with the full set of observations during the sample period

2 Nominal values are deºated using industry-speciªc ex-factory price indices obtained from

the China Statistical Yearbook 2006.

6

Asian Economic Papers

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

as estimation of TFP growth and its components using DEA analysis, which requires
balanced data sets. Mesa 1 reports the ownership structure of ªrms for each indus-
intentar. Of the total 29 SIC two-digit manufacturing industries under study, 7 de ellos
are dominated by FIEs that produce more than 50 percent of the total outputs. En
cultural, educational and sports goods, electronic and telecommunications, and in-
struments and meters, foreign ªrms produce even 70 a 80 percent of the country’s
total output. This is phenomenal given the size of the Chinese industry. In these
three categories and the apparel industries, más que 20 percent of the ªrms are in-
vested through foreign capital mainly from OECD countries. Mesa 2 sees a steady
modest growth in the portion of ªrms that invest in R&D across all industries over
the sample period. The medical and pharmaceutical, tobacco, and electronic and
telecommunications industries top the chart with 50, 45, y 33 percent innovative
ªrms, respectivamente.

International industry-speciªc R&D stock is linked explicitly to the Chinese ªrm-
level data. The estimates of international R&D capital stocks are based on R&D ex-
penditure data from the OECD’s Main Science and Technology Indicators. Real R&D
expenditures are nominal expenditures deºated by an R&D price index (PR).3 Fol-
lowing Coe and Helpman (1995), research and development capital stocks (S),
which are deªned here as beginning of period stocks, were calculated from R&D ex-
penditure (R) based on the perpetual inventory model as

Calle

(cid:2) (1 (cid:3) (cid:4)) Calle(cid:3)1

(cid:5) Rt .

Aquí (cid:4) is the depreciation or obsolescence rate, which was assumed to be 5, 10 y
15 por ciento, alternatively. The benchmark for S was calculated following the proce-
(cid:2) R0/(gramo (cid:5) (cid:4)), where g is the average an-
dure suggested by Griliches (1979), as S0
nual logarithmic growth of R&D expenditures over the period for which published
R&D data were available, R0 is the ªrst year for which the data were available, y
S0 is the benchmark for the beginning of the year. The domestic R&D capital stocks
were converted into euros at 2000 constant price. The R&D stocks of the 22 OECD
countries are then summed to proxy the world R&D stock.

3.2 Metodología
For the empirical test, we ªrst estimate TFP growth using the Malmquist index, y
decompose it into technical progress and efªciency change. En segundo lugar, we use econo-

3 PR is deªned as PR (cid:2) 0.5PAG (cid:5) 0.5W., where P is the implicit deºator for business sector output
and W is index of average business sector wages (the same source as for Y). This deªnition of
PR implies that half of R&D expenditures are labor costs, which is broadly consistent with
available data on the composition of R&D expenditures.

7

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

r
mi
v
oh
norte
r
tu
t

y
b
%

s

metro
r
ª
norte
gramo
i
mi
r
oh
F

C
i
norte
h
t
mi
–
norte
oh
norte

s

metro
r
ª
C
i
norte
h
t
mi

norte
gramo
i
mi
r
oh
F

yo
yo
a

s

metro
r
ª

r
mi
v
oh
norte
r
tu
t

y
b
%

r
mi
b
metro
tu
norte
y
b
%

s

metro
r
ª
F
oh

s

metro
r
ª
F
oh
r
mi
b
metro
tu
norte
y
b
%

s

metro
r
ª
norte
gramo
i
mi
r
oh
F

s

metro
r
ª
s
tu
oh
norte
mi
gramo
i
d
norte
I

8
2

8
4

0
4

0

6
2

2
5

3
6

5
3

2
6

6
3

7
3

0
7

2

0
2

7
2

8
2

2
4

8
4

1
2

7

3
1

6
4

8
2

8
1

5
2

9
3

9
7

7
7

4
5

3
3

2
1

7
1

3
1

0

9

1
2

9
1

4
1

8
1

7

5

8

0
2

0
1

1
1

0
1

1
1

3
1

6

4

6

3
1

0
1

8

0
1

2
1

8
2

0
2

5
1

2
1

9

6
1

0
1

3

7
1

0
3

3
3

7
1

7
2

3
1

3
1

9
3

5

0
1

0
1

6
1

6
1

7
2

8

5

7

7
1

6

8

9

6
1

2
3

2
2

0
3

5
1

1
2

3
3

2
2

3

6
2

2
5

2
5

0
3

6
4

0
2

8
1

8
5

3
1

0
2

1
2

7
2

7
2

9
3

4
1

9

3
1

0
3

6
1

6
1

9
1

7
2

0
6

2
4

6
4

7
2

2
7

2
5

0
6

0
0
1

4
7

8
4

8
3

6
6

8
3

4
6

3
6

0
3

8
9

1
8

3
7

2
7

8
5

2
5

9
7

3
9

7
8

4
5

2
7

2
8

5
7

1
6

2
2

3
2

6
4

7
6

9
7

8
6

8
7

7
9

4
7

8
4

8
4

0
7

5
5

0
8

2
8

2
4

7
8

1
8

0
8

3
7

4
7

1
6

6
8

1
9

7
8

0
7

4
8

4
8

1
8

3
7

0
4

8
5

5
5

3
7

2
4
3
,
4
1

9
3
6
,
9
3

s
t
C
tu
d
oh
r
pag

yo
a
C
i
metro
mi
h
C
d
norte
a

s
yo
a
i
r
mi
t
a
metro

yo
a
C
i
metro
mi
h
C
w
a
R
–
6
2

s
t
C
tu
d
oh
r
pag

yo
a
C
i
t
tu
mi
C
a
metro
r
a
h
pag
d
norte
a
yo
a
C
i
d
mi
METRO
–
7
2

norte
oh
i
t
C
tu
d
oh
r
pag
mi
r

i

metro
tu
d
mi
metro
d
r
oh
C
mi
r
d
norte
a
gramo
norte
i
t
norte
i
r
PAG
–
3
2

s
d
oh
oh
gramo
s
t
r
oh
pag
s
d
norte
a
yo
a
norte
oh
i
t
a
C
tu
d
mi

,
yo
a
r
tu
t
yo
tu
C
–
4
2

s
t
C
tu
d
oh
r
pag
r
mi
pag
a
pag
d
norte
a
gramo
norte
i
k
a
metro
r
mi
pag
a
PAG
–
2
2

gramo
norte
i
k
oh
C
d
norte
a
gramo
norte
i
norte
ª
mi
r

metro
tu
mi
yo
oh
r
t
mi
PAG
–
5
2

s
t
C
tu
d
oh
r
pag
d
mi
t
a
yo
mi
r
d
norte
a

,

norte
w
oh
d

,
s
r
tu
F

,
r
mi
h
t
a
mi
l
–
9
1

s
t
C
tu
d
oh
r
pag
r
mi
b
ª
r
mi
h
t
oh
d
norte
a

s
t
norte
mi
metro
r
a
GRAMO
–
8
1

gramo
norte
i
r
tu
t
C
a
F
tu
norte
a
metro
mi
r
tu
t
i
norte
r
tu
F
–
1
2

gramo
norte
i
s
s
mi
C
oh
r
pag
r
mi
b
metro
t
–
0
2

i

s
yo
a
t
mi
metro

s
tu
oh
r
r
mi
F
norte
oh
norte
F
oh
gramo
norte
i
s
s
mi
r
pag
&
gramo
norte
i
t
yo
mi
metro
S
–
3
3

s
yo
a
t
mi
metro

s
tu
oh
r
r
mi
F

F
oh
gramo
norte
i
s
s
mi
r
pag
d
norte
a
gramo
norte
i
t
yo
mi
metro
S
–
2
3

s
t
C
tu
d
oh
r
pag

yo
a
r
mi
norte
i
metro

yo
a
t
mi
metro
norte
oh
norte
–
1
3

y
r
mi
norte
i
h
C
a
metro
d
norte
a

t
norte
mi
metro
pag
tu
q
mi

i

C
i
r
t
C
mi
yo
mi
–
9
3

s
norte
oh
i
t
a
C
i
norte
tu
metro
metro
oh
C
mi
yo
mi
t
d
norte
a
C
i
norte
oh
r
t
C
mi
yo
mi
–
0
4

gramo
norte
i
r
tu
t
C
a
F
tu
norte
a
metro

r
mi
h
t
oh
d
norte
a
t
C
a
F
i
t
r
A
–
2
4

s
r
mi
t
mi
metro
d
norte
a

s
t
norte
mi
metro
tu
r
t
s
norte
I
–
1
4

t
norte
mi
metro
pag
tu
q
mi

i

s
mi
s
oh
pag
r
tu
pag

yo
a
i
C
mi
pag
S
–
6
3

t
norte
mi
metro
pag
tu
q
mi

i

t
r
oh
pag
s
norte
a
r
t
–
7
3

y
r
mi
norte
i
h
C
a
metro
y
r
a
norte
i
d
r
oh
–
5
3

s
t
C
tu
d
oh
r
pag

yo
a
t
mi
METRO
–
4
3

s

metro
r
ª
F
oh
r
mi
b
metro
tu
norte
yo
a
t
oh
t

yo
a
t
oh
t

s
t
C
tu
d
oh
r
pag
r
mi
b
b
tu
R
–
9
2

s
t
C
tu
d
oh
r
pag
C
i
t
s
a
yo
PAG
–
0
3

r
mi
b
ª

yo
a
C
i
metro
mi
h
C
–
8
2

y
r
t
s
tu
d
norte
i

t
i
gramo
i
d
–
2
C
I
S

gramo
norte
i
s
s
mi
C
oh
r
pag
oh
C
C
a
b
oh
t
–
6
1

y
r
t
s
tu
d
norte
i

mi
gramo
a
r
mi
v
mi
B
–
5
1

y
r
t
s
tu
d
norte
i

mi
yo
i
t
X
mi
t
–
7
1

gramo
norte
i
s
s
mi
C
oh
r
pag
d
oh
oh
F
–
3
1

norte
oh
i
t
C
tu
d
oh
r
pag
d
oh
oh
F
–
4
1

pag
tu
oh
r
gramo
y
r
t
s
tu
d
norte
i
h
C
a
mi
norte
i

mi
r
tu
t
C
tu
r
t
s
pag
i
h
s
r
mi
norte
w
oh

.
1

mi
yo
b
a
t

.
gramo
norte
i
d
a
h
s

y
b
d
mi
t
h
gramo
i
yo
h
gramo
i
h
mi
r
a

s
t
tu
pag
t
tu
oh

yo
a
t
oh
t

mi
h
t

F
oh

%
0
5
norte
a
h
t

mi
r
oh
metro

mi
C
tu
d
oh
r
pag

oh
h
w

)
s
mi
I
F
(

s
metro
r
ª
d
mi
t
s
mi
v
norte
i
norte
gramo
i
mi
r
oh
F

y
b

d
mi
t
a
norte
i
metro
oh
d

s
mi
i
r
t
s
tu
d
norte
I

:
mi
t
oh
norte

8

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

Mesa 2. Proportion of ªrms investing in R&D

SIC 2-digit industry

13-Food processing
14-Food production
15-Beverage industry
16-Tobacco processing
17-Textile industry
18-Garments and other ªber products
19-Leather, furs, abajo, and related products
20-Timber processing
21-Furniture manufacturing
22-Papermaking and paper products
23-Printing and record medium reproduction
24-Cultural, educational and sports goods
25-Petroleum reªning and coking
26-Raw chemical materials and chemical products
27-Medical and pharmaceutical products
28-Chemical ªber
29-Rubber products
30-Plastic products
31-Nonmetal mineral products
32-Smelting and pressing of ferrous metals
33-Smelting and pressing of nonferrous metals
34-Metal products
35-Ordinary machinery
36-Special purposes equipment
37-Transport equipment
39-Electric equipment and machinery
40-Electronic and telecommunications
41-Instruments and meters
42-Artifact and other manufacturing

2001

0.1022
0.1711
0.1692
0.4271
0.0857
0.0494
0.1049
0.0756
0.0930
0.0748
0.0654
0.1143
0.1818
0.2126
0.4451
0.2189
0.1921
0.0991
0.1131
0.1388
0.1683
0.1092
0.2168
0.3237
0.2738
0.2476
0.3114
0.3159
0.0926

2002

0.1080
0.1940
0.1897
0.4583
0.0909
0.0589
0.1258
0.0913
0.1027
0.0834
0.0592
0.1212
0.2102
0.2285
0.4819
0.2090
0.1952
0.1071
0.1304
0.1490
0.1707
0.1221
0.2365
0.3323
0.2855
0.2604
0.3293
0.3368
0.1157

2003

0.1169
0.1743
0.1888
0.5000
0.0962
0.0631
0.1096
0.1041
0.0969
0.0753
0.0724
0.1322
0.2216
0.2294
0.5092
0.2239
0.2127
0.1100
0.1215
0.1521
0.1851
0.1178
0.2514
0.3253
0.3035
0.2671
0.3332
0.3851
0.1278

2004

0.0786
0.1680
0.1645
0.4479
0.0769
0.0536
0.0841
0.0628
0.1008
0.0601
0.0675
0.1006
0.1875
0.1848
0.5016
0.1791
0.1746
0.0832
0.0822
0.0976
0.1599
0.0892
0.1950
0.2856
0.2646
0.2218
0.3266
0.3133
0.0775

2005

0.0786
0.1680
0.1645
0.4479
0.0769
0.0536
0.0841
0.0628
0.1008
0.0601
0.0675
0.1006
0.1875
0.1848
0.5016
0.1791
0.1746
0.0832
0.0822
0.0976
0.1599
0.0892
0.1950
0.2856
0.2646
0.2218
0.3266
0.3133
0.0775

metric techniques to estimate the impact of technological spillovers of FDI on TFP
growth, technical change, and efªciency improvement. TFP is estimated for each in-
dustry separately, allowing for different technology and production functions.

Due to the limitations of the traditional parametric approach, this paper estimates
TFP growth by using a non-parametric programming method developed by Fare
et al. (1994). Following Fare et al.’s approach, a production frontier is constructed
based on all the existing observations. The distance of each of the observations from
the frontier is estimated by using non-parametric programming methods. Técnico
efªciency is deªned as the distance of each observation relative to the frontier. TFP
growth is deªned as a geometric mean of two Malmquist productivity indexes,
which is to be estimated as the ratios of distance functions of observations from
the frontier. This approach has the advantage in that it allows for the decomposition
of productivity growth into two mutually exclusive and exhaustive components:
(1) efªciency change in movements toward (or away from) the frontier, which is a
measurement of catching-up; y (2) technical change measured by shifts in techno-
logical frontier (Fare et al. 1994). This decomposition of TFP growth enables us to in-
vestigate the impact of foreign and indigenous innovation efforts on technical prog-
ress and technological catch-up.

9

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

Assuming a production technology that produces a vector of outputs, y
using a vector of inputs, X
Malmquist productivity change index is deªned as the geometric mean of two
Malmquist productivity indices as follows:

N∈ + , for each time period t (cid:2) 1, . . . , t, the output-based
Rt

M∈ + , por
Rt

t
y
M0(xt(cid:5)1, yt(cid:5)1, xt, yt) (cid:2) D x
(
0
t
D x y
(

,
,

1

t

t
0

⎡
⎢
⎣

⎛
⎜⎜
⎝

+

+

t

1

t

)

)

⎞
⎟⎟
⎠

⎛
⎜⎜
⎝

D

(
+

+

t
0
D

1

t
0

t

X
1
(

+

1

y
,
t
x y
,

+

t

1

)

t

)

1
2

.

⎞
⎟⎟
⎠

⎤
⎥
⎦

A value greater than 1 indicates positive TFP growth in period t(cid:5)1. When perfor-
mance deteriorates over time, the Malmquist index will be less than 1. Rewriting
equation (1), tenemos

y
M0(xt(cid:5)1, yt(cid:5)1, xt, yt) (cid:2) D x
,
(
t
t
t
D x y
(
0

,

+

1

t
0

+

t

1

+

t

1

)

×

)

⎡
⎢
⎣

⎛
⎜⎜
⎝

t

t
D x
(
0
+
t
1
D x
(
0

+

1

+

t

+

1

+

t

,
1

t

y
y
,

)
1

⎞
⎟⎟
⎠

⎛
⎜⎜
⎝

)

t

t
0
+

t
D x y
(
t
1
D x y
,
(
0

,
t

1
2

,

⎞
⎟⎟
⎠

⎤
⎥
⎦

)
t

)

dónde

y

y
efªciency change (EFFCH) (cid:2) D x
,
(
t
t
t
D x y
(
0

,

+

1

t
0

+

t

1

)
,

+

t

1

)

t
technical change (TECHCH) (cid:2) D x
(
0
+
t
1
D x
(
0

t

⎡
⎢
⎣

⎛
⎜⎜
⎝

+

1

+

t

,
1

t

y
y
,

+

1

+

t

)
1

⎞
⎟⎟
⎠

⎛
⎜⎜
⎝

)

t

t
0
+

t
D x y
,
(
t
t
1
x y
D
,
(
0

1
2

.

⎞
⎟⎟
⎠

⎤
⎥
⎦

)
t

)

(1)

(2)

(3)

(4)

TFP change is thus decomposed into two components: efªciency change and techni-
cal change. Efªciency change measures whether production is getting closer to or
farther away from the frontier, reºecting the changes in x-efªciency. Técnico
change captures the shift in technology between the two periods. A value greater
than one indicates catch-up with the frontier or technical progress. A value less than
1 indicates deterioration in performance. Scale efªciency is deªned as the ratio of
technical efªciency calculated under the assumption of constant returns to scale
(CRS) to technical efªciency calculated under the assumption of variable returns to
escala (VRS) (Fare et al. 1994). It measures how close a ªrm is to the most productive
scale size. en este documento, output is measured by total output of ªrm, inputs are capi-
tal measured by net ªxed assets, labor measured by number of employees, and in-
termediate inputs measured by variable costs. We use the output-oriented model
under VRS for estimation.

Following the decomposition of productivity growth, the next step is to estimate the
determinants of TFP growth, technical change, and efªciency improvement and to
investigate the impact of indigenous innovation efforts and foreign R&D spillover
effects on productivity growth and its components. R&D variables and FDI vari-

10

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

ables are two sets of variables that are of interest. In our speciªcation, there are three
types of innovation efforts: ªrm level, industry level, and international level. Nosotros
construct the variables as follows: (1) at the ªrm level, R&D intensity is used as the
direct effect of innovation on a ªrm’s growth performance; (2) at the industry level,
innovation effect in each of the 171 three-digit industries and 31 provinces are con-
structed as the proportion of R&D expenditure accounted for by different owner-
ship types in the same industry and region; (3) the interaction terms of international
industry speciªc R&D stock and FDI share at both the ªrm and industry level are
adopted to measure the international innovation effect through FDI spillover effects.
FDI is represented by two variables—the share of foreign and ethnic capital at the
ªrm level. Por lo tanto, the empirical analysis of the indigenous and foreign technol-
ogy spillovers on the technology upgrading of indigenous ªrms are based on the
model as follows,

Dp

él

=

+

a

j

F
r
él

+

s
X
r
él

+

yo

w
r
él
1

∗

F

s
él

+

yo

w
r
él
2

∗

F

F
él

+

s

X

él

+

d

D

él

+ mi

él

(5)

where the dependent variable p represents TFP growth, technical change, y
efªciency improvement, respectivamente. rf is ªrm R&D intensity, rs is a vector of
industry-level R&D spillovers variables measured by industry average R&D inten-
sity by different ownership types, rw is world R&D stock constructed from OECD
STAN database as discussed earlier, and f f and f s are FDI intensity at the ªrm level
and industry level, respectivamente. FDI intensity is further divided into foreign and
ethnic capital intensity and enters the regression at the same time. X is a vector of
control variables; D is the full set of time, sector, and year dummies, and e is a ran-
dom error term.

The choice of control variables is guided by the existing empirical literature on the
determinants of TFP growth (p.ej., Bernard and Jensen 1999; Aw, Chung, and Roberts
2000; Fu 2005, 2008b). It includes the initial level of technology efªciency, edad, ªrm
tamaño, export, intangible assets, labor training, and market concentration that are hy-
pothesized to affect the dependent variable. Smaller ªrms and ªrms with exports
and more labor training are more likely to have faster TFP growth, technical change,
as well as efªciency improvement. Firms standing at the frontier are less likely to
grow as fast as other ªrms. There is no conclusive relationship between ªrm age, en-
tangible assets,4 and market concentration.

4 According to Accounting System for Business Enterprises, costs to develop intangible assets are
regarded as R&D costs of self-created products that are registered for a legal right to the as-
colocar, such as a patent (Pacter and Yuen 2001).

11

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

Huang (2003) argues forcefully that a sizable portion of FDI (especially joint venture
and acquisition FDI) in China has grown in response to the insolvency problems fac-
ing state-owned enterprises (SOEs). This suggests that industry-regional speciªc
FDI might be endogenous in the sense that foreign-invested enterprises might be
attracted to sectors or regions in which the performance of SOEs is weakest. El
use of industry and region dummies in the regressions is designed to mitigate this
potential endogeneity problem. Sin embargo, in order to guard against further endo-
geneity problems unaccounted for by these (time-invariant) dummies, and to en-
sure the robustness of our results, we implement the FDI variables with their lagged
valores, the growth of total industry sales and changes in the output share of SOEs,
and the proportion of loss-making SOEs, calculated at the industry and region5
nivel.

There are good reasons to suspect that R&D, labor training, and foreign capital par-
ticipation are potentially endogenous, even after controlling for ªxed effects. por ejemplo-
amplio, ªrms with relatively many R&D activities are more likely to have higher TFP
growth and faster technical change than others. Sin embargo, it is possible that ªrms
with a higher growth rate will invest more in R&D activities to keep their technol-
ogy advantages. Another example is the foreign share of a ªrm. Firms with a higher
foreign share could have better access to foreign technology and therefore have
higher growth rates. Sin embargo, there also might be a “cherry-picking” effect (Huang
2003) where foreign ªrms choose the faster-growing ªrms to invest in. Similar argu-
ments can also be made in the case of export and ethnic capital participation.

We employ the ªxed effects generalized method of a moments regression technique
(ver, inter alia, Hansen 1982 and Arellano and Bond 1991) to deal with the endo-
geneity problem. Lagged values of the potentially endogenous variables are used
as instruments. Además, the shares of foreign and ethnic ªrms in the industry
and region are used as extra instruments. We assume that a sector might be more
efªcient than others if there are more foreign ªrms or ethnic ªrms participating
in it, given the low level of competition from state-owned ªrms. We formally
test whether the assumption of endogeneity is borne out by the data at hand and
whether our instruments are relevant in that they exhibit sufªciently strong correla-
tion with the potential endogenous variables. We also carefully test for the appropri-
ateness of the instrumental variable candidates using Hansen’s J test for over identi-
fying restrictions and the validity of the instruments with Sargan test. Reassuringly,
we ªnd that our instruments are appropriate on all counts.

5 These instruments, along with the FDI indices, are constructed for the 171 industries in each

del 31 provinces. Por lo tanto, there is plenty of variability in them.

12

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

4. Resultados

Estimated TFP growth and its decomposed components for each industry are re-
ported in Table 3. From all industries, the Chinese ªrms have experienced consider-
able TFP growth over the 2001–05 period at an average annual rate of 4.5 por ciento.
The growth is mainly due to technical change at an average annual growth rate of
4.3 percent rather than efªciency change. The average annual growth rate of efªci-
ency improvement was only 0.7 percent over the sample period, which suggests
limited catch-up process of the followers to the innovation leaders in the Chinese
manufacturing sector.

The growth is widely spread across different sectors. The industries in which foreign
ªrms have obvious dominance include electronic and telecommunications, instru-
ments and meters, cultura, educational and sports goods, as well as the garment and
leather products industries where ethnic ªrms have a clear lead. Indigenous ªrms
have a dominant presence in the food-processing, beverage, tobacco, timber process,
paper-making, printing, petrol reªning, chemical ªber, smelting and processing of
ferrous and nonferrous metals, metal products, and transport equipment industries,
although the lead may be contributed to by different indigenous sectors in different
industries.

Following Table 3, it is interesting to see that industries in which foreign ªrms pro-
duce more than 50 percent of the total outputs have a higher efªciency change rate
but a lower technical change rate, as well as lower TFP growth rate than those in-
dustries dominated by indigenous ªrms. Por ejemplo, among those industries dom-
inated by FIE, the garment and other ªber products industry enjoy the highest efªci-
ency rate (14.4 por ciento) and the lowest technical change rate (10 por ciento). It might
suggest that foreign ªrms are more likely to keep their technical advantage in their
home countries, and are more reluctant to improve their technical efªciency than
their Chinese competitors are. Mientras tanto, they focus on adapting their technology
to the local technological frontier.

4.1 Determinants of productivity growth
Mesa 4 reports both the OLS and the General Method of Moments (GMM) estimados
of effects of technological spillovers from foreign innovation efforts on the TFP of in-
digenous ªrms. Results from the Wu–Hausman speciªcation test suggest signiªcant
endogeneity between R&D, exports, and FDI on one hand and the dependent vari-
able on the other. The GMM estimation results is therefore preferred to the OLS esti-
compañeros. For a robustness check, estimated results of the basic model and models with
industrial and international R&D spillovers at three alternative depreciation rates

13

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

Mesa 3. Technical change, efªciency improvement, and productivity change, 2001–05

SIC 2-digit industry

13-Food processing
14-Food production
15-Beverage industry
16-Tobacco processing
17-Textile industry
18-Garments and other ªber products
19-Leather, furs, abajo, and related products
20-Timber processing
21-Furniture manufacturing
22-Papermaking and paper products
23-Printing and record medium reproduction
24-Cultural, educational and sports goods
25-Petroleum reªning and coking
26-Raw chemical materials and chemical products
27-Medical and pharmaceutical products
28-Chemical ªber
29-Rubber products
30-Plastic products
31-Nonmetal mineral products
32-Smelting and pressing of ferrous metals
33-Smelting and pressing of nonferrous metals
34-Metal products
35-Ordinary machinery
36-Special purposes equipment
37-Transport equipment
39-Electric equipment and machinery
40-Electronic and telecommunications
41-Instruments and meters
42-Artifact and other manufacturing

Total

Malmquist
índice
(tfpch)

Técnico
cambiar
(techch)

Efªciency
cambiar
(effch)

Scale
cambiar
(sech)

1.0542
1.0321
1.0510
1.0507
1.0438
1.0370
1.0379
1.0427
1.0275
1.0535
1.0213
1.0277
1.0461
1.0486
1.0339
1.0450
1.0514
1.0380
1.0698
1.0566
1.0358
1.0420
1.0732
1.0555
1.0578
1.0541
1.0460
1.0262
1.0257

1.0450

0.9698
0.9787
1.0828
1.0432
1.1728
0.9074
1.1274
1.0584
0.9247
1.0664
1.0440
0.9890
1.1272
1.1372
1.0634
1.0510
1.0094
0.9891
1.0005
1.0646
0.9690
0.9536
1.0207
1.1878
1.2346
1.0149
1.1088
0.9497
0.9667

1.0432

1.0879
1.0604
0.9723
1.0074
0.8916
1.1440
0.9273
0.9885
1.1135
0.9899
0.9793
1.0393
0.9302
0.9272
0.9735
0.9934
1.0411
1.0529
1.0693
0.9992
1.0807
1.0958
1.0519
0.8895
0.8622
1.0438
0.9446
1.0849
1.0661

1.0074

1.0347
1.0156
0.9864
0.9963
0.9476
1.0040
0.9378
0.9951
1.0160
0.9902
0.9831
1.0206
0.9824
0.9825
0.9871
1.0011
1.0773
0.9848
1.0368
1.0177
1.0838
1.0043
1.0189
0.9838
0.9215
1.0343
0.9737
0.9947
1.0374

0.9963

Nota: Industries dominated by foreign invested ªrms (FIEs) who produce more than 50% of the total outputs are highlighted by

shading.

were applied. The estimated coefªcients from different model speciªcations are con-
sistent suggesting the robustness of the estimated results. We have only reported the
results with a 10 percent R&D depreciation rate due to space limitation.6

In terms of the coefªcients of the control variables, we ªnd that they all turn out as
esperado. Firms with better initial technical efªciency tend to grow more slowly.
Smaller ªrms appear to be more productive. Firms with high export-intensity, alto
FDI-intensity, more training, and greater intangible assets have higher TFP growth
than those who lack these characteristics. These estimated results are robust and sta-
tistically signiªcant across industry sectors and different model speciªcations. Firm
age does not appear to be a signiªcant factor. Curiosamente, industry concentration
and low levels of competition seem to increase ªrm productivity although the esti-
mated coefªcient loses its statistical signiªcance when international R&D spillovers
are controlled.

6 Results of all the estimations are available from the authors.

14

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

The variables of most interest to us are, as we mentioned previously, the sets of R&D
and FDI variables. Indigenous R&D efforts have a signiªcant positive impact on
ªrm-level TFP growth. The estimated coefªcients bear the expected positive sign,
and are statistically signiªcant across different model speciªcations. R&D spillovers
from domestic ªrms in the same industry exert a signiªcant positive effect on the
TFP growth of indigenous ªrms. Sin embargo, it is interesting to see that innovation ef-
forts from the FIEs in the same industry show a negative and signiªcant impact on
Chinese manufacturing ªrms, and there is no signiªcant impact from ethnic ªrms at
the same industry. This is likely because of either the competition effects of foreign
R&D on the indigenous ªrms, or because the technologies developed by these sec-
tors are not appropriate for the current technology frontier.

As we expected, both foreign and ethnic capital participations foster Chinese do-
mestic ªrms’ productivity growth. Foreign capital has a bigger inºuence on ªrm’s
growth than ethnic capital and both magnitudes of the coefªcients have been
doubted when international R&D spillover effects are controlled. Por ejemplo, el
estimate result based on the industrial level indicates that a 10 percent increase in
the share of foreign capital leads to a 2.5 percentage point increase in the rate of pro-
ductivity growth, whereas a 10 percent increase in the share of ethnic capital only
leads to 1.7 percentage point growth. Sin embargo, both growth rates increase to 5.6
y 4.9 puntos de porcentaje, respectively at the speciªcation with international R&D
spillovers. It might suggest that domestic ªrms with foreign capitals have better
channels to foreign advanced technical and management knowledge compared to
their competitors with only ethnic capital participations.

The estimated coefªcients on international R&D spillover variables are statistically
negatively signiªcant for ªrms with foreign and ethnic capitals at the ªrm level, pero
insigniªcant for ªrms with ethnic capitals at the industry level. It is likely to be ex-
plained by their inappropriate nature in the developing country context and the
strong intellectual property rights protection in the high-technology industries. En
addition, we should bear in mind that the international R&D spillover data are from
the OECD countries, which might have a limited knowledge transfer through the
ethnic channel.

4.2 Determinants of technical change and scale efªciency improvement
Mesa 5 reports the estimated result on the impact of indigenous innovation efforts
and foreign R&D spillovers on technical change and efªciency improvements of the
indigenous ªrms. We ªnd no systematic relationship between the age of the ªrm
and both technical change and efªciency improvement. Bigger ªrms appear to have
experienced faster rates of technical change across all speciªcations. Sin embargo, el

15

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

yo
a
norte
oh
i
t
a
norte
a
r
t
norte
I

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

yo
a
norte
oh
i
t
a
norte
a
r
t
norte
I

METRO
METRO
GRAMO

S
l
oh

s
r
mi
v
oh
yo
yo
i
pag
s

*
*
*
3
9
1
0
(cid:3)

.

)
2
7
0
0
0
(

.

5
1
0
0
0

.

)
8
1
0
0
0
(

.

*
*
*
6
1
2
0
0
(cid:3)

.

)
3
1
0
0
0
(

.

*
*
5
3
1
0

.

)
8
5
0
0
(

.

*
*
*
1
5
0
0
0

.

)
9
0
0
0
0
(

.

*
*
*
6
2
7
0
0

.

)
4
8
0
0
0
(

.

*
*
*
4
1
0
0

.

0

)
4
0
0
0
0
(

.

*
*
*
3
4
9
0

.

)
9
2

.

0
(

*
*
*
1
6
5
0

.

)
2

.

0
(

*
*
1
9
4
0

.

)
2
2

.

0
(

*
*
*
4
3
0
0
0

.

)
1
1
0
0
0
(

.

s
r
mi
v
oh
yo
yo
i
pag
s

*
*
*
3
0
2
.
0
(cid:3)

)
1
7
0
0
.
0
(

4
0
0
0
.
0

)
8
1
0
0
.
0
(

*
*
*
0
2
2
0
.
0
(cid:3)

)
3
1
0
0
.
0
(

*
*
*
1
6
1
.
0

)
6
5
0
.
0
(

*
*
*
7
4
0
0
.
0

)
9
0
0
0
.
0
(

*
*
*
7
2
7
0
.
0

)
3
8
0
0
.
0
(

*
*
*
1
1
0
0
.
0

)
4
0
0
0
.
0
(

*
*
*
7
3
9
.
0

)
8
2
.
0
(

*
*
*
6
4
2
.
0

)
3
7
0
.
0
(

*
*
1
7
1
.
0

)
5
7
0
.
0
(

*
*
*
2
4
0
0
.
0

)
1
0
0
.
0
(

yo
mi
v
mi
yo

metro

r
i
F

*
*
*
3
0
2
.
0
(cid:3)

)
1
7
0
0
.
0
(

6
0
0
0
.
0

)
8
1
0
0
.
0
(

*
*
*
5
1
2
0
.
0
(cid:3)

)
3
1
0
0
.
0
(

*
*
*
7
8
1
.
0

)
4
5
0
.
0
(

*
*
*
7
4
0
0
.
0

)
9
0
0
0
.
0
(

*
*
*
1
3
7
0
.
0

)
3
8
0
0
.
0
(

*
*
*
0
1
0
0
.
0

)
4
0
0
0
.
0
(

*
*
*
6
3
9
.
0

)
8
2
.
0
(

*
*
*
3
4
2
.
0

)
3
7
0
.
0
(

*
*
1
6
1
.
0

)
5
7
0
.
0
(

s
r
mi
v
oh
yo
yo
i
pag
s

*
*
*
0
6
1
.
0
(cid:3)

)
7
4
0
0
.
0
(

2
1
0
0
.
0

)
1
1
0
0
.
0
(

*
*
*
0
9
1
0
.
0
(cid:3)

)
9
0
0
0
.
0
(

*
*
*
1
2
1
.
0

)
2
4
0
.
0
(

*
*
*
3
4
0
0
.
0

)
7
0
0
0
.
0
(

*
*
*
2
8
6
0
.
0

)
9
6
0
0
.
0
(

*
*
*
5
2
0
0
.
0

)
2
0
0
0
.
0
(

*
*
*
9
5
4
.
0
(cid:3)

1
9
1
0
.
0

)
5
3
0
.
0
(

1
7
0
0
.
0

)
3
2
0
.
0
(

)
8
0
.
0
(

*
*
*
1
3
0
0
.
0

)
8
0
0
0
.
0
(

s
r
mi
v
oh
yo
yo
i
pag
s

*
*
*
4
6
1
.
0
(cid:3)

)
7
4
0
0
.
0
(

0
1
0
0
.
0

)
1
1
0
0
.
0
(

*
*
*
4
8
1
0
.
0
(cid:3)

)
9
0
0
0
.
0
(

*
*
*
8
4
1
.
0

)
1
4
0
.
0
(

*
*
*
3
4
0
0
.
0

)
7
0
0
0
.
0
(

*
*
*
7
7
6
0
.
0

)
7
6
0
0
.
0
(

*
*
*
4
2
0
0
.
0

)
2
0
0
0
.
0
(

*
*
*
4
4
4
.
0
(cid:3)

)
5
7
0
.
0
(

*
6
7
3
0
.
0

)
2
2
0
.
0
(

8
2
0
0
.
0

)
6
1
0
.
0
(

*
*
*
9
3
0
0
.
0

)
7
0
0
0
.
0
(

yo
mi
v
mi
yo

metro

r
i
F

*
*
*
4
6
1
.
0
(cid:3)

0
7
4
0
0
.
0
(

3
1
0
0
.
0

)
1
1
0
0
.
0
(

*
*
*
9
7
1
0
.
0
(cid:3)

)
9
0
0
0
.
0
(

*
*
*
0
6
1
.
0

)
9
3
0
.
0
(

*
*
*
3
4
0
0
.
0

)
7
0
0
0
.
0
(

*
*
*
2
8
6
0
.
0

)
7
6
0
0
.
0
(

*
*
*
4
2
0
0
.
0

)
2
0
0
0
.
0
(

*
*
*
7
3
4
.
0
(cid:3)

)
5
7
0
.
0
(

*
6
6
3
0
.
0

)
2
2
0
.
0
(

1
2
0
0
.
0

)
6
1
0
.
0
(

y
C
norte
mi
i
C
ª
F
mi

yo
a
C
i
norte
h
C
mi
t

yo
a
i
t
i
norte
I

t
norte
mi
i
C
ª
F
mi
oh
C

mi
r
tu
t
i
d
norte
mi
pag
X
mi
gramo
norte
i
norte
i
a
r
t

t
mi
s
s
a

mi
yo
b
i
gramo
norte
a
t
norte
I

y
t
i
s
norte
mi
t
norte
i

t
r
oh
pag
X
mi

y
t
i
s
norte
mi
t
norte
i

D
&
R

mi
r
a
h
s
norte
gramo
i
mi
r
oh
F

mi
r
a
h
s

C
i
norte
h
t
mi

mi
r
a
h
s
D
&
R
C
i
t
s
mi
metro
oh
D

t
norte
mi
metro
y
oh
yo
pag
metro
mi

mi
z
i
s

t
mi
k
r
a
METRO

mi
gramo
A

h
t
w
oh
r
gramo
PAG
F
t
mi
h
t

F
oh
s
t
norte
a
norte
i
metro
r
mi
t
mi
D

.
4

mi
yo
b
a
t

16

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

s
r
mi
v
oh
yo
yo
i
pag
s

)
6
0
0
0
0
(

.

0
0
0
0
0
(cid:3)

.

*
*
*
9
1
0
0
0
(cid:3)

.

)
7
0
0
0
0
(

.

*
*
5
4
1
0
0
(cid:3)

.

)
9
6
0
0
0
(

.

*
*
8
9
1
0
0
(cid:3)

.

)
5
9
0
0

.

0
(

*
*
*
3
5
0
0
0
(cid:3)

.

)
8
1
0
0
0
(

.

)
8
1
0
0
0
(

.

6
0
0
0
0
(cid:3)

.

*
*
*
5
2
3
1

.

)
2
2
0
0
(

.

5
8
8
5
5
1

,

1
0

.

0

0

1
5
4
6
0

.

2
7
2
,
8
5
1

1
0
.
0

0

2
6
9
5
.
0

*
*
*
7
7
2
.
1

)
1
1
0
.
0
(

*
*
*
5
9
2
.
1

)
8
8
0
0
.
0
(

2
7
2
,
8
5
1

1
0
.
0

0

2
9
7
6
.
0

METRO
METRO
GRAMO

yo
a
norte
oh
i
t
a
norte
a
r
t
norte
I

s
r
mi
v
oh
yo
yo
i
pag
s

)
5
0
0
0
.
0
(

7
0
0
0
.
0
(cid:3)

*
*
*
4
2
0
0
.
0
(cid:3)

)
7
0
0
0
.
0
(

yo
mi
v
mi
yo

metro

r
i
F

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

s
r
mi
v
oh
yo
yo
i
pag
s

)
5
0
0
0
.
0
(

1
0
0
0
.
0
(cid:3)

*
*
9
0
0
0
.
0
(cid:3)

)
5
0
0
0
.
0
(

8
0
0
0
.
0

)
1
1
0
0
.
0
(

3
0
0
0
.
0
(cid:3)

)
1
0
0
.
0
(

*
*
3
5
0
0
.
0
(cid:3)

)
3
1
0
0
.
0
(

)
3
1
0
0
.
0
(

5
0
0
0
.
0
(cid:3)

*
*
*
9
8
1
.
0

)
4
1
0
.
0
(

5
3
1
,
3
9
1

6
5
.
0

S
l
oh

yo
a
norte
oh
i
t
a
norte
a
r
t
norte
I

s
r
mi
v
oh
yo
yo
i
pag
s

*
8
0
0
0
.
0
(cid:3)

)
4
0
0
0
.
0
(

*
*
4
1
0
0
.
0
(cid:3)

)
4
0
0
0
.
0
(

yo
mi
v
mi
yo

metro

r
i
F

8
0
9
,
7
9
1

7
5
.
0

*
*
*
4
3
1
.
0

)
3
7
0
0
.
0
(

*
*
*
3
5
1
.
0

)
4
5
0
0
.
0
(

8
0
9
,
7
9
1

7
5
.
0

y
r
t
s
tu
d
norte
i
_
norte
gramo
i
mi
r
oh
F
_
D
&
R

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

y
r
t
s
tu
d
norte
i
_
C
i
norte
h
t
mi
_
D
&
R

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

metro
r
ª
_
norte
gramo
i
mi
r
oh
F
_
D
&
R

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

metro
r
ª
_
C
i
norte
h
t
mi
_
D
&
R

yo
a
norte
oh
i
t
a
norte
r
mi
t
norte
I

d
mi
tu
norte
i
t
norte
oh
C

.
4

mi
yo
b
a
t

mi
r
a
h
s
D
&
R
C
i
norte
h
t
mi

t
norte
mi
i
C
ª
F
mi
oh
C

mi
r
a
h
s
D
&
R
norte
gramo
i
mi
r
oh
F

s
norte
oh
i
t
a
v
r
mi
s
b
oh

2
R

t
s
mi
t

s
tu
oh
norte
mi
gramo
oh
X
mi

t
s
mi
t

j
norte
mi
s
norte
a
h

t
norte
a
t
s
norte
oh
C

17

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

.
yo
mi
v
mi
yo

t
norte
mi
C
r
mi
pag
1

mi
h
t

t
a

t
norte
a
C
ª
i
norte
gramo
i
s
y
yo
yo
a
C
i
t
s
i
t
a
t
S
*
*
*

.
yo
mi
v
mi
yo

t
norte
mi
C
r
mi
pag
5

mi
h
t

t
a

t
norte
a
C
ª
i
norte
gramo
i
s

y
yo
yo
a
C
i
t
s
i
t
a
t
S
*
*

.
yo
mi
v
mi
yo

t
norte
mi
C
r
mi
pag
0
1
mi
h
t

t
a

t
norte
a
C
ª
i
norte
gramo
i
s

y
yo
yo
a
C
i
t
s
i
t
a
t
S
*

.
s
mi
i

metro
metro
tu
d
y
r
t
s
tu
d
norte
i

t
i
gramo
i
d
–
oh
w

t
d
norte
a

mi

metro

i
t

F
oh

t
mi
s

yo
yo
tu
F

mi
h
t

mi
d
tu
yo
C
norte
i

s
norte
oh
i
t
a
C
ª
i
C
mi
pag
s

yo
yo

A

.
s
mi
s
mi
h
t
norte
mi
r
a
pag
norte
i

s
r
oh
r
r
mi

d
r
a
d
norte
a
t
s

t
s
tu
b
oh
R

:
mi
t
oh
norte

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

Mesa 5. Determinants of the technical change and efªciency improvement

Coefªcient

Initial technical efªciency

Age

Employment

Market competition

Intangible asset

Training expenditure

Export intensity

R&D intensity

Foreign share

Ethnic share

Domestic R&D share

Ethnic R&D share

Foreign R&D share

International R&D_foreign_ªrm

International R&D_ethnic_ªrm

International R&D_foreign_industry

International R&D_ethnic_industry

Constant

Observaciones
R2

Exogenous test

Hansen J test

TECH

Intranational
spillovers
(cid:3)0.183***
(0.0076)
0.0029
(0.002)
0.0161***
(0.0014)
(cid:3)0.0951
(0.064)
(cid:3)0.0039***
(0.001)
(cid:3)0.0414***
(0.008)
(cid:3)0.0001
(0.0004)
(cid:3)0.223
(0.17)
(cid:3)0.0957
(0.075)
(cid:3)0.124*
(0.071)
0.0064***
(0.0013)
(cid:3)0.0005
(0.0007)
(cid:3)0.0085***
(0.0008)

0.958***
(0.013)

158,272

0.69

0

0.9587

Internacional
spillovers
(cid:3)0.189***
(0.0077)
0.0019
(0.002)
0.0154***
(0.0015)
(cid:3)0.199***
(0.068)
(cid:3)0.0046***
(0.001)
(cid:3)0.0447***
(0.0081)
(cid:3)0.0001
(0.0004)
(cid:3)0.239
(0.17)
(cid:3)0.369*
(0.2)
(cid:3)0.2760
(0.2)
0.0029**
(0.0014)
0.0031***
(0.0008)
(cid:3)0.0083***
(0.0009)
0.0125*
(0.007)
0.0112
(0.0088)
0.0064***
(0.0022)
(cid:3)0.0193***
(0.0021)
1.096***
(0.025)

155,885

0.69

0

0.3326

EFFCH

Intranational
spillovers
(cid:3)0.493***
(0.011)
0.0029
(0.0027)
0.0009
(0.0019)
0.0472
(0.094)
(cid:3)0.0007
(0.0013)
0.0532***
(0.012)
0.0011*
(0.0006)
1.395***
(0.33)
0.139
(0.11)
0.0086
(0.091)
(cid:3)0.00551**
(0.0016)
0.0000
(0.0009)
0.0031***
(0.0011)

1.545***
(0.018)

158,272

0.41

0

0.4248

Internacional
spillovers
(cid:3)0.482***
(0.011)
0.0048*
(0.0027)
0.0020
(0.0019)
0.153
(0.097)
0.0002
(0.0014)
0.0559***
(0.012)
0.0009*
(0.0006)
1.445***
(0.34)
0.46
(0.29)
0.0887
(0.26)
(cid:3)0.0017
(0.0018)
(cid:3)0.0038***
(0.001)
0.0027**
(0.0011)
(cid:3)0.016
(0.0098)
(cid:3)0.0051
(0.011)
(cid:3)0.0031
(0.003)
0.0182***
(0.0027)
1.375***
(0.035)

155,885

0.41

0

0.4048

Nota: Robust standard errors in parentheses. All speciªcations include the full set of time and two-digit industry dummies.

*Statistically signiªcant at the 10 nivel porcentual. **Statistically signiªcant at the 5 nivel porcentual. ***Statistically signiªcant at the

1 nivel porcentual.

effect of size is not statistically signiªcant to efªciency improvement. Exports con-
tribute to efªciency change and catch-up, but not the shift of the technology frontier.
This result is consistent with the ªndings in Fu (2005), which uses Chinese industry-
level panel data and suggests that the focus on low-cost competitiveness based on
cheap unskilled labor and the dominance of process trading in the export structure
provide no effective incentive for ªrms to innovate.

18

Asian Economic Papers

yo

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

/

yo

F
/

/

/

/

/

8
2
1
1
6
8
2
5
6
2
a
s
mi
pag
2
0
0
9
8
2
1
pag
d

.

.

.

.

.

F

b
y
gramo
tu
mi
s
t

t

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

Derrames tecnológicos internacionales e intranacionales y crecimiento de la productividad en China

The estimated coefªcients on the intangible assets variable are negative and statisti-
cally signiªcant in the technical change equation. This is likely because intangible
assets include, according to Chinese accounting practices, R&D investment in the
development stage but not the research stage. For ªrms in technology-intensive in-
Industrias, novel research activities may play a more important role in keeping them
on the frontier and promoting TFP growth. The second reason may be that intangi-
ble assets are correlated with the ªxed assets that we used for the TFP estimation.
Por lo tanto, technically, there is a negative association between TFP and intangible as-
conjuntos. Training exerts a signiªcant positive impact on efªciency improvement as ex-
pectado, but surprisingly bears a signiªcant negative impact on technical change.
This fact suggests that training is likely to occur for teaching new or advanced prac-
tices, but not for the creation of frontier technology. The estimated coefªcients of
market concentration are mostly insigniªcant.

The indigenous R&D of individual ªrms has no signiªcant impact on technical
cambiar. Sin embargo, it has contributed signiªcantly to efªciency improvement, cual
reºects the catch-up process. This is not surprising given the fact revealed from the
First National Economic Census in 2004 that about 95 percent of total business R&D
expenditure was spent on development and only 5 percent was spent on basic
scientiªc research. Curiosamente, R&D activities in the domestic ªrms at the industry
level have shown signiªcant and robust positive spillovers on the technical progress
of indigenous ªrms. This evidence suggests that it is collective indigenous R&D ac-
tivities, a saber, R&D at the industry level, which push up the technology frontier
and drive the technology upgrading of indigenous ªrms. R&D activities of foreign
invested ªrms at the industry level have shown a negative spillover effect on the
technical change of indigenous ªrms but a positive spillover effect on technological
catch-up. This may be explained by the competition effects from the foreign R&D
activities and the ªndings of recent studies that the core technology development of
MNEs still remain at the headquarters, while applied research and adaptation are
the main tasks of its afªliates in foreign countries. Por lo tanto, these R&D activities
may not contribute to technical change but their impact on catching up is positive
and statistically signiªcant. The spillover effects of R&D investment in ethnic ªrms,
sin embargo, only show signiªcances when international R&D spillover effects are con-
sidered and have opposite signs from those of FIEs. It suggests that the R&D activi-
ties from ethnic ªrms at the industry level can help to foster technical upgrading,
but not technological catch-up.

In terms of FDI spillover effects, there is no direct FDI impact on efªciency change,
but a signiªcant negative impact is detected on technical change. De nuevo, this is
likely for ªrms with foreign capital that might tend to keep their core technology re-

19

Asian Economic Papers

yo

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