Data

4.1 Description of the data set and variables

Firms in the Z-Park must file an annual report with the Administrative Committee of the ZPark containing balance sheet information as well as information on their ownership, human resources, and R&D activities. The data set used in this paper is compiled from the annual reports for the period 2000–2003.¹⁰ The advantages of this data set are that most firms in the Z-Park are high-tech firms and that the annual report includes detailed information on firm employment by educational level. Therefore, the use of this data set enables us to estimate the size of knowledge spillovers from FDI through employment of educated labor in MNEs more accurately than other data sets. In addition, since firms in the data set are located close to each other in the science park, we do not have to take into account estimation issues that may arise as a result of the effect of distance on spillovers from FDI. Firms from both the manufacturing and the non-manufacturing sector are located in the ZPark. Non-manufacturing firms include those involved in hardware consultancy, software consultancy and supply, and data processing. Although knowledge spillovers from FDI in these activities would be of great interest, we focus on the manufacturing sector since the value of intermediate goods for non-manufacturing firms is not available for the period 2000– 2001.

In the data set, we define MNEs as firms that have a foreign ownership ratio of 30% or more. Since each respondent firm reported nationality of the largest investor of the firm, we can identify the home country of each MNE. Investments from Hong Kong, China; Macau, China; or Taipei,China are not defined as foreign capital, since characteristics of those investments are different from investments from other countries, mostly developed countries such as Japan and the US.¹¹ When we construct the aggregate of FDI variables within each "industry," "industries" are categorized according to the Industrial Classification and Codes for National Economic Activities of China at the two-digit level. Details on the construction of variables used in the estimation are presented in the Appendix.

4.2 Summary statistics

The sample for the regression in this paper consists of domestic firms that are defined as firms with a zero foreign ownership share. In addition, since we calculate capital and R&D stocks by the perpetual inventory method and use the GMM system to calculate our estimate, we include only domestic firms that reported the necessary data for at least three consecutive years during the four-year period 2000–2003. This selection process results in a sample of 798 firms and 1,504 firm-year observations. Table 1 [ PDF 56.5KB | 1 page ] reports summary statistics for the firm-level variables for these domestic firms and the industry-level variables relating to MNE capital and educated labor used in the regression. Table 2 [ PDF 56.5KB | 1 page ] shows the mean of the industry-level FDI variables by year, pointing to a drastic increase of MNEs in the Z-Park from 2001 to 2002. This increase is probably because of the PRC's accession to the World Trade Organization in December 2001.

Table 3 [ PDF 57.6KB | 1 page ] presents the extent of FDI penetration in the Z-Park by year, by industry, and in total, represented by the MNE share of sales, total employment, and employment of workers, with a master's or higher degree. The average share of MNEs in the total employment of workers with a master's or higher degree is smaller than the share of MNEs in sales or total employment in every year, indicating that MNEs are less likely to hire educated labor than domestic firms. In addition, we find a large variation in the extent of FDI penetration across industries. The MNE share of total sales mean value average is 10.3%, but more than 20% in four industries (rubber, electrical machinery, communication and computing equipment, and precision and optical instruments). The foreign share of educated labor mean value average, however, is 4.6% and exceeds 10% in the pharmaceuticals and medicinal chemicals, and the communication and computing equipment industries.

4.3 Differences in MNEs across home countries

Since our data set contains information on the nationality of the largest shareholder of each firm, we use that information to classify MNEs by home country. Table 4 [ PDF 38.4KB | 1 page ] presents the number of MNEs by year and in total for selected home countries. This table indicates that in terms of the number of MNEs, the United States is the largest home country of FDI to the ZPark, followed by Japan.

Further, Table 5 [ PDF 38.6KB | 1 page ] presents the amount of capital stock and the number of workers for Japanese, US, and other MNEs in the Z-Park, as well as the share of each group of total MNEs. The table indicates that Japanese MNEs have the largest shares in both capital stocks and employment. The share of Japanese MNEs in foreign capital stock mean value average exceeded 60% during the period 2000–2003, whereas their share in foreign employment was about 40%. The table also shows that Japanese MNE share in foreign capital stock is larger than their share in foreign employment, while the relation is the opposite for US and other MNEs. This evidence suggests that Japanese MNEs are more capital intensive than non-Japanese MNEs.

Table 6 [ PDF 57.5KB | 1 page ] shows the share of Japanese, US, and other MNEs in total employment by industry. This indicates that MNE employment patterns differ substantially across industries. The Japanese MNE share is large in the machinery, the electrical machinery, communication and computing equipment industries, as well as in the other (i.e. not steel or iron) basic metal industry.. Japanese and US MNE shares are similar in some industries, such as the plastics, communication and computing equipment, and precision and optical instruments industries. The share for either Japanese or US MNEs, however, is high while the other is low in some other industries, such as the food processing, machinery, transport equipment, and other basic metals industries.

Differences between Japanese and non-Japanese MNEs can also be seen in their utilization of educated labor. Table 7 [ PDF 38.3KB | 1 page ] presents the number of employees with a master's or higher degree, those with a bachelor's or higher degree, and those with overseas education (i.e., outside the PRC) for Japanese, US, and other MNEs. The three right hand columns in the table show the share of the total number of educated workers in the Z-Park for MNEs from Japan, the US and elsewhere. Table 7 clearly indicates that the share of both educated and overseas workers is much smaller at Japanese MNEs compared with non-Japanese MNEs. For example, the mean value average of the ratio of Japanese MNE employees to total MNE employees with a master's or higher degree was only 8.7%, in contrast with the share of the total capital for MNEs operating in the Z-Park that Japanese MNEs provided—the mean value average of which exceeded 60% (Table 5). The mean value average of the share of total educated employees for US MNEs was above 30%, as compared with the US share of total foreign capital of 19%. Table 8 [ PDF 56.7KB | 1 page ] looks at the utilization of educated workers by home country from another perspective, showing the ratio of educated employees to total employees for the three types of MNE. For Japanese MNEs, the mean value average of all workers held a bachelor's, or higher, degree was only 12%, while in the case of non- Japanese MNEs, the share ranged from 47% to 79%.Similarly, Japanese MNEs employ many fewer PRC returnees from overseas than non-Japanese MNEs.

Download this Paper [ PDF 194.5KB| 27 pages ].

Comment(s)

There are [0] comment(s) for this entry. Post a comment.