Determinants of the Location of Foreign Direct Investment in China

We are ready to represent the best custom paper writing assistance that can cope with any task like Determinants of the Location of Foreign Direct Investment in China even at the eleventh hour. The matter is that we posses the greatest base of expert writers. Our staff of freelance writers includes approximately 300 experienced writers are at your disposal all year round. They are striving to provide the best ever services to the most desperate students that have already lost the hope for academic success. We offer the range of the most widely required, however, not recommended for college use papers. It is advisable to use our examples like Determinants of the Location of Foreign Direct Investment in China in learning at public-education level. Get prepared and be smart with our best essay samples cheap and fast! Get in touch and we will write excellent custom coursework or essay especially for you.

Determinants of the Location of Foreign Direct Investment in China.


Dingshan WAN


CCER 01 101607


1





¢ñ.INTRODUCTION


Undoubtedly large inflow of foreign direct investment (FDI) has greatly contributed to the rapid economic growth in China since 180¡¯s. From the advent of reform in 178 to 15 China has received 18.1 billion U.S. dollars in FDI. Recent FDI inflows to China account for 40 percent of such flows to all developing countries combined. In fact, China emerged as the largest recipient of FDI among developing countries beginning in 1, and has been the second largest recipient in the world since 1.


While the gross volume of FDI is much impressive, however, potential problems exist. The pattern of FDI in China is highly geographically concentrated. Of the total amount of FDI that China has received since 18, the coastal areas¡¯ share has been over 0 percent. In contrast, the inland provinces, which are considerably less developed and poorer, and in greater need of capital investment, have not played host to FDI to any significant degree. These huge differences in regional distribution of FDI, in return, might be partly responsible for the increasing gap in economic performance among various areas in China. Consequently, it will make sense to find out those significant determinants of regional distribution of FDI in China. This is what the current paper is exactly aimed to.


¢ò. LITERATURE REVIEW


There has been large literature on determinants of geographical distribution of FDI overseas, especially on inter-country distribution. In contrast, research in determinants


of regional distribution of FDI in one certain country is rather less. As to this topic in China, there is even much less literature. Moreover, the majority of previous research is qualitative but not quantitative. According to my knowledge, Broadman&Sun (17), Head&Ries (16), Chen (16), Kwan&Cheng (000) overseas, and Lu (17) and Sun (001) domestically, have discussed this problem.


Broadman and Sun (17) empirically analyzed the geographic determinants of FDI in China. Their econometric model is much simple. The dependent variable employed is the accumulated stock of FDI in each province at year-end 1. The explanatory variables are gross national product, labor costs, human capital, infrastructure, and geographical location. All data are cross-sectional at year-end 1, no matter stock or flow. All the reported coefficient estimates bear the expected sign and are statistically significant at the 10 percent or above confidence level (except the intercept). However, such analysis using cross-sectional data might be problematic in some way, considering that province-specific effects unobserved will create correlation between explanatory variables and disturbances, since we can¡¯t eliminate them by difference.


According to Head and Ries (16), Chinas Open Door policy has created a natural experiment for studying agglomeration externalities and the role of incentives designed to attract foreign direct investment. They built a model that predicts that foreign firms will prefer cities where other foreign firms are located. They estimate the model using data on 1 foreign ventures. They then use simulations to explore the effect policies favoring particular cities had on the distribution of investment.


They find that attractive cities-those with good infrastructure and an established industrial base-gained most and that agglomeration effects greatly magnified the direct impact of policy.


Chen (16) examines how the locational choice of foreign direct investment (FDI) is influenced by regional characteristics in Mainland China, such as the potential for market share extension, labor cost differences, allocative efficiency, transportation infrastructure, and research and development capability. The conditional logit model using pooled cross-section and time-series data conducts empirical testing. Empirical findings for the 187-11 period indicate that the variable for market share extension potential only affects FDI in the middle region. Surprisingly, labor cost differences do not affect the location of FDI. Interregional railroad connections are found to be positively related to the choice of location of FDI, but FDI may not necessarily locate near innovative Chinese industries.


Contrast to the above, Kwan and Cheng (000)¡¯s method is much different. They first built a partial stock adjustment model, which naturally reduced to a dynamic econometric model. Then they estimated this model using panel data from 185 to 15. In their model, the stock of FDI in region j at time t is dependent variable. This is different from the current paper. They find that large regional market, good infrastructure, and preferential policy had a positive effect but wage cost had a negative effect on FDI. The effect of education was positive but not statistically significant. In addition, there was also a strong self-reinforcing effect of FDI on itself. There was no convergence in the equilibrium FDI stocks of the regions between 185


4


and 15, but there was convergence in the deviations from the equilibrium FDI stocks.


Lu (17) made research on this topic too. Different from all the other authors, Lu substituted relative wage rate for wage rate level as an explanatory variable, since he had come to the same unexpected conclusion as others by using wage rate level. He then found that relative wage rate is negatively related to FDI and the result is significant statistically. This is what¡¯s new in Lu¡¯s paper.


Sun (001)¡¯s analysis also used panel data, from 185 to 1. But his econometric model is much simple than Kwan and Cheng¡¯s, since he didn¡¯t take the dynamic effect into account. He first conducted a Hausman¡¯s test to decide which model should be adopted, fixed effects or random effects. Then he estimated the fixed effects model by LSDV methods. He also made a Chow-test to check if there are structural changes before and after 1. The result is statistically insignificant. According to Sun¡¯s analysis, there is clear evidence that factors as government policy, industrial structure, openness and the level of marketization have played important roles to affect the location of FDI in China. The effect of government¡¯s preferential policy has been very strong all the time, but industrial structure has also been playing an increasingly significant role.


From the above, we can see that most of previous studies have arrived at the same conclusion on the effects of some factors, such as market size, preferential policy from government, etc. Though there is difference between various studies, all these estimates are consistent with economic theory in direction. On the other way, on the


5


effects of some factors, disagreement is more popular than agreement, such as wage rate, human capital, etc. The huge divergence might be attributed to varied data, model, and methods.


¢ó. ECONOMIC THEORY AND ECONOMETRIC MODEL


According to economic theory, expected net income of investment in varied areas decides the regional choice made by one certain investor, which can be divided into two parts, expected benefits and costs. Expected benefits from one certain investment are correlated to such factors as market size, etc. Costs are mainly from inputs, such as labor, and energy, etc., and from transactions, such as transportation and communication, etc. Also, institutional factors as development of service industry and market have influence on transaction costs.


In addition, I suppose there might be self-enhancing effects in FDI, which should probably be the case for some reasons. First, foreign corporations usually geographically concentrated in some areas in one province. In these areas foreigners are more densely distributed than other areas. Cultural considerations might be a determinant when a new foreign investor makes a decision. Second, due to uncertainty about the political environments in China, potential foreign investors may hesitate on their decision. Increasing FDI itself will be a good signal for those hesitating.


So, our analysis will point to the importance of seven sets of variables. Conceptually, they are market size, wage costs, human capital, development of service industry,


6


availability and quality of infrastructure, development of market and economies of agglomeration. In view of the difficulty of measurement, some proxy variables have been used. GDP and per capita GDP have been supposed to express market size alternatively. Average annual wage of employers denotes wage costs. Illiteracy rate is used as proxy for human capital. Development of service industry could be denoted by the weight of the third industry in the whole national economy. Measurement for infrastructure is rather difficulty since there are so many kinds of infrastructure, including transportation and communication. As a result, to accurately denote the whole scope of infrastructure in one single variable is nearly impossible and can never bring unanimity. For convenience, we have selected three variables as proxies for infrastructure. They are railway in operation, road and high-grade roads, averaged by corresponding geographical area of that region. Development of market exists only conceptually. So a proxy is necessary. We suppose the development of private economy is an appropriate substitute for it in a transition economy like China. We have calculated the weight of private components of the whole economy for each region in each year.


Let be the stock of FDI in region at time t and itYiiy¦�be the flow of FDI in region at time i¦�. Then,


1titiYy¦�¦�==¡Æ (1.1)


Collecting the above-mentioned explanatory variables in a vector , we postulate a panel formulation for , itXity


1itititiityXYu¦Â¦Ã¦ÁW=+++ (1.)


7


where ¦Â is a vector of parameters; ¦Ã is parameter for 1itYW. i¦Á is unobserved, region-specific effects; and is a random disturbance. That is, itui¦Á captures time-invariant, regional effects such as geographic location and culture. Equation (1.1) is a dynamic model. For such model, we can¡¯t estimate it by OLS or other methods directly for the unobservability of i¦Á. So, we should first transform the original equation into the first-differenced version. It is,


111()(ititititititityyXXyuu¦Â¦ÃWWWW=W++W1)W (1.)


to estimate the above equation, an instrumental variable is needed to ensure the consistency of the estimate, since 1ityWityWW is correlated to the differenced residual . According to related literature, what are usually utilized as instrument for in such models are or 1itituuWW1ityWityWityW, e.g., y lagged two periods or the first difference of y lagged two periods and y lagged three periods. Compared with the later, the former demands one-period information less than the later.1 So we will use the former as instruments here. Under the assumption of serially uncorrelated level residual, values of y lagged two periods qualify as instruments in the first-differenced system, since


,1()itititCovyuuWWW=01,()itityyWW¡Ù Cov0 (1.4)


However, there remain potential methodological problems. The IV estimates based on (1.4) may be highly inefficient since it has omitted much useful information. To solve this problem, more instrument variables should be found out and then we can use


1 For a comparison of asymptotic efficiency of the two instruments, see Anderson and Hsiao (181). Also refer to Hsiao (186, Chapter 4).


8


GMM method. But we will not take this efficiency problem into account here.


¢�. DATA AND ESTIMATION PROCEDURE


DATA


We use panel data for 8 provinces, administrated municipals or self-governed districts from 186 to 001, except for Hainan, Chongqing, Xizang (Tibet). All data are from the Chinese Statistical Yearbook, Chinese Labor Statistical Yearbook and Chinese Foreign Trade Statistical Yearbook for relevant years.


As to the data, we have much thing to clarify that may create problems in an unknowable way. There are some typical kinds of problems. First, some variables in the yearbooks are not conceptually consistent. For example, there are no records for GDP before 10. The substitute is national income. Before 10 the records for production value of the third industry is also absent. We have to use the sum of architecture, transportation and business as its substitute. Second, some variables are absent for most of the whole periods. In China, the illiteracy rate is not recorder annually. The census is conducted every ten years or even longer. Except that, only sampling data is reported in each single year. Even though this data is not reported in every year during the period of our interest. We only get reported data of the forth census in 10 and the fifth census in 000. Also we have get two years data from sampling, e.g., the one percentage sampling survey of population in 187 and 15 respectively. For the other years, we have get the corresponding values for illiteracy


For detailed exposition, refer to Holtz-Eakin et al. (188), Arellano and Bond (11), Ahn an Schmidt (15,17), Arellano and Bover (15),and more recently, Blundell and Bond(18).


rate in each year by interpolation.For further illustration for all explanatory variables, please refer to appendix A.


When conducting estimation, considering that IV is consistent with the two-stage OLS method, we have first regressed 1ityW on ityW. Then we used the predicted value for as substitutes for 1ityW1ityW to regress.


¢õ. ESTIMATION RESULTS


Six sets of regression results using different sets of explanatory variables have been reported in Table1. Much to our surprise, most of the explanatory variables are not statistically significant, and part of estimates doesn¡¯t have the expected sign.


As can be seen in Table 1, the coefficient for lagged FDI stock is significant at five-percentage level when GDP is adopted to be explanatory variable, however its sign is contrary to our intuition. The three coefficients are much approximate to each other, indicating that an increase of one thousand U.S.D in a region¡¯s lagged would tend to reduce its FDI by about thirty-three U.S.D. Moreover, the coefficient assumes opposite sign when per capita GDP is instead taken as explanatory variable, but none of the three is statistically significant at ten-percentage level or lower.


As to the coefficient for average annul wage rate, the case is reversed. In case of regression with per capita GDP the coefficients are positive and statistically significant at five-percentage level. The estimates all indicate about an increase of more than sixty thousand U.S.D in FDI if there is an increase of one Yuan in wage


The formulation is1111()iktiktikiktxxxxttt¦�¦�W=+WW.


10


costs. However, when GDP proxies for market size the coefficient for it takes negative sign and statistically insignificant at ten-percentage level or lower.


Table 1 Estimation results4


(1) () () (4) (5) (6) Lagged FDI stock -0.075 (-.6) -0.06 (-.5) -0.075 (-.6) 0.016


(1.6) 0.016


(1.45) 0.0178


(1.) Wage -1.651


(-0.70) -.7110


(-0.7) -1.651


(-0.70) 6.140 (.56) 6.1


(.1) 6.0


(.0) Service industry 41.076


(0.8) 45.054


(0.8) 41.076


(0.8) 6.8568


(1.1) 61.6878


(1.11) 6.667


(0.6) Illiterate rate 75.685


(0.1) 76.6


(0.1) 75.685


(0.1) 57.540


(0.1) 4.0786


(0.10) -10.8577


(-0.0) GDP 66.51


(7.0) 65.46


(7.17) 66.51


(7.0) Per capita GDP 0.45


(1.1) 0.001


(1.1) 0.85


(1.0) Railway 7.8571


(1.0) 65.6


(0.65) Roads 5.00


(0.8) 84.6044


(1.) High-grade road 7.8571


(1.0) 7.65


(1.6) Private sector 58.48 (.) 601.8774 (.40) 58.48 (.) 78.66 (4.1) 77.81841 (4.0) 74.5415


(4.0)


All regression results denote that FDI is highly positively related to the development of service industry, indicating that foreign investors might pay much importance to this factor when making a investment decision. Unfortunately, all the six estimates are statistically insignificant at ten-percentage level or lower.


4 t-statistics in parentheses. and denote statistically significant at one-percentage level and five-percentage level, respectively.


11


Five of the six coefficients for illiteracy rate are positive, varied from forty to seventy or more, and indicate that FDI will greatly increase if illiteracy is higher. This is much contrary to our expectation too. We suppose that foreign investors might have paid little attention to the quality of human capital in most of the past periods, since the majority of foreign investments until now has concentrated on those labor-intensive and export-oriented industries, such as weaving, shoes and the appliances, etc. All estimates are statistically insignificant at ten-percentage level or lower.


In contrast, all the coefficients for FDI are positive and statistically significant at one-percentage level. When GDP increases by one billion Yuan FDI will increase by more than six hundred thousand U.S.D. Keep in mind this is a very strong effect. So foreign investor have attached great importance to market size of one region. When makes a locational decision he is aimed to take possession of the regional market. However, the coefficients for per capita GDP are not statistically significant at ten-percentage level or lower, indicating that GDP as a whole has more attractions for foreign investors than per capita GDP.


Using the density of railway in operation, all roads and high-grade as proxies for infrastructure respectively, the reported coefficients are not statistically significant, though the sign is consistent with our intuition.


According to Table 1, the development of market in a region is highly positively correlated to FDI in that region. All estimates are statically significant at one-percentage level, though the exact value has many variations.


1


¢ö. CONCLUDING REMARKS


By taking FDI stock lagged one year into account, we have build a dynamic econometric model on decision of FDI in one region. To estimate such a model, we first difference the original model and try to estimate it by IV. The IV we have employed is dependent variable lagged two years, which is supposed to satisfy the two necessary conditions for IV under the assumption of serially uncorrelated level residuals. There have some potential problems with our estimation method, however. The estimates should be inefficient because IV has omitted much useful information. To solve this problem, we should found out more instruments and then we can use GMM method, which is supposed to be more efficient.


Due to cause of method and data, our estimation results is contrary to our intuition or expectation to much extent. Only small part of the coefficients of explanatory variable is statistically significant. According to six sets of regression results, lagged FDI stock and GDP are significant explanatory variables at five-percentage or one-percentage level while FDI enters into regression equation. However, the coefficient for the former doesn¡¯t assume expected sign. On the other hand, the coefficient for the former doesn¡¯t consistent with estimation result when per capita GDP as a substitute for GDP. The coefficient for development of market, weight of private sector as a proxy, is always positive and statistically significant at one-percentage level. Then, what we can conclude is that when making a locational decision, a foreign investor will pay much attention to the development of market. In areas where private sector has assumed important role in economic activities, FDI seems to increase at more rapid


1


rate. For foreign investors, excellent performance in marketization may be a rather instructive signal for more efficient government, more transparent policies and looser business environment, etc.


There is much to improve on this current paper, especially methodologically.


14


References


Broadman, Harry G. and Xiaolun Sun, ¡°The distribution of foreign investment in China,¡± World Bank, Policy Research Working Paper 170, 17.


Chen, Chien-Hsun, 16, ¡°Regional determinants of foreign direct investment in Mainland China,¡± Journal of Economic Studies, vol. , no. , pp. 18-0.


Greene, William H., Economic Analysis, Third edition, Prentice-Hall, 17.


Head, Keith and John Ries, ¡°Inter-City Competition for Foreign Investment Static and Dynamic Effects of Chinas Incentive Areas,¡± Journal of Urban Economics, 16, vol. 40, pp. 8-60.


Hsiao, Cheng, Analysis of Panel Data, New York Cambridge University Press, 186.


Kwan, Yum K. and Leonard K. Cheng, ¡°What are the determinants of the location of foreign direct investment? The Chinese experience,¡± Journal of International Economics, 000, vol. 51, pp. 7-400.


³Ã÷ãü£¬¡°Íâ¹úÖ±½�Ͷ×ÊÇø�ò·Ö²¼�ÚÖйúͶ×Ê�·¾³ÆÀ¹À¡±£¬¡¶¾­¼Ã�о¿¡·£¬£¬17µÚ1ÆÚ¡£


Ëï¿¡£¬¡°ÖйúFDIµØµã�¡�ñµÄ�òËØ·ÖÎö¡±£¬¡¶¾­¼Ã�§£¨¼¾¿¯£�¡·£¬00ÄêµÚ1¾íµÚÆÚ£¬ µÚ687-68�³¡£


¹ú¼�ͳ¼Æ¾Ö£¬¡¶Öйúͳ¼ÆÄê¼ø¡·£¬186-00Äê¡£


¹ú¼�ͳ¼Æ¾Ö£¬¡¶Öйú¶�Íâó�×ͳ¼ÆÄê¼ø¡·£¬184-185Äê¡£


Appendix A. Definition of variables


1. FDI Inflow of FDI in one year, measured by ten thousand nominal U.S.D . GDP /Per capita GDP Nominal GDP or per capita GDP for one region in one year. . Wage Average nominal annual wage of employer. 4. Service industry The ratio of the third industry to the whole economy. 5. Illiteracy rate Percentage of population with no schooling years 6. Railway Length of railway in operation averaged by geographical area of one region () /mkm


7. All roads Length of all roads averaged by geographical area of one region (). /mkm


8. High-grade road. Length of high-grade road averaged by geographical area of one region (). /mkm


. Private sector Percentage of the economy that is owned by private.


15


16


Mind that the sample papers like Determinants of the Location of Foreign Direct Investment in China presented are to be used for review only. In order to warn you and eliminate any plagiarism writing intentions, it is highly recommended not to use the essays in class. In cases you experience difficulties with essay writing in class and for in class use, order original papers with our expert writers. Cheap custom papers can be written from scratch for each customer that entrusts his or her academic success to our writing team. Order your unique assignment from the best custom writing services cheap and fast!