Patterns of China’s industrialization
Cheryl Long
Colgate University
Xiaobo Zhang
International Food Policy Research Institute
September 2008
Introduction
In the past three decades China has experienced the same degree of industrialization that
took two centuries to occur in Europe (Summers, 2007). Its rapid industrialization has
been accompanied by the emergence of numerous “specialty cities” of a particular kind.
Thousands of firms, large and small, each specialized in a finely defined production step,
are lumped together in a densely populated region, where some particular manufactured
consumer good is churned out in millions (if not billions) annually. Many formerly rural
towns in the coastal areas have become so specialized, boasting themselves as the world’s
Socks City, Sweater City, Kid’s Clothing City, Footwear Capital, and so on. Despite
numerous popular media reports on this phenomena, 1 few studies have rigorously
documented these patterns using data covering a large sample and a long period.
Each of the specialty cities we describe above fits Porter’s concept of an industrial cluster,
which is “a geographically proximate group of inter-connected companies (and
associated institutions) in a particular field” (Porter 1998, Porter 2000). Is increasing
clustering part of the general patterns of China’s industrialization and economic growth
in the past three decades? We will document the patterns of China’s industrialization by
proposing and applying multi-dimensional measures for clustering. We will also provide
some preliminary results on how these patterns relate to economic growth in China.
Our paper is unique on several fronts. First, we have access to firm level data in two time
periods for China as a whole, which is more disaggregate and updated than previous
studies. Second, we have adopted five different measures to capture the evolving patterns
of clustering, a key feature of China’s industrialization. Our results suggest that China’s
rapid industrialization is marked with increased industrial concentration, regional
specialization, and a growing number of small and medium enterprises (SMEs), as well
as closer interactions among firms within industry and within region. This development
fits particularly well with China’s comparative advantage, which is marked with
abundant labor with limited capital, during its initial stage of taking off.
1
For example, see http://www.nytimes.com/2004/12/24/business/worldbusiness/24china.html for a New
York Times report.
1
China’s miraculously rapid industrialization and commercialization (Lin, Cai, and Li,
2003; Summers, 2007) provide a unique laboratory enabling us to observe and
understand the process of industrialization. While industrialization in Western Europe
and North America at the early stages of the Industrial Revolution can now be studied
only through the relatively dim mirror of history, they can be viewed directly in the
ongoing economic revolution in China. A clearer understanding of these processes will
have great value in helping propagate them to the world’s less fortunate regions.
The structure of this paper is as follows: Section 2 will review the literature and discuss
how to measure clusters. Section 3 describes data. Section 4 presents the patterns of
China’s industrialization, while a short conclusion is offered in Section 5.
Literature Review and Measurement of Clustering
Industrialization is often accompanied by spatial agglomeration of industrial activities.
Two types of agglomeration have been observed during the industrialization process of
developed countries. In the U.K., the decentralized production system scattered in
different family workshops was replaced by a large integrated factory system during the
Industrial Revolution (Landes, 1998). The trend was similar and more evident in the
United States during its industrialization (Chandler, 1977). For example, the auto industry
is highly concentrated in the Detroit metropolitan area with several dominant large firms.
One strand of theoretical literature argues that spatial agglomeration is a key feature of
industrialization (Krugman, Fujita, and Venables, 2001). In this type of industrial cluster
(or spatial agglomeration), firm size is generally very large. As highlighted by Marshall
(1920), agglomeration generates several positive externalities for firms: better access to
the market and suppliers, labor pooling, and easy flow of technology know-how.
Italy, Japan and other East Asian countries and regions experienced a different path of
spatial agglomeration during the course of industrialization, which was led by small and
medium enterprises (SMEs). In this business model, a large number of SMEs often
cluster together with comprehensive vertical division of labor. One noted example is the
putting-out system in which a merchant obtained market orders and subcontracted the
production to nearby farmers or skilled workers who usually finished the work in their
homes or family workshops (Hounshell, 1984). The putting out system was popular in the
U.K. prior to its Industrial Revolution and was widely observed in nineteenth-century
Japan (Nakabayashi, 2006). Outsourcing (or subcontracting), the modern variant of the
traditional putting-out system, remains a major feature of industrial production
organization in contemporary Japan and Taiwan (Sonobe and Otuska, 2006). Industrial
districts where different workshops and factories cluster together were ubiquitous in
France and Italy until the mid-twentieth century and are still viable in some regions of
Italy (Piore and Sabel, 1984; Porter, 1998). When the SMEs cluster, they also form
agglomeration. The key difference of this type of agglomeration from the first one is the
finer vertical division of labor in the production process. By dividing a production
process into incremental stages, a large lump sum investment can be transformed into
many small steps, therefore lowering the capital entry barriers (Schmitz, 1995; Ruan and
2
Zhang, 2008). Therefore, this mode of industrial organization may fit better to countries
or regions with scarce capital and less developed financial sector.
One strand of literature on China’s industrialization debates whether China’s economic
sectors have become more or less concentrated since economic reform began in the late
1970s. Previous studies seem to have provided mixed conclusions. Young (2000) found
that China’s regions had become less specialized among the broadly defined sectors
(agriculture, industry, construction, transportation, and commerce, or primary, secondary,
and tertiary) up to the early 1990s. Using data from the second and third national
industrial censuses, Wen (2004) showed that manufacturing industries have become
increasingly geographically concentrated up to 1995. Based on a panel data set of 32
industries at the two-digit level of aggregation in 29 provinces, Bai et. al (2004) found
that regional industrial production has become more specialized. The findings from both
studies are in contrast to Young’s findings. Using similar data to Young (2000) but with a
longer and more recent period, Zhang and Tan (2007) showed that the regional
distribution of industries initially became more dispersed before growing more
concentrated. The difference among these findings may be primarily due to their different
levels of aggregation and time periods covered. With access to firm level data nationwide,
we will be able to extend the previous analysis and shed more light on the debate in the
current study.
Another limitation of this strand of literature on China’s industrial concentration is its
negligence to distinguish the two different types of agglomeration. To our knowledge,
there have been no quantitative studies to empirically test which type of agglomeration
China has followed during its rapid industrialization course. Several in-depth case studies
and popular media reports (as mentioned in the beginning of the paper) seem to suggest
that China followed the cluster-based industrialization path (Sonobe, Hu and Ostuka,
2002 and 2004; Ruan and Zhu, 2008; Huang, Zhang, and Zhu, 2008). For example,
Sonobe, Hu and Ostuka (2002) studied how a garment cluster formed in a rural town in
Zhejiang Province starting with small scale production. Huang, Zhang, and Zhu (2008)
detailed how the footwear cluster in Wenzhou helped overcome financial, institutional,
and technological barriers. Ruan and Zhu (2008) in particular demonstrated that
clustering lowers capital entry barriers and enables more entrepreneurs to participate in
the production process. These studies provide insight on how clusters work. However, it
is not clear whether these case studies can be generalized broadly or can shed light on
how China’s industrialization has evolved over time.
Most of these studies use regional specialization or industrial concentration as the
measure for agglomeration or clustering. The market share of a certain number of the
largest, say, three firms, in an industry or region is often used as a concentration measure.
The advantage of this measure is that it is easy to calculate and interpret, but when the
distribution of firms is relatively spread out, it may miss those firms below the cut-off
lines. To overcome this problem, the Gini coefficient is often used to calculate the
regional variation of output or employment shares for all the firms in an industry.
Krugman (1991) modifies the Gini coefficient by accounting for the discrepancy between
a region’s share of output/employment in an industry and its share in all manufacturing
3
industries in calculating the Gini coefficient. In this paper, we use the market
concentration index, the Gini, and the Krugman Gini as measures of regional
concentration.
Oftentimes these concentration measures do not distinguish between the following two
kinds of “agglomeration”: one where a small number of large firms with minimal interfirm connections are located, versus the other where a large number of small firms
congregate and interact closely with one another. While the first type of agglomeration
characterizes cities such as Detroit, the second type of agglomeration is the one that
better fits the patterns observed in coastal China.
To better identify clusters as defined in this paper, which closely follow the definition
given by Porter (Porter 1998, Porter 2000), we adopt a multi-dimensional measure of
clustering. In addition to the industrial concentration and regional specialization
mentioned above, we also study the distributions of the average number and size of firms
for each industry in the various regions in China, the average value added/output ratio of
firms for each industry in various regions, and the industry proximity of firms in each
region.
The distribution of the number of firms and firm size allows us to explore the geographic
distribution of firms operating in each industry. The ratio of value added/output enables
us to examine the degree of vertical division of labor among firms in a location. If there
are many firms engaged in a wide range of production steps for a homogenous product,
then the gross output value will be large as it is counted many times along the chain of
production. Therefore, the smaller the value added/output ratio, the higher degree of
vertical division of labor among firms.
Porter (1998) has used input-output tables to calculate the connection of industries within
a location as a proxy of clustering. Along the same line, Hausmann-Klinger (2006) put
forward a proximity matrix to illustrate the technological affinity between different
products. Using the matrix, one can calculate the weighted average of industry proximity
between each pair of industries located in the same region. In this paper, we use the
Hausmann-Klinger measure as one of the indicators for clustering to study the
technological closeness of industries in a region.
Data Description and Processing Procedures
We utilize firm level data from the China Industrial Census 1995 and the China
Economic Census 2004 for analysis in this paper. Tables 1 and 2 present summary
statistics of gross industrial output value from the censuses by industry and by region,
respectively. Table 3 compares the sample of our data sets with the published national
aggregate statistics for China in 1995 and 2004. As shown in the table, our data sets
capture the whole universe of Chinese industrial firms in these two years. Compared to
data sets used in previous studies (Young, 2000; Bai et. al, 2004), our data sets have more
4
comprehensive coverage and include industrial firms of all sizes (not only those above a
certain scale).
Since the data is at the firm level, we can calculate the degree of clustering at any level of
our choice, such as township, county, prefecture, or province, for regional aggregation,
and 2-, 3-, or 4-digit industry level for sectoral aggregation. For the main part of the
analysis, we choose province and 2-digit CIC (China Industry Code) as the levels of
aggregation. But for robustness tests, we also use prefecture and county levels for
geographic aggregation, and 3-digit and 4-digit CIC for industrial aggregation.
Because China modified its industry coding system in 2002 (switching from GB1994 to
GB2002), we match industry codes that have changed from 1994 to 2002 as follows: for
industry codes that have become more disaggregate, we use the 1994 codes as the
standard; for those that have become more aggregate, we use the 2002 codes as the
standard. In other words, we use the more aggregate codes to group and compare
industries between 1995 and 2004. During the period between the two censuses (19952004), some counties have also been elevated to cities and have changed their names. We
have carefully tracked these changes to match the counties throughout the time period.
When constructing the various measures, we first determine the level of aggregation to
convert firm level data to cell-level totals, where each cell is a combination of region and
industry. For example, the most detailed cell is the 4-digit-CIC-industry-county
combination. Five sets of measures are then created using the cell-level data:
concentration indices (both for industries and for regions), distributions of firm number
and firm size, value added/output ratio, and an industry proximity measure.
Concentration indices include the total share of gross industrial output value contributed
by the three largest producers (referred to as CR3 henceforth), the Gini coefficient, as
well as the Krugman-Gini coefficient. When these measures are computed for each
industry (thus with regions as producers), they are referred to as industrial concentration
indices. Similar measures become indicators for regional specialization when they are
computed for each region (thus with industries as producers). Using the cell-level
aggregates, we also compute the average number of firms and the average firm size in the
cell, as well as the value added/output ratio. For firm number and firm size, we further
compute their concentration indices (CR3, Gini, and Krugman-Gini) for each industry.
For the value-added/output ratio, we compute its average value for each industry. To
measure industry proximity, we use the proximity matrix from Hausmann-Klinger (2006).
Because the proximity matrix is computed for products at the SITC 4-digit level, we have
made concerted effort to convert the CIC code first to ISIC code and then to SITC code
based on the manuals obtained from China’s National Bureau of Statistics as well as
correspondence tables from EuroStat and the United Nations. This measure shows both
how specialized China’s regions have grown and how firms within each region interact
with one another. In summary, we have three sets of measures describing concentration,
distribution, and interaction of firms within the same industry, and two measures
describing specialization and interaction of firms within the same region.
5
Patterns of China’s Industrialization
We obtain several mutually consistent patterns when using the above measures to explore
China’s industrialization process between 1995 and 2004. In addition, we find some
preliminary linkage between some of these patterns and regional growth in China. We
begin with patterns along the sector dimension, then present those along the regional
dimension, and finish with discussions on how these patterns relate to one another and
how they affect regional growth in China from 1995 to 2004.
(1) Increasing industrial concentration:
Our data show that between 1995 and 2004, most industries in China have become more
concentrated geographically. Various measures provide consistent evidence for this
pattern. Table 4 compares three concentration measures between 1995 and 2004: the total
output share of the top three producing provinces, the Gini coefficient of output across all
provinces, and the Krugman-Gini coefficient of output across all provinces, all computed
for each of the 2-digit industries. As shown in the table, the three measures mostly yield
consistent results of industrial concentration: most industries have become more
concentrated geographically during this time period. The only exception is the tobacco
industry, where the decline in concentration may be explained by continued local
protectionism (Bai et. al, 2004).
The last two rows provide the weighted sample means of the measures as well as the
1995-2004 differences and their t-statistics. These results indicate that the differences
between 1995 and 2004 are statistically significant. Figure 1 illustrates the same pattern
graphically, using 4-digit CIC industries and CR3 as the concentration index. As the
majority of the observations lie above the 45 degree line, most industries were more
concentrated in 2004 than in 1995.
Table 5 lists the top three provinces ranked by output for each of the industries.
Explanations offered in Wen (2004) for the geographic patterns of industries in 1995 still
largely apply for the time period of 1995-2004. Briefly, affinity to natural resources,
availability of labor and infrastructure, as well as policy initiatives, all explain the output
rankings of Chinese regions in different industries. In addition, as the trend of industrial
concentration has intensified between 1995 and 2004, the initial conditions listed above
seem to have become more important in determining the geographic allocation of
industries over time. It is also apparent from the table that several coastal provinces, such
as Guangdong, Zhejiang, Jiangsu, and Shandong, have become increasingly dominant in
more industries.
(2) Greater concentration of more equally sized firms:
As discussed previously, the industrial concentration measures do not distinguish
between the Detroit-style agglomeration and the Chinese-style cluster, since the
6
concentration measure is an aggregate measure masking most firm-level details. To
explore the distinction between these two types of agglomeration, we study an additional
set of measures, the number of firms and average firm size in a certain region,
constructed for each industry. These measures suggest that the higher geographic
concentration of industries observed above seems to have been achieved mainly through
a larger number of similar-sized firms (some of which are necessarily new firms)
combined with a more skewed distribution of these firms, rather than through the
disproportionate expansion of existing firms.
For each 2-digit industry, the average number of firms in a province has increased from
543 to 1,550 from 1995 to 2004, while the average gross industrial output value of firms
has increased from RMB 43.6 million to RMB 63.2 million. Although both changes are
statistically significant, the change in firm size is much less significant. But it is more
important to study the distributions of firm number and firm size. Table 6 presents the
Gini coefficients for both the average number and the average size of firms, while Table
7 lists the number of firms and average firm size by industry. Clearly, the average
number of firms has become significantly less evenly distributed, while the average size
has become slightly more evenly distributed (although not statistically significant). In
other words, while the average firm size has become more equal across regions, the
number of firms has become less equally distributed across regions, with the top
producing regions in each sector having a larger share of the total number of firms in that
sector in 2004 than in 1995. This suggests that the increased industrial concentration
across regions is mainly driven by the increasing number of similar firms located in the
leading production regions.
Again, these patterns are robust to different levels of aggregation. Figure 2 presents the
Gini-coefficients in 1995 versus 2004 for firm number at the 4-digit CIC industry level,
while Figure 3 presents the corresponding Gini-coefficients for firm size. As
demonstrated in the figures, while the number of firms has become significantly more
concentrated during this period, the average firm size has become slightly more evenly
distributed, although the change over time is not significant.
(3) Increasing interaction among firms:
The greater concentration of more equally sized firms that we observed in 2004 suggests
a trend of clustering in Chinese industries that differs from the Detroit-style
agglomeration. But is there much interaction among the large number of firms located in
the same region and industry? Using the value added/output ratio as a measure for firm
interaction, we found that generally there have been increasing interactions among firms
located in the same industry in 2004 than in 1995. Table 8 presents the average value
added/output ratio for the 2-digit industries in 1995 and 2004, where the ratio is
computed by dividing the total value added in the 2-digit industry by the total gross
industrial output value in the same industry. As shown in the table, the value
added/output ratio has generally decreased from 1995 to 2004 and the change is
statistically significant.
7
Using 3-digit or 4-digit industries gives similar results. Figure 4 uses 4-digit CIC
industries to compare the value added/output ratio in 1995 with that in 2004, showing the
same pattern. In other words, Chinese firms are not only increasingly lumped together
physically, but they are also increasingly interacting with firms that are technologically
close by. In particular, a lower value added/output ratio indicates finer division of labor
among firms in the same industry.
(4) Increasing regional specialization:
As shown in Table 5, Chinese industries have become more concentrated geographically
and several provinces appear as the top producers in multiple industries. Does this mean
that a small number of provinces have allocated increasing amounts of resources to
multiple industries? In other words, have regions become less specialized or more
specialized while industries have become more concentrated?
To answer this question, we use measures of regional specialization. They are similar to
measures for industrial concentration. The only difference is that a region instead of an
industry is used as an observational unit. Table 9 compares the following measures
between 1995 and 2004: the total output share of the top three producing industries, the
Gini coefficient of output across all industries, and the Krugman-Gini coefficient of
output across all industries, computed for each of the Chinese provinces. 2
These different measures give largely consistent evidence: within each region, the top
industries account for a significantly larger share of the total regional output in 2004 than
in 1995. In other words, Chinese regions have also become more specialized during this
time period. Although Table 9 uses province as the regional unit to facilitate presentation,
measures at the prefecture and county levels show the same trend of increasing regional
specialization. Figure 5 illustrates the comparison between CR3 in 1995 and that in 2004
using prefecture level data, showing the same pattern. 3 Our results based on more
updated and comprehensive data support the view that Chinese regions have become
increasingly specialized since the mid-1990s. In other words, the product market has
become more integrated over time.
(5) Increasing product proximity within each region:
The results presented above have established the pattern that Chinese regions have
become more specialized. In other words, they have increasingly allocated more
resources to the top few industries. But how are these industries chosen? Do they tend to
be closely related or diversely distributed industries? We use the measure of product
proximity to answer this question. Using the product proximity measure defined in
2
Measures computed for the prefectural and county levels give similar results.
County level data gives the same results but has too many observations to give effective graphic
illustration.
3
8
Hausmann-Klinger (2007), we found that within each region, the proximity among
industries has increased significantly between 1995 and 2004.
Table 10 presents the industry proximity measure for each of the Chinese provinces in
1995 and 2004. The measures constructed at the county and the prefecture levels give the
same pattern of higher industry proximity in each region. Figure 6 shows the pattern
using measures computed at the prefecture level. 4
(6) Consistency among patterns of China’s industrialization:
In the previous sections, we use multiple measures to examine the patterns of
industrialization in China between 1995 and 2004. Tables 11 and 12 also present the
correlations among different measures in industry concentration and regional
specialization. For the industries, the more concentrated their gross outputs are, the more
concentrated the number of firms and the average firm size, and the finer the division of
labor among firms. It is especially interesting that the correlation coefficient between the
value added/output ratio is negative with other measures. This suggests that in areas with
more concentrated industries, the vertical division of labor among firms is finer. This
provides supportive evidence that China’s industrial agglomeration is largely driven by
SMEs, rather than by the emergence of large firms.
For a particular region, the more specialized it is in industrial allocation, the greater
proximity will be among industries, though the patterns hold more for 2004. Interestingly,
the correlation between the proximity measure and other measures is insignificant in
1995, probably suggesting the existence of local protectionism at the time. At that time,
many regions may still have been developing their industries based on something other
than their comparative advantages in terms of the linkage between upstream and
downstream industries. This may explain why Young (2000) found a serious problem of
protectionism in the early 1990s. Largely in response to the prevalent market
fragmentation, since the mid-1990s China has implemented a series of policy reforms to
integrate the product market (Zhang and Tan, 2007). Reassuringly, our industry
proximity measure became significantly correlated with the CR3 and Gini measures by
2004. Regions that are more specialized in their industry allocations are also more likely
to have closer linkages among upstream and downstream industries, a key feature of
industrial clusters.
(7) Relationship with Regional Growth:
Table 13 presents estimation results from regressing the growth rate of regional gross
industrial output (log[output_2004/output_1995]) on the growth rate of regional
specialization (log[CR3_2004/CR3_1995] or log[Gini_2004/Gini_1995]), while Table 14
gives results from regressing output growth on industry proximity growth
4
Again, county level data gives the same results but has too many observations to give effective graphic
illustration.
9
(log[proximity_2004/proximity_1995]). To control for regional variation to the greatest
extent, all regressions include the most detailed region dummies possible. As province,
prefecture, and county correspond to 2-digit, 4-digit, and 6-digit area codes, respectively,
we use regional dummies corresponding to the one-digit area codes for provincial level
data, the three-digit area codes for prefecture level data, and the five-digit area codes for
county level data.
The results from regressions controlling for region dummies are largely consistent. There
is a positive correlation between output growth and regional specialization growth as well
as between output growth and industry proximity growth. Rigorously determining the
causality of such correlation is beyond the scope of investigation in this paper, but these
results do suggest the potentially positive role of regional specialization and industry
proximity in promoting regional growth. We plan to further explore this issue in the
future.
Conclusion
Using census data at the firm level from 1995 and 2004, we have shown in this paper that
China’s industrialization has been accompanied by more spatial concentration and more
regional specialization of industrial activities as well as more interactions among firms
within industry and within region. The pattern of increasing spatial concentration of
industrial activities resembles the industrialization path in other countries. In addition,
our results indicate that the number of firms is growing faster and firm size is not
significantly larger in clustered areas than non-clustered regions, while at the same time
there is finer division of labor and closer technological affinity among firms. This pattern
is similar to the East Asian cluster-based industrialization model led by numerous SMEs
but differs from the observed patterns in the US where regional agglomeration and
industrial districts were mainly driven by the presence of large firms.
This cluster-based industrialization may have fit well with China’s comparative
advantage at the onset of its reform that was marked with limited capital and abundant
labor. This business model makes use of more entrepreneurs and labor and less capital
compared to the non-clustered large factories, thus may have emerged as the choice of
Chinese firms over time, leading to more clustered industries in China.
Although the patterns are clear and robust, many important issues remain to be explored.
For example, has clustering really caused economic growth or is it a result of growth?
Why have such patterns emerged during China’s industrialization in the past decade? Are
they simply a reversal of the planned-era distortions, or are they a response to the
resource constraints still in place? More generally, is there a natural tendency toward
clustering in the process of industrial growth? And what roles do government
intervention and increased engagement in international trade play in the process of
clustering? A related issue is why clusters have emerged in some places and industries
10
but not in others? In other words, what factors tend to promote the development of
industrial clusters?
Furthermore, what are the mechanisms through which clustering affects firm growth and
economic development? Two crucial problems facing Chinese firms are credit constraints
and weak protection of property and contracting rights. Does clustering play any role in
helping resolve these difficulties? Some characteristics suggest a potentially important
role of industrial clusters in promoting economic growth. Close proximity and intense
competition among firms within a cluster may reduce the temptation of cheating, and
finer division of labor and frequent trade credit among firms within a cluster may reduce
the need for external finances.
Even if clustering helps increase the productivity of firms within the cluster, what effects
does clustering have on neighboring regions and areas that are more distant? If the main
effect of clustering is stealing away business from other regions instead of nurturing new
firms and creating new business opportunities, then should policies to promote clustering
still be adopted by the national government? Finally, innovation is the engine for growth
in the long run. But what role does clustering play in upgrading the local economy’s
structure and bringing about technological, managerial, and institutional innovations? In
other words, will clustering become a long-term feature of China’s economic growth or
will it fade away as a temporary arrangement waiting to be replaced by some other more
efficient system? Our empirical work is just the first step in identifying the regularities of
China’s rapid industrialization. More research is needed to address these questions. A
better understanding of these issues will help shed light on whether the lessons drawn
from China’s experience in industrialization can be applied to other developing regions in
the world.
11
Reference
Bai, Chong-En, Yingjuan Duan, Zhigang Tao, and Sarah T. Tong. 2004. “Local
Protectionism and Regional Specialization: Evidence from China’s Industries.”
Journal of International Economics 63 (2): 397–417.
Chandler, Alfred Jr., 1977. The Visible Hand: The Managerial Revolution in American
Business. Cambridge, Harvard University Press.
China National Bureau of Statistics, various years. China Statistical Yearbooks. Beijing:
China Statistical Publishing House.
China National Bureau of Statistics, 1995. China Industrial Census.
China National Bureau of Statistics, 2004. China Economic Census.
Fujita, Masahisa, Paul Krugman, Anthony J. Venables, 2001. The Spatial Economy:
Cities, Regions, and International Trade. Massachusetts: MIT Press.
Hausmann, Ricardo and Bailey Klinger, 2007. “The Structure of the Product Space and
the Evolution of Comparative Advantage,” Center for International Development
Working Paper No. 146, Harvard University.
Huang, Zuhui, Zhang, Xiaobo, and Zhu, Yunwei, 2008. “The Role of Clustering in Rural
Industrialization: A Case Study of Wenzhou’s Footwear Industry.” China
Economic Review (Forthcoming).
Krugman, Paul, 1991. Geography and Trade. Cambridge, MA: MIT Press.
Landes, David S., 1998. The Wealth and Poverty of Nations, W.W. Norton and Co., New
York.
Lin, Justin Yifu, Fang Cai, and Zhou Li, 2003. The China Miracle: Development Strategy
and Economic Reform. Hong Kong: The Chinese University Press.
Nakabayashi, Masaki. 2006. “Flexibility and Diversity: The Putting-Out System in the
Silk Fabric Industry of Kiryu, Japan,” Discussion Paper in Economic and
Business No. 06-10, Osaka University.
Piore, Michael J. and Charles F. Sabel, 1984. The Second Industrial Divide: Possibilities
for Prosperity. Basic Books.
Porter, E. Michael, 1998. “Clusters and the New Economics of Competition.” Harvard
Business Review, 76 (6): 77-90.
Porter, E. Michael, 2000. “Location, Competition, and Economic Development: Local
Clusters in a Global Economy.” Economic Development Quarterly 14(1): 15-34.
Ruan, Jianqing and Xiaobo Zhang, 2008. “Finance and Cluster-based Industrial
Development in China,” Economic Devleopment and Cultural Change
(Forthcoming).
Schmitz, H., 1995. “Collective Efficiency: Growth Path for Small-Scale Industry,”
Journal of Development Studies, 31 (4): 529-566.
Sonobe, Tetsushi and Keijiro Otsuka. 2006. “The Division of Labor and the Formation of
Industrial Clusters in Taiwan.” Review of Development Economics, 10 (1), 71-86,
2006.
Sonobe, Tetsushi, Dinghua Hu, and Keijiro Otsuka, 2002. “Process of Cluster Formation
in China: A Case Study of a Garment Town.” Journal of Development Studies,
39(1), 118–139.
12
Sonobe, Tetsushi, Dinghuan Hu, and Otsuka, Keijiro, 2004. “From Inferior to Superior
Products: and Inquiry into the Wenzhou Model of Industrial Development in
China.” Journal of Comparative Economics 32, 542-563.
Sonobe, Tetsushi, and Keijiro Otsuka, 2006. Cluster-Based Industrial Development: An
East Asia Model. New York: Palgrave MacMillan.
Summers, Lawrence, 2007. “The Rise of Asia and the Global Economy.” Research
Monitor (the bi-annual newsletter of the Global Development Network), Special
Issue: 4-5.
Wen, Mei, 2004. “Relocation and agglomeration of Chinese industry,” Journal of
Development Economics, 73 (1), 329-347.
Young, Alwyn. 2000. “The Razor’s Edge: Distortions and Incremental Reform in the
People’s Republic of China.” Quarterly Journal of Economics 115 (4): 1091–35.
Zhang, Xiaobo and Kong-Yam Tan, 2007. “Incremental Reform and Distortions in
China’s Product and Factor Markets,” World Bank Economic Review, 21(2): 279299.
13
Table 1: Summary statistics of gross industrial output by 2-digit industry
1995
industry
Coal Mining & Dressing
Petroleum & Natural Gas Extraction
Ferrous Metals Mining & Dressing
Nonferrous Metals Mining & Dressing
Non-Metallic Mineral Mining & Dressing
Other Minerals Mining & Dressing
Foodstuff Processing Industry
Foodstuff Manufacturing Industry
Beverage Manufacturing Industry
Tobacco Processing
Spinning Industry
Manufacturers of Clothes & Other Fibre Products
Leather, Fur, Feather & Other Products
Timber Processing & Bamboo, Cane, Palm, Straw
Products
Furniture
Paper Makers & Paper Products
Printing & Record Medium Reproduction
Teaching & Sport Products for Daily Use
Oil Processing & Refining
Chemical Material & Products
Pharmaceutical & Medicine Manufacturing
Chemical Fibres
Rubber Products
Plastic Products
Non-Metallic Mineral Products
Smelting & Pressing of Ferrous Metals
Smelting & Pressing of Non-Ferrous Metals
Metal Products
Common Machines
Special Equipment
Traffic Equipment
Manufacture of Electrical Machinery and Apparatus
Electrical Machines & Equipment
Electronic & Communication Equipment
Instruments, Culture & Office Devices
Recycling of Material Waste and Scrap
Electricity, Steam, Thermal Power Production &
Supply
Coal Gas Production & Supply
Tap Water Production & Supply
Total
Mean (in
1000s of
RMB)
9,664.20
1,066,011.00
5,228.05
8,553.58
3,087.27
2,515.33
10,042.10
5,855.71
7,851.64
237,406.30
18,002.44
7,452.73
9,308.45
2004
96,604.69
4,254,463.00
25,018.69
36,978.28
11,113.81
4,604.39
40,142.00
34,955.28
43,375.50
902,941.80
53,656.19
25,219.11
30,235.77
Number of
firms
11,953.00
134.00
2,141.00
3,766.00
11,820.00
149.00
30,962.00
16,313.00
14,719.00
423.00
24,459.00
18,937.00
10,468.00
Mean (in
1000s of
RMB)
17,643.27
962,612.70
9,553.81
14,918.50
3,292.75
3,948.45
13,718.98
11,025.56
10,739.79
885,793.50
14,029.43
9,670.91
13,815.78
sd
224,515.30
5,758,893.00
46,736.95
104,297.60
19,150.26
23,920.48
105,608.60
87,350.86
96,639.89
2,275,005.00
100,348.10
58,139.79
59,022.05
Number
of firms
26,822
481
10,256
6,075
34,945
263
69,521
29,811
25,485
281
83,011
48,250
22,677
2,619.68
2,580.30
7,303.50
2,552.74
8,574.78
73,924.94
13,750.11
16,527.35
76,277.40
13,293.67
5,856.40
4,925.68
44,107.93
29,696.83
5,534.26
7,719.31
10,804.52
17,008.77
13,206.29
34,343.10
12,551.72
4,576.05
2,798.89
12,019.98
9,564.34
27,231.72
9,898.23
31,218.84
716,203.60
116,761.10
61,056.75
474,156.30
74,614.20
20,254.42
18,678.22
479,425.00
143,611.90
24,406.73
41,075.43
69,346.77
216,986.40
74,917.41
250,415.80
75,699.33
17,210.64
10,226.52
15,480.00
8,760.00
13,890.00
16,763.00
5,356.00
2,744.00
26,872.00
6,051.00
1,034.00
4,663.00
19,255.00
61,278.00
8,429.00
4,621.00
26,744.00
31,474.00
18,391.00
19,522.00
18,928.00
5,489.00
9,735.00
12,127.00
4,440.00
5,072.08
6,255.15
10,005.26
4,234.09
9,702.21
126,788.60
19,174.92
29,861.06
59,127.60
13,489.79
7,573.37
6,305.85
84,283.87
41,173.88
7,849.46
9,032.19
10,556.24
27,663.70
20,680.07
74,793.30
40,515.75
6,966.12
4,490.53
32,941.00
39,320.08
83,103.28
21,719.22
42,731.84
1,373,127.00
239,355.00
148,193.90
329,805.00
128,383.70
44,340.95
32,082.12
961,068.50
259,277.20
52,709.39
80,814.59
84,887.41
475,230.00
249,967.10
1,048,297.00
532,204.90
38,783.04
28,719.71
37,028
23,892
39,669
44,070
14,711
7,146
69,120
11,271
3,372
15,178
69,729
157,734
20,494
15,162
80,976
113,691
55,095
51,844
54,979
15,211
35,203
26,627
6,156
19,369.47
20,474.37
3,544.61
10,762.71
107,369.60
82,662.64
29,941.79
134,409.90
12,600.00
372.00
5,147.00
506,409.00
60,652.89
30,310.35
5,057.70
16,198.02
1,052,128.00
129,999.50
34,172.95
310,685.90
24,568
1,445
11,035
1,363,284
sd
Note: The gross industrial output is reported in 1000s of RMB at current price.
14
Table 2: Summary statistics of gross industrial output by province
1995
Province
Beijing
Tianjin
Hebei
Shanxi
Neimeng
Liaoning
Jilin
Heilongjiang
Shanghai
Jiangsu
Zhejiang
Anhui
Fujian
Jiangxi
Shandong
Henan
Hubei
Hunan
Gongdong
Gongxi
Hainan
Chongqing
Sichuan
Guizhou
Yunnan
Tiebet
Shaanxi
Gansu
Qinghai
Ningxia
Xinjiang
Total
Province
code
BJ
TJ
HEB
SX
NM
LN
JL
HLJ
SH
JS
ZJ
AH
FJ
JX
SD
HEN
HUB
HUN
GD
GX
HAIN
CQ
SC
GZ
YN
TB
SAX
GS
QH
NX
XJ
Mean (in
1000s of
RMB)
15,067.10
13,638.07
9,216.23
8,538.13
5,558.71
10,183.97
7,750.83
8,523.55
23,259.75
15,815.09
10,362.87
6,911.61
8,080.35
4,527.56
17,466.32
9,702.62
10,432.29
5,737.98
17,715.48
7,719.27
9,932.23
6,676.02
6,674.90
5,500.25
13,970.34
2,342.60
6,182.41
8,259.79
8,192.80
9,011.04
9,990.26
10,724.71
sd
256,246.90
155,565.30
80,343.00
91,722.85
83,739.84
171,331.30
167,291.90
308,215.90
273,997.00
106,861.10
58,129.79
64,939.67
47,063.31
45,442.61
179,589.60
75,317.70
139,562.00
66,417.34
114,051.60
42,786.33
52,780.98
82,140.76
74,866.45
48,962.10
223,903.50
6,553.74
59,999.89
109,848.00
77,821.31
56,606.26
187,434.60
134,909.30
2004
Number of
firms
9,623.00
10,735.00
23,592.00
11,416.00
9,432.00
29,435.00
13,100.00
18,745.00
16,690.00
41,582.00
32,725.00
23,474.00
19,038.00
18,253.00
26,980.00
23,119.00
20,881.00
23,720.00
34,536.00
12,312.00
1,278.00
11,456.00
26,380.00
7,450.00
6,267.00
295.00
12,950.00
7,140.00
1,446.00
1,706.00
5,077.00
500,833.00
Mean (in
1000s of
RMB)
18,926.12
23,948.82
15,789.47
14,490.05
19,688.51
16,844.39
22,085.22
19,613.45
26,262.88
15,617.85
11,235.65
10,807.79
15,126.40
9,331.14
20,476.78
12,064.91
18,191.13
9,668.12
22,968.96
11,870.34
21,085.97
12,676.51
12,136.63
13,831.37
16,156.97
7,003.53
12,250.77
14,647.74
17,523.91
15,131.59
28,812.50
16,198.02
sd
418,527.50
500,129.20
201,284.30
216,239.90
223,746.50
365,898.70
542,689.40
738,370.10
499,885.90
262,799.90
173,579.80
189,113.60
225,394.00
146,980.50
303,313.60
164,347.40
359,618.60
145,046.90
473,908.10
155,917.80
198,546.60
149,313.00
168,971.30
178,497.20
239,845.30
22,096.48
209,434.00
305,596.80
218,482.40
151,483.00
441,176.00
310,685.90
Number
of firms
31,364
25432
64,062
28,641
11,759
54,115
16,037
20,101
55,315
187,212
187,588
38,827
49,532
29,144
119,699
76,292
28,937
43,529
136,606
18,753
2,025
20,359
43,325
10,996
14,271
354
25,573
11,549
2,168
3,984
5,735
1,363,284
Note: The gross industrial output is reported in 1000s of RMB at current price.
15
Table 3: Comparing sample with aggregate data
1995
2004
1995
2004
Gross Industrial Output (Trillions of RMB, at current price)
Sample (1)
Statistical Yearbook (2)
5.495
5.528
20.174
18.722
industrial value added (Trillions of RMB, at current price)
sample (3)
Statistical yearbook (4)
1.536
1.545
5.678
5.481
(1)/(2)*100%
99.438
107.754
(3)/(4)*100%
99.445
103.604
Note: The official figures for gross industrial output and industrial value added are from
China Statistical Yearbook in 1996 and 2005. However, the official figures in the China
Statistical Yearbook 2005 do not include non-state owned small enterprises below a
certain scale. Therefore, the ratio of the tabulated to official figures exceeds one in 2004.
16
Table 4: Industrial concentration
Industry
Coal Mining & Dressing
Petroleum & Natural Gas Extraction
Ferrous Metals Mining & Dressing
Nonferrous Metals Mining & Dressing
Non-Metallic Mineral Mining & Dressing
Other Minerals Mining & Dressing
Foodstuff Processing Industry
Foodstuff Manufacturing Industry
Beverage Manufacturing Industry
Tobacco Processing
Spinning Industry
Manufacturers of Clothes & Other Fiber Products
Leather, Fur, Feather & Other Products
Timber Processing & Bamboo, Cane, Palm, Straw Products
Furniture
Paper Makers & Paper Products
Printing & Record Medium Reproduction
Teaching & Sport Products for Daily Use
Oil Processing & Refining
Chemical Material & Products
Pharmaceutical & Medicine Manufacturing
Chemical Fibers
Rubber Products
Plastic Products
Non-Metallic Mineral Products
Smelting & Pressing of Ferrous Metals
Smelting & Pressing of Non-Ferrous Metals
Metal Products
Common Machines
Special Equipment
Traffic Equipment
Manufacture of Electrical Machinery and Apparatus
Electrical Machines & Equipment
Electronic & Communication Equipment
Instruments, Culture & Office Devices
Recycling of Material Waste and Scrap
Electricity, Steam, Thermal Power Production & Supply
Coal Gas Production & Supply
Tap Water Production & Supply
Weighted sample mean
Difference
CR3_2004 CR3_1995
0.495
0.390
0.486
0.570
0.486
0.422
0.488
0.369
0.348
0.322
0.612
0.488
0.419
0.302
0.341
0.307
0.319
0.305
0.367
0.457
0.629
0.434
0.591
0.511
0.593
0.431
0.414
0.290
0.500
0.373
0.486
0.297
0.443
0.352
0.659
0.559
0.338
0.457
0.420
0.349
0.308
0.326
0.710
0.598
0.493
0.432
0.544
0.448
0.371
0.320
0.378
0.391
0.293
0.285
0.524
0.443
0.493
0.592
0.407
0.359
0.290
0.389
0.553
0.440
0.720
0.555
0.638
0.588
0.610
0.406
0.348
0.453
0.306
0.267
0.404
0.499
0.391
0.381
0.460
0.387
0.073 (0.015)***
yr2004_gini yr1995_gini
0.610
0.574
0.632
0.651
0.607
0.574
0.638
0.582
0.524
0.540
0.746
0.598
0.574
0.470
0.530
0.512
0.487
0.483
0.525
0.614
0.748
0.641
0.770
0.716
0.771
0.665
0.617
0.526
0.693
0.559
0.666
0.481
0.609
0.504
0.803
0.693
0.529
0.632
0.570
0.512
0.472
0.503
0.777
0.706
0.678
0.580
0.706
0.636
0.561
0.501
0.527
0.558
0.465
0.418
0.705
0.612
0.668
0.683
0.607
0.532
0.541
0.570
0.718
0.654
0.813
0.727
0.792
0.723
0.774
0.658
0.569
0.645
0.446
0.419
0.545
0.633
0.508
0.521
0.622
0.563
0.060 (0.010)***
yr2004_Kgini yr1995_Kgini
0.556
0.567
0.712
0.655
0.612
0.772
0.542
0.579
0.382
0.483
0.750
0.592
0.326
0.244
0.337
0.336
0.352
0.259
0.626
0.688
0.438
0.306
0.595
0.465
0.596
0.441
0.441
0.531
0.395
0.292
0.365
0.226
0.323
0.308
0.658
0.445
0.517
0.493
0.192
0.215
0.406
0.276
0.555
0.462
0.483
0.397
0.374
0.299
0.318
0.233
0.321
0.357
0.493
0.494
0.339
0.241
0.350
0.330
0.300
0.338
0.524
0.418
0.392
0.371
0.658
0.588
0.634
0.497
0.559
0.428
0.545
0.337
0.250
0.258
0.408
0.418
0.357
0.390
0.424
0.370
0.054 (0.013)***
17
Table 5: Top producing provinces for each 2-digit industry
Industry
Coal Mining & Dressing
Petroleum & Natural Gas Extraction
Ferrous Metals Mining & Dressing
Nonferrous Metals Mining & Dressing
Non-Metallic Mineral Mining & Dressing
Other Minerals Mining & Dressing
Foodstuff Processing Industry
Foodstuff Manufacturing Industry
Beverage Manufacturing Industry
Tobacco Processing
Spinning Industry
Manufacturers of Clothes & Other Fiber
Products
Leather, Fur, Feather & Other Products
Timber Processing & Bamboo, Cane, Palm,
Straw Products
Furniture
Paper Makers & Paper Products
Printing & Record Medium Reproduction
Teaching & Sport Products for Daily Use
Oil Processing & Refining
Chemical Material & Products
Pharmaceutical & Medicine Manufacturing
Chemical Fibers
Rubber Products
Plastic Products
Non-Metallic Mineral Products
Smelting & Pressing of Ferrous Metals
Smelting & Pressing of Non-Ferrous Metals
Metal Products
Common Machines
Special Equipment
Traffic Equipment
Manufacture of Electrical Machinery and
Apparatus
Electrical Machines & Equipment
Electronic & Communication Equipment
Instruments, Culture & Office Devices
Recycling of Material Waste and Scrap
Electricity, Steam, Thermal Power
Production & Supply
Coal Gas Production & Supply
Tap Water Production & Supply
Prov1_04
SX
HLJ
HEB
SD
SD
HLJ
SD
SD
SD
YN
JS
Prov2_04
SD
SD
LN
HEN
HEN
HEN
HEN
GD
GD
SH
ZJ
Prov3_04
HEN
XJ
SX
HUN
ZJ
SD
JS
HEN
SC
JS
SD
Prov1_95
SX
HLJ
HEB
SD
SD
GX
SD
SH
SD
YN
JS
Prov2_95
SD
SD
LN
JX
JS
ZJ
JS
GD
GD
HUN
GD
Prov3_95
HEB
XJ
AH
HEN
AH
HEN
GD
SD
JS
SH
SH
JS
ZJ
GD
GD
ZJ
FJ
SH
GD
JS
FJ
GD
ZJ
JS
GD
SD
GD
GD
LN
JS
JS
ZJ
SD
GD
SD
HEB
JS
GD
JS
SD
SH
SD
ZJ
GD
ZJ
ZJ
SD
SD
SD
JS
JS
ZJ
HEN
JS
ZJ
JS
ZJ
JS
JL
ZJ
SD
ZJ
JS
JS
GD
GD
ZJ
SD
ZJ
JS
GD
LN
HEN
ZJ
SD
GD
GD
JS
GD
SD
SH
GD
ZJ
JS
SH
SH
SH
GD
JS
SH
JS
SH
SD
JS
SH
GD
SD
GD
GD
SH
LN
SH
JS
JS
SD
JS
SD
LN
SH
JS
JS
SD
JS
HLJ
JS
ZJ
JS
JS
SD
GD
GD
ZJ
JS
ZJ
GD
HEB
LN
GD
SH
SH
HUB
GD
GD
GD
GD
SAX
JS
SH
JS
ZJ
JS
ZJ
JS
TJ
SD
SD
SH
GD
GD
SD
GD
JS
SH
SH
GD
SH
HEN
JS
JS
JS
JS
GD
GD
GD
ZJ
JS
JS
JS
SC
ZJ
GD
SH
GD
JS
GD
SH
SD
LN
JS
18
Table 6: Distributions of number of firms and average firm size
Industry
Coal Mining & Dressing
Petroleum & Natural Gas Extraction
Ferrous Metals Mining & Dressing
Nonferrous Metals Mining & Dressing
Non-Metallic Mineral Mining & Dressing
Other Minerals Mining & Dressing
Foodstuff Processing Industry
Foodstuff Manufacturing Industry
Beverage Manufacturing Industry
Tobacco Processing
Spinning Industry
Manufacturers of Clothes & Other Fiber Products
Leather, Fur, Feather & Other Products
Timber Processing & Bamboo, Cane, Palm, Straw
Products
Furniture
Paper Makers & Paper Products
Printing & Record Medium Reproduction
Teaching & Sport Products for Daily Use
Oil Processing & Refining
Chemical Material & Products
Pharmaceutical & Medicine Manufacturing
Chemical Fibers
Rubber Products
Plastic Products
Non-Metallic Mineral Products
Smelting & Pressing of Ferrous Metals
Smelting & Pressing of Non-Ferrous Metals
Metal Products
Common Machines
Special Equipment
Traffic Equipment
Manufacture of Electrical Machinery and Apparatus
Electrical Machines & Equipment
Electronic & Communication Equipment
Instruments, Culture & Office Devices
Recycling of Material Waste and Scrap
Electricity, Steam, Thermal Power Production & Supply
Coal Gas Production & Supply
Tap Water Production & Supply
Weighted sample mean
Difference
gini_n_04
0.580
0.533
0.660
0.541
0.426
0.512
0.512
0.444
0.432
0.411
0.728
0.700
0.732
gini_n_95
0.531
0.680
0.541
0.475
0.404
0.435
0.433
0.321
0.427
0.432
0.526
0.507
0.446
0.569
0.492
0.551
0.364
0.566
0.373
0.514
0.341
0.738
0.515
0.507
0.486
0.524
0.394
0.391
0.366
0.740
0.556
0.610
0.428
0.650
0.468
0.477
0.405
0.474
0.403
0.552
0.395
0.635
0.401
0.643
0.446
0.630
0.439
0.548
0.385
0.673
0.469
0.711
0.579
0.698
0.562
0.701
0.461
0.517
0.378
0.583
0.521
0.472
0.447
0.531
0.485
0.586
0.461
0.125 (0.014)***
gini_size_04
0.668
0.524
0.524
0.718
0.305
0.648
0.316
0.341
0.308
0.548
0.318
0.358
0.350
gini_size_95
0.526
0.643
0.776
0.418
0.381
0.414
0.364
0.408
0.419
0.461
0.233
0.430
0.412
0.281
0.395
0.311
0.261
0.459
0.455
0.230
0.194
0.391
0.476
0.256
0.317
0.321
0.415
0.286
0.233
0.271
0.437
0.275
0.493
0.420
0.345
0.467
0.619
0.495
0.508
0.375
-0.016 (0.020)
19
0.402
0.410
0.253
0.316
0.373
0.528
0.264
0.255
0.405
0.423
0.302
0.308
0.508
0.365
0.347
0.285
0.323
0.380
0.349
0.457
0.295
0.418
0.432
0.595
0.751
0.585
0.391
Table 7: Number of firms and average firm size
Industry
Coal Mining & Dressing
Petroleum & Natural Gas Extraction
Ferrous Metals Mining & Dressing
Nonferrous Metals Mining & Dressing
Non-Metallic Mineral Mining & Dressing
Other Minerals Mining & Dressing
Foodstuff Processing Industry
Foodstuff Manufacturing Industry
Beverage Manufacturing Industry
Tobacco Processing
Spinning Industry
Manufacturers of Clothes & Other Fiber Products
Leather, Fur, Feather & Other Products
Timber Processing & Bamboo, Cane, Palm, Straw Products
Furniture
Paper Makers & Paper Products
Printing & Record Medium Reproduction
Teaching & Sport Products for Daily Use
Oil Processing & Refining
Chemical Material & Products
Pharmaceutical & Medicine Manufacturing
Chemical Fibers
Rubber Products
Plastic Products
Non-Metallic Mineral Products
Smelting & Pressing of Ferrous Metals
Smelting & Pressing of Non-Ferrous Metals
Metal Products
Common Machines
Special Equipment
Traffic Equipment
Manufacture of Electrical Machinery and Apparatus
Electrical Machines & Equipment
Electronic & Communication Equipment
Instruments, Culture & Office Devices
Recycling of Material Waste and Scrap
Electricity, Steam, Thermal Power Production & Supply
Coal Gas Production & Supply
Tap Water Production & Supply
n_04
925
21
342
203
1,165
9
2,243
962
822
10
2,678
1,556
732
1,194
771
1,280
1,422
525
238
2,230
364
116
506
2,249
5,088
683
489
2,612
3,667
1,777
1,672
1,833
507
1,173
859
199
793
48
356
n_95
386
6
71
126
369
7
968
510
460
14
764
592
327
484
274
448
524
173
91
840
189
34
150
602
1,915
263
149
836
1,015
575
610
592
177
314
379
143
394
12
161
size_04
17,643
962,613
9,554
14,919
3,293
3,948
13,719
11,026
10,740
885,793
14,029
9,671
13,816
5,072
6,255
10,005
4,234
9,702
126,789
19,175
29,861
59,128
13,490
7,573
6,306
84,284
41,174
7,849
9,032
10,556
27,664
20,680
74,793
40,516
6,966
4,491
60,653
30,310
5,058
size_95
9,664
1,066,011
5,228
8,554
3,087
2,515
10,042
5,856
7,852
237,406
18,002
7,453
9,308
2,620
2,580
7,303
2,553
8,575
73,925
13,750
16,527
76,277
13,294
5,856
4,926
44,108
29,697
5,534
7,719
10,805
17,009
13,206
34,343
12,552
4,576
2,799
19,369
20,474
3,545
20
Table 8: Value added/Output ratio (2004 v. 1995)
Industry
Coal Mining & Dressing
Petroleum & Natural Gas Extraction
Ferrous Metals Mining & Dressing
Nonferrous Metals Mining & Dressing
Non-Metallic Mineral Mining & Dressing
Other Minerals Mining & Dressing
Foodstuff Processing Industry
Foodstuff Manufacturing Industry
Beverage Manufacturing Industry
Tobacco Processing
Spinning Industry
Manufacturers of Clothes & Other Fiber Products
Leather, Fur, Feather & Other Products
Timber Processing & Bamboo, Cane, Palm, Straw Products
Furniture
Paper Makers & Paper Products
Printing & Record Medium Reproduction
Teaching & Sport Products for Daily Use
Oil Processing & Refining
Chemical Material & Products
Pharmaceutical & Medicine Manufacturing
Chemical Fibers
Rubber Products
Plastic Products
Non-Metallic Mineral Products
Smelting & Pressing of Ferrous Metals
Smelting & Pressing of Non-Ferrous Metals
Metal Products
Common Machines
Special Equipment
Traffic Equipment
Manufacture of Electrical Machinery and Apparatus
Electrical Machines & Equipment
Electronic & Communication Equipment
Instruments, Culture & Office Devices
Recycling of Material Waste and Scrap
Electricity, Steam, Thermal Power Production & Supply
Coal Gas Production & Supply
Tap Water Production & Supply
Weighted sample mean
Difference
vad2out_2004
vad2out_1995
0.353
0.508
0.617
0.555
0.336
0.394
0.292
0.348
0.202
0.385
0.053
0.400
0.207
0.165
0.260
0.216
0.347
0.312
0.683
0.533
0.238
0.184
0.205
0.237
0.192
0.198
0.176
0.257
0.160
0.264
0.217
0.234
0.229
0.310
0.160
0.185
0.194
0.257
0.276
0.260
0.373
0.293
0.189
0.189
0.205
0.221
0.185
0.189
0.242
0.318
0.264
0.266
0.218
0.221
0.194
0.249
0.241
0.292
0.226
0.216
0.230
0.252
0.233
0.243
0.237
0.228
0.281
0.277
0.200
0.276
0.164
0.219
0.356
0.530
0.308
0.081
0.426
0.478
0.262
0.282
-0.021 (0.013)*
21
Table 9: Regional specialization
Province
Beijing
Tianjin
Hebei
Shanxi
Neimeng
Liaoning
Jilin
Heilongjiang
Shanghai
Jiangsu
Zhejiang
Anhui
Fujian
Jiangxi
Shandong
Henan
Hubei
Hunan
Gongdong
Gongxi
Hainan
Chongqing
Sichuan
Guizhou
Yunnan
Tiebet
Shaanxi
Gansu
Qinghai
Ningxia
Xinjiang
Weighted sample average
Difference
CR3_04
CR3_95
0.357
0.351
0.402
0.304
0.400
0.307
0.590
0.485
0.394
0.337
0.386
0.318
0.571
0.437
0.499
0.432
0.322
0.320
0.293
0.309
0.280
0.284
0.259
0.269
0.244
0.205
0.286
0.231
0.239
0.235
0.294
0.231
0.428
0.334
0.255
0.238
0.385
0.234
0.383
0.318
0.443
0.371
0.451
0.406
0.279
0.283
0.426
0.308
0.478
0.510
0.614
0.453
0.308
0.252
0.490
0.374
0.612
0.497
0.389
0.331
0.610
0.549
0.340
0.293
0.047 (0.019)**
gini_04
gini_95
0.664
0.627
0.656
0.560
0.625
0.565
0.789
0.713
0.679
0.614
0.640
0.599
0.757
0.647
0.719
0.639
0.589
0.553
0.633
0.564
0.552
0.551
0.559
0.529
0.600
0.489
0.562
0.525
0.531
0.506
0.551
0.541
0.666
0.597
0.550
0.544
0.676
0.537
0.661
0.573
0.670
0.626
0.683
0.665
0.560
0.583
0.698
0.644
0.728
0.708
0.734
0.635
0.584
0.563
0.733
0.660
0.802
0.711
0.716
0.666
0.782
0.742
0.620
0.563
0.057 (0.014)***
Kgini_04
0.483
0.477
0.402
0.689
0.570
0.325
0.534
0.699
0.289
0.310
0.403
0.284
0.357
0.407
0.274
0.361
0.400
0.378
0.381
0.431
0.577
0.467
0.342
0.601
0.709
0.697
0.551
0.587
0.705
0.591
0.690
0.386
Kgini_95
0.414
0.366
0.359
0.548
0.542
0.520
0.519
0.578
0.315
0.280
0.312
0.322
0.555
0.534
0.276
0.335
0.279
0.329
0.367
0.496
0.608
0.476
0.309
0.503
0.662
0.740
0.383
0.500
0.614
0.599
0.588
0.368
0.018 (0.015)
22
Table 10: Product proximity
Province
Beijing
Tianjin
Hebei
Shanxi
Neimeng
Liaoning
Jilin
Heilongjiang
Shanghai
Jiangsu
Zhejiang
Anhui
Fujian
Jiangxi
Shandong
Henan
Hubei
Hunan
Gongdong
Gongxi
Hainan
Chongqing
Sichuan
Guizhou
Yunnan
Tiebet
Shaanxi
Gansu
Qinghai
Ningxia
Xinjiang
Weighted sample average
Difference
2004
0.220
0.208
0.219
0.208
0.214
0.205
0.220
0.197
0.219
0.210
0.220
0.211
0.202
0.206
0.205
0.209
0.216
0.210
0.215
0.214
0.207
0.197
0.202
0.196
0.197
0.238
0.192
0.205
0.217
0.220
0.199
0.211
0.005 (0.001)***
1995
0.206
0.194
0.212
0.207
0.198
0.204
0.206
0.186
0.222
0.210
0.211
0.204
0.208
0.200
0.200
0.201
0.207
0.201
0.209
0.208
0.201
0.206
0.198
0.188
0.187
0.223
0.191
0.199
0.197
0.215
0.190
0.206
23
Table 11: Correlation matrix for industry concentration measures and firm interaction
measure
2004
CR3
CR3
Gini_output
K-Gini_output
Gini_firm number
Gini_firm size
Value added/output
1
473
Gini_output
K-Gini_output
Gini_firm number
Gini_firm size
Value added/output
1995
CR3
0.955
0
463
1
463
0.6723
0
463
0.6867
0
463
463
0.7234
0
463
0.7895
0
463
0.4393
0
463
463
0.3342
0
463
0.272
0
463
0.4743
0
463
0.0507
0.2761
463
463
-0.3742
0
473
-0.4432
0
463
-0.1291
0.0054
463
-0.4008
0
463
-0.0514
0.2699
463
1
1
1
1
473
1
455
Gini_output
K-Gini_output
Gini_firm number
Gini_firm size
Value added/output
0.9259
0
455
1
456
0.7168
0
455
0.7315
0
456
456
0.5039
0
455
0.542
0
456
0.4591
0
456
456
0.4163
0
455
0.3614
0
456
0.6398
0
456
0.2312
0
456
456
-0.3198
0
455
-0.4183
0
456
-0.2138
0
456
-0.0233
0.6195
456
0.099
0.0345
456
1
1
1
1
456
24
Table 12: Correlation matrix for regional specialization measures and product proximity
measure
2004
Gini
CR3
CR3
Gini
K-Gini
Proximity
K-Gini
1995
Gini
CR3
1
1
345
346
0.9515
0
345
1
345
0.73
0
345
0.7896
0
345
0.1797
0.0008
345
0.1286
0.0168
345
K-Gini
0.9487
0
345
345
345
0.7391
0
345
0.7668
0
345
345
0.0072
0.8934
345
0.0981
0.1137
261
0.0698
0.2614
261
-0.0432
0.4869
261
1
1
1
25
Table 13: Regional growth rate and regional specialization
growth1_CR3
growth1_gini
Constant
Observations
R-squared
Provincial level
(1)
(2)
0.658*
(1.86)
0.471
(0.63)
1.331***
1.403***
(11.57)
(11.07)
31
31
0.40
0.32
Prefecture level
(3)
(4)
0.464***
(4.11)
1.295***
(5.23)
1.285***
1.249***
(4.81)
(4.76)
345
345
0.54
0.55
County level
(5)
0.725***
(9.26)
1.011***
(4.06)
2732
0.54
(6)
1.285***
(14.35)
0.993***
(4.06)
2759
0.56
Note: The dependent variable is log(output_2004/output_1995).
Table 14: Regional growth rate and industry proximity growth rate
growth1_prox
Constant
Observations
R-squared
Provincial level
(1)
-1.613
(1.02)
1.495***
(11.20)
31
0.29
Prefecture level
(2)
-0.172
(0.44)
0.293
(0.91)
261
0.48
County level
(3)
0.337***
(3.11)
1.174***
(4.88)
2746
0.51
Note: The dependent variable is log(output_2004/output_1995).
26
27
28
29
30
31
32