The views expressed in the papers are solely those of the authors and do not represent the views of DRC or RIETI. The views expressed in the articles are solely those of the author(s) and do not represent the views of the Research Institute of Economy, Trade and. However, due to the local nature of knowledge spillovers [10], potential entrants operating near the established exporters are more likely to benefit from the externality.
However, the complex nature of the impact of agglomeration becomes apparent once we look at the ownership of established exporters and potential new entrants to the export market. The agglomeration of industrial activity in China has continuously increased since the mid-1990s, driven by the globalization of the Chinese economy. While such studies deepen our understanding of the nature of multinationals' knowledge spillovers in China, they do not fully explore the contribution of industrial agglomeration.
We investigate this formally by comparing the productivity distribution of the two types of exporters by a non-parametric Kolmogorov-Smirnov (KS) test. The KS test allows us to compare the overall productivity distribution of companies according to their export status and location based on the concept of first-order stochastic dominance. We perform the KS test separately for each year to avoid a possible violation of the independence assumption.
We also assess whether export spillovers differ depending on the ownership of established exporters and the location of agglomeration, by observing the following four agglomeration variables: the number of domestic exporters located in the same county (agg2_1), the number of domestic exporters located in the same province , but in different counties.
Empirical results
Column (1) corresponds to the model that includes the industrial aspect of agglomeration, while column (2) corresponds to the model that includes the ownership of the established exporters and the location of the agglomeration. In column (2), we see how the export spillover from agglomeration varies according to the ownership and location of established exporters. Regarding agglomeration of native exporters, it is found that the numbers of exporters located in the same country (agg2_1) and those in the same province but different provinces (agg2_2) have positive and significant coefficients.
The marginal effects indicate that spillover from the urban area by domestic exporters is greater when the urban area is within the same county (agg2_1) than when it is located in different counties (agg2_2). Conversely, an agglomeration of FIF exporters in the same province but different counties (agg3_2) has a negative sign, while in the same county (agg3_1) it has a positive sign. One way to interpret these results is that the agglomeration of exporting FIFs is associated with significant congestion in the labor and intermediate goods markets.
On the other hand, the positive impact of the agglomeration of domestic exporters suggests that they either generate a larger knowledge effect or face less negative congestion effects than exporting FIFs. Columns (3) and (4) indicate that the agglomeration of domestic exporters, whether located in the same county or not, contributes significantly to the exports of both domestic firms and FIFs. However, the agglomeration of FIF exporters within the same county has a negative impact on the export input of the native firms.
On the contrary, the agglomeration of executive FIFs provides an opportunity for indigenous firms to access global demand by supplying their products to the concentrated area instead of invading their own foreign markets. If such “indirect exports” did flourish in China, the agglomeration of executive FIFs may be negatively correlated with the export entry of indigenous firms. Columns (1) and (2) indicate that only the agglomeration of exporters in the same country and industry (agg1_1) contributes to the export entry of firms in high-tech industries.
However, the results in Table 6 show that the spillover resulting from agglomeration of exporters is significant in both areas, and even stronger inland. Furthermore, domestic firms enjoy a positive and significant spillover from the clustering of FIF exporters located in different counties (agg3_2). Therefore, spillover from agglomeration of exporters plays a greater role in the export entry of inland firms than coastal firms.
Since Probit is a non-linear model, the marginal spillover effects from exporter agglomeration depend on the values of all other explanatory variables. This exercise also suggests that larger, more productive and more skill-intensive domestic firms with prior export experience are rewarded with greater spillovers from the exporter agglomeration.
Concluding remarks
It can be seen that the sign of the marginal effects is consistent with the results in Table 4, both at the minimum and maximum values of each variable. Interestingly, column (1) indicates that the marginal effects of agglomeration spillover for indigenous firms were significantly larger when the firm previously exported. Furthermore, in the case of indigenous firms, the marginal effects are larger when a firm's TFP, employment size, and wage take their maximum rather than minimum value.
However, if the potential entrant is a FIF, the marginal effects on entry are larger when employment size and wages are minimal. While it is difficult to provide a concrete explanation for such contrasting effects of export spillovers on indigenous firms and FIFs, large FIFs could be well aware of trends in global markets through their parent companies, and thus use less of knowledge spillover. Our non-parametric analysis shows that the productivity distribution of exporters in agglomeration regions is stochastically first-order dominated by that of domestic firms in the region.
The parametric analysis, which measures the contribution of the agglomeration of established exporters to the probability of export entry, found that the agglomeration of exporters in the same region and in the same industry has the greatest positive impact. While the agglomeration of indigenous exporters contributes to the export entry of both indigenous firms and FIFs, that of exporting FIFs has a negative impact on the export entry of indigenous firms. Other fixed characteristics such as previous export status, productivity, employment size and skill intensity also shape the spillover impact of agglomeration.
In the case of domestic firms, larger, more productive, and more capable firms with prior exporting experience are likely to enjoy greater spillover from the exporter cluster. As China seeks economic growth driven by innovation and domestic consumption, access to foreign markets by a greater number of Chinese firms—especially private enterprises—remains important. Our study shows that policies that support the clustering of domestic exporters can be effective in promoting internationalization, especially when firms face high costs to export.
However, as observed in many studies on FDI spillovers, and also suggested by our study, the absorption of knowledge spillovers requires skills. It is desirable that small local firms receive support that enables them to make the best use of this knowledge spillover. Poncet, Chinese firms' entry into export markets: the role of foreign export spillovers, World Bank Policy Research Working Paper 6398, 2013.
Tomiura, Regional variations in exporters' productivity premium: evidence from factory-level data, RIETI Discussion Paper Series 13-E-005, Research Institute of Economics, Trade and Industry (RIETI), 2013. Notes: Table shows only summary statistics of the regressors during the two years 2001 and 2007.
