The effect of inward foreign direct
investment on China’s high-tech product
export and differences across regions,
based on data from 1995-2014
Name: Yike Ding
Student number: 11373652
Supervisor: C.W. Haasnoot
Word count: 5602
Date: 31/01/2018
Faculty of Economics and Business
University of Amsterdam
Statement of Originality
This document is written by Student [Yike Ding] who
declares to take full responsibility for the contents of
this document.
I declare that the text and the work presented in this
document are original and that no sources other than
those mentioned in the text and its references have been
used in creating it. The Faculty of Economics and
Business is responsible solely for the supervision of
completion of the work, not for the contents.
1. Introduction
In the past 3 decades, China has become one of the world’s largest exporters. In 1980, China only had 18 billion dollars exports while in 2015, the exports reached 2273.5 billion dollars. With the policy of “Transformation and openness” in the late 1970s, China opened its markets to the world and has attracted large amount of foreign direct investment (FDI) from OECD countries (Zhang 2006). As the major form of FDI in recipient country, foreign invested enterprise accounted for increasing number of export during the period. Until 1999, exports by foreign invested enterprises already took up 45.5% of total exports (Zhang and Song 2000).
Many past literatures have proved that FDI is not only an engine for capital formation, but also helps transferring technology and skills. Therefore, along with the amount of exports, the export structure has also changed. Initially, China was favored by OECD countries due to its cheap and abundant labor force, huge potential market and various resources. At that time, China’s export mainly focused on manufacturing and agricultural products that require less in technology level. With the FDI technology spillover and skill transfer afterwards, China started to produce and export more high-technology goods. In 2001, high-tech product export was 46.5 billion dollar, accounting for 17.45% of all export. This number increased to
29.33% in 2011, in which year the high-tech export amount was 600.9 billion dollar. This vast increase in both FDI and high-tech export gives a good topic to investigate.
In addition to favorable conditions, China’s attraction towards FDI was also contributed by its supporting policies. China had relatively liberal FDI regime than other East Asia countries such as Japan and Korea. To make the best of FDI to improve high-tech industry production, many high-tech industry development zones were set in most provinces across all the country since 1992, which to some extent has led to increase in high-tech product export (Zhang 2011; Wang and Wei 2010).
There are limited amount of existing literatures discussing the determinant of high-tech export. In terms of high-tech industry, it is an industry requires specially trained labor force. Borensztein et. al (1998) argued that the effect of FDI is dependent on the human capital level in the economy and there exists a positive interaction between FDI and educational attainment. Their argument was based on a general theoretical level, but due to the
particularity of high-tech industry, educational attainment seems to be even more important. Once technology are transmitted to the recipient country, labor force with capability to use it will be needed. Therefore the human capital level, which could also be measured as high college enrollment rate, can be considered a determinant of high-tech export.
In this paper, panel data at provincial level from 1995 to 2014 will be used to examine the impact FDI has on high-tech product export. Additionally, 28 provinces selected will be
divided into 3 regions and the differences between regions will be another interest of this paper. The main contribution of this paper is to investigate the relationship between FDI and high-tech export which has not be widely discussed before. The time series will be more up-to-date as well. A little overview and conclusion of results would be that FDI, as expected, has positive and significant effect on high-tech export in China both for total data and regional data. The Eastern region, which is located next to coast, has a larger coefficient of FDI than Middle and Western region do. For the effect of education and the presence of high-tech industry incentive zone, it diverse across regions. These results could be meaningful for policy makers. According to the results, attracting FDI would be better off for high-tech export. Moreover, the distribution of high-tech industry development zone and educational resources should also be taken into account
The rest part consists of 4 section. Section 2 is a review of previous literature, data and empirical methodology will be introduces in section 3. In section 4 the regression results will be displayed and analyzed. Last section will be the conclusion.
2. Literature review
Some theoretical literatures will be introduced and then literatures focusing on China as well as literatures concerning other variables of my model will also be discussed.
For the relation between FDI and export, plentiful researches proved that the relation is positive. Harding and Javorcik (2011) did tests on 150 more countries over period of 1984-2000. Under such a wide range, they argued that export quality is positively related to FDI. This finding indicates that entry of foreign capital in developing countries could contribute to the increasing quality of export both in absolute terms as well as bridging the distance to quality frontier. Beugelsdijk et. al (2008) discussed the different impact of horizontal FDI and vertical FDI have on host country’s economic growth. They distinguished between horizontal and vertical FDI and brought up that vertical FDI are efficiency seeking while horizontal FDI are market seeking. Using data of both developed and developing countries from 1983 to 2003, they found that in developing countries, there is no significant connections existing between horizontal FDI respectively vertical FDI and economic growth. It is not necessary to distinguish FDI in this paper as finding appropriate data will be a difficult task, but it is good to know that different types of FDI have different impact.
Other authors had researches of the same topic but specialized in different countries and areas. Enimola (2011), Temiz and Gokmen (2009), Nguyen and Xing (2007), Kutan and Vuksic (2007) and Asirvathamtest et.al (2017) test the relation between FDI and export performance of Nigeria, Turkey, Vietnam, central and eastern Europe and Association of Southeast Asian nations (ASEAN) countries respectively. Same as China, these countries and areas are all developing ones, so the results might have some referential value to the study of China. Enimola (2011) and Temiz and Gokmen (2009) employed Granger causality test and found that there exists a one-way causal connection between FDI and export, notably FDI has
positive impact on export but export does not have the same effect the other way around. The others only proved that the relation between FDI and export is positive and significant. However, Liu et. al (2002), who conducted the test based on data f China, had different results that in the long run, there exists two-way causal connections between FDI and export by using China’s quarterly data of export, import, FDI and approximate quarterly GDP from January 1981 to April 1997.
Some literatures mentioned FDI’s function of transferring technology to recipient countries. Zhang (2006) stated that FDI is the most powerful engine for export and makes huge
contribution to capital formation in host countries. Pack et. al (1997) also pointed out that compared to other form of technology transfer channel such as licensing contract and market transactions, FDI is the dominant channel and has most outstanding effect. Borensztein et. al (1998) had similar opinion of FDI transferring technology. Their research was about how FDI affect economic growth and their conclusion was that the influence of FDI in economic growth is highly rely on the level of human capital available. Economic growth has a strongly positive link with education attainment.
In terms of human capital stock, the finding of Borensztein et. al (1998) makes sense because high-tech industry involves research, development and innovation that require special trained labor force. Although this paper will be talking about the link between FDI and export rather than economic growth, the channel through which FDI can be made good use of are the same. The training and education to get labor force prepared for new technologies is an important factor indicating how FDI affect host countries’ economic growth as well as export. In line with Borensztein et. al (1998), Pack and Saggi (1997), Xu and Lu (2009) and Wang and Wei (2006) had similar conclusion that higher level of human capital will lead to better export performance and economic growth.
Large number of literatures also shared the view that government policies and support are a vital factor that could accelerate the development of high-tech industry both from a broad perspective (Pack and Saggi 1997, Storeya and Tetherb 1998, Rodrick 2006) and based on China specifically (Zhang 2006, Wang and Wei 2010, Xu and Lu 2009). Certain policies that support and provide favorable developing environment for high-tech oriented firms would reinforce the level of technology and therefore increase a country’s high-tech export. Rodrick (2006) argued that support from Chinses government is one of the major factor contributing to China’s high-tech industry development and increase of high-tech export. Storeya and Tetherb (1998) investigated the public policies in EU countries for the high-tech industry firms in different stage of development and different locations during 1980s and early 1990s. Their judgement is that policy should differ as different development stage requires different provision and this is what most policy makers fail to do. In China, there are high-tech
industry development zones across all the nation. Wang and Wei (2010) had two regression analysis on what contributes to the rising export sophistication in China. They included data of GDP, FDI, population and college enrollment of 140 cities in China, which cover all provinces across China. These data were linked with another set of data containing export, designation of policy zone, firm ownership and transaction type. One of their main
conclusion that the existence of certain policy zone have worked to incentivize firms to upgrade their product to obtain a higher unit value and also lead to more sophisticated export. Zhang (2006) provided with a qualitative analysis of why China attracted large amount of FDI and how China utilize FDI. He illustrated China has an open FDI regime and also an export-oriented and technology-oriented FDI strategy, which enables China to attract and make good use of FDI and improve its technology level as well as export.
There are several literatures regarding FDI and export in China. Zhang and Song (2000) first provided some theoretical and qualitative analysis of the direct and indirect effect of FDI on host countries’ export. Then they used provincial data of FDI of last year, export, import, GDP, share of manufactured goods of GDP and exchange rate of CNY to US dollar from 1986-1997. The results indicated that the coefficient of FDI in the previous year is significant and positive, which means FDI is an important factor prompting export. Since their test was based on provincial data, they also concluded some provincial features. Both FDI and export were more concentrated in coastal areas rather than inland areas, with more than 85 percent of export come from coastal areas and more than 85 percent of inward FDI went to coastal areas. However, one thing they did not investigate, and is one of the interest of this paper is that whether the FDI affect export to the same extent in different provinces. Sun (2001) had a similar research which was about FDI and regional export performances. Implementing Zhang and Song (2000)’s conclusion, he compared the difference in effect of FDI, notably FDI has strongest and significant effect in coastal region. The effect in middle region is less strong but still significant while the effect is insignificant in western region.
Generally, previous literatures have already provided an explicit and comprehensive framework for what will be done in this paper. FDI has a positive and significant effect on export in developing countries and also has the function of transferring technology. It makes sense to take education attainment and the presence of high-tech industry development zone into consideration. I would also like to add GDP to my model because it is an important and common measurement of economic performance. However, most literatures used relatively out of date time series. The focus of this paper will be narrowed down to the effect of FDI export of high technology goods rather than the whole export as many of literature suggest FDI can transmit technology. Moreover, all provinces selected will be divided into 3 regions according to their locations, which are Easter, Middle and Western regions. The differences of FDI’s effect between regions will also be an interest. Time series will be more up-to-date in this paper, to be specifically, from 1995-2014. Hypothesis are that the coefficient of FDI is positive and significant in total and every region and Eastern region have higher coefficient than Middle and Western regions.
3. Data and method
As mentioned in previous part, data of provincial level from 1995 to 2014 will be used. Unfortunately, high-tech products export amount for each province of 1996, 1997 and 2009 could not be found. As the 3 years are not continuous, omitting them would not have very big influence to the results. Therefore, time series would be exactly 1995, 1998-2008 and 2010-2014. 28 out of 31 provinces in mainland of China are included. 3 provinces, namely Chongqing, Qinghai and Tibet, are left out due to different reasons. For Chongqing, it is a municipality directly under Central Government set up in 1997, so the data of Chongqing before 1997 is unavailable. For Qinghai, export data of several years are somehow missing. Tibet is a special case as most of its export amount are zero or very low, which can be considered as an outlier and should be left out. The data needed for the model includes high-tech export, FDI, GDP, proportion of population that being enrolled into college and the number of high-tech incentive zone a certain province has in a certain year.
The data of high-tech products export are found in the Yearbook of High Technology
Industry of China of each year. Because the data are classified by categories instead of being integrated, I add up all the data of different categories of each province in each year.
According to the yearbook, 5 sector are considered as high-tech industry, namely
pharmaceutical and medical products, aircrafts, spacecraft and related equipment, electronic and communication equipment, computer and information technology equipment and measuring instrument and medical equipment. GDP of each province are available on the website of National Bureau of Statistics of China. In Beugelsdijk et. al (2008)’s paper,
horizontal and vertical FDI are found to have different impact on economic growth and trade. I would like to distinguish FDI in my data as well, but the statistic provided are all gross FDI and to my limit, I cannot find any distinct FDI data. Hence, gross FDI will be employed here. The FDI data, as well as the proportion of population being enrolled into college, can be found in the yearbook of each province. In some certain year, for example, 2000, 2005 and 2010, the college enrollment rate can also be found in Report of Demographic Census. For the number of high-tech industry development zone, I searched all the zones and see which year were they set up and which province were they belong to, then adding up the amount respectively.
In consideration of fairness, per capita data of high-tech export, FDI and GDP seem to be more convincing than gross data. It is not hard to find GDP per capita on the Internet as it is a common measurement of economic. Nevertheless, for the rest variables, high-tech export and FDI inflow, this is not easy. Hence I decided to calculate per capita data of high-tech export and FDI in the following way.
𝐹𝐷𝐼𝑝𝑐 =𝐹𝐷𝐼 ∗ 𝐺𝐷𝑃𝑝𝑐
𝐺𝐷𝑃
𝐻𝑇𝐸𝑋𝑃𝑝𝑐 =𝐻𝑇𝐸𝑋𝑃 ∗ 𝐺𝐷𝑃𝑝𝑐
𝐺𝐷𝑃
It is a bit tricky, but it is the best way I can get. Firstly, I divide GDP by GDP per capita and get a number. Theoretically, this number can be considered as population. Then I divide high-tech export and FDI by that number as well. In this way, per capita data can be obtained.
To get a straight forward impression of what the data looks like, I first drew a scatter plot of high-tech export and FDI and of high-tech export and GDP respectively during the whole time period. It is shown in the scatter plot that most plots gather around in the left corner while there are two sets of data has relatively high high-tech export and FDI at the same time. In the high-tech export and GDP scatter plot, the same thing happens and the gap is even bigger. From the plots at the right bottom part we can tell that some provinces have high GDP but export very little, whereas some provinces, indicated at the top of the scatter plot, has export oriented economic structure.
The graphs above include data of all province during 1995-2014, which contain variations both between provinces and time. To make it more explicit and clear, I also made two scatter plot of high-tech export and FDI and GDP respectively in the year of 2014. It is very
straightforward in the graphs that 2 provinces are out of the ordinary, which I assume are Jiangsu and Guangdong again. Also, the data of high-tech export, FDI and GDP are likely to have large standard deviations so taking logarithm of them might be a good solution. .
In addition to the overall situation described above, the distinction between provinces is also a good thing to know. In the panel line graph of high-tech export, FDI and GDP by province, Jiangsu and Guangdong are two province acting really differently, with much higher high-tech export, FDI and GDP. One thing should be mentioned is that there are 4 provinces, namely Jiangsu, Guangdong, Henan and Shandong, have FDI inflow more than their GDP in most of the time period. Some provinces, such as Beijing, Tianjin and Shanghai, have high GDP but low high-tech export. This is interesting because they are 3 out of 4 municipality directly under Central Government (Recall that the other one is Chongqing and are left out due to data incompleteness). These provinces are more advanced in tertiary industry and do not devote themselves to manufacturing industry. Although the technology level in these provinces are probably higher as most top universities centered here, they do not produce very much of high-tech products. Actually from these graphs, the distinctions between provinces are already revealed.
The data of EDU are all percentage and its standard deviation is quite low, therefore it would be hard to read if it is in the same graph with high-tech export. In order to have a more explicit impression, I had a line graph shows its tendency during the years by province. It is not surprised to see that college enrollment rate is upward sloping in all provinces. 2
provinces stand out in the graph, which are Beijing and Shanghai. As mentioned in previous part, these 2 provinces are both municipality directly under Central Government. Compared to other common provinces, they have more educational resources but less population.
In terms of HTIDZ, the same way of drawing graph is adopted. Jiangsu and Guangdong again become the provinces have relatively higher data. From the graph of high-tech export, FDI and GDP we are able to infer that Jiangsu and Guangdong have much more high tech export and are export oriented, which is in line with them having more development zones.
Therefore, a preliminary inference is that HTIDZ has a positive effect on high tech export.
0 5 1 0 0 5 1 0 0 5 1 0 0 5 1 0 0 5 1 0 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015 1995 2000 2005 2010 2015
Anhui Beijing Fujian Gansu Guangdong Guangxi
Guizhou Hainan Hebei Heilongjiang Henan Hubei
Hunan Inner mongolia Jiangsu Jiangxi Jilin Liaoning
Ningxia Shaanxi Shandong Shanghai Shanxi Sichuan
Tianjin Xinjiang Yunnan Zhejiang
H T ID Z Year Graphs by Province
3.2 Methodology
Table 1. Summary of Variables
e(count) e(sum_w) e(mean) e(Var) e(sd) e(min) e(max) e(sum)
HTEXP 476 476 550.853 4555261 2134.306 .0015243 18322.41 262206
FDI 476 476 16499.37 9.32e+08 30532.67 1.287896 186593.4 7853701
GDP 476 476 22574.16 4.15e+08 20370.05 1826 105231 1.07e+07
EDU 476 476 .0744307 .0033979 .0582911 .0043 .4121 35.429
HTIDZ 476 476 2.766807 5.244454 2.290077 0 11 1317
Table 2. Summary of Log Variables
e(count) e(sum_w) e(mean) e(Var) e(sd) e(min) e(max) e(sum)
LHTEXP 476 476 2.745114 9.17447 3.028939 -6.486237 9.81588 1306.674 LFDI 476 476 7.970458 5.349344 2.312865 .2530102 12.13669 3793.938 LGDP 476 476 9.644158 .793518 .8907963 7.509883 11.56391 4590.619 EDU 476 476 .0744307 .0033979 .0582911 .0043 .4121 35.429 HTIDZ 476 476 2.766807 5.244454 2.290077 0 11 1317
Table 1 above shows the summary of original data. The high-tech export, FDI and GDP data have extremely high standard deviation, which could cause skewness in regression and makes it important to use the logarithm. Another two variables, education attainment and high-tech industry development zone, have standard deviation in relatively normal range. Also several years of high-tech industry development zone’s data are zero and makes no sense in
logarithm. Therefore, only high-tech export, FDI and GDP will be in logarithm while educational attainment and high-tech industry development zone remain unchanged. After transforming high-tech export, FDI and GDP to logarithm, their standard deviation turn back to a normal level.
To examine the difference across provinces, I would like to divide the 28 provinces into 3
regions. Thisway to classify provinces is to sort them by administrative division prescribed
by country. In China, there are altogether 3 regions presented as the following table. The following map shows the geographical distribution. In the map, red part indicates Eastern region, blue part indicates Middle region and yellow part indicates Western region. Almost every province in the Eastern region are in coastal area or has a port while the other two regions lies inland.
Table 3. Division of provinces
Eastern area Middle area Western area
Beijing Shanxi Sichuan
Tianjin Inner Mongolia Guizhou
Hebei Jilin Yunnan
Liaoning Heilongjiang Shaanxi
Shanghai Anhui Gansu
Jiangsu Jiangxi Ningxia
Zhejiang Henan Xinjiang
Fujian Hubei
Shandong Hunan
Guangdong Guangxi Hainan
The complete model will be:
LHTEXPit= β0+β1LFDIit+β2LGDPit+β3EDUit+β4HTIDZit+εit
where i represents the province, t represents the year and ε is the error term. In the model, LHTEXP, LFDI and LGDP are the natural logarithm of high-tech products export amount per capita, FDI per capita and GDP per capita respectively. Since the variables are in logarithm,
this model will be measuring elasticity.Besides the five variable mentioned before, a group
if the province is in Eastern region, 2 if in Middle region and 3 if in western region.In this way, total effect and cross region effect can both be obtained. One of the main focus of this paper is whether the coefficient of LFDI is larger than zero, in other words, whether FDI has a positive effect on high-tech product export. Then it is also interesting to compare the coefficient of other variables of different regions.
Because the data set are panel data that has variations between both year and province, it is necessary to test what model should be employed. Practically, either random effect model or fixed effect model will be used to process estimation. The assumption of fixed effect model is that the slope coefficients remain no change for all units cross groups, and the intercepts only varies over individual cross-group units not time series. The random effect model has same assumption that slope coefficients are constant cross sections but in this model the
interception is a random variable, which shows the unit distinctions in the intercept value of each cross group unit Greene (2003). To figure out which model should be used, I did the Hausman test to see whether random effect or fixed effect is better for these data. With a p-value equal to zero it turns out that fixed effect model will provide more precise results. The next step is to run all the regressions needed.
Table 4. Hausman test, Fixed and Random Effect
Variables Coefficient Values
Fixed Effects Random Effects Difference S.E.
LFDI 0.9847 0.9335 0.5012 0.0155
LGDP o.3256 0.2755 0.0501 0.0083
EDU 4.3759 6.5137 -2.1378 0.3383
HTIDZ 0.1621 0.2127 -0.0506 0.0074
Constant -9.0183 -8.4258 -0.5925 0.1850
Hausman test (Recommended) Prob>chi2=0.0000 (Fixed Effect)
4. Results and analysis
4.1 Total
First of all, I ran a regression of all data to see the overall effect FDI has on high-tech export. In this regression, the variable Region is left out because it is intended to measure the whole effect. From the results of regression, the coefficient of LFDI is 0.326 with a p-value equal to zero. This coefficient is statistically positive significant and means that 1% increase in FDI would lead to 0.326 % increase in high-tech export. Taking a look at other variables, GDP and education attainment are positive and significant but high-tech industry development zone has insignificant impact. The coefficient of education attainment is relatively larger than others, this is probably because education attainment is between 0 and 1. From the value, it is smaller than other variables. Therefore to have impact on high-tech export, its coefficient should be larger. In terms of high-tech industry development zone, this results is out of expectations as in the previous literatures, government’s supporting policies are supposed to
increase export. In the sample I choose, high-tech industry development zone do have a positive effect. Maybe in the long run, with increasing number of high-tech industry
development zone, the existence of high-tech industry development zone are not special any more will not play a influential role as before. Also the overall R square is 0.724, which means 72.4% of the dependent variable are explained by this model.
Table 5. Regression results of total data
(1) VARIABLES LHTEXP LFDI 0.326*** (0.0681) LGDP 0.920*** (0.110) EDU 4.719*** (1.791) HTIDZ 0.0456 (0.0476) Constant -9.204*** (0.750) Observations 476 Number of province 28 R-squared 0.724
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
4.2 Easter, Middle and Western regions
Consistent with expectations and previous findings, FDI has positive and significant impact on high-tech export in all three regions. Among these regions, Eastern region’s high-tech export respond most to FDI with the coefficient equal to 1.494, followed by 0.875 of Western region and 0.846 of Middle region. All 3 regions have high R square above 0.8. Eastern region, which can also be referred to as coastal area because most of the provinces in this region are near to coast, is a relatively more prosperous area in China. From data and method section we can see that province from coastal area obtain more FDI and export more. Zhang and Song (2000) also made comparison between coastal and inland area. During 1986 and 1998, coastal area attracted 87.7% of total FDI that flowed into China and produced 85.45% of goods exported. In their research of the determinants of FDI, they concluded that location is an important factor that affects the amount of FDI inflow. This result can be seen
consistent with theirs.
The results for other variables are more or less unexpected. GDP has positive effect on high-tech export for all regions, but the effect is statistically significant only in Eastern region. The rest two variables, education attainment and high-tech industry development zone, show big regional distinctions. In Eastern region, coefficient of education attainment is positive and
coefficient of high-tech industry development zone is negative, however, neither of them is significant. In Middle area, education attainment has relatively large influence to high-tech export, with its coefficient equal to 10.902. High-tech industry development zone has positive impact as well. Unfortunately, both of them are insignificant again. In Western area, both the coefficient of education attainment and high-tech industry development zone are significant, but EDU has negative effect on high-tech export. Possibly Explanation for this is that although China experiences large increase in high-tech export, the most of it are still from processing industry (Rodick 2006, Nataraj and Tandon 2011, Zhang and Song 2000). In purely processing manufacture, education is not required. Then the reason of differences of tech industry development zone between regions may be that in 1995, majority of high-tech industry development zone were located in Eastern region. In the following years, the increase of number of zones in Eastern region was smaller than that in Middle and Western region. Therefore, the impact of high-tech industry development zone is stronger in Middle and Western region.
Table 6. Regression results of regions
Eastern Middle Western
(1) (2) (3)
VARIABLES LHTEXP LHTEXP LHTEXP
LFDI 1.494*** 0.846*** 0.875*** (0.1000) (0.0429) (0.0699) LGDP 0.671*** 0.186 0.261 (0.191) (0.206) (0.279) EDU 0.902 10.90* -14.69** (1.870) (5.969) (7.072) HTIDZ -0.0579 0.0899 0.342** (0.0511) (0.0941) (0.136) Constant -15.78*** -7.381*** -6.565*** (1.524) (1.707) (2.085) Observations 204 153 119 R-squared 0.848 0.836 0.830 Number of Region 1 1 1
Standard errors in parentheses *** p<0.01, ** p<0.05, *p<0.1
To see whether high-tech export has increased under the effect of FDI, another measurement, namely the percentage of high-tech export out of all export may also be taken into account. However, this will not be the main model of this paper and is only an auxiliary evidence. Also, because percentage is between 0 and 1, it seems better to use beta regression. The regression results are as follows.
Table 7. Regression results for high-tech export percentage
(1) (2)
LFDI 0.326*** 0.234*** (0.0681) (0.0261) LGDP 0.920*** -0.146* (0.110) (0.0810) EDU 4.719*** 5.084*** (1.791) (1.030) HTIDZ 0.0456 0.0376 (0.0476) (0.0231) Constant -9.204*** -2.618*** (0.750) (0.681) Observations 476 476 R-squared 0.724 Number of province 28 28
Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1
The coefficient of LFDI is still positive and significant. Nevertheless the coefficient of LGDP is surprisingly negative and insignificant. The education attainment also has positive and significant while high-tech industry development zone’s effect is insignificant. Apart from the effect of GDP, the other variables all have similar results as the model using high-tech export amount as dependent variable, which makes the hypothesis more strongly proved.
5. Conclusion
In this paper, the relationship between inward FDI and high-tech product export in China and how differently high-tech export are affected in different provinces are examined. The
previous literatures relating to this topic mostly focused on the connection between FDI and total export, as well as with GDP. Based on findings of previous literatures, I found it would be meaningful to test the impact FDI have on exactly high-tech export as FDI has the
function of transferring technology. The data collected are at provincial level, so the differences between provinces or regions can also be researched. Additionally, time series used in previous literatures are old, so new and more up-to-date time series are employed in this paper.
In the empirical part, totally 4 regressions are conducted. Before conducting regressions, because panel data are used, it is necessary to test whether fixed effect model should be employed. I did the Hausman test and had the result that fixed effect model performs better than a random effect model with these data set. Firstly a regression involving all data are run to investigate the effect of FDI on the whole China. The results are as expectations that FDI has a positive and significant relation with high tech export. Then I do another 3 regressions with respect to 3 different regions, namely Eastern region, Middle region and Western region.
In terms of connection between FDI and high tech export, all the 3 regions had similar results that this connection is positive and statistically significant. Among them, Eastern region’s high tech export seems to be more sensitive to FDI change compared to other 2 regions. Regarding college enrollment rate and number of high tech industry incentive zone, the results diversify a lot. In both Eastern and Middle region, college enrollment rate has positive but insignificant effect on high tech export while that effect in Western region is negative and significant. For number of high tech industry development zone, results are on the contrary of college enrollment one. In Middle region and Western region, number of high tech industry development zone both has positive effect on high tech export. However, in Eastern region the effect is negative and insignificant. From these results, a rough conclusion can be made that provinces closer to the coast have more opportunities and better resources, so their high tech export level are more sensitive to FDI and educational attainment plays a more important role also. In contrast, high tech export of provinces in Middle and Western area respond positively but less to the change of FDI. They rely more on the government policies such as setting up development zones to expand the volume of high tech export.
These results may bring some inspiration for policy makers. Considering the function of FDI, it is of great help for those countries in developing phases to attract more FDI and make the best of them. For China government, according to the results, setting up more high-tech industry development zone in Western region may be a good choice. Storeya and Tetherb (1998) pointed out that to set differentiated policies for high-tech firms in different stage of development is also important. Therefore government should better recognize special requirements and qualities of different firms and formulate individualized policies for them. Increasing the educational investment in Middle region and helping transfer China’s export structure towards less processing trade but more research and development will also be decent implications for policy makers.
In sum, the most robust finding of this paper is FDI has a positive and significant effect on China’s export of high tech products and provinces in Eastern region of China are more sensitive to the change of FDI. There are also some limitation of this paper. The most important one is the incompleteness of data. Data of 1996, 1997 and 2009 are missing. In addition, Beugelsdijk et. al (2008) made the argument in their paper that it is vertical FDI in host country that can contribute to economic growth and export. However, existing data available did not make distinguishment between vertical and horizontal FDI. The time series include 2008 and 2010 in which the big financial crisis took place. Economic situation may change significantly during recession period, yet data of these two years are not analyzed separately. Maybe further researches could make improvement on these limitations.
Reference
Asirvatham, J., Rasiah, R., Thangiah, G., & Naghavi, N., (2017) “Impact of foreign direct investment, imports and tariff deregulation on exports among pioneering ASEAN members: Panel data analysis”, International Journal of Business and Society, Vol. 18 No. 1, 2017, pp.1-12
Beugelsdijk, S., Smeets, R., & Zwinkels, R., (2008) “The impact of horizontal and vertical FDI on host’s country economic growth”, International Business Review 17 (2008), pp. 452– 472
Borensztein, E., De Gregorio, J., & Lee. J-W., (2008) “How does foreign direct investment affect economic growth?”, Journal of International Economics 45 (1998), pp.115–135 Enimola, S. S., (2011) “Foreign direct investment and export growth in Nigeria”, Journal of Economics and International Finance Vol. 3(11), pp. 586-594
Greene, W. H., (2003) Econometric Analysis, Supper Saddle River, New Jersey: Prentice Hall.
Harding, T., & Javorcik, B.S., (2011) “FDI and export upgrading”, University of Oxford
Department of economics discussion paper series
Kutan, A. M., & Vuksic, G., (2007) “Foreign Direct Investment and Export Performance: Empirical Evidence”, Comparative Economic Studies, 2007, 49, pp. 430–445
Liu, X., Burridge, P., & Sinclair, P. J. N., (2002) “Relationships between economic growth, foreign direct investment and trade: evidence from China”, Applied Economics, 34:11, pp.1433-1440
Nataraj, G., & Tandon, A. (2011) “China's Changing Export Structure: A Factor-Based Analysis”, Economic and Political Weekly, Vol. 46, No. 13 (MARCH 26-APRIL 1, 2011), pp. 130-136.
Nguyen, T. X., & Xing, Y., (2007) “Foreign direct investment and exports, the experiences of Vietnam”, Economics of Transition Volume 16(2) 2008, pp.183–197
Pack, H., & Saggi, K., (1997) “Inflows of Foreign Technology and Indigenous Technological Development”, Review of Development Economics 1(1), 1997, pp.81–98
Rodrik, D., (2006) “What’s So Special about China’s Exports?”, China &World Economy,
Vol. 14, No. 5, 2006, pp.1 – 19
Storey, D. J., & Tether, B. S., (1998) “Public policy measures to support new technology-based firms in the European Union”, Research Policy 26 1998, pp.1037–1057
Sun, H., (2001) “Foreign direct investment and regional export performance in China”,
Journal of Regional Science, Vol. 41, NO. 2, 2001, pp.317-336
Temiz, D., & Gokmen, A., (2009) “Foreign Direct Investment and Export in Turkey: The Period of 1991-2008”, EconAnadolu 2009: Anadolu International Conference in Economics
June
Wang, Z., & Wei, S., (2010) “What Accounts for the Rising Sophistication of China's Exports?”, China's Growing Role in World Trade, 0-226-23971-3, pp.63-104.
Xu, B., & Lu, J., (2009) “Foreign direct investment, processing trade, and the sophistication of China's exports”, China Economic Review 20 (2009) pp.425-439.
Zhang, K. H., (2006) “Foreign direct investment in China”, Canadian Foreign Policy
Journal, 13:2, pp. 35-50
Zhang, K.H., & Song, S., (2000) “Promoting exports, the role of inward FDI in China”,
