The regional differences of receiving
FDI in China
Ruan Jiajia (s1621335)
University of Groningen
Faculty of Economics
Thesis for master IE&B
ABSTRACT
1 INTRODUCTION
Foreign Direct investment （FDI）in short is the investment of foreign assets into domestic structures, equipment, and organization but dose not include foreign financial investment. Different from the investment in portfolio, which can be divested easily and does not have significant influence on the management of the firm, FDI to a country is durable with the objective of establishing a ‘lasting interest’ in the country, focuses on a long-term relationship and will have a significant influence on the management of the firm (Navaretti and Venables，(2004 )).
China has started an ‘open door’ policy to encourage FDI since the 1980s as a strategy to stimulate domestic competition and as a way to import the latest technology and managerial experience. Nowadays, FDI is already an important part of the Chinese economy and will help develop and contribute to Chinese economic growth as well as the profit that investors obtain in the Chinese market.
However, behind the promising development of FDI in China, we can not ignore the important phenomenon of the regional differences in Chinese receiving FDI. If we divide China into three regions—East, Middle and West, it is very obvious that FDI in China shows a basic structure of “High in East, Low in Middle and West’. For example, during the period of 1991—1996, China received 155900 million dollars totally, in which East China received 135669 million dollars, occupied 88.17% of all the Chinese receiving FDI. On the contrary, Middle China and West China received 13170 million dollars and 5034.4 million dollars, which occupied 8.56% and 3.27% respectively. The absolute difference between East China and Middle China in a sum increased to 161.64 billion dollars till the end of 2001.
Table 1 Chinese economic indicator in three regions (2005)
Region East Middle West Population (10000 persons) 46388 35202 35976 Area (1000 sq. km) 91.6 102.8 686.7 GDP (100 million Yuan) 109924.6 37230.3 33493.3 Primary Industry 8681.8 6204.6 5924.6 Secondary Industry 56673.2 17412.7 14331.6 Tertiary Industry 44569.7 13613.1 13237.1 Per Capita GDP (Yuan) 23768 10608 9338 Employment At the Year-end 24810 19065 19448 Unemployment Rate 3.0 3.8 4.1 Import & Export 12781.6 415.1 451.3 Exports 6798.0 244.2 257.6 Imports 5983.5 170.9 193.8 FDI 11597 1039 870 Source: Chinese Statistics Yearbook (2006)
No matter East, Middle or West of China, they are all the regions in China, but what makes FDI in these regions so different from each other? The answer is complex, but there must be some unique factor specific to each region. As there are so many different characteristics in Chinese different regions, as shown in table 1, I wonder which factors are the determinants of the regional differences of receiving FDI in China and how the FDI location is influenced by regional characteristics. As a result, the purpose of this paper is to find the causes of regional differences of receiving FDI in China.
research to find what the ultimate causes of the differences are. Therefore, in my thesis, I will develop the relevant fields of inquiry ( Determinants of FDI and regional differences of receiving FDI) into the determinants of FDI in Chinese regions.
2 LITERATURE AND HYPOTHESES
When considering China as a whole, there is abundance literature studying why China can receive so much FDI. The study named “Main determinants and impacts of foreign direct investment on China’s economy” published from OECD(2000) pointed out that the main determinants of FDI in China are size and growth prospects of the Chinese economy, human resource endowments-cost and productivity of labor, infrastructure, and openness to international trade and so on.
The relationship between economic growth and FDI is almost the first thing when we consider about the FDI in China. To see how FDI flows are driven by host economic growth, it is necessary to distinguish two types of FDI based on its motives. The first one is the horizontal FDI, also named market-seeking FDI which is in order to access host countries’ markets more successfully by efficient utilization off resources and exploitation of economies of scale (Markusen et al., 1996). And another one is vertical FDI, also named export-oriented which is motivated by factor-cost saving (such as cheap labor in host countries) along with human capital and infrastructure conditions ( Zhang and Markusen, 1999). The general implication is other things being equal, a country’s market size (usually measured by GDP) rises with economic growth, encouraging foreign firms to increase their investment. Moreover, better economic performances in host countries provide a better infrastructural facilities and greater opportunities for making profits, and so greater incentive for FDI. Even for export-oriental FDI which is prevalent in China, the market size is also important because lager economies can provide larger economies of scale and spill-over effect.
was conducted by Parantap, Basu, Chakraborty and Reagle (2003).
However, the situation when we consider the country’s perspective and the regional perspective should be different. Using the factor method, Mi (2006) finds that the factors affecting the regional efficiency of FDI allocation can be grouped in three views. For the country as a whole, the market and public capital play a major role compared with technology and cost of labor. As for the regions, the cost of labor, the location and technology are more important than the market. In his research, the regional differences of GDP seem not to play a substantial role in attracting FDI in Chinese regions.
Another very important thing when we consider Chinese receiving FDI is the absolute abundant human resource endowments and the low cost of labor in China. Helpman(1984) postulated that the formation of MNEs is driven by factor endowment differences. The geographical separation of high-skilled labor intensive headquarter services and low-skilled labor intensive production activities, leads to cost savings for MNEs Especially the vertical FDI can be observed in countries sufficiently abundant in low-skilled labor.
In the Knowledge Capital Model (KCM), developed by Markusen, et al. (1996), FDI is driven by both factor costs and market access and, thus, the KCM model incorporates both vertical and horizontal FDI. Vertical FDI is found in two areas in which the relative endowments are very different; while horizontal FDI is found in two areas in which both the endowments and country size are similar.
There are also many examples in reality revealing the relationship between factor endowments and FDI, such as the surge of North-South regional investments in the 1990s, between the US and Mexico or between the EU and Central and Eastern European countries, all confirm this prediction.
their costs are minimized. Finally, factor costs are assumed to be determined by factor endowments, meaning that skilled labor would be relatively cheap in skilled-labor abundant countries, and vice versa.
On the other hand, many studies, such as Kravis & Lipsey (1982); Wheeler & Mody (1992); Braunerhjelm (1994); Hatzius(1997) had included measures of labor cost differences as a source of attracting FDI, but not all of them find the positive relationship between labor cost differentials and FDI, while some even find opposite relationship. With the world’s largest population, China has rich resources of labor, with average salaries of workers remaining at a relatively low level.
From the regional perspective, on one side, Mi (2006) pointed that when we do research for the regions, the cost of labor plays relatively important role. On the other hand, Ting (2005) pointed out that within the country’s border, a very important point which should be considered in China is the labor quality. Ting (2005) has examined the effect of labor quality on the location of foreign direct investment in China and found that labor quality plays a significant and positive role in attracting FDI. In his research, quality is more important than pure cost advantage or the abundance of labor.
A similar conception of labor quality should be labor skills and there is also evidence to support that labor skills have a positive effect, see Ekholm (1997) and Carr et al. (2001).
perspective, which is opposite to the substitution theory of Mundell, suggests that foreign investment will increase trade and under certain condition, trade and FDI will mutually encourage and strengthen each other. With regard to the Chinese regions, the second theory may work better because inward FDI in China is deemed to be more export oriented. The annual report of the Ministry of Commerce shows that foreign enterprises account for 63.3% of the total trade surplus in the year 2005.
Considering the relationship between FDI and trade openness from the regional perspective, many authors such as Qiu, Tang and Sun (2006) and Cheng and Kwan (1998) believe that this relationship also exists from the regional perspective in China. Their explanation is that East was the first region to carry out the open door policy and this region receives most of the FDI in China.
However, Chengdu office (2007) which is a organization located in Chengdu city and does the economic research of China has the different point. It argues that during these years, the openness policy is pursued in West China; Nevertheless, the per capita receiving FDI in this region is still low. In that situation, trade openness seems not to be a determinant of FDI. Another example in China is provided by some cities such as Lian Yungang in the Jiangsu province which started the trade openness policy very early but still has a low level of receiving FDI.
There is little literature available about the regional differences FDI in China. Chen (1996) has examined how the location choice of FDI is influenced by regional characteristics in mainland China. In his paper, he found that the variable for market share is not important and labor cost differences do not affect the location of FDI. On the contrary, interregional railroad links which is an index of infrastructure is found to be positively related to the choice of location of FDI.
Cheng and Kwan (1998) have estimated the effects of the determinants of FDI in 29 Chinese regions from 1985 and 1995 and found that a large regional market, good infrastructure, and preferential policy such as tax refund policy had a positive effect and wage cost had a negative effect on FDI. They pointed to another important determinant of FDI from regional perspective, namely agglomeration.
FDI from a regional perspective, the political stability, infrastructure of the region, capital stock and the economies of agglomeration are important.
From this literature, we can identify two determinants, infrastructure and economies of agglomeration, that are important when we consider FDI from the Chinese regional perspective. These determinants seem to be the special characteristics which influence the receiving FDI in Chinese regions.
It can be easily understood that the availability of physical infrastructure such as the traffic convenience, communication convenience and so on affect the decision of selecting the place of FDI. The conveniences of infrastructure can reduce the total costs of investment which will attract FDI.
Table 2 The determinants of receiving FDI in China
Perspective Regional (relationship) Market Size + Cost of labor － Productivity of labor + Quality of labor + Trade openness + Infrastructure + Economies of agglomeration +
(“+” represents there is positive relationship between the independent variables and FDI;
“－”represents there is negative relationship between the independent variables and FDI)
Another important concept here is agglomeration here which was first mentioned by Wheeler and Mody (1992). Proximity to other firms may also play a role in location of FDI because it can provide benefits though knowledge spillovers, larger markets for specialized factors, and forward and backward linkages between customer and supplier firms.
regional perspective can be summarized in Table 2.
As shown in Table 2, there are 7 hypotheses in my paper:
Hypothesis 1: Market size is positively related to FDI.
As show in the literature review, there are 2 different streams about the relationship between market size and FDI. So, the impact of market size of investing country is not easy to predict. China has a big market size which attracts more and more foreign investors who want to sell their commodities in China. For example, in 2005, Chinese GDP was 2.29 trillion Dollars which ranked the 4th in the world.
Although the main type of FDI in China is still the vertical FDI, also named export-oriented FDI, we can not ignore that China is attracting more and more foreign investors who want to sell their commodities in China, that is, do the market-seeking FDI. As supported in Table 1, the market sizes which are measured by GDP have a positive relationship with FDI in different Chinese regions.
Similar research by Cheng and Kwan (1998) found that a large regional market, good infrastructure, and preferential policy had a positive effect.
On the other hand, in Mi (2006)’s research, he pointed out that as for the regions, the cost of labor, the location and technology are more important than the market size. But he still found the positive relationship between GDP and FDI in provinces of Middle China.
Hypothesis 2: Labor abundance is positively related FDI.
If we divide China into regions, then this theory should also apply to these regions, which means the greater the low-skilled labor endowment, the more the FDI a region should receive. As a result, a have positive relationship between the labor abundance and FDI are expected when the vertical FDI can be observed in regions where low-skilled labor is sufficiently abundant. The dominant FDI in China is still vertical FDI and the positive relationship is expected to exist in different regions of China.
Hypothesis 3: Wage rate is negatively related to FDI.
As shown in the literature part, there are three parts to the labor cost theory. The second and third one can apply to China because low-skilled labor would be relatively cheap in China which is a low-skilled labor abundant country. And MNEs can reduce their cost if they locate downstream activities in China. In regions of China, the wage differences exist, although the wage rate’s differences in regions are not very large compared to the differences between two countries. In order to minimize the cost, foreign investors will choose the regions in China with the lowest labor cost.
Meanwhile, the similar research by Cheng and Kwan (1998) found labor cost has a negative effect. And Mi (2006) also pointed that when we do research for the regions in China, the cost of labor plays a relative important role.
However, there is always a negative relationship between labor abundance and labor cost, which is the more the abundance of labor, the cheaper the labor. But the analysis of Braconier, Norbackand and Urban (2004)shows that relative factor costs and relative factor endowments are significantly, but only weakly, correlated.
Hypothesis 4: Labor quality is positively related to FDI.
Both the skilled labor and the quality theory both suggest that labor quality is also an important determinant attracting FDI.
China. So, both the low-skilled labor and high-skilled labor are the determinants of FDI in Chinese regions these years.
Some similar research such as Gao (2005) pointed out that the labor quality plays a significant and positive role in attracting FDI. In his research, quality is more important than pure labor cost advantage or the abundance of labors.
Hypothesis 5: Trade openness is positively related to FDI.
As show in the literature review, with regard to the type of FDI in China, the second theory-- the mutual complement theory may work better because inward FDI in China is deemed to be more export oriented. That means, in China, trade is the complement of FDI. The area which is more trade openness that is more convenience to import or export commodities, the more vertical VFDI it will receive. The trade openness level in these 3 regions of China is so different because of many reasons such as the location, the convenience of transportation, promotion policy and so on which finally cause the regional differences of receiving FDI in China.
There is some evidence such as the example of the city Lian Yungang which shows that trade openness is not always important when we consider FDI in regions. But in my opinion, this is only an example which is not common in these regions.
Hypothesis 6: Infrastructure is positively related to FDI.
It can be easily understood that the availability of physical infrastructure affects the decision of selecting the place of FDI. The more developed of infrastructure in a region is, the lower the other costs (except the labor cost) such as the communication cost, the transportation cost, then the more FDI will inflow in this region. Many authors such as Leonard K. Cheng and Yum K. Kwan (1998), Chien-Hsun Chen (1996) and the Chengdu office (2007) all pointed out that there is a positive relationship between a good infrastructure and FDI.
.
Hypothesis 7: Economies of agglomeration is positively related to FDI.
role in the location of FDI because it can give rise to benefits though knowledge spillovers, thick markets for specialized factors, or forward and backward linkages between customer and supplier firms. In that situation, the more economies of agglomeration in a region, the more FDI received in this region.
In East China, many industries such as computer assembly, and textile weaving are agglomerated in this region. Foreign investors prefer to establish the same industrial activities near these established enterprises to benefit from the existence of positive externalities.
3 METHODOLOGY
Description of variables
In this paper, I use information about the regional choices of foreign direct investment. We extend the regional perspective into 28 provinces of the 3 main regions in order to get more exact result.
The economic model of my analysis is:
it it it it it it it it it c X X X X X X X y = + 1β1+ 2β2 + 3β3+ 4β4 + 5β5 + 6β6 + 7β7 +ε
Y: the annual FDI in each province;
X1: the annual market size in each province;
X2: the annual labor quantity in each province;
X3: the annual labor cost in each province;
X4: the annual quality in each province;
X5: the annual trade openness in each province;
X6: the annual infrastructure in each province;
X7: the annual economies of agglomeration in each province;
Dependent variable. The dependent variable Y is the annual FDI in each province,
and as I showed in the hypotheses part, seven independent variables can be used to describe the regional characteristics.
Independent variables. The independent variables are market size X1, labor
economies of agglomeration X7. The measurement of these variables is discussed as
follows:
Market size. The most common method to measure market size is GDP. Many
authors, such as Asiedu (2000), Chowdhury and Mavrotas (2003) and Basu et al.( 2003）all used GDP as the measure of market size in their research.
Therefore, in my paper, I will choose GDP as the measure of regional market size. The data of each province’s annual GDP will be obtained from the China Statistical Yearbook directly.
In this paper, market size is assumed to be positively related to FDI in regions of China.
Labor abundance (quantity). The measurement of labor abundance is indirect in
my paper. First I will collect the data of the number of staff and workers at the year-end of region from China Statistical Yearbook. But the number of staff and workers is not equal to the labor abundance in a region because the un-employers should be a part of labor supply in this year. Therefore, next, the yearly employment rate of each province yearly can be derived from the Chinese Statistical Yearbook. Then, the labor supply=the number of staff and workers/(1-unemployment rate). The different regional areas in these provinces should be taken into consideration. Generally, one province will have a larger population if it is bigger than another one. That means one province may have a bigger population or more workers only because it has bigger area. As a result, the density of the labor supply will be a more exact measurement here. And, the density of the labor supply=labor supply/area of a region. In this paper, the labor abundance is assumed to be positively related to FDI in regions of China.
Labor cost. A motivation for FDI is to lower the production costs through the
utilization of low-cost factors of production in the host country (Cushman, (1987)). The main cause of why China can receive so much FDI is that China offers low-cost labor in comparison to many other countries.
Labor cost will be measured by average annual wage per staff and workers by
to the average wage in money terms per person during a certain period of time for staff and workers in enterprise, institutions, and government agencies. It reflects the general level of wage income during a certain period of time. It is calculated as follows:
average wage=total wages of staff and workers at reference time/average number of staff and workers at reference time.
In this paper, labor cost is predicted to be negatively related to FDI in regions of China.
Labor quality. This paper focuses on the average labor quality at the regional
level. The China Statistical Yearbook publishes information on the composition of employed persons by education level and province. There are four labor quality measures according to the different levels of education: the percentage of workers who have at least primary school, junior secondary school, senior secondary school, and college education, respectively.
Labor quality is measured as the percentage of people who have at least college education. The reason is that in China, there is a policy called “9-years’ compulsory education” which is in order to ensure that every person can finish the education of primary school (6 years) and junior secondary (3 years). In that situation, the differences of the percentage of who at least has junior secondary school will not be very different per region. Thus, the percentage of the population who have at least
college education school is preferred to measure labor quality. The number of people
who have at least a college education can be obtained from the China Statistical Yearbook. The calculation is:
The percentage of the population who at least have a college education=the people who at least have the college education/population.
In this paper, labor quality is assumed to be positive related to the FDI in regions.
Trade openness. One standard measure of trade openness is defined as the sum of
exports and imports divided by GDP (Liberati, 2007). Another is defined as exports as a percentage of GDP.
seeking type which implies a substitution relationship between trade and FDI (Mundell, (1957)) and the other is export-oriented FDI which suggests that reasonable foreign investment will increase trade and under certain condition, trade and FDI will mutually encourage and strengthen each other (Markuson and Svensson (1985), Kiyoshi, (1987)).
In this thesis, trade openness is expected to be positively related to FDI in regions of China because most FDI in China is export-oriented FDI in which FDI is a complement of trade. Thus, it’s better to focus on the exports of regions instead of both exports and imports. Therefore, the measure of trade openness in this paper is the
exports and the data can be collected from China Statistical Yearbook directly.
Infrastructure. The Commission on Engineering and Technical Systems (1996)
pointed out that as a multifunctional system, infrastructure provides a range of specific services that differ substantially from one region to another (e.g., transportation, telecommunication, power generation). But judging infrastructure performance is really a complex matter.
As a result, how to measure infrastructure here in this paper is also a complex matter. One frequently mentioned consideration is local accessibility to regional and national markets via transport linkages (Bartik, (1985) (1989), Helms, (1985)). Overall, in this paper, I will use the density of the lengths of highways, railways and
navigable inland waterways to measure the infrastructure.
Here, the length of highways refers to the length of highways which are built in conformity with the grades specified by the highway engineering standard formulated by the Ministry of Communications and have been formally checked and accepted by the departments of highways and put into use. Length of railways refers to the total length of the trunk line for passenger and freight transportation. The length of navigable inland waterways is an indicator reflecting the size and development of inland water network. It refers to the length of the natural rivers, lakes, reservoirs, and canals open to navigation during a given period, and which enables transport by ships (the Chinese Statistical Yearbook 2006, p.671)
compose the transport linkages in a region which reflects the convenience of the transportation in this region.
However, as was the case with the measure of the labor abundance, I should also consider the different regional areas in constructing this variable.
Therefore, the measurement is used the density of the lengths of highways, railways and navigable inland waterways which equals the total lengths of highways, railways and navigable inland waterways divided by the area of each province.
In this paper, infrastructure is expected to be positively related to FDI in regions of China.
Economies of agglomeration. Two potential agglomeration phenomena will be
employed in this paper. One is international agglomeration which arises due to foreign investment and the other is local agglomeration which is due to domestic investment. A number of variables have been used in previous research to measure agglomeration economies. Head and Ries (1996) used industrial output and count the number of industrial enterprises to represent agglomeration economies. This paper will use the number of foreign enterprises in each province as the indicator of international agglomeration. With regard to the local agglomeration, the number of Chinese domestic enterprises in each province is adopted. Industrial output is inappropriate as a measurement here because it has a high co-linearity relationship with GDP which is the first independent variable in this paper.
Therefore, the measurement of this dependent variable will be separated into two components which are the number of domestic enterprises in each province and the number of foreign enterprises in each province. In addition, I will also consider the different areas in these provinces. Therefore, the chosen measurement of agglomeration will be the density of domestic enterprises and the density of
international enterprises with both equal to the number of relevant enterprises divided
by the area of each province.
Sample and data source
of China’s 3 big regions—East, Middle and West, and cover 9-year period from 1997 to 2005.
I selected this period for several reasons. First, Chongqing has been an independent province from 1997. The period starting from 1997 can make the observations consistent. Second, the regional composing of receiving FDI in China changes during this period. Although Middle and West China still represent a small part of receiving FDI, the percentage rises during these years. The reason is that during this period, in 2001, China joined into WTO. Consequently, FDI is expected to increase, with some independent variables increasing too, and others changing only a bit. With this time effect, we can get more exact result.
As my thesis focuses on FDI in regions of China, almost all of the data, including FDI data and other independent data such as GDP, exports and so on are collected from the Chinese Statistical Yearbooks of individual provinces. Since the Chinese Statistical Yearbook publishes information on its website from 1997, the information I need in my paper is available for the period 1997-2005.
Therefore, the data for the independent variables can be gotten from the Chinese Statistical Yearbooks directly by some calculation after obtaining the raw data as explained above.
The methodology
There are 3 approaches in general to deal with panel data, namely the pooled model, the random effect model and the fixed effect model.
The biggest difference between the former model and the later two models is that the pooled model assumes that the errors are homoscedastic. In other words, it can be expressed in the following formula:
it
X
y = β+ε ,
approach as the classic linear model with OLS.
In Chen’s (1996) paper “Regional Determinants of Foreign Direct Investment in Mainland China”, the author used the pooled model. The advantage of using this model is that it is the simplest way to deal with cross-section and time data. However in the case of this paper, the pooled model seems not to be appropriate because by assuming each observation is iid, we have essentially ignored the panel structure of the data, where the errors term are inferred to be heteroscedastic. Heteroscedastic here refers to the unobserved variables, which means, the error terms are not iid, but change across the individuals. For example, location can be taken as an unobserved variable which change across provinces. There is a lot of empirical evidence that the shorter the distance between origin and destination of FDI, the higher the level of FDI. As an example, Hong Kong which accounts for nearly half of the total FDI in China has most of its FDI in Guangdong province just because Guangdong province is adjacent to Hong Kong and shares the similar dialect and local customs. Thus, a well-located province seems like attracting more actual FDI than provinces with relatively disadvantaged locations. Therefore, it’s reasonable to assume that the errors are heteroskedastic here (as location, local customs etc,)
However, at this stage, this is only an inference. In order to decide if the pooled estimator is appropriate here, a homoscedastic test on OLS model should be performed. Therefore, the White Heteroskedasticity test in Eviews will be used, which tests for the existance of heteroskedasticity.
If the White Heteroskedaticity test tell us that the panel data are not poolable, I turn to the other two approaches which assume the existence of heteroskedastic errors. The distinction between the fixed effects model and the random effects model shows as: Fixed effects regression is the model to use when you want to control for omitted variables that differ between cases but are constant over time; Random effects regression is the model to use when some omitted variables may be constant over time but vary between cases, and others may be fixed between cases but vary over time (Gelbach, 2005).
further and specify the error structure for the disturbance term: εit =αi +ηit
where I assume that η is uncorrelated with Xit, and the first term of the
decomposition, αi, is called an individual effect. The other two models can be
distinguished in terms of the correlation between αi and independent variables X. If there is no correlation between the two (cov(Xit, αi)=0), then the random effect
model is consistent and efficient, if not a fixed effect model is preferred.
Here, the selection of the two models should be done with care and precision, because the random effect model is usually upward-biased as a result of the omission of the fixed effect. However, the fixed effect model helps get rid of or rule out the fixed effect by either assigning dummy variable for each individual or by transforming all the variables.
The choice between these two estimators depends on the correlation between
ηand the explanatory variables. Specially, if the effects are uncorrelated with the explanatory variables, the random effects estimator is consistent and efficient. The fixed effects estimator is consistent but not efficient; if the effects are correlated with the explanatory variable, the fixed effects estimator is consistent and efficient but the random effects estimator is now inconsistent (Johnston and Dinardo, (1997)).
In order to decide which model is appropriate, random effect or fixed effect, a WU-Hausman test is employed to test our data to see which model fits the data set better.
The WU-Hausman test is a test with the purpose of testing for model mis-specification. Typically, it tests a null hypothesis of the form: H0 : θ = θ0, where
θ0 is some specified value that may or may not be taken by the parameter θ, against
the alternative hypothesis that θ takes some other value. When the null is violated, we have a clear way to define why it is violated: because θ takes on some other value (Hausman, (1978)).
estimates, one of which is consistent under both the null and the alternative hypothesis and another which is consistent only under the null hypothesis. Here, the null hypothesis under a Hausman test is that the random effects estimator is correct. Or in other words, whether the omission of fixed effects in the random effects model has any effect on the consistency of the random effect estimates. If the result of the test is significant, that is, if the omission of the fixed effect has a significant effect on the consistency of the estimates, then the null hypothesis is rejected and it’s better to use the fixed effect model, and vice verse.
Regardless of model chosen depending on the result of the Hausman test, the model can be expressed in the following equation:
yit =c+β Xit +β Xit +β Xit +β Xit +β Xit +β Xit +β Xit +β Xit +εit 8 8 7 7 6 6 5 5 4 4 3 3 2 2 1 1 and, εit =αi +ηit,
where i (i=1, 2, 3,……n=28) stands for individual, here in the case the province; t (t=1, 2, 3,…..n=9) stands for the number of years involved in the data set; Xitn stands
for the seven independent variables with 28 individuals across 9 years for each one including market size X1, labor quantity X2, wage cost X3, labor quality X4, trade
openness X5, infrastructure X6, economies of local agglomeration X7 and economies
of international agglomeration X8 (the measurements or operationalization of those
variables have been discussed in the previous paragraph); αi (one part of the error
term) stands for the individual effects; ηit stands for the error (ηit~iid(0,σ
2)); and y it
stands for regional level of FDI. This procedure will be carried out by the application of the statistic software of Eveiws after the correct model is selected.
Because all the hypotheses involve a one-tailed t-test, the process to test the hypotheses is:
Step 1: H0: ß=0, H1: ß<0 (hypothesis 3) or ß>0 Step 2: Use Eviews to test the hypotheses.
null hypothesis.
Step 4: In the case of the p-value approach, reject H0 if p-value < 0.05, the level of significance.
If the p-value of the coefficient between explanatory variable and FDI is insignificant, the hypotheses will be rejected; otherwise, if the p-value of the coefficient between the explanatory variable and FDI is significant, the hypotheses will be accepted.
In addition, the adjusted R square of the model will also be evaluated to test whether the model provides a strong explanation of the dependent variable. If not, more control variables are expected to be taken into model to get a stronger model.
4 THE RESULT
The result of estimating the random effect model is shown in Appendix 1. We can not sure if this result is acceptable before doing the Hausman test. Therefore, the next step is to run Hausman test in Eviews to see if the result of random effect model is desirable. The result of Hausman test is shown in Appendix 2 and in this table, the p-value is 0.00. It means the omission of fixed effect has significant effect on the consistency of the estimates. So, the null hypothesis, here, the random effect model, is rejected and it’s better to use fixed effect model.
The result of estimating the fixed effect model is shown in Appendix 3 and it is the estimator used for the final result of my research.
The results show that there are three variables named labor abundance, labor cost
and labor quality whose results are not significant. In that situation, I can obtain the conclusion that the relationships between FDI and labor abundance, FDI and labor cost, FDI and labor quality do not exist.
On the other hand, the left 5 variables all get a significant results at 05.05 level which means the other 5 assumptions are proved in my test.
Another indication is the Adjusted R square of the model. It is 0.985878 which is
5 DISCUSSION AND CONCLUSION
Discussion 1: Relationship between FDI and GDP
The positive relationship between FDI and GDP which is the measurement of market size also exists and is shown to be important among Chinese regions.
As a whole country, one of the most important reason for China to attract FDI is its rapid growing economy. In the world, most FDI happens because of market-seeking, so, we can not deny that horizontal FDI should also exist in China, although the dominant FDI in China is still vertical FDI. And the positive relationship between FDI and GDP in the test supported this point.
These years, because of the vigorous growing economy of China, the market size in China becomes bigger and bigger. In order to enlarge the market share, it seems reasonable for foreign investors to conduct more horizontal FDI in China. It’s obvious that these companies invest into China to focus on China’s huge market instead of Chinese abundant labor.
And if we divide China into regions, the story is the same. It is reasonable that this FDI will flow into the bigger market as its purpose is to seek market. Eastern China with a high GDP is the most developed region in China. In this region, people have a high level of consumption because of the better economy, so the market-seeking purpose is easier to realize in this region.
similar FDI and GDP, so it is not surprising that the relationship dose not exist in each region. But in my paper, I try to find the relationship among these 3 regions using 28 provinces data and the relationship is found because the GDP and FDI are so different among these 3 regions.
Another issue which should be mentioned here is the direction of causality between the two variables. Many authors did research on this issue, such as Hansen and Rand (2004) and Chowdhury and Mavrotas (2003) who both found the bi-directional causality between the two variables. But in my paper, I only consider the positive relationship between FDI and GDP in Chinese regions, but can not state which causes which.
Discussion 2: Relationship between FDI and labor abundance
The positive relationship between FDI and workers which is the measurement of labor abundance dose not exist in Chinese regions.
As shown in literature review part, the geographical separation of high-skilled labor intensive headquarters services and low-skilled labor intensive production activities, leads to cost savings for MNEs. Especially, the vertical FDI can be observed in locations with sufficiently abundant low-skilled labor. So, it’s not surprising to find the positive relationship which exists between FDI and labor quantity when we consider China as a whole country. But the story is not the same when we divide China into different regions. The reasons are the following:
First, no matter in which province of China FDI takes place, the labor is abundant compared to the developed countries although we can not deny the fact that labor is not distributed evenly in every region in China. In that situation, foreign investors will not consider labor abundance as a determinant of FDI in regions of China, because no matter which region they choose, the number of labor is extremely high compared to their countries.
leads to cost savings for MNEs. But if the migration is easy between locations, the labor cost differences will not exist, or at least will not be as large as it would otherwise be, because investors can employ foreign labors as many as they want until the labor abundance is similar in every country. On the other hand, the situation is dissimilar among regions in China. China is a single country and the population can migrate easily among different provinces in China. Suppose there are two provinces, one of which has a higher labor demand than labor supply and another province has the opposite situation. As in a single country, the labors will migrate from the province with higher labor supply to another one with higher labor demand. Therefore, foreign investors will not pay attention to the labor abundance of a province even the labor in this province is not abundant.
Discussion 3: Relationship between FDI and wage
The result shows that the relationship between labor cost and FDI in regions of China is insignificant so I have to reject this assumption and accept the result that there is no relationship between these two variables.
There are several reasons to explain this result. First, it may be because of the measurement of labor cost. As I showed in the methodology part, the measurement of labor cost in my paper is the annual wage per worker in each province. If we examine attention to the data, we can find that the wage increased every year during the period 1997-2005. For example, in 1997, the average wage per person in Beijing is 11,019, but in 2005, it is 34,191 which is triple that of 1997. The same happened in every province. On the other hand, the growth rate of FDI is not as fast as the growth in wages. In my test, I used the nominal wage of every year without getting rid of the time effect which may explain the result.
FDI. Other authors such as Chen. (1996) who have done similar research on regions of China, also found that cost differences do not affect the location of FDI in China. Therefore, it is obvious that as a whole, low labor cost is a predominant reason of Chinese huge receiving FDI, but the low labor cost no longer seems a determinant of the regional differences of FDI. Because just like the absolute high labor abundances in all provinces in China, the wages in all provinces are absolutely low comparing to the other countries that they will not care the slight differences of wages among different regions. This point can also be explained by the migration among provinces. Furthermore, labor cost is not the only cost. Besides it, foreign investors should also consider other costs such as transportation cost, management cost and so on. In the West of China, although the wage is lower than in Eastern China, it is short of efficient management, good transportation and so on which will mean other high costs for foreign investors. In this situation, foreign investors prefer to invest in the area with a bit of higher wage but more convenient infrastructure under which situation overall cost will be less totally.
Discussion 4: Relationship between FDI and labor quality
The result shows that there is no relationship between FDI and labor quality used in Chinese regions.
wage, it can attract many workers from the “floating population” with a high education level. In conclusion, this result is also explained by the migration of population among different provinces in China.
Discussion 5: Relationship between FDI and trade openness
The positive relationship between FDI and trade openness also exists among Chinese regions. As I showed in the literature part, there are two opposite theories with respect to the impact of trade openness, namely the substitution theory of trade and the complement theory of trade. The first one is used for horizontal FDI while the other one is suitable for vertical FDI.
The relationship existing between FDI and trade openness in Chinese regions indicates that vertical FDI is still the dominant type of FDI in China.
The low openness of Western and Middle China is a direct barrier for attracting FDI in these regions. From the provinces’ perspective, Guangdong which is one of the Eastern provinces, has exports which are well higher than other provinces. For example, in 2005, its export amounted to 1974068 million RMB which was more than 33% of the total export in China in 2005. The next largest exporting provinces are Jiangsu (17.1%), Shanghai (11.9%), Zhejiang (11.2%), Shandong (6.5%), Fujian (4.9%), Tianjin (3.6%), Beijing (2.5%), Hebei (1.65%). These Eastern coastal provinces accounted for 92.47% of the total exports in 2005 because of their superior location. More directly, form the regions’ perspective, in 2005, as shown in Table 1, Eastern China, Middle China and Western China had exported of 679800, 24420 and 25760 Million USD separately, and they received 1159700, 103900 and 87000 USD of FDI separately. It’s very obvious that with a higher value of exports, Eastern China got a relatively higher value of FDI.
Discussion 6: Relationship between FDI and infrastructure
The positive relationship between infrastructure and FDI indicates that in a single country, the convenience of infrastructure in a region plays an important role to attract FDI.
As I showed, foreign investors choose to invest in a region (especially the vertical FDI) in order to cut cost, which dose not only include labor cost, but also many other costs such as the efficiency of infrastructure. Most western provinces in China (except Tibet) have a relatively low labor cost but have very underdeveloped infrastructure such as transportation because of its location and topography. If commerce is attracted to these provinces, while the labor cost is saving, foreign investors will pay more to transport this commerce, and this may not cover the labor cost saving. Therefore, foreign investors prefer to invest in the area with a bit of higher wage or other costs, but more efficient infrastructure under which situation total cost will be less totally.
However, transportation is only one of the aspects of infrastructure. There are several aspects of physical infrastructure which complement each other, such as telecommunication, transport, information or energy availability (Kumar (2001)). In my paper, I just used one of the measurement, transportation, which is the most important one in Chinese regions. Maybe making the indicator more complicated, by incorporating all of the aspects mentioned, will provide better results. This is a suggestion for further research.
Discussion 7: Relationship between FDI and economies of agglomeration.
The last two hypotheses which assumed the positive relationship between economies of agglomeration as shown to exist in Chinese region. Both the economies of local agglomeration and the economies of international agglomeration are important in Chinese regions.
especially in Western China which has low GDP, relatively poor location, and a low value of exports. It is obvious that it is not easy to increase GDP, or the value of exports in a short time, or change the location of western provinces, but it is much easier to enhance the role of economies of agglomeration to attract more FDI. As the economies of agglomeration now really exist in China, we can further agglomerate similar industries in a region, in order to increase the knowledge spillovers, extend markets for specialized factors, and lower the effective distance between suppliers and customers. Foreign investors will then prefer to choose these regions which have the similar industries.
Conclusion
The main contribution of this paper is the finding that all the characteristics of labor which include the labor abundance, labor cost and labor quality are no longer the determinants of FDI at the level of Chinese regions. As a whole, the most important determinants explaining why China receives so much FDI are the abundant labor and low cost of labor. But if we divide China into regions, the situation is different. There are two reasons. First, it is the high degree of mobility of population among different regions in China. Because migration is easy among provinces, labor will flow from the province with high labor supply and low wage to the province with high labor demand and high wage. In that situation, the labor quantity and wage would be almost similar in different regions in China. Migration also helps explain why labor quality dose not play a role in Chinese regions. It is because both high-skilled and low skilled labor can move to the region which is suitable for them mostly.
On the other hand, GDP, trade openness, good infrastructure, and the economies of agglomeration are all proved positively related to FDI in regions of China. Especially the good infrastructure and the economies of agglomeration are the unique reasons for Chinese regions attracting FDI.
In conclusion, the most important determinants of FDI to China at the national level are no longer important when we consider it at the regional level. On the other hand, good infrastructure and agglomeration economies play important roles when Chinese province receives FDI.
REFERNCE
Asiedu, E. 2000. The Role of Natural Resources, Market Size, Government Policy, Institutions and Political Instability.
(http://papers.ssrn.com/sol3/papers.cfm?abstract_id=717361)
Bartik, T. J. 1989. Small Business Start-ups in the United States: Estimates of the Effects of Characteristics of States. Southern Economic Journal, Vol. 55 pp.1004-18.
Basu, P., D.Chakraborty, & R.,Derrick 2003. Liberalization, FDI, and Growth in Developing Countries: A Panel Co-integration Approach.
(http://ideas.repec.org/a/oup/ecinqu/v41y2003i3p510-516.html)
Brainard, S. L. 1993. An Empirical Assessment of the Factor Proportions Theory.
NBER working paper. No. 4269 (1993).
Braconier,. H., P.J., Norback, & D., Urban. 2004. Multinational Enterprises and Wage Costs: Vertical FDI Revisited. Journal of International Economics, Vol. 23 pp 446-2004.
Braunerhjelm P. 1994. Regional Integration and the Locational Response of Knowledge Intensive Multinational Firms: Implications for Comparative Advantage and Welfare of Outsiders and Insiders. Dissertation, IUI, Stockholm, Sweden, and the Graduate Institute for International Studies, Geneva, Switzerland.
Chen, C. H. 1996. The Determinants of Foreign Direct Investment in Mainland China”, Journal of Economic Studies, Vol. 23 pp. 18-30(13)
China Statistical Yearbook. 2006. China Statistics Press.
Chowdhury, A., & G., Mavrotas. 2003. FDI & Growth: What Causes what.? The
World Economics, Vol. 29, No. 1 pp. 9-19
Commission on engineering and technical systems (CETS). 1996. Measuring and Improving Infrastructure Performance.
Cushman, D.O. 1987. The Effects of Real Wages and Labor Productivity on Foreign Direct Investment. Southern Economic Journal, Vol. 54 pp. 174-85
Gelbach, J.B. 2005. The Hausman Test. Econ 626 Lecture Notes.
Hansen H. & J., Rand 2004. On the Causal Links between FDI and Growth in Developing Countries. The World Economy, Vol. 29(1) pp. 21-41, 01.
Hatzius, J. 1997. Domestic Jobs and Foreign Wages: Labor Demand in Swedish Multinationals. CEP Discussion Papers Vol. 337.
Hausman, J. 1978. Speciation Tests in Econometrics. Econometrica 46(6), 1251-71.
Head K., & J., Ries 1996. Inter-city Competition for Foreign Investment: Static and Dynamic Effects of China’s Incentive Areas. J Urban Econ Vol. 40(1) pp. 38–60.
Helpman, E. 1984. Simple Theory of International Trade with Multinational Corporations. The journal of political economy, Vol. 92, No. 3, pp. 451-471.
Johnston, J., & J., Dinardo. 1997. Panel Data. Econometric Methods. Fourth edition. New York [etc.] : McGraw-Hill. pp. 388-409.
Kojima, K. 1987. A Model of Trade-oriented Direct Foreign Investment. Direct Foreign Investment and Export Promotion: Policies and Experiences in Asia (1987), pp. 159-83.
Kravis, I.B., & R.E., Lipsey. 1982. The Location of Overseas Production and Production for Exports by U.S. Multinational Firms. Journal of International
Economics, Vol. 12 pp. 201-409.
Liberati, P. 2007. Trade Openness, Capital Openness and Government. Fnl Publ. Pil, Vol. 27, 2 pp. 215-47.
Markusen, J., A., Benables, D., Konan, & K., Zhang. 1996. A Unified Treatment of Horizontal Direct Investment, Vertical Direct Investment, and the Pattern of Trade in Goods and Services. NBER Working Paper, No. 5696.
Markusen, J.R., & K.E.,Maskus. 2001. General-Equilibrium Approaches to the Multinational Firm: A Review of Theory and Evidence. NBER Working Paper Number 8334.
Markusen, J. R., & L.E.O., Svensson. 1985. Trade in Goods and Factors with International Differences in Technology. International Economic Review, Vol. 26 pp. 175-92.
(http://cms.cass.cn/show_News_e.asp?id=12087)
Mundell, R. A. 1957. International Trade and Factor Mobility. American Economic
Review, Vol. 47(3) pp. 321-35
Navaretti, G.B., & A.J., Venables. 2004. Multinational Firms in the World Economy. Princeton, N.J., [etc.]: Princeton University Press.
Organization for Economic Co-operation and Development. 2000. Main Determinants and Impacts of Foreign Direct Investment on China’s Economy. OECD: working
papers on international investment, No. 2000/4.
Qiu, B., B.Q., Tang & S.Q., Sun. 2006. A Retest of Linkages between International Trade and FDI in China. Nankai Economic Studies, No. 4 pp. 32-46, 70
Ting G. 2005. Labor Quality and the Location of Foreign Direct Investment: Evidence from China. China Economic Review, Vol. 16, Issue 3 pp. 274-92.
Wheeler, D., & A., Mody. 1992. International Investment Location Decisions: the Case of U.S. Firms. Journal of International Economics, Vol. 33 pp. 233-261.
Zhang, K., & J., Markuse. 1999. Vertical Multinational and Host-country Characteristics. Journal of Development Economics, Vol. 59, pp. 233-52.
Website:
http://www.gddpc.gov.cn/common_file/show_file.asp?id=21841&lanmu=14
http://www.china-changjiang.net/news/details.php?info_id=13561
