Third, spillover effects are greater in a special economic zone designed to attract foreign capital to accelerate the development of China's own high-tech industry. We find that, on average, domestic companies participating in joint ventures with a foreign partner, export plants and plants in special economic zones are more productive. First, spillovers are less likely to occur in ... wholly foreign-owned firms than in joint ventures.
This led to the growing recognition of China's economic potential and led to a surge in the number of FDI projects and their value at the beginning of the 90s (See Figure 1). At the beginning of the liberalization of the Chinese economy, the central government strategically directed FDI to the Special Economic Zones (SEZs) in the Guangdong and Fujian provinces. On average, foreign sales represent 48% of the total sales of large and medium-sized …rms in Chinese manufacturing.13.
Productivity estimation
If …rms have some pricing power, the estimates of the 's will still be biased, as inputs are likely to be correlated with the price charged by a …rm. The estimated input coefficients of the Cobb-Douglas production function for domestic …rms are listed by sector in Appendix F.1 and F.2. Since we are unable to control for the selection bias due to data limitations, the estimates for all three approaches are likely to still be biased downward.
To visualize the impact of the different methodologies, the core densities for two different sectors are plotted in appendix F.5.
Productivity spillovers
However, as discussed in Section 2, different forces (or of a different magnitude) may be at work for exporting plants compared to their non-exporting counterparts. Most foreign companies consider China to be a low-cost production base, especially the investments originating from overseas Chinese in Hong Kong, Macau and Taiwan are mainly export-oriented (Whalley and Xin, 2006). We construct an additional dummy variable Exporterit, which has a value of one if the plant is exporting at time t.
In particular, we want to examine whether spillovers can be attributed exclusively to these areas or whether they occur throughout the economy. To do this, we construct a new dummy variable SEZi that takes the value of one when the plant is located in a city where a development zone is established.19 The dummy variable is constructed based on a list provided by the China Development Association. Zone. We include this dummy in our regression to control for possible higher productivity levels in these areas due to special benefits not granted to ...rms outside the areas.
Additionally, we interact the dummy with our spillover measure Spilloverjt SEZi to investigate whether e¢ ciency spillovers are more likely to occur within these zones. 19Unfortunately, we do not know the exact location of the facility in a city, which means that we cannot distinguish whether the …rm is located in the zone or in the surrounding area of the same city. 2Spillover_W F OEjt+dj +dc+dt+ it: (9) The reason behind this distinction relates to the expected difference in spillovers related to WFOEs versus JVs, as discussed in Section 2.
Because these companies are less afraid to use their latest technological innovations, they are more likely to outperform local manufacturers. We therefore hypothesize that the competitive effect will be greater in the case of WFOEs, while the imitation spillovers could dominate for JVs.
Sectoral spillovers
First ... we find that exporting factories are slightly more productive than their non-exporting counterparts. However, regardless of their higher productivity, export-oriented factories seem to suffer from the presence of foreign ... companies in their sector. The negative coe¢ cient on Spilloverjt Exporterit sets the positive average spillover effect.20 In section 2 we summarized the various positive and negative effects of foreign investments for exports of …rms.
This shows that the scope for learning from incoming FDI is indeed quite limited for those … companies that export and already possess the necessary skills to compete in the world market. For domestic markets, this positive externality is more likely to offset the negative effect of competition in factor markets. The labor pooling effect can be particularly important in the case of exporting goods as they need highly qualified technical and managerial personnel to survive in the export market.
The existence of Special Economic Zones (SEZs) is a distinctive feature of Chinese foreign policy and a third source of consideration in our analysis. Although ...rms located in a SEZ are on average more productive than other plants, they do not seem to react differently to the presence of foreigners. In fact, the F-statistic of the joint hypothesis is 43.93, so we are able to reject the null hypothesis at the 10% level.
In ... the first column, the dependent variable is the logarithm of labor productivity22, while in column 2 and column 3 TFP is calculated using the OLS estimate of the Cobb-Douglas production function and the Olley-Pakes methodology to control simultaneity bias. . Across all approaches, there is evidence of positive FDI spillovers on average in China's manufacturing industry, while the effect is on exports.
Spillovers and the role of ownership structure
EPZs) and national border and economic cooperation zones (BECZs) are established primarily for the development of trade and the implementation of processing for re-export. On the other hand, ETDZs and national high-tech industrial development zones (HIDZs) are created to attract foreign capital and technology to accelerate the development of China's own high-tech industry. These results are consistent with ... the finding of negative export spillover effects and show that attracting export-oriented investments is not necessarily a useful strategy for generating positive externalities in the domestic market.
In Table 4 we reestimate equation (7) using three different methodologies to check the sensitivity of our results to the chosen Olley-Pakes approach in Table 3. Overall, the three approaches show very similar results, with only the magnitude of the effects differing slightly. Consistent with the summary statistics in Chapter 4, the coefficient on Exporterit becomes negative when we consider the spillover effects on labor productivity caused by FDI (columns 1 and 4).
In order to avoid an abundance of results and to keep a clear overview, we will henceforth discuss only the results of the Olley-Pakes specification. However, this should not be a limitation, as we have just shown that spillover effects are robust to different approaches. When we dig deeper into the impact on different subgroups in column 2, we see that WFOEs mainly compete with local …rms within an SEZ or with those …rms that are engaged in exporting.
Given that exporters are already more productive than the average... company, the opportunity to learn from foreign companies is less. This result demonstrates that technology leakage is likely to be the driver of the large differences in spillovers originating from the two ownership structures.
Impact according to the origin of FDI
Consequently, information leakage is less important for this group, so the presence of both JVs and WFOEs has a similar effect on exporters. Moreover, the relatively high capital intensity and advanced and complex technology used in the production of industrialized countries. Third, the negative impact on domestic export …rms … is greater in the case of HMT investment than non-HMT investment.
For example, Hong Kong's role as an export entrepôt between China and the rest of the world encouraged HK …rms to move their labor-intensive activities to the Chinese mainland as labor costs began to rise at home. As such, HMT investors view China as a low-cost manufacturing base and their investment is mainly export-driven.
Robustness checks
From the results of this regression in the third column of Table 7, we can see that…rms does indeed benefit…from foreign investors located nearby. In cities with a 10 percentage point higher share of foreign turnover, domestic companies will be on average 1.7 percentage points more productive. Our results tell us that in a sector with less competition…RMS are on average significantly more productive.
By doing so, we correct for the bias caused by imperfect competition in output markets and the related markups earned by …rms. We see that the level of productivity of joint ventures compared to domestic …rms is somewhat lower once we control for market power. This confirms our earlier conclusions that exporting ... firms suffer increased pressure on factor markets due to the presence of foreign firms, but this results in lower profit margins rather than a lower level of productivity per se.
In this paper, we used a rich panel data set of ... rms manufacturing in China to analyze in detail the impact of inward foreign investment on the performance of Chinese manufacturing plants. China's central government is introducing all kinds of mechanisms to increase the likelihood of positive spillover effects on domestic ... rms. The basic result of our spillover analysis is that, on average, the positive spillover effects on China's local ... rms.
Moreover, this study made it clear that the extent to which domestic firms are able to absorb technological knowledge largely depends on the origin of foreign direct investment. Finally, the robustness checks in the last section reemphasize the importance of studying within-group heterogeneity of…rms when analyzing policy questions. Nevertheless, this unique dataset at the …rm level allows us to reveal the heterogeneity among …rms in their response to the presence of foreign …rms.
In order to be able to assume pro...t maximizing behavior, we exclude state...rms from our analysis. De Loecker (2007) then extends their framework by controlling for simultaneity bias using OP and allowing multi-product …rms.26 Suppose that each …rm produces a product for which it faces the following demand. Expressing revenue net of sales as a function of demand and supply factors allows us to take into account the degree of competition in the production market, and in particular the price change in ...rms.
