Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Does Foreign Direct Investment Crowd out Domestic Investment in China? Sumei Tang* School of International Business and Asian Studies, Griffith University, Nathan, Queensland, AUSTRALIA April 2005 Abstract This paper investigates the causal link between foreign direct investment (FDI), domestic investment and economic growth in China at the aggregate level for the period 1988-2003 by using an investment error correction model (ECM) and the innovation accounting (variance decomposition and impulse response function analysis) techniques. We find that there is bi-directional causality between domestic investment and economic growth and only single-directional causality from FDI to domestic investment and economic growth. There is no evidence to support that FDI crowd out domestic investment but instead FDI has complimentary effects on domestic investment in China in the period 1988-2003. FDI has played a great role in China’s economic growth and development. While FDI has supplementally provided somewhat solutions to the shortage of capital formation in China’s economic reform era, it has also stimulated domestic investment through the channel of technology diffusion and the rapid expansion of the export industries. * Email: [email protected] I am grateful to Professor G. H. Wan, World Institute for Development Economic Research, United Nations University, and Professor Saroja Selvanathan, DAFE, Griffith University, Australia for helpful comments. 1 Does Foreign Direct Investment Crowd out Domestic Investment in China? 1. INTRODUCTION China, the one of the world’s fastest growing economies, has attracted a large amount of FDI over the last two decades and became the largest FDI recipient country in the world since 2002. The amount of FDI inflows has totalled $487.686 billion US dollars during the period 1988 - 2003 in China. Does this enormous amount of FDI crowd out domestic investment in China or whether FDI has substitution or complementary effects on domestic investment? China is a typical developing country short of capital resources. FDI has been certainly becoming a great capital flow of choice comparing with portfolio flows in the wake of the Asian and South American financial crises. It is commonly believed that FDI not only assist in overcoming a pressing need of capital resources, it also could contribute to the economic growth and development of the host developing countries through technology, management skills transfer and spillover effects on domestic industries, as well as increasing demand for exports. Despite the role of FDI in the developing economy is highly emphasized and the developing countries aggressively compete for FDI inflows, the impact of FDI on economic growth is still highly controversial. de Mello (1999) argues that the impact of FDI on economic growth depends on the relationship between FDI and domestic investment. This is, the impact of FDI on economic growth will be positive/negative if FDI has complementary/substitution effects on domestic investment, respectively. The impact of FDI on economic growth and development has been discussed extensively but the direct discussion of the casual link between FDI and domestic investment hence economic growth is relatively insufficient, particularly in the context of China. Traditionally, those empirical studies most use either cross-sectional/time series regressions or panel data framework. However, this study, by applying quarterly time series data of China from 1988 to 2003, uses an investment error correction model (ECM) combining with recently developed time series concepts of cointegration, 2 innovation accounting (impulse response function and variance decomposition) and Grange causality testing procedures to investigate: (1) whether FDI has complementary/substitution effects on domestic investment in China; (2) the causal relationships between FDI, domestic investment and economic growth; (3) the role of FDI in China’s economic growth and development. This study is important. Firstly, China currently is a largest FDI recipient country in the world, which provides a rich contest of studying and research. Understanding the dynamics of FDI and its impact on its economy in China is essential in understanding the dynamics of FDI in the world economy as a whole. Secondly, China is one of the fastest growing economies in the developing world, its central and local governments have seen attracting FDI is an important strategy to overcome capital shortages and to promote economic growth and development, thus, intensive investment policies such as tax rates deductions and low regulations were encouraged. Understanding the impacts of FDI on domestic investment hence economic growth and development in China would provide a great lesson for other developing countries. The remained paper is divided into four sections. First, it is a theoretical and empirical review on previous studies. Second, it overviews FDI inflows, domestic investment and economic growth in China. Third, attempts to use econometrical analysis to find out whether FDI has crowded out/in domestic investment in China, the causal links between FDI, domestic investment and economic growth and the role of FDI in China’s economic development. The final section presents the conclusion. 2. A REVIEW OF PREVIOUS STUDIES Theories of FDI, primarily from a manufacturing perspective, study the determinants of FDI under taken by multinational enterprises (MNEs) outside their national boundaries. FDI occurs as MNEs possess ownership-specific advantages, market internalization advantages and location advantages (Dunning, 1981). The ownership-specific advantages include advanced technologies, management know-how skills, intangible advantages and transaction cost minimizing advantages etc. The market internalization advantages explains that firms engage in FDI is to avoid market imperfections such as government 3 interventions on quotas, tariffs and price controls, and seek controls on supplies of inputs and conditions of sale in a host country. The local advantages describe how the factors of host nations determine the siting of production of MNEs. The factors include preferential policies such as tax deduction and low regulation, resource endowments, input prices and markets etc are crucial elements for attracting FDI inflows. In the presence of FDI inflows, an open developing economy may achieve economic growth substantially through technology diffusion (Broensztein et al. 1998) and human capital development (de Mello, 1999). While FDI by MNEs is a major channel for the access to advanced technologies by developing countries as most MNEs possess ownership-specific advantages. FDI plays a vital role in economic growth (Broensztein et al. 1998, Kim and Seo 2003, Liu et al. 2002 and Shan 2002) and the long-run growth, as indicated by the new endogenous growth theory, can be achieved because the stock of FDI increases if the marginal product of capital can be bounded away from the rate of time preference (de Mello, 1999). In this regard, complementary effects of FDI on domestic investment are presented in several ways (Noorzoy, 1979). First, FDI in host country may result in an increase in demand for exports, hence stimulating investment in the export industries. Second, FDI could be made in high risk areas or establishes new industries where domestic investment will be limited, therefore, FDI will supplement domestic capital formation. Third, when FDI occurs in resource industries it is likely to expand investment in related industries. The ownership-specific advantages, somehow, also imply that FDI as an aggressive global strategy by MNEs to advance monopoly power over and above indigenous firms of the host economy (Hymer, 1960 and Caves, 1971). While the market internalization and location advantages, for instance, controls on supplies of inputs in the host economy and tax subsidy benefits provided by the host country, further reinforce the competitive advantages of MNEs over indigenous firms. Eventually, domestic firms will be forced to exit their industries so that FDI substitutes for domestic investment. The substitution effects of FDI may also occur when investment opportunities are limited in a host economy and MNEs compete for scarce factors in the economy. In addition to an adverse 4 effect of FDI on domestic investment in the same industry, FDI may also disrupt backward linkages through substitution of imports for domestic commodities (Noorzoy, 1979). However, it is argued that economic growth could still be achieved when substitution effects of FDI are presented. Because when substitution effects overwhelm complementary effects, FDI may contribute more to growth than domestic investment in an economy if repatriation of profit by MNEs can be under control, but predatory behaviour of MNEs could create a neo-colonial economy (Bornschier and Chase-Dunn 1985, Gowan 1999). However, as an economy, complementary effects of FDI are always preferred in the context of growth. Applying province-level panel data set for the years 1986-99 to a regression model, Braunstein and Epstein (2002) analyzed the long-term effects of FDI on domestic investment and found that FDI crowded out domestic investment during the late 1980s and 1990s in China. This was, Braunstein and Epstein argued, contributed by the decentralization of the FDI bidding process in China. Braunstein and Epstein (2002) viewed that the social benefits of FDI have been dissipated at least at the provincial level due to the intensive competition for FDI among the regions in China. The competition, in turn, forces the regions to lower regulations, taxes, environmental protections, wages and working conditions in order to attract and retain capital. However, the result produced by applying a set of the combined provinces panel data to a single regression model is arguable. First, the panel data is very heterogeneous hence hardly compatible, for instance, the geographical impact on FDI is prominent and the distribution of FDI in 28 regions and provinces is very unbalanced. Second, the estimation from the single regression equation is used to interpret some causal long-run relationship may be subject to a simultaneity bias and failed to consider the feedback among the macroeconomic variables hence the dynamics within the system. In Contrast to the study of Braunstein and Epstein, Sun (1998) found a strong significant positive correlation between FDI and domestic investment in China in the period from 1979 to 1996, but it was argued that the result of Sun’s study is unconvinced due to the simple regression approach and omitted important variables (Braunstein and Epstein, 2002). In regard to FDI and domestic investment in China, Huang (2003) argued that the Chinese investment policies are more 5 ‘‘friendly’’ to FIEs than to domestic private firms, as an incentive, Chinese partners are egger to form a FIE with foreign investors. This type of investment, in fact was invested by Chinese partners, occurs just for the motivation to qualify for FIE status (Huang, 1998). As a consequence, those types of FIEs not only gain benefits from the preferential policies and even possess privileges in competing for local scarce resources. Thereby, FDI crowds out domestic investment (Huang, 1998). However, their study is only a qualitative analysis lacking of statistical background. In comparison to the above cited studies, using a more advanced time series analysis approach, a vector auto-regression (VAR) system, and a set of quarterly time series data from 1986 to 1998, Shan (2002) examined the inter-relationship among FDI, output growth and other economic variables in China. Shan found that FDI has a significant impact on the Chinese economy when FDI shares in China’s output rises. The shortcoming is that Shan ignored the important issue existing in quarterly data, the seasonality. In addition, the measurement of economic growth in Shan’ study is not appropriately defined, a better and more accurate variable reflecting economic growth, e.g. GDP should be considered. Overall, the empirical studies cited above are only limited in the context of China, thus, the following section provides some countries’ empirical studies other than China. Following Caves and Reuber (1971) and Van Loo (1977), Noorzoy (1979) developed an accelerator-flow of funds model of investment to estimate the effects of FDI outflows/inflows on domestic investment in Canada for the period 1957-1971. Noorzoy introduced FDI into the traditional neoclassical investment model as the flow of external funds. In Noorzoy’s study, the accelerator term was measured by lagged changes in output (GNP) and the internal funds of investment by firms was given by the sum of depreciation and undistributed corporate profits. The dependent variable, investment, in the accelerator-flow of funds model of investment was given by gross capital formation. Noorzoy found that FDI inflows had complementary effects on domestic investment, whereas, FDI outflows had negative impact on domestic investment. However, the model used by Noorzoy is a single regression model which failed to consider the strong causal links and the feedback among the macroeconomic variables in the system. In contrast to Noorzoy’s study, applying the bivariate VAR model and the time series concepts of 6 cointegration to a set of time series and panel data, de Mello (1999) analyzed the impact of FDI on economic growth in 32 host countries and found that FDI is growth-enhancing via technological upgrading and knowledge spillovers but depends on the degree of complementarity and substitution between FDI and domestic investment. This study by de Mello is more theoretical rather than empirical analysis. It is inconvincible that FDIenhancing growth depends on the degree of complementarity and substitution between FDI and domestic investment. Because, in de Mello’s study, 1) the time series methodology has hardly ever been utilized to explore dynamic interactions (Kim and Seo (2003) among economic growth, FDI and domestic investment, 2) only 22 annual observations were used which might be subject to a small sample bias and (3) the explanation of complementary and substitution effects of FDI was very unclear. 3. AN OVERVIEW OF THE FDI INFLOWS, DOMESTIC INVESTMENT AND ECONOMIC GROWTH IN CHINA: 1978 - 2003 Attracting FDI has been a key pillar of China’s “opening up” policies and economic reforms since 1978. The economic reforms started with the establishment of special economic zones in selected provinces and regions, while the FDI policies were basically formed with preferential policies including tax concessions and special privileges for foreign investors. During the reform period, the Chinese government also has developed various FDI and ownership legislations, laws, property rights and contract laws to improve investment condition and business environment in order to attract FDI. Figure 1 presents domestic investment (DI), GDP and FDI inflows in China from 1978 to 2003. At the initial “opening up” period, FDI inflows were quite low varying between RMB 0.0468 to RMB 1.2645 billion from 1978 to 1983, but from 1984 till early 1990’s, FDI increased on an average rate of over 30% per annum. However, the total amount of FDI was still small and remained as low as RMB 40 billions until 1992†. † Calculated from various issues of China Statistical Yearbook. 7 In 1992, the famous “South Tour” by the Chinese leader, Deng XiaoPing, resulted in further economic reforms in China which led to a new phase of FDI liberalization policies. As a result, FDI to China entered into a stage of rapid growth and placed China as the largest FDI hosting country in the developing world since 1993. In 1998, during the Asian financial crisis period, the Chinese government further liberalized FDI policy. One such change is the abolishment of the FDI project approval requirement. However, as the Asian financial crisis was very strong, FDI inflows still decreased slightly in China during the year of 1999 and 2000. In December 2001, China joined the World Trade Organization (WTO), which marked the FDI liberalization entering into a new era. In the same year, China’s FDI inflows (see Figure 1) have increased dramatically from RMB337.0965 billions in 2000 to RMB 388.0092 billions in 2001 and by 2002 China became the largest FDI host country in the world attracting RMB436.5538 billions of FDI. As can be seen from Figure 1, China’s economic growth has also shown a remarkable growing strength at an average annual rate of 9 per cent since the economic reform in 1978. China has become one of the fastest growing economies of the world, for which, many argue, FDI has played an important role in contributing to the growth. For example, the national total export average annual growth rate from 1978 to 2002 has reached 12.56%, whereas, the share of export by foreign -invested enterprises (FIEs) in the national total export was almost 50% in 2001. Meanwhile, the domestic investment‡ in China, in line with the rapid growth in GDP and FDI, has demonstrated a significant increase with an average rate of 20.11% per annum from 1978 to 2003 as indicated in Figure 1. The investment system in China is very different from that in the western economies, which has been continuously following plan-era patterns in the last quarter of the country. The Chinese National People’s Congress approves annual plans for investment. The State-Owned-Enterprises (SOE), in general, benefit the most from the plan while private businesses have been largely excluded from the domestic capital market. ‡ The domestic investment in this study is the aggregate investment which excludes all types of foreign investments. 8 Figure1. Trends of Domestic Investment, GDP and FDI in China§ Figure 2 shows the growth rates of FDI and domestic investment in China in the period from 1978 to 2003. As indicated in Figure 2, in the period 1978 -1983, the amount of FDI was relatively small but during 1984 -1992, FDI had a large increase. During the two periods, domestic investment in China had a very similar annual average growth rate of 18.11% and 18.53%**, respectively. From 1993 till the eve of the Asian financial crisis, FDI showed a sharp rise in China, while the domestic investment also entered into a rapid growth at an average rate of 23.91% per annum. Domestic investment in China was only slightly affected during the Asian financial crisis period. Since joining the WTO, FDI in China has surged to a historical high making China the largest host country of FDI in the world since 2002, whereas domestic investment also increased dramatically. Overall, Figures 1 and 2 clearly demonstrate that both FDI and domestic investment have an upward trend well matching the strong economic growth trend of GDP during the period of 1978 – 2003. § FDI* equals FDI but scaled differently on the right side of Y axis for easier illustration and visual view. All figures in this paragraph are calculated from various issues of China Statistical Yearbook. ** 9 Figure2. The growth rate of FDI and domestic investment in China Table 1 presents the ratios of FDI to GDP, domestic investment to GDP and FDI to domestic investment from 1978 to 2003. As can be seen from Table 1, the relationships among FDI, domestic investment and economic growth have further confirmed. The proportions of FDI to GDP (column 2) were quite low and less than 1% from 1978 to 1990, and then it increased gradually but still relatively low at an average rate of 4.18% from 1991 to 2003. Comparatively, the proportions of domestic investment to GDP (column 3) were higher than the proportions of FDI to GDP during the same period. In 1978, the proportion of domestic investment to GDP (column 3) was 18.45% and increased steadily. However, when FDI surged in 1993, the ratio of domestic investment to GDP also increased to 34.99%, and in 2003 it further increased to 47.56%. However, the last column of Table 1 indicates that the proportion of FDI to domestic investment has increased dramatically from 0.07% in 1978 to 1.66% in 1984, and by 1994 it has reached an all time high of 18.12% but it then gradually decreased to 8.04% in 2003. The above analysis demonstrates that the dramatic increase in FDI inflows were associated largely with the rapid rise of domestic investment as well as economic growth during the period of 1978 to 2002. 10 Table1. The Ratios of FDI to GDP, Domestic Investment to GDP and FDI to Domestic Investment (DI) (percent) Year FDI/GDP DI/GDP FDI/DI (1) (2) (3) (4) 1978 0.01 18.45 0.07 1979 0.06 17.32 0.36 1980 0.06 20.16 0.28 1981 0.14 19.02 0.74 1982 0.15 22.10 0.70 1983 0.21 22.98 0.93 1984 0.41 24.57 1.66 1985 0.54 27.35 1.99 1986 0.63 29.24 2.16 1987 0.72 30.18 2.38 1988 0.80 29.33 2.71 1989 0.75 24.36 3.10 1990 0.90 22.82 3.94 1991 1.08 24.40 4.41 1992 2.28 28.57 7.98 1993 4.58 34.99 13.09 1994 6.22 34.34 18.12 1995 5.36 31.17 17.19 1996 5.11 30.36 16.83 1997 5.04 30.32 16.62 1998 4.80 33.31 14.42 1999 4.07 33.81 12.03 2000 3.77 35.12 10.73 2001 4.04 37.79 10.70 2002 4.17 41.00 10.16 2003 3.82 47.56 8.04 Sources: China Statistical Yearbook various issues. Figure 3 shows the plots of FDI against domestic investment (DI) and GDP, respectively. The findings in Figures 1 and 2, as well as Table 1 are also reflected in Figure 3, which clearly indicates that there is an upward trend and positive relationship between FDI, domestic investment and GDP. Overall, it appears that FDI inflows to China have had a 11 complementary effect on domestic investment hence spurring economic development and growth. However, results from a causal investigation on FDI, domestic investment and economic growth is far from conclusive in drawing an empirical relationship for China as discussed. Thus, a formal econometric analysis is required which is performed in the following section. Figure3. Observations on FDI, domestic investment (DI) and GDP 4. THE EMPIRICAL EVIDENCES Based on the theoretical argument and empirical study, an investment error correction model is developed to incorporate the accelerator-flow of funds model of investment into the error correction model (ECM) form. This model can enable the empirical investigation to find out the three questions proposed in the first section. The investment of error correction model We consider the following three equations as part of the investment error correction model: ∆FDIt = α1 + αfdi êt-1 + k k k i 1 i 1 i 1 α11 (i) ∆FDIt-i + α12 (i) ∆DIt-i + α13 (i) ∆GDPt-i + β1Dt + εfdit (1) 12 ∆DIt = α2 + αdi êt-1 + k k k i 1 i 1 i 1 α21 (i) ∆FDIt-i + α22 (i) ∆DIt-i + α23 (i) ∆GDPt-i + β2Dt + εdit ∆GDPt = α3 + αgdp êt-1 + k k k i 1 i 1 i 1 (2) α31 (i) ∆FDIt-i + α32 (i) ∆DIt-i + α33 (i) ∆GDPt-i + β3Dt + εgdpt (3) Where FDI = DI = GDP = êt-1 = Dt = αi, αij (i) and βi = εfdit, εdit and εgdpt = FDI inflows in China gross capital formation represents domestic investment but exclude any forms of foreign investment gross domestic product represents economic growth the error-correction term; the centered seasonal dummy variable. the parameters; white-noise disturbance terms that may be correlated with each other; The investment error correction model used in this study can reflect the lagged changes in the output (GDP) hence well representing the accelerator term by differencing the variable. In addition, this model can capture the interactions of all the variables including the lagged variable itself in the system, which is the advantage over the accelerator-flow of funds model of investment used by Noorzoy (1979), Van Loo (1977), Caves and Ruber (1971). Moreover, including both the first differences and the levels of the three variables in the system will enable the revealing of the short and long run effects of the all variables. However, a variable for internal fund is not present in the model which is different from Noorzoy but as same as Van Loo. Data The data of the three variables, FDI, DI and GDP are all quarterly time series in terms of Chinese RMB currency for the period 1988:1 to 2003:4, which are compiled from China Monthly Statistics (1987:1- 2004:3), Comprehensive Statistical Data and Materials for 50 years of New China and various issues of China Statistical Yearbook all published by the National Bureau of Statistics (NBS). GDP quarterly time series is specially constructed on the basis of the monthly gross industrial output (GIO) and the yearly GDP statistics because of lacking quarterly or monthly GDP statistics for China. It is found 13 that the annual growth pattern of GDP is similar to that of GIO, and following Liu, Song and Romilly (1997) and Liu, Burridge and Sinclair (2002), the estimated GDP is given as: GDPt,q = gt ×GIOt,q q =1,…, 4 t = 1988, 1989,…, 2003 where gt is the annual GDP/GIO ratio and GIOt,q is the quarterly value of GIO. Also, a special attention has been given to the seasonality of the three series, FDI, DI and GDP. China’s central planned economic regime has been a major force in generating a regular pattern of large seasonal fluctuations (Rawski, 2002) in its economy. It can also be clearly seen in Figures 3(a), 3(b) and 3(c), the three original time series indeed display some seasonal patterns. To minimize the effect of seasonal fluctuations, a centered (orthogonalized) seasonal dummy variable is applied when conducting in cointegration analysis and model estimation. The reason to use the centered seasonal dummy variable is that a standard 0-1 seasonal dummy variable will affect both the mean and the trend of the level series in VAR system but the centered seasonal dummy variable can shift the mean without contributing to the trend (Johansen, 1995). In addition, given the deficiency of Chinese official statistics, extra care has been taken to ensure the comparability and consistency of the data over time. (a) 14 (b) (c) Figure 3 The original series of FDI, DI and GDP in China, 1988 (1) – 2003 (4) Unit root test As indicted in Figures 3(a), 3(b) and 3(c), there is a clear upward trend in each of the variables. In order to ensure all time series in this study are stationary hence avoiding any potential problem of spurious regression, the augmented Dickey-Fuller (ADF) test is Table2. ADF test for unit root 1. ADF test for unit root on the level series Variables FDI DI GDP 2. With constant -1.12(2) -0.98(2) -0.87(1) With trend & constant No trend & constant -1.04(1) -4.40(2)*** -2.29(1) 1.07(2) 0.27(2) 3.46(1)*** ADF test for unit root on the first differenced series FDI DI GDP -2.29(2)** -7.43(2)*** -3.55(1)*** -6.45(1)*** -7.45(2)*** -3.69(1)*** - 2.22(2)** -7.26(2)*** -1.99(1)*** Notes: (1). ** and *** denotes significance at the 5% and 1% levels, respectively. (2). The figures in the parentheses are the number of lags used. 15 conducted in this study. The ADF tests for unit roots are performed on both levels and first differences for all three variables in the models of (1) with constant, (2) no trend, (3) with trend. The results of the ADF tests are reported in Table2. As is apparent from the table, the null hypothesis of a unit root is not rejected by the level series of FDI in all three models but rejected by the level series of GDP in the model without constant and trend, and DI in the model with trend and constant. First differencing of all the three time series FDI, DI and GDP indicate that they are stationary integrated of order one, I (1), for all models. Cointegration test Based on the results in Table 2, a cointegration test is performed to find out whether there have been any long-run equilibrium relationships among the three variables of FDI, DI and GDP. After a careful search and test, a model with 6 lags, constant, drift and centered seasonal dummy variable was chosen. The result of Joansen cointegration rank test is summarized in Table 3 and it indicates that there are two presences of cointegrating vectors at 1% and 5% levels respectively of significance (i.e., the null hypotheses of no cointegration are rejected for rank of zero and less than or equal to 2). This means that there exists a long-run stationary relationship among the three variables. Table3. Johansen co-integration tests Null (H0) Alternative (H1) λmax Rank = 0 Rank ≤ 1 Rank ≤ 2 r≥1 r≥2 r≥3 30.02** 10.06 5.94* 95% CV 23.78 16.87 3.74 λtrace 46.01** 15.99 5.94* 95% CV 34.55 18.17 3.74 Note: * and ** denote rejection of the null hypothesis at the 5% and 1% significance levels, respectively. Model selection test A model with constant, centered seasonal dummy and 6 lags of length is selected after a general-to-specific search starting with a system with 4 lags, constant, trend and seasonal dummies. The model was already shown in the above, the three equations of (1), (2) and 16 (3). The fits of the all three equations are quite high, 76%, 83% and 96% for (1), (2) and (3), respectively and there are no signs of residual autocorrelation either. The model diagnostics are summarised in Table 4. Table 4. Model diagnostics Adj-R2 ∆FDIt ∆DIt ∆GDPt 0.76 0.83 0.96 Autocorrelation test LM (1), CHISQ (1) = 15.06, P-value = 0.09; LM (6), CHISQ (6) = 15.84, P-value = 0.07 Normality test CHISQ (6) = 5.55, P-value = 0.0623 Innovation accounting analysis and Granger causality test Within this near VAR (ECM) system, the accounting innovation (variance decomposition and impulse response function) technique can be utilised to examine the relationships among economic variables. In our case, this technique would be a useful tool enabling us to pick up any complementary or substitution effects of FDI on domestic investment and simultaneously to reveal the impact on economic growth. The forecast error variance decomposition allows inference over the proportion of the movements in a time series due to its own shocks versus shocks to the other variables in the system (Enders, 1995). The impulse response function analysis is a practical way to visually represent the behaviour of a time series in response to the various shocks in the system (Enders, 1995). In other words, the forecast error variance decomposition explains all the forecast error variance effects on each endogenous variable at short horizons and smaller proportions at longer horizons, and the impulse response function analysis traces out the time path of the effects of the various shocks on each endogenous variable to determine how each endogenous variable responds over time to a shock in that variable and in every other endogenous variable. 17 Table 5. Decomposition of ten-year forecast error variance (%) Percent of forecast error variance in (Quarters) FDI DI GDP (1) (2) (3) FDI 1 4 8 12 16 20 24 28 32 36 40 100.0000 90.82013 77.02271 64.17280 56.72252 53.52018 51.69749 49.27031 46.78047 45.03550 43.87666 0.000000 3.468023 5.266766 6.136423 8.290025 9.805131 10.40132 10.38458 10.06185 9.710682 9.375944 0.000000 5.711849 17.71052 29.69078 34.98746 36.67469 37.90118 40.34511 43.15767 45.25382 46.74739 (4) DI 1 4 8 12 16 20 24 28 32 36 40 0.456898 2.371060 31.67926 49.15293 49.35877 46.67251 46.32242 47.64437 48.79696 49.43196 49.74517 99.54310 79.98403 56.38692 39.86692 39.01755 40.75727 40.29452 39.14215 38.42148 38.01316 37.65754 0.000000 17.64491 11.93382 10.98015 11.62368 12.57022 13.38306 13.21348 12.78156 12.55489 12.59729 GDP 1 4 8 12 16 20 24 28 32 36 40 2.825318 1.884889 31.06297 63.21354 61.82440 56.48597 56.70327 58.09982 59.20184 59.04479 58.87606 37.36442 40.81599 26.99503 15.28816 15.38310 16.15336 16.32276 16.63251 16.79938 16.47414 15.82107 59.81026 57.29912 41.94200 21.49830 22.79250 27.36067 26.97397 25.26767 23.99877 24.48107 25.30287 With a ten-year forecasting horizon used in this study, the variance decompositions in FDI, DI and GDP of the ECM by using the Choleski decomposition method are reported in Table 5. As indicated by the result that each time series explains the preponderance of its own past values. FDI explains 90.82% (column 2) of its forecast error variance at the 4th quarter of the year one, whereas, DI and GDP explain only 3.47% and 5.71% of the forecast error variance of China’s FDI inflows. DI and GDP explain near 80.00% 18 (column 3) and 57.30% (column 4) of their forecast error variances at the 4th quarter of the year one, respectively, in contrast with: a). FDI and GDP explain 2.37% and 17.64% of the forecast error variance of China’s total domestic investment, respectively; b). FDI and DI explain 1.88% and 40.82% of the forecast error variance of China’s economic development, respectively. Importantly, Granger causality tests in Table 6 indicate that: a). the effects of DI and GDP on FDI are not statistically significant at conventional levels; b). the effects of FDI and GDP on DI are statistically significant at 1% level; c). the effects of FDI and DI on GDP are statistically significant at 1% and 5% levels, respectively. Thus, FDI affects DI and GDP but not the reverse, whereas, the causal links between GDP and DI is bidirectional. Table 6. Results of Granger causality Test among FDI, DI and GDP in China 1988-2003 Dependent Wald test statistics Causality Variables ______________________________ inference ∆FDIt ∆DIt ∆GDPt ∆FDIt ∆DIt 3.98 24.63** 4.38 FDI DI GDP DI 26.06** FDI GDP DI GDP ∆GDPt 20.15** 16.20* Two-way Links GDP DI Note: * and ** reject null hypothesis at 5% and 1%, respectively. Given the results of the variance decompositions and Granger causality tests, the impulse response functions of FDI, DI and GDP to three types of shocks are proceeded. The illustrations of the response of FDI, DI and GDP to a shock in FDI, DI and GDP, respectively, are given below: 1. Figure 4 impulse responses to a shock in FDI. The effect of a shock in FDI is to cause an immediate decrease in DI, shortly, DI begins to increase even surpassing the 19 original level and then declines again. This has been a visible permanent pattern that DI fluctuates always around its mean level. FDI itself drops rapidly immediately and around the second quarter of the third year its movement maintains at a new lower mean long-run level. GDP responses in an immediately increase and converges to its new higher level permanently by the third quarter of the fourth year. It is interesting to note that the peak and trough of FDI are almost at opposite positions of DI; whereas, the peak and trough of GDP occur almost simultaneously with DI. In general, in the case of China from the period of 1988 to 2003, the result from Figure 4 evidently indicates that the FDI inflows do not crowd out the domestic investment in the long-run. In regard to the interesting character that the peak and trough of FDI inflows always have been at the opposite positions to the total domestic investment in the short-run, it, perhaps, can be described metaphorically that shortage of domestic investment in China urges for FDI inflows in the study period. The larger FDI inflows the greater domestic investment was stimulated. It is true that FDI indeed has resulted in an increasing demand for exports in China (Zhang and Felmingham, 2001; Liu, Burridge and Sinclair, 2002), consequently, as the market demand increases, more investment was stimulated in the export industries. In this regard, FDI has complementary effects in China, which confirms the result from the initial analysis proposed in the third section and the findings of Sun (1998) and Noorzoy (1979). The movement of FDI in the Figure 4 also demonstrates that the FDI inflows in China largely are influenced and determined by its past values. The upward movement of GDP indicates that FDI has positive impact on China’s economic development. Overall, Figure 4 provides robust evidence confirming that FDI plays a crucial role to complement domestic investment in China hence leading economic growth and development. 2. Figure 5 impulse responses of a shock in DI. In response to a shock in DI, DI itself immediately drops rapidly and then reverts back but it never reaches its original level. GDP temporarily decreases and shortly in an expansion position even surpassing its original level but its mean long-run position is slightly below its original level. 20 Whereas, FDI is not affected much by the shock in DI initially, and it’s long-run mean level remains around its original level. It seems that the domestic investment in China during the study period has not had much influence on the FDI inflows, which has confirmed by the result from the Granger causality test as well. In contrast, the economic growth seems sensitive to the shock in DI and both are positively correlated to each other. On the other hand, the levels of domestic investment in China very much likely depend on its past values. This is true, especially in a highly planned economy. 3. Figure 6 impulse responses of a shock in GDP. After a shock in GDP, DI begins to decline continuously reaching the trough point in the 1st quarter of the fourth year. Also, it is noticeable that the fluctuation pattern of DI is nearly the same as that of the GDP, the higher level of DI and the greater GDP would be. In response to the shock in GDP, FDI only slightly declines initially, and then increases converging to a lower long-run new mean level. The results show that there is a strong causal link between economic growth and domestic investment in China, which supports the use of an investment error correction model for the study period. This also confirms the accelerator principal of investment, the larger the change in GDP, the greater the desired change in the capital stock and, therefore, the greater the level of domestic investment. However, GDP seem to have a slight immediate and medium term impact on the FDI inflows. 5. CONCLUSION This study has used the ECM techniques of innovation accounting (variance decomposition and impulse response function analysis) to search for an answer whether (1) FDI in China crowd out domestic investment or whether FDI has substitution/ complementary effects on domestic investment; (2) the causal links between FDI, 21 domestic investment and economic growth; and (3) the role of FDI in China’s economic development. Fig 4. Impulse responses to a shock in FDI 160 120 80 40 0 -40 5 10 15 20 FDI 25 DI 30 35 40 GDP Fig 5. Impulse responses to a shock in DI 2000 1500 1000 500 0 -500 -1000 -1500 -2000 5 10 15 20 FDI 25 DI 30 35 40 GDP Fig 6. Impulse responses to a shock in GDP 1200 800 400 0 -400 -800 5 10 15 FDI 20 25 DI 30 35 40 GDP 22 We find that FDI does not crowd out domestic investment but instead it has complementary effects on domestic investment in China during the period 1988-2003. Secondly, the causal links between GDP and DI is two-ways, whereas, only a one-way directional causality runs from FDI to DI and GDP. Overall, FDI indeed has had significant positive impact on China’s economic growth and development during the period 1988-2003. FDI not only has supplementally provided somewhat solutions to the shortage of capital formation in China’s economic reform era, it also has stimulated domestic investment through the channel of technology diffusion and the rapid expansion of the export industries. This study further suggests that China’s economic growth, the determinants of FDI inflows and the level of domestic investment, somehow, are also explained by other forces either than variables used in this study. This implies that more empirical research and studies are required to include variables such as trade, productivity, new technologies and economic policies. Moreover, further study and research on different types of FDI and their effects on domestic investment and economic growth in individual industry will be interesting and useful for FDI host nations and their economic development. REFERENCES Borensztein, E., De Gregorio, J. and Lee, J-W. 1998, How does foreign direct investment affect economic growth? Journal of International Economics, Vol. 45, pp115-35. Bornschier, V. and Chase-Dunn, C. 1985, Transnational Corporations and Underdevelopment, New York: Praeger. Braunstein,Elissa and Epstein, Gerald. 2002 Bargaining power and foreign direct investment in China: can 1.3 billion consumers tame the multinationals? Working Paper, University of Massachusetts Amherst, MA. Caves, R. E. 1971, International corporations: the industrial economics of foreign investment, Economica, 38, pp1-27. Caves, R.E. and C. Reuber et al., 1971, Capital transfers and economic policy, Canada 1951-62, Harvard University Press, Cambridge, MA. 23 de Mello, L. 1999, Foreign direct investment-led growth: evidence from time series and panel data, Oxford Economic Papers, Vol. 51, pp 133-51. Dunning, J. H.1981, International production and the multinational enterprise, London and Boston: Allen & Unwin. Enders, W. 1995, Applied econometric time series, John Wiley & Sons, Inc New York. Gowan, P. 1999, The global gamble, New York: Verso. Huang, Yasheng. 1998, FDI in China: an Asian perspective, Singapore: the Institute of Southeast Asian Studies. Huang, Yasheng. 2003, One country, two systems: foreign-invested enterprises and domestic firms in China, China Economic Review, 14, pp404-416. Hymer, S. 1960, On multinational corporations and foreign direct investment, Selected by Dunning, J. H., in Dunning, J. H and United Nations eds. 1993 The theory of transnational corporations, London : Routledge. Johansen, S. 1995, Likelihood-based inference in Cointegrated Vector Autoregressive models, Oxford University Press. Kim, David Deok-Ki and Seo, Jung-Soo, 2003, Does FDI inflow crowd out domestic investment in Korea?, Journal of Economic Studies, Vol. 30, No. 6, pp605-622. Liu, X., Burridge, P. and Sinclair, P. J. N. 2002, Relationships between economic growth, foreign direct investment and trade: evidence from China, Applied Economics, 34, pp1433-1440. Liu, X., Song, H. and Romilly, P. 1997, An empirical investigation of the casual relationship between openness and economic growth in China, Applied Economics, 29, pp1679-86. Noorzoy, M. S. 1979, Flows of direct investment and their effects on investment in Canada, Economic Letters, 2, pp 257-261. Rawski, Thomas G. 2002, Will investment behavior constrain China’s growth? China Economic Review 13, pp361–372. Razin, Assaf. 2003, FDI flows and domestic investment, CESifo Overview Economic Studies, 49, 3, pp415. Shan, J. 2002, A VAR approach to the economics of FDI in China, Applied Economics, (7) 34, pp885-893. 24 Song, H., Liu, Z. and Jiang, P. 2001, Analysing the determinants of China’s aggregate investment in the reform period, China Economic Review 12, pp227–242. State Statistical Bureau of China (SSBC). China Statistical Yearbook 1980-2003, China Statistics Press, Beijing. Sun, Haishun. 1998, Macroeconomic impact of direct foreign investment in China: 197996, The World Economy, 21(5), pp675-94. Van Loo, F. 1977, The effect of foreign direct investment on investment in Canada, Review of Economics and Statistics 59, pp474-481. Zhang, Qing and Felmingham, Bruce. 2001, The relationship between inward direct investment and China’s provincial export trade, China Economic Review, Vol 12, Issue 1, Spring, pp 82-99. 25