Briefing Series Issue 30 FOREIGN DIRECT INVESTMENT AND GROWTH DIFFERENTIALS IN THE CHINESE REGIONS Kailei WEI Shujie YAO Aying LIU Copyright China Policy Institute November 2007 China House University of Nottingham University Park Nottingham NG7 2RD United Kingdom Tel: +44 (0)115 846 7769 Fax: +44 (0)115 846 7900 Email: CPI@nottingham.ac.uk Website: www.chinapolicyinstitute.org The China Policy Institute was set up to analyse critical policy challenges faced by China in its rapid development. Its goals are to help expand the knowledge and understanding of contemporary China in Britain, Europe and worldwide, to help build a more informed dialogue between China and the UK and Europe, and to contribute to government and business strategies.
Summary High economic growth and increasing disparities between China s regions have attracted serious attention in academic as well as policy circles. The spatial income divergence trend has coincided with a skewed distribution of Foreign Direct Investment (FDI). FDI has been a key driver of economic growth in China over the past three decades, but FDI has been blamed by some researchers for the widening inequalities between Chinese regions. It is difficult to reconcile the positive effect of FDI on economic growth with its potential negative effect on equality. The empirical studies across countries on the linkage between FDI and regional inequality have produced controversial results. FDI is regarded as a powerful driver of economic growth for China, enabling it to catch up with the advanced countries in the world. Consequently, it is expected that with more FDI, the less developed regions of China such as the Western and Central provinces might be able to catch up with their richer eastern counterparts. Economic reforms over the past three decades have brought about exciting growth prospects throughout China, especially for some provinces along the eastern coast. Among many factors, the uneven distribution of resources and preferential policies given to the East are widely regarded as the dominant causes of regional inequality in China. FDI inflows into China started from a very low level in the 1980s but rose dramatically after Deng Xiaoping s famous southern tour in 1992. The patterns of growth for FDI and GDP suggest a strong correlation between them. GDP and FDI were highly concentrated in the East during the reform period. Radical economic liberalisation enabled the Eastern region to have the advantages of offering foreign investors preferential policies, and of large and rapidly expanding markets. The income gap, measured by coefficient of variation in the whole country during the period 1979 2003, experienced three phases: declining from 1979 to 1989, expanding in 1998, and declining again in 2003. At the regional level, income gaps were found to be the highest within the East and the lowest within the Central among the three macro geographical regions. The initially poorer regions have failed to grow faster than the initially richer ones due to their diversified economic backgrounds and the biased policies faced by these economies. Neither the poor provinces nor poor regions have managed to grow faster than their richer counterparts. As a result, regional inequality rose rather than declined during the data period. Setting aside the issues of investment ratios and effective population growth rates, FDI and exports are found to have significant and positive effects on regional growth differences. In addition, the effect of FDI on economic growth is weak among the western provinces. FDI is an important factor in economic growth, and it cannot be blamed for widening regional inequalities in China. It is the uneven distribution of FDI that has led to higher inequality. Consequently, containing regional inequality by restricting inflows of foreign capital into China is not advisable. China should continue to open its door to more FDI but efforts should be made to improve the spatial distribution of foreign investments. The reason why the poor regions have received less FDI than their richer counterparts is their lower absorption capability. Hence, government policy should be directed towards improving the absorption capability of the poor regions by implementing preferential policies and other incentives, such as the development of education and infrastructure. 2
Foreign Direct Investment and Growth Differentials in the Chinese Regions Kailei Wei, Shujie Yao and Aying Liu The Debate on FDI and Regional Inequality 1.1 High economic growth and increasing disparities between China s regions have attracted serious attention in academic as well as policy circles because fast economic growth over the last 30 years has been associated with widening inequalities between the coastal, central and western regions. 1 Although different studies have produced controversial findings on whether regional inequality in China has increased or declined, the general view is that China has been clearly divided into three main geo economic regions (coastal, central and west) and per capita incomes among them have diverged significantly under economic reforms. 2 The Gini coefficient that measures inequality rose from 0.3 in the mid 1980s to more than 0.45 in recent years. Up to one third of China s total inequality has been caused by inter regional inequality. 3 1.2 The spatial income divergence trend has coincided with a skewed distribution of FDI, with the coastal region receiving over 86% of FDI and the other two regions less than 14% in all the years throughout the 1978 2006 data period. As a result, many researchers have blamed FDI for causing the widening gaps in regional incomes. 4 1.3 Compared to other FDI related empirical studies of China, relatively few studies have provided a detailed assessment on FDI and regional economic inequality. However, there are some scholars who have attempted to do so. Most of the arguments in earlier studies claim that FDI leads to increased poverty, isolation, a neglect of local capabilities and a larger inequality. 5 1.4 More recently, Fu (2004) investigated the spillover and migration effects of exports and FDI and estimated their impact on regional income inequalities in China. The study found that exports and FDI played an important role in increasing regional disparities. In contrast, conceptually opposite evaluations of FDI have been proposed by other studies. 6 They show that FDI is not only growth promoting but also equality enhancing since FDI and exports were found to accelerate the convergence process in the Asian newly industrialized economies and Japan. Dollar and Kraay (2002) argued that the current wave of FDI from the 1980s promoted equality and reduced poverty. Zhang (2001) investigated the role of trade and FDI in a cross country convergence analysis, Kailei Wei, Middlesex University Business School, Hendon, London NW4 4BT, UK. Prof. Shujie Yao, Head of School, School of Contemporary Chinese Studies, University of Nottingham, Nottingham NG7 2RD, UK. Dr. Aying Liu, Senior Lecturer, Middlesex University Business School, Hendon, London NW4 4BT, UK. 1 In this paper, the coastal (eastern) region covers Beijing, Tianjin, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong (Hainan), Guangxi and Hebei. Central region includes the central region, covering Shanxi, Inner Mongolia, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei and Hunan. The western region consists of Sichuan (Chongqing), Guizhou, Yunnan, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang. 2 Yao S. and Zhang Z. (2001), On regional inequality and diverging clubs: a case study of contemporary China, Journal of Comparative Economics, 35, 211 234. 3 Yao S. and Zhang Z. (2001), Regional growth in China under economic reforms, Journal of Development Studies, 38(2), 167 186. 4 Fleisher, B.M., and Chen, Jian (1997) The coast noncoast income gap, productivity, and regional economic policy in China, Journal of Comparative Economics, 25:220 36. Fu X. (2004), Limited linkages from growth engines and regional disparities in China, Journal of Comparative Economics, 32, 148 164. 5 Mazur, J. (2000), Labor s new internationalism, Foreign Affairs, 79(1): 79 93. 6 Fu X. (2004), Limited linkages from growth engines and regional disparities in China, Journal of Comparative Economics, 32, 148 164. 3
which indicated that exports and FDI tend to accelerate the convergence process in the Asian newly industrialized economies and Japan. 7 1.5 The controversial empirical studies on the linkage between FDI and regional inequality require further research. The most recent study by Yao and Wei (2007) concludes that FDI plays a dual role in economic growth as a mover of production efficiency and a shifter of production frontiers. FDI is hence regarded as a powerful driver of economic growth for China, enabling it to catch up with the most advanced countries in the world. Consequently, it is expected that with more FDI, the less developed regions of China such as the Western and Central provinces might be able to catch up with their richer east counterparts. 8 Regional Inequalities in China after Economic Reform 2.1 Economic reforms over the past three decades have brought about exciting growth prospects throughout China, especially for some provinces along the coastal region. From 1979 to 2003, real per capita GDP increased more than eight fold, registering an average annual growth of 9.41%, while growth rates in the East, Central and West were 10.17%, 8.5% and 8.05% respectively. The provinces with the highest growth were concentrated on the eastern coast, possessing advantages of geography, resources and preferential policies. Such policies included the establishment of Special Economic Zones (SEZs) and Open Coastal Cities, as well as other incentives to attract foreign investment. At the same time, the industrialization policies for the Central and Western regions were removed. All of these have allowed the coastal region to grow much faster than the other regions in the country. Consequently, the Chinese economy experienced unprecedented rapid and steady growth with increasing interregional disparities, which particularly deteriorated after the 1990s. The ratio of East Central West per capita real GDP was 1.71:1.23:1 in 1979, 2.03:1.15:1 in 1992, and rose to 2.98:1.56:1 in 2005. 2.2 The regional disparities in China can be attributed to many factors such as different natural resources, human capital endowments, infrastructure and transportation, geographical location, proximity to foreign markets and investors, economic structures, coast oriented regional policy and foreign direct investment. Among these factors, the uneven distribution of resources and preferential policies given to the East are widely regarded as the major causes of regional inequality in China. 2.3 Figure 1 shows that the gap of real GDP per capita between the East and the inland region (the Central and West regions combined) widened dramatically from 1992. Per capita incomes between the Central and the West used to be very close but started to differ gradually in the following years as well. 7 Dollar, D. and Kraay, A. (2002), Growth is good for the poor, Journal of Economic Growth, Springer, 7(3), 195 225. Zhang Z. (2001), Trade liberalization, economic growth and convergence: Evidence from east Asian economies, Journal of Economic Integration, 16(2), 147 164. 8 Yao S. and Wei K. (2007), Economic growth in the presence of FDI: the perspective of newly industrialising economies, Journal of Comparative Economics, 29, 466 484. 4
Figure1 Real per capita GDP in comparison, 1979 2005 (yuan) 12000 RGDP per capita (yuan) 10000 8000 6000 4000 2000 0 East Central West 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 year 1997 1998 1999 2000 2001 2002 2003 2004 2005 Note: Per capita GDP is measured in 1990s prices. Sources: China Statistical Data for 50 Years 1949 98 (NBS, 1999) and Statistical Yearbook of China (NBS, 1998 2006, various issues). 2.4 The inflows of FDI into China started from a very low level in the 1980s but rose dramatically after Deng s famous southern tour in 1992. Table 1 FDI and GDP in China 1983 2005 Year FDI ($ billion) (1) FDI (RMB billion) (2) GDP (RMB billion) (3) FDI/GDP (%) (4) 1983 0.92 2.69 596.27 0.45 1984 1.42 4.17 720.81 0.58 1985 1.66 4.87 901.60 0.54 1986 1.87 6.47 1027.52 0.63 1987 2.30 8.61 1205.86 0.71 1988 3.19 11.89 1504.28 0.79 1989 3.39 12.77 16992.3 0.08 1990 3.49 16.68 1866.78 0.89 1991 4.37 23.24 2178.15 1.07 1992 11.01 60.70 2692.35 2.25 1993 27.52 158.54 3533.39 4.49 1994 33.77 291.03 4819.79 6.04 1995 37.52 313.33 6079.37 5.15 1996 41.73 346.91 7117.66 4.87 1997 45.26 375.17 7897.30 4.75 1998 45.46 376.39 8440.23 4.46 1999 40.32 333.83 8967.71 3.72 2000 40.72 337.06 9921.46 3.40 2001 46.88 388.01 10965.52 3.54 2002 52.74 436.55 12033.27 3.63 2003 53.51 442.86 13582.28 3.26 2004 60.63 501.82 15987.83 3.14 2005 60.32 494.12 18308.48 2.70 Notes: FDI in Column (1) and GDP in column (3) are measured in current price. Column (2) and (4) are calculated by the authors. Column (1) figures are converted to those in column (2) by using period average exchange rates. Column (4) equates Column (2) divides Column (3). Sources: China Statistical Data for 50 Years 1949 98 (NBS, 1999) and Statistical Yearbook of China (NBS, 1998 2006, various issues). 2.5 Before 1978, China virtually closed its door to foreign investment as a result of the Maoist ideology of self sufficiency. Since China s pursuit of reform with its 5
opening up policy in 1978, FDI has gradually blossomed. 9 FDI inflow into China increased sluggishly before 1992 and was mainly concentrated in a few coastal cities. As indicated in Table 1, FDI in China was only $0.92 billion in 1983 and grew slowly to $4.37 billion in 1991. However, FDI inflows expanded dramatically to $11 billion in 1992, rising to $60.32 billion in 2005, making China the largest recipient of FDI in the developing countries after 1996 and then the biggest in the world in 2003. Meanwhile, China s GDP expanded sharply from RMB 0.60 trillion in 1983, to RMB 2.69 trillion in 1992, RMB18.31 trillion in 2005, and over RMB 21 trillion in 2006. The patterns of growth for FDI and GDP suggest a strong correlation between them. This can be further confirmed by the ratio of FDI to GDP, as shown in column 5 in Table 1, which increased slightly from 0.45% in 1983 to 1.07% in 1991, but doubled in 1992 at 2.25% and reached a peak at 6.04% in 1994. However, this ratio gradually declined to only 2.7% in 2005. Figure 2 Shares of real GDP by region, 1979 2005 100% 80% 60% 40% West Central East 20% 0% 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Note: real GDP is measured in 1990s prices Source: 1.Comprehensive Statistical Data and Materials on 50 Years of New China, NBS, 1999 2. China Statistical Yearbook, NBS, 2000 2006 Figure 3 Shares of real FDI by region, 1979 2005 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 West Central East Note: real FDI is actually used FDI measured in 1990s prices Source: 1.Comprehensive Statistical Data and Materials on 50 years of New China, NBS, 1999 2. China Statistical Yearbook, NBS, 1985 2006 9 Zhang, X. and Zhang, K.H. (2003), How does globalization affect regional inequality within a developing country? Evidence from China, Journal of Development Studies, 39(4), 47 67. 6
2.6 GDP and FDI were highly concentrated in the East during the reform period (Figures 2 and 3). The Eastern region accounted for over 52% of China s GDP in 1979 and the share increased in the following consecutive 25 years. The West accounted for a small and declining share of China s GDP over the same period. The distribution of FDI across regions is far more skewed than that of GDP. Over 86% of China s FDI inflows were concentrated in the East. The other two regions received just 14%. 2.7 Several reasons explain this geographic polarization of FDI. First of all, the early reform was focused on the eastern provinces. Guangdong, Fujian, 14 coastal cities, Hainan and Pudong were gradually opened to foreign investors in the form of designated Special Economic Zones, Development Zones, Economic and Technology Development Zones. All of them were given preferential policies to attract foreign capital and promote exports. Since the mid 1980s, the openingup policy has been extended northward. Only in the early 1990s were inland cities and border areas encouraged to open up. In the late 1990s, the Chinese government announced a Western Development Programme, aiming to restore balance to regional development, and decided to implement preferential policies to attract more FDI to the inland areas. The Western Development Programme may have helped the West to accelerate its economic growth but failed to narrow the gaps between its growth and foreign investment, and those of the Eastern region. 2.8 In fact, preferential policies have only been one of the advantages that the Eastern region offered to foreign investors. The East also has better economic resources which give it comparative advantages over the Central and West regions: geographic proximity to international markets, better transport infrastructures, and higher skilled labour. Furthermore, many coastal provinces have advanced rapidly with economic liberalisation, developed a dynamic nonstate sector, and thus provided a more favorable environment to foreign investors. Finally, because of higher economic growth, they have also provided foreign businesses with larger and rapidly expanding markets. What Does the Empirical Evidence Say? 3.1 In the economic literature, we can use both the concepts of σ convergence and β convergence to study whether regional incomes have converged or diverged and to explain why. If they converge, regional inequality declines. If they diverge, regional inequality rises. 3.2 The concepts of σ convergence can be defined as a group of economies are converging in the sense of σ if the dispersion of their real per capita GDP levels tends to decrease over time 10. It is used to reflect the static disparities in per capita income. It can be regarded as evidence of σ convergence between China s regions if regional income disparity declines over time. It is commonly measured by the coefficient of variation (CV), which is the ratio of standard deviation to the mean. 3.3 Figure 4 presents the indices of CV at national and regional levels based on real per capita GDP. The income gap in the whole country during the 1979 2003 period experienced three phases: declining from 0.6924 to 0.6028 in 1979 89, expanding to 0.6680 in 1998, and declining again to 0.6180 in 2003. This pattern reflects the process of economic reform and policies adopted during the 10 Sala i Martain X. (1996), Regional cohesion: evidence and theories of regional growth and convergence, European Economic Review, 40, 1325 1352. Sala i Martain X. (1996), The classical approach to convergence analysis, The Economic Journal, 106, 1019 1036. 7
past decades. At the beginning, the whole country benefited from economic reform, achieving impressive economic development. Some initially poorer economies took advantage of their backwardness and performed more rapidly than some initially richer ones, leading to a contraction of income gap within the country. In the second stage, the coastal provinces benefited greatly from the preferential policies granted by the central government. For instance, the Eastern region was allowed to adopt a market system and to open its door to foreign investors before the rest of the country. Consequently, FDI largely flowed into the coastal cities, greatly accelerating export activities and local development. As a result, the income disparity between the Coastal and Inland regions began to deteriorate during this period. Over the three stages, the income gap appeared to have declined thanks to the government s Western Development Programme and the Rebuilding Programme of the Northeast Region. 11 However, the reduction in the CV in the third phase may be too little to be statistically significant and whether the income inequality between the Chinese regions really declined has to be tested using a more robust parametric approach as shown below. 3.4 At the regional level, three macro geographical regions (East, Central and West) are found to have different CVs, based on their values and trends. The East had the highest CV values and the Central the lowest, meaning that income gaps were highest within the East and lowest within the Central. The CVs have a clearly declining tendency in all regions, especially in the East. Oscillation in regional CVs implies that intra regional inequality declined, especially among the Eastern provinces. The sluggishness of the national CV and the reduced regional CVs indicate that inter regional inequality must have risen. This is consistent with the conclusion drawn by Yao and Zhang (2001a) on the formation of three geo economic clubs in China under economic reforms. 12 0.8000 0.7000 0.6000 Figure 4 Coefficient of variation in Comparison 0.5000 0.4000 0.3000 0.2000 Nation East Central West 0.1000 0.0000 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Note: CV is calculated according to equation 1. Source: 1.Comprehensive Statistical Data and Materials on 50 years of New China, NBS, 1999 2. China Statistical Yearbook, NBS, 1985 2004 3.5 The definition of absolute β convergence can be described as we say that there is absolute β convergence if poor economies tend to grow faster than rich ones. In other words, this is to test if an initially lower income group has a higher rate of income growth, and convergence is a process in which the poorer economies 11 Zheng F., Xu, L. D. and Tang B. (2000), Forecasting regional income inequality in China, European Journal of Operational Research, 124, 243 254. Chang G.H. (2002), The cause and cure of China s widening income disparity, China Economic Review, Vol. 13, 335 340. 12 Yao S. and Zhang Z. (2001a), On regional inequality and diverging clubs: a case study of contemporary China, Journal of Comparative Economics, 35, 211 234. 8
catch up with the richer ones. This argument is based on the neoclassical model, which predicts that initially poor countries will grow faster than initially rich ones if the only difference across countries lies in their initial levels of capital. 13 However, in the real world, economies may differ in other respects such as technological progress, population growth, investment, infrastructure and political stability. If these differences are considered, the neoclassical models predict that the growth of an economy will be positively related to the distance that separates it from its own steady state. This concept is known in classical literature as conditional β convergence. 3.6 In absolute convergence, initial income level is the only factor of concern and the catching up process will take place if the initially poorer economies have higher growth than the initially richer ones. 3.7 As for conditional convergence, income convergence and the catching up process can only be initiated given the presence of additional control factors, such as investment ratio, population growth, openness, FDI ratio, human capital and infrastructure, etc. If absolute β convergence is observed, then conditional β convergence is also implied. In conditional β convergence, the abovementioned growth related factors determine the steady state income level of an economy and if an economy is far from its steady state income level, it will tend to have a higher rate of economic growth until it arrives at its steady state. However, in the process of conditional β convergence, the initially poorer economies will have a tendency to move just towards their own steady state income levels. 3.8 Since FDI is considered to be one of the engines of economic growth in China, this research applies similar specifications by adding FDI to examine conditional β convergence at national and regional levels. Regressions are run on both cross sectional and pooled basis to estimate the pace of unconditional or conditional β convergence, which is the speed at which different economies return to their respective steady state output levels. The estimations also aim at addressing the growth of these economies, as well as the contributions of different factors, such as investment ratio, population growth, openness ratio, FDI ratio, human capital and transportation to the growth pace and speed of income convergence. 14 3.9 The mathematical model is derived from the Solow growth model with a Cobb Douglas production function as the basis. Data is based on a panel of 29 provinces and municipalities (Tibet is excluded and Chongqing is merged with Sichuan) for the 1979 2003 period. Two principal data sources are available: China Statistical Data 50 Years 1949 98 (NBS, 1999) and China Statistical Yearbook (NBS, various years, 1999 2004). Key Findings and Conclusions 4.1 Compared with previous studies, this article makes some new contributions to the understanding of the impact of FDI on regional growth and inequality in China after economic reforms. It uses a more recent dataset for all the Chinese regions from 1979 to 2003. It employs both cross section and panel data approaches to study the same question. It examines regional inequality from three different perspectives: inter province, intra region and inter group (each 13 Solow, R. M. (1956), A contribution to the theory of economic growth, Quarterly Journal of Economics, 70:65 94. Sala i Martain X. (1996), The classical approach to convergence analysis, The Economic Journal, 106, 1019 1036. 14 Yao and Wei (2007) find that the factors of FDI ratio, export ratio, human capital and transportation have been proved to present significant impact on the GDP growth in China at both national and regional levels. 9
pair of two regions). It employs more determinants of income growth such as FDI and transportation into the β convergence estimation, with special attention on FDI and its role in the economic convergence process across the country and within each geo economic region as well as between regional groups. The purpose of examining the same topic of the effect of FDI on spatial growth differences and income inequality, based on various model specifications and estimations, is to provide a comprehensive anatomy of whether FDI has caused regional income inequality, which is a controversial topic in the literature with significant policy implications for economic growth and the development of China, and any other similar less developed economy in the world. 4.2 Apart from β convergence, σ convergence with the coefficient of variation is used to assess whether there is σ convergence between China s regions. The results show that the country experienced three phases of the process of income inequality over the period 1979 2003, declining in the first decade, expanding in the second and then starting to decline again in the third decade. However, this slight decline in CV may not be statistically significant as CV is a non parametric approach which is not subject to statistical testing. Furthermore, a declining CV does not necessarily imply economic convergence if the reduced CV is not caused by the poorest regions catching up with the richest regions, but by the catching up of the medium income regions with high income regions, or by the convergence among the medium income regions. As a result, CV is not an ideal measurement for income convergence for all regions within a country although it can be used to indicate the trend of overall inequality. The last disadvantage of CV is that it cannot show why regional per capita incomes converge or diverge. 4.3 In contrast, β convergence is a more useful tool in measuring income convergence as it can testify whether poor regions are catching up with rich ones. It can also explain why regions converge or diverge in per capita incomes. In this paper, the β convergence test indicates no evidence of absolute convergence, based on different estimations in both the cross sectional and pooled analyses. This implies that the initially poorer regions have failed to grow faster than the initially richer ones due to their diversified economic backgrounds and biased policies. It implies that neither the poor provinces nor poor regions have managed to grow faster than their richer counterparts. As a result, it can be concluded that regional inequality rose, rather than declined during the data period. 4.4 But, what explains the rising inequality? This question can be answered through the analysis on conditional convergence. The analysis in this paper provides some striking findings on income inequality. First of all, apart from investment ratio and effective population growth rate, FDI and exports are found to have significant and positive effects on regional growth differences. In addition, the effect of FDI on economic growth is weak among the western provinces. These two findings could easily lead to the conclusion that FDI is an important cause of regional inequality, especially if one considers that the skewed distribution of FDI among the three large geo economic regions in China is coincided with a similar spatial pattern of real per capita GDP. If such a conclusion is based on the correct interpretation of the results, and is logical as many previous studies have argued and suggested, then the reduction of FDI inflows into China would be able to curb the trend of widening regional income inequalities. Such a policy implication would also make it difficult to reconcile the positive effect of FDI on economic growth and its negative effect on income distribution. As such a conclusion and its potential policy implications are obviously controversial, there needs to be a better understanding and a more accurate interpretation of the econometric results, which show a positive and significant relationship between FDI and economic growth, according to all kinds of model specifications presented in this paper. 10
4.5 If we summarize the results of our modeling exercises, we have the following findings: (1) regional income inequality rose in the data period; (2) regions can converge to their own steady states only after controlling for the differences in savings rate, population growth, human capital endowment, transportation, and above all, FDI and exports; (3) the same factors that have a significant effect with national level data have similar effect with regional (or groups of regions) level data; (4) FDI is singled out to have played a consistent and positive effect on growth differences according to all specifications, except for the Western region and the combined West/Central regions; and (5) FDI is highly unevenly distributed among the regions, with the West accounting for a very small share. 4.6 All these findings should point to the following conclusion, which is very different from that drawn by many other authors: FDI is an important factor in economic growth but it is unevenly distributed across regions. As a result, it is the uneven distribution of FDI, rather than FDI itself, which has been a cause of regional income inequality. This conclusion gives rise to the following policy implication: in order to reduce regional inequality, FDI should be encouraged, rather than discouraged. However, FDI has to be directed to the Western and Central regions through preferential policies and government intervention so as to create a more conducive environment for absorbing FDI in these relatively backward areas. 11