UNIVERSITY OF AMSTERDAM
The effects of foreign direct investment on Chinese economic
growth in the period 1999 – 2014
Bachelor Thesis
David Bodar 10270663
Supervisor: Rui Zhuo
January 2016
Abstract
In this thesis the effects of foreign direct investment (FDI) on economic growth in China are investigated. A methodology inspired by Borensztein et al. (1998) is followed and two OLS
estimations on quarterly data on Chinese growth indicators are executed. Some insights are given on the change regarding this effect when two periods are compared, namely 1999-2009 and 1999-2014. Since the latter includes consecutive declining growth numbers, the positive effects commonly agreed upon in the literature are expected to show a reduction. To some extent, this is confirmed. However, firm conclusions are difficult to draw because of some biases and ambiguities that were apparent during the implementation of the research method. The main finding in literature is confirmed; FDI contributes positively to economic growth when a sufficient level of human capital is present.
2
Statement of Originality
This document is written by Student David Bodar who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is 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.
3
Contents
1. Introduction 4-5
2. Literature Review 5-8
2.1 The effects of FDI on economic growth 6-7
2.1.1 FDI and human capital 6
2.1.2 FDI and export promoting policy 7
2.2 The effects of other factors on economic growth 7-8
2.2.1 Export 7 2.2.2 Government Consumption 7-8 2.2.3 Domestic investment 8 3. Methodology 8-11 3.1 Data 8-9 3.2 Model specification 10-11 4. Empirical results 11-13 5. Conclusions 14 6. References 15
4
1. Introduction
Over the last decades, the Chinese economy has been a topic of interest in economic research. After the reform and the opening-up policy in 1978, the economy became more receptive to foreign countries and foreign money. These developments contributed to China being one of the fastest growing economies in the world, often facing double digit growth numbers. An indication is given by the World Bank graph below, which shows the GDP growth numbers for China from 1980 to 2015.
Graph 1: GDP growth
The transition of a planned economic system towards a more open market system involved the increase of foreign direct investors willingness to invest in China (Zhang, 2001). This is supported by Graph 2 showing the amount of FDI to be increasing heavily from 1982 onwards. According to the latest World Investment Report by the United Nations Conference on Trade and Development, China has become the world’s second largest recipient of FDI and is only exceeded by the United States (UNCTAD, 2015). The growing inflow of FDI and the considerable economic growth are investigated by several researchers (Berthélemy and Demurger, 2000, Zhang, 2001, Whalley and Xian, 2010) that conducted their research on questions such as: what are the effects of FDI on GDP growth in China? As can be concluded from Graph 1, from 2009 onwards the economic growth faced a decline. Most of the research on this topic agrees on a direct or indirect significant effects of FDI on economic growth, which makes it interesting to investigate this further (Berthélemy and Demurger, 2000, Zhang, 2001, Whalley and Xian, 2010, Balasubramanyam, 1996). This study attempts to come to conclusions on the effects of FDI on GDP growth in the more recent history. It seems relevant to make a distinction between the period before and after 2009, due to the decline of economic growth in the latter. Questions that raise from this are: are there any significant changes visible compared to older research on this topic? And, are there changes visible when the research period is extended with
5 recent data on GDP growth? The above questions are summarized in the following research question, which will play a central role in this thesis: To what extent have the effects of foreign direct
investment on Chinese GDP growth changed when more recent data is investigated than in existing studies?
Graph 2: Net FDI inflow
In the attempt to form a well-considered answer to this research question, two research periods are introduced. The first runs from 1999 to 2009 and the second from 1999-2014.
This thesis starts with a literature overview regarding the indicators of economic growth in China, with an emphasis on FDI. Next, the dataset used for the empirical analysis is described. After that an adapted version of the regression model used by Borensztein et al. (1998) is introduced and discussed. Then, the results of the Ordinary Least Squares(OLS) regressions on the two research periods are presented. Finally, conclusions are drawn from the insights this research has brought about.
2. Literature review: The indicators of economic growth
The effects of FDI on economic growth have been studied by various researchers in the last decades. Especially for the case of China, sufficient research is available. In this literature review a selection of the most influential papers on this topic will be discussed. To provide a solid foundation for the model used in this thesis, research on other indicators of economic growth that are commonly considered important will also be reviewed. A brief overview of the literature follows below.
Borensztein et al. (1998) provide a general framework to test for the effects of FDI and other economic variables on economic growth. Berthélemy and Demurger (2000) conducted research on the effects of FDI for the specific case of China. Balasabramanyam et al. (1996) investigated the
6 differences in the effect of FDI on economic growth between import substituting and export
promoting countries. Park and Prime (1997) analyzed Chinese export performance and its effect on economic growth. Barro and Lee (1994) wrote an influential paper on the sources of economic
growth. Barro (2001) elaborates on the effect of human capital on economic growth. The model that is used comes from Borensztein et al. (1998) and will be discussed in detail in the methodology section. The outcome of their panel data study on the effects of FDI on economic growth in 69 developing countries is discussed below.
2.1 The effects of FDI on economic growth
Most of the papers discussed in this section, conclude that the strongest and most significant effect of FDI on economic growth is not a direct effect. Some found out that it holds positive relations with human capital (Borensztein et al., 1998, Berthélemy and Demurger, 2000). Others focus on the relation between FDI and the export behavior of a country, like Balasubramanyam et al. (1996). The focus of this section will be therefore on the combined effects between FDI and human capital and FDI and export policy, rather than the effects of FDI on its own.
2.1.1.FDI and human capital
The most robust finding in Borensztein et al. is that the influence of FDI on economic growth depends on the level of education, their proxy for human capital (1998). They define education as Barro and Lee (1993) did: the initial-year level of average years of the male secondary schooling, since this is the related the most significant with economic growth, Barro and Lee find (1993). Borensztein et al. suggest that FDI is an indicator of economic growth because it is well able to transfer technology (1998). A prerequisite for this feature is that a recipient country holds a certain minimum stock of human capital. This is required to absorb of technological spillovers brought by multinational corporations (MNCs) that possess valuable technological knowledge. The stock of human capital therefore forms a limitation to the absorptive capacity of the recipient country. The more people that are well educated, the higher the absorptive capacity of a country of technological spillovers. Hence, the research model emphasises the effects of two determinants of economic growth. One is the inflow of more advanced technology via FDI and the other is the condition of absorptive capabilities in the recipient country (Borensztein et al., 1998).
Berthélemy and Demurger (2000) conducted research on the differences in the effects of FDI on growth in 24 Chinese provinces and the findings of Borensztein et al. They conclude the effect of FDI on economic growth to be positive and higher when an interaction term between FDI and human capital is included:
“This result suggests that the marginal effect of foreign investment on growth is increased where the
share of educated people is higher. Thus, provinces with a higher level of human capital seem to have benefited more from foreign investment than others.” (Berthélemy and Demurger , p. 151)
7
2.1.2. FDI and export promoting policy
In their 1996 paper, Balasubramanyam et al. agree upon FDI being an important vehicle for the transfer of new technology to the recipient county. They argue that the size of the effect of FDI strongly depends on the type of trade strategy the recipient country is adopting. The two types
distinguished are the policies of import substitution and export promotion. The reasoning followed by Balasubramanyam et al. (1996) could be summarized as follows: export substitution allows for and focuses on the attraction of FDI to benefit the country’s economic growth via the export sector. By contrast, the intention of import substitution policy is to protect the domestic market and attract domestic investments to realize this. The strength of FDI’s contribution to economic growth in an import substituting country is therefore subject to the country’s attitude towards international trade, and is therefore expected to be smaller. Export promotion induced FDI is expected to contribute more to growth since it benefits from a distortion-free international setting (Balasubramanyam et al. 1996). In the case of export promotion, the effects are stronger since it emphasizes the forces of the free market and competition. These are factors considered important for the economic environment in which FDI can support economic growth. This works in two steps. First, the export promoting country is provided with a larger amount of FDI and second, larger social benefits such as economic growth are obtained from it (Balasubramanyam et al., 1996).
2.2 The effects of other factors on economic growth
In this section, the effects of other factors on economic growth will be discussed. There is plenty of literature available on these variables because of the general macro-economic character this topic. Since the focus of this thesis is on the effects of FDI on GDP growth, these papers will be reviewed more concisely.
2.2.1. Export
Park and Prime (1997) run a regression on the effects of export on economic growth in China, and conclude on a direct influence. To assess the effect of exports on growth three variables are used; the annual growth rate of export, the percentage share of export in GDP and the percentage share of changes in export in GDP. In the regression results all three variables are proven to be significantly positive. This implies that China as a large country does not suffer from the so-called large country constraint. Several researches elaborate on this topic and conclude the larger a country, the smaller the export share in GDP (Park and Prime, 1997). This implies less competitive pressure from abroad when opening up to foreign trade, investment and technology transfer. In the case of China, the shortage of competitive pressure has not slowed down its GDP growth (Park and Prime, 1997). Therefore the research discussed above seems to provide sufficient evidence for the significant and positive effects of exports on GDP growth in China.
2.2.2 Government Consumption
In his 1989 paper, Barro argues that government consumption does not have a direct effect on the productivity level of an economy. However, he shows negative influences of the government
8 consumption to GDP ratio. Its relation to both GDP per capita growth and the ratio of private
investment to GDP is proven to be negative. An interpretation could be that government consumption provokes distortions like high taxes, but does not compensate for this by providing a compensating stimulus for investment and per capita growth (Barro, 1989).
The findings presented in the above analysis are supported by Barro and Lee (1994). One of the variables they investigate is government size which is expressed as the level of government consumption over GDP. The effect of a larger government is shown to be of significant negative influence on economic growth. Another, more recent paper by Barro provides evidence for the negative influence of government consumption on GDP growth too (2001).
2.2.3 Domestic investment
Domestic investment is an indicator of economic growth used and discussed in more than one of the papers mentioned in the previous paragraphs. The Borensztein et al. (1998) paper uses the measuring technique for domestic investment constructed by Barro and Lee (1994). Their conclusions about domestic investment are not straightforward. In line with the FDI variable, an interaction term for domestic investment and human capital is introduced. Its coefficient turns out to be negative but insignificant. An explanation provided for this tells us that while FDI flows merely to sectors of technological innovation, the domestic investment is implemented in more traditional sectors. That could be a reason why the interaction effect between domestic investment and human capital is not large enough to be statistically significant. Traditional sectors do not require as much level of schooling as the technologically innovative ones. This shows a connection in the literature on the indicators of economic growth of which domestic investment is considered to be one.
In their attempt to use and adapt the Borensztein et al. model to the specific case of China, Berthélemy and Demurger find a negative but nonsignificant coefficient (2000). However, the Barro and Lee study concludes economic growth to be positively depending on the ratio of domestic investment to GDP (1994). Because of this ambiguity in results, the effects domestic investment will be investigated in this thesis.
3. Methodology
In this section the empirical research will be discussed and explained. First off, the origin of the data used for the variables and the measuring methods are discussed. Afterwards, the adaptations necessary to use the model provided by Borensztein et al. (1998) for China are showed and discussed.
3.1 Data
The dataset for this study is constructed by data from several institutes. The National Bureau of Statistics of China supplies data for GDP growth, FDI, human capital, government consumption and domestic investment. The data for export is used from the General Administration of Customs of People’s Republic of China and found on Datastream. The values of initial GDP per capita are from Oxford Economics and also obtained via Datastream.
9 The research period is to a certain extent determined by the availability of data. FDI data was not available before 1999 and after 2014 data was unavailable for some other variables used. To create enough data points in this period, quarterly data is used. For the first research period (1999-2009) the total of observations is 44 and for the second period (1999-2014) it is 64.
In order to make the dataset useable for this research, some adaptations had to be made. The GDP growth is measured as the year on year percentage growth between the same quarter in a year. The data used for FDI are described as “value of foreign direct investment actually utilized” and are expressed in millions of U.S. dollars. The data was available on monthly basis only, so it is summed per three months to turn in into quarterly observations. Human capital is defined following Zhang (2001): the share of secondary school enrolments in the total population. This was necessary since the data required for the Borensztein method (the initial-year level of average years of the male secondary schooling) was not available (1998). Data provided by the National Bureau of Statistics of China concerns the enrolments of new students in senior secondary schools in 10000 persons. Since enrolment in education mostly takes place on one certain moment in a year, the ratio can expected to stay equal over the four quarters. A possible drawback is the population growth over the year, so this number is different from the case where quarterly population data is available. Government
consumption data are expressed in billions of Chinese Yuan and were available on quarterly basis at Oxford Economics. For domestic investment monthly investment actually completed in fixed assets is used, provided by the Chinese Statistical Bureau expressed in 100 million Yuan. This is added as a control variable as was done in the research by Berthélemy and Demurger (2000) since data on gross domestic investment as constructed by Borensztein et al. (1998) and Barro and Lee (1994) was not available. Another control variable added is export because of its positive significant effect on GDP growth showed by Park and Prime (1997). Export is measured in 100 million U.S. dollar. The last variable is initial GDP per capita which is measured in USD.
To make the dataset consistent, all variables expressed in US dollars are converted to Chinese Yuan via historical exchange rates and expressed in 100 million. Following Borensztein et al. (1998) FDI, government consumption and domestic investment are divided by GDP to make them
meaningful and put them into perspective. This has also been executed regarding the export variable. Finally, to correct for the non-stationarity of the initial GDP per capita variable, the Hodrick-Prescott filter(HP-filter) is applied. This is elaborated on in section 3.2.
An important note to all data obtained from the National Bureau of Statistics of China, as holds for all data, involves reliability. As stated in literature (Graham and Wada, 2001) the Statistical Bureau tends to publish higher FDI numbers than statistical institutes in the United States and Japan do. A possible explanation for this can be that some U.S. and Japanese companies, both substantial investors in China (UNCTAD 2015), invest in China via subsidiaries in Hong Kong. In the U.S. and Japan these investments therefore do not count as FDI in China. Other possible explanations are also possible, this is the most likely according to Graham and Wada (2001).
10
3.2 Model specification
To investigate the effects of FDI on economic growth in China, an adapted version of the model provided by Borensztein et al. (1998) will be used. In the original study, the model was used as a general regression framework in a panel data study for 69 different countries. Therefore the model is adapted and extended with variables relevant for the specific case of China, according to studied literature. The basic framework of Borensztein (1998) is given by the following equation:
g = c0 + c1FDI + c2FDI*H + c3H + c4Y0 + A
in which g represents the GDP growth, FDI is foreign direct investment over GDP, H is the stock of human capital, Y0 is the initial GDP per capita and A is a set of control variables that influence economic growth. These control variables differ per country and include for example dummies for Sub Saharan or Asian countries that are not relevant in this study. Therefore, A is left out and replaced by control variables that do show relevance for the case of China. The first variable added is export because of its significant and positive effects in China (Park and Prime, 1997).
One of the variables covered by A was government consumption. This variable remains included in the model since the effects of government consumption on GDP growth are shown to be significantly negative in Borensztein et al. (1998), Barro (1989), Barro and Lee (1994) and Barro (2001).
Another control variable added to the model is domestic investment. Berthémely and Demurger used it in their adaptation of the Borensztein framework and found a non-significant coefficient (2000). However, Barro and Lee did find significant evidence for its influence on
economic growth (1994). To make sure this research is not suffering from an omitted variable bias the variable DOMI is added. The above adaptations result in the following regression function that will be used in the analysis:
g = c0+ c1FDIt + c2FDIt*Ht + c3Ht + c4Y0t+ c5Et + c6CCONSt + c7DOMIt
To tackle the endogeneity problem that this regression model may suffer from, lagged FDI is used instead of current period value, as in Borensztein et al. (1998). The current FDI can affect GDP growth, but can be reversely influenced by the GDP as well. One may think that a faster growing economy attracts more FDI than a stagnating economy. This reverse causality leads to an endogeneity problem, and using lagged FDI may mitigate this problem. One can also expect that the economic situation not immediately changes after FDI is made. As suggested by Berthélemy and Demurger (2000), when foreign investors consider the possibility of FDI in China, they use reference data from the year before. With this knowledge, they introduce a lag in the FDI variable of one year, which decision will be followed in this thesis.
11 Another modification to be made is on the level of GDP per capita. To have reliable results by running OLS, all variables in this regression equation need to be stationary. The level of GDP per capita exhibits a trend, which is a sign of non-stationarity. So a correction must be made to it. Simply leaving out the GDP per capita leads to the omitted variable problem. A better approach is to apply a HP-filter to this series, which disentangle a time series into a stationary cyclical part and a trend. Here, a HP-filter with a smooth parameter of 1600 is applied to the natural logarithm of GDP per capita and the cyclical part is used in this regression.
Finally, two different ways of measuring GDP growth are tested in the data analysis. Since almost all variables on quarterly base, it seems to make sense to use quarter on quarter GDP growth numbers for the dependent variable. However, all fourth quarter GDP observations used to construct this growth variable are larger than the first quarter observations of the next year. This causes al first quarter growth numbers to be negative which can result in skewed output results. Therefore, also year on year growth numbers between the same quarters in a year are used as dependent variable in another regression.
4. Empirical results
The results of the OLS estimations with a lagged FDI and FDI interaction term and a HP-filtered GDP per capita are displayed in Table 1 and 2 below.
Table 1 - OLS output 2000 – 2009: FDI and interaction term lagged
Note 1. Variables used: gdpgrowth: year on year GDP growth, fdi: FDI over GDP percentage, fdih: FDI over GDP percentage multiplied by human capital over population percentage, h: human capital over population percentage, hp_percap_ln: hp-filtererd natural logarithm of GDP per capita, export: export over GDP percentage, gcons: government consumption over GDP percentage, domi: domestic investment over GDP percentage
Note 2. Because of the lagged effect of FDI introduced, four observations are lost. This reduces the periods to 2000-2009 and 2000-2014 _cons -54.16202 50.9587 -1.06 0.296 -157.9615 49.63747 domi -.2523223 .0486148 -5.19 0.000 -.3513474 -.1532972 gcons -2.81988 .8349541 -3.38 0.002 -4.520626 -1.119134 export 1.165637 .2142788 5.44 0.000 .7291659 1.602109 hp_percap_ln 32.71281 9.76763 3.35 0.002 12.8168 52.60882 h -67.58143 43.36119 -1.56 0.129 -155.9053 20.74242 fdihlag 86.19849 23.54179 3.66 0.001 38.24543 134.1516 fdilag -52.52112 14.32761 -3.67 0.001 -81.70551 -23.33673 gdpgrowth Coef. Std. Err. t P>|t| [95% Conf. Interval] Robust Root MSE = 2.934 R-squared = 0.6819 Prob > F = 0.0000 F( 7, 32) = 18.54 Linear regression Number of obs = 40
12 Table 2 - OLS output 2000-2014: FDI and interaction term lagged
When the results in Table 1 and Table 2 are compared, one can notice that the direct effect of the lagged FDI variable is significant in both cases. This is in line with the findings of Borensztein et al. (1998). However, the negative effect in the second regression shows a decrease in comparison with the first regression. Although the coefficients are significant, firm conclusions about the effects of FDI come with great care. There is a possibility that the analysis suffered from reversed causality problems, which are difficult to correct for. In Borensztein et al. (1998), FDI was considered exogenous, but Berthémely and Demurger conclude FDI to be also dependent on economic growth (2000). This can be explained by the following. Foreign funded firms contribute positively to the increase of technological progress, which is seen as a key determinant of GDP growth (Berthémely and Demurger, 2000). Technological progress on its turn, together with a human capital threshold, is a prerequisite for a country to profit from FDI, according to Borensztein et al. (1998). This suggests economic growth and FDI inflow could be reversely related and that could have done harm to the analysis output, in which a non-reversed effect is assumed.
The negative human capital coefficient shows a rather remarkable insignificant value. The most straightforward reaction to this is thinking an increase in the number of enrolments in secondary education has a negative effect on GDP growth. However, this seems highly unlikely and is in contradiction with the findings of the literature (Borensztein et al., 1998, Barro, 2001, Zhang, 2001). A possible explanation could be in the construction of the variable. Because of the unavailability of data on a more frequent basis, annual data was used under the assumption that it was equal for the four quarters in a year. This involves a lack of variance in observations that could be a problem here. The population is assumed to grow over time and the number of enrollments could change over year due to school dropout. Therefore the ratio of these two loses accuracy compared to the Borensztein et al. method (1998). In the second regression the coefficient is still negative but it turns out to be significant. In an attempt for a more meaningful human capital variable, lagged effects of one, two and three years implemented. One can assume it takes some time for a school student to find a job and
_cons 1.02288 15.90978 0.06 0.949 -30.90243 32.94819 domi -.0870449 .0526975 -1.65 0.105 -.1927902 .0187004 gcons -.46742 .4692927 -1.00 0.324 -1.409125 .4742847 export .9474011 .1758933 5.39 0.000 .5944453 1.300357 hp_percap_ln 9.970562 3.779294 2.64 0.011 2.386854 17.55427 h -65.65304 23.1833 -2.83 0.007 -112.1737 -19.13235 fdihlag 13.32419 5.135535 2.59 0.012 3.018982 23.62939 fdilag -7.798388 2.893179 -2.70 0.009 -13.60398 -1.992801 gdpgrowth Coef. Std. Err. t P>|t| [95% Conf. Interval] Robust Root MSE = 2.946 R-squared = 0.6826 Prob > F = 0.0000 F( 7, 52) = 29.09 Linear regression Number of obs = 60
13 contribute significantly to GDP growth. This unfortunately did not show any improvement. This could make one think the secondary schooling enrollment ratio is not the best indicator for human capital and better can be replaced by for example the ratio of secondary school graduates over population, or university students over population. As a stand-alone variable human capital does not contribute to economic growth as expected. However, when it is included in the interaction term with FDI, its effects seem to make more sense. These are discussed in the next paragraph.
One conclusion that came about in the papers of Borensztein et al. (1998) and Berthélemy and Demurger (2000) can be assumed to be still holding. For a certain level of FDI the effectiveness of human capital is certain, since the coefficient of the interaction term is both positive and significant. The coefficient of FDI itself is negative and only if it interacts with human capital, its contributions are positive. From this can be concluded that a certain threshold of human capital is needed to make it cause a positive effect. An explanation for this could be familiar to the one in the discussed literature. The foreign invested companies often bring considerable amounts of technological know-how to the country in which they are settled (Balasubramanyam et al., 1996, Borensztein et al., 1998). For China to profit from these technologies, the relevant workforce needs to be educated to a certain degree to be able to adopt and understand them (Berthélemy and Demurger 2000). With this conclusion the
findings of Borensztein et al. (1998) on the interaction effect are supported to a certain extent. Also a new insight is obtained. The coefficient of the interaction term is smaller in the second regression which implies the interaction effects on GDP growth have decreased. This provides some proof for the expectation of the GDP growth decline being influenced by FDI; in years of high economic growth levels its contributions are larger and more significant than in the recent years with decreasing economic growth. This effect only holds under the assumption of the presence of a minimum
threshold of human capital. A possible explanation for this can be the reverse causality effect as discussed before; a county tends to attract more FDI when economic growth is higher (Berthémely and Demurger, 2000).
The initial GDP per capita shows an unexpected sign. For both periods it has a large coefficient and is significantly positive. The problem here could be related to the human capital variable, since GDP per capita was available on quarterly basis. Therefore it did not have to be constructed from annual data. This resulted in an unavoidable use of inconsistent population numbers. For GDP per capita more accurate population numbers are used than for human capital over
population, due to availability.
The export coefficient is positive and significant and shows a more realistic value which is in line with the research of Park and Prime (1997). Government consumption’s effects are significant and negative, as are the coefficients of domestic investment. The first effect is supported by Park and Prime (1997) and the second by Borensztein (1998) and Barro and Lee (1994). These last three variables seem to follow more or less the expected course, which can be seen as a confirmation of their role of useful control variables.
14
5. Conclusions
This thesis started with a brief introduction of the Chinese economic situation since the opening-up policy in 1978. The high GDP growth numbers and high levels of FDI inflow were and still are a widely studied topic in economic science. The recent decline of economic growth formed the base for the leading research question of this thesis: To what extent have the effects of foreign direct
investment on Chinese GDP growth changed when more recent data is investigated than in existing studies?
To form sufficient background, literature on the indicators of economic growth in general and in the specific case of China was studied. Also more specified papers on the effects of FDI on GDP growth in both cases were reviewed.
In this thesis, an attempt is made to find any changes in the effects of FDI in China by two OLS regressions. One for the period 1999-2009 in which GDP growth in China increased almost every year, and one on the period 1999-2014 in which the added five years show declining growth numbers. This was done by using an adapted version of the regression framework provided by Borensztein et al. (1998). These adaptations were necessary to make it suitable for China. The adaptations were made following other papers on the effects of FDI in China.
The OLS estimation on quarterly data of the most important indicators of Chinese GDP growth had some expected and some unexpected outcomes. The unexpected outcomes were possibly due to endogeneity effects and a possible reversed causal relation between FDI and GDP growth. Both problems seemed hard to correct for. One conclusion supported more by the outcomes is the significant change of the interaction effect of FDI and human capital between the two investigated periods. This is in line with the regression outcomes of Borensztein et al. (1998) whose research played a major role in this thesis. The interaction term shows a decline in its positive effect, which could have had a contribution in the decline in GDP growth. The main finding in literature is confirmed; to let the effects of FDI be beneficial for economic growth a threshold level of human capital has to be present.
Although the effort of this thesis was to make it as complete as possible, this topic definitely requires more extended and advanced research. For example, a panel data study on the Chinese provinces could help avoiding possible small sample bias problems, since it is easier to create more observations. This also contributes to the identification of provincial differences, since their economic situations are subject to various influences. To decrease the chances of omitted variable bias, the research could be extended with more control variables, such as inflation, exchange rate, the number of natural disasters per period and a black market premium as in Borensztein et al. (1998). Due to the unavailability of accurate data this unfortunately was not possible. Also the procedure of calculating human capital could be improved when better data on secondary schooling is available on a more regular basis. Overall, the disparities in FDI numbers and other data should be investigated in high detail to assess their accuracy and trustworthiness.
15
6. References
Balasubramanyam, V. N., Salisu, M., & Sapsford, D. (1996). Foreign direct investment and growth in EP and IS countries. The economic journal, 92-105.
Barro, R. J. (1989). Economic growth in a cross section of countries (No. w3120). National Bureau of Economic Research.
Barro, R. J., & Lee, J. W. (1993). International comparisons of educational attainment. Journal of
monetary economics, 32(3), 363-394.
Barro, R. J., & Lee, J. W. (1994). Sources of economic growth. InCarnegie-Rochester conference
series on public policy (Vol. 40, pp. 1-46). North-Holland.
Barro, R. J. (2001). Human capital and growth. American Economic Review, 12-17.
Berthélemy, J. C., & Demurger, S. (2000). Foreign direct investment and economic growth: theory and application to China. Review of development economics, 4(2), 140-155.
Borensztein, E., De Gregorio, J., & Lee, J. W. (1998). How does foreign direct investment affect economic growth?. Journal of international Economics, 45(1), 115-135.
Graham, E. M., & Wada, E. (2001). Foreign direct investment in China: effects on growth and economic performance. Institute for International Economics Working Paper, (01-03).
Park, J. H., & Prime, P. B. (1997). Export performance and growth in China: A cross-provincial analysis. Applied Economics, 29(10), 1353-1363.
United Nations Conference on Trade and Development. (2015). World Investment Report 2015: Foreign direct investment and the challenge of development. Geneva
Whalley, J., & Xian, X. (2010). China's FDI and non-FDI economies and the sustainability of future high Chinese growth. China Economic Review, 21(1), 123-135.
Zhang, K. H. (2001). How does foreign direct investment affect economic growth in China
