Why China’s cities are clusters of innovation

Spatial agglomeration of economic activities is generally assumed to improve productivity and spur firms’ innovation through localization economies and urbanization economies. There is an extensive empirical literature investigating the effects of localization and urbanization on firm-level productivity. Despite its economic importance, there are few empirical studies focusing on agglomeration and product innovation.

Feldman and Audretsch (1999) and De Beule and Van Beveren (2010) are two of the few exceptions. Feldman and Audretsch found a tendency for innovative activity in complementary industries sharing a common science base to cluster together in a city. In the US, diversity has a strong positive effect and specialization has a negative effect on new product introductions. De Beule and Van Beveren found a positive impact of own-industry employment concentration on the product innovation of Belgian firms. However, they considered localization only.

Does agglomeration account for product innovation in developing countries such as China? If it does, how do the innovation activities of Chinese firms benefit from agglomeration economies, from localization economies, and/or from urbanization economies? To address the above questions, in my latest paper (Zhang 2014) I empirically analyzed the effect of agglomeration economies on firm-level product innovation using Chinese firm-level data from 1998 to 2007. Although the deepening of industrial agglomeration in China and its effect on productivity are well documented (see for example Lu and Tao 2009, Lin et al. 2011), whether agglomeration contributes to product innovation has not been explored.

Measures of agglomeration and product innovation

I followed Martin et al. (2011) to construct variables for agglomeration economies at the industry-city level. Localization economies are measured by the number of other workers working for neighbouring firms in the same industry and in the same city. Urbanization economies are measured by the number of workers in other industries in the same city and by the diversity of industries in the same city.

The criteria used by the National Bureau of Statistics for measuring the variable ‘new products’ are as follows: new products refer to brand new products produced with new technology and new design, or products that represent a noticeable improvement in terms of structure, material, or production process, significantly improving the character or function of the older versions. They include new products certified by relevant government agencies within the period of certification, as well as new products designed and produced by enterprises within a year without certification by government agencies.

Geographical distribution of product innovation

Table 1 shows that there is wide variation across provinces regarding new product output, new product intensity and the share of new product firms. The coastal provinces – especially Beijing, Tianjin, Shanghai, Jiangsu, Zhejiang, Shandong and Guangdong – account for a large amount of new product output. Meanwhile, inland provinces, such as Inner Mongolia, Guizhou, Qinghai and Ningxia have very limited new product output. In terms of new product intensity, compared with the average of all provinces (3.6%), Beijing (19.2%), Tianjin (10.5%), Sichuan (9.1%), Chongqing (8.3%), and Zhejiang (7.4) are the most innovative, whereas Hainan (0.4%), Inner Mongolia (0.5%), Tibet (0.8%), and Xinjiang (1.3%) are the least innovative. Furthermore, Beijing, Tianjin, Zhejiang, Chongqing and Sichuan have the highest percentage of firms with new product introduction, at about 20%. On the other hand, Inner Mongolia, Fujian, and Hainan have the lowest percentage of such firms at 1.1%–2.9%. These findings suggest that there are large disparities in product innovation across regions in China.

Table 1. Product innovation across regions (2007)

Note: New product intensity is the ratio of new product output to total output.

Urbanization economies, not localization economies, contribute to product innovation in China

From the estimation of a model of new product introduction, I find that urbanization economies – both size and diversity of neighbouring industries – make statistically significant contributions to the probability of new product introduction. Conversely, I find that there are no positive effects of localization economies. Using an alternative measure of product innovation for the dependent variables, I find that the new product output and new product intensity are also benefits from urbanization economies rather than localization economies. All estimations are carried out after controlling for the local strength of competition and firm-specific characteristics, including total factor productivity (TFP), firm size, average wage, ownership, and production subsidy. So it is obvious that product innovation is determined not only by firms’ internal capacity, but also by external economies (urbanization economies here).

In addition, I divide all of the samples into several sub-samples – domestic firms and foreign-invested firms, high-tech industries and other industries. The aim is to investigate whether there are differences in factors affecting product innovation by ownership or sector. Urbanization, size, and diversity are still strongly positively associated with the probability of new product introduction and new product output. The results show that the contributions of urbanization economies are stronger for domestic firms than for foreign-invested firms, and stronger in high-tech industries than in other industries.

Policy implications

Clusters with industrial diversity can be an effective means of fostering product innovation. Policymakers could promote product innovation by encouraging firms to locate in industrially diversified areas. Chinese authorities have established more than 100 economic zones (such as economic and technological development areas and high-technology industry development areas) in over 60 cities since 1995. Note that most of these economic zones – such as Zhongguancun Science Park (Beijing) and Suzhou Industrial Park (Jiangsu Province) – are not clusters of firms from the same narrowly-defined industry (such as a 2-digit or 3-digit industry). The deepening of inter-industry outsourcing, knowledge spillovers, and procurement of intermediate goods are likely to lower the sunk costs of introducing new products, and thereby to promote product innovation.

Published in collaboration with VoxEU.

Author: Hongyong Zhang, Fellow, RIETI; Lecturer at the School of Economics, Senshu University

Image: Employees work on a solar panel production line at Suntech Power Holdings headquarters in Wuxi, Jiangsu province, April 3, 2013. REUTERS/Stringer