Taiwan investment in MainlandChinaover the last 20 years has shown uneven distribution characteristics in the eastern, central and western regions. What factors led to this type of investment pattern? Did this characteristic continue or change after ECFA? Based on statistical data and the investment conditions, this paper used Moran’s I index analysis method to measure the heterogeneity of the spatial distribution ofTaiwaninvestment in MainlandChinaover 20 years. Then the study applied ordinary and spatial panel analyses to discuss impacting factors of nearly three years ofTaiwaninvestment in MainlandChina. The outcome shows thatTaiwaninvestment in MainlandChinahad a double-fixed timing effect and special effect during the last three years, which impactedTaiwaninvestment in different regions. Finally, this study proposed the regional strategy choices of theTaiwaninvested enterprise in MainlandChinain the future at four levels.
Over the past 20 years, cross-strait trade and offshore investment of Taiwan firms in Mainland China have bloomed rapidly. With the improvement of economic infrastructure and situations of China, the pattern and mechanism of industrial cooperation for cross-strait have changed a lot. As of 2010, there were 83,133 Taiwaninvested projects in Mainland China, and the actual use of Taiwan investment reached US$52.02 billion. The domains of investment have shifted from down-stream manufacturing to up-stream manufacturing and modern services as well as from labor-intensive industries and capital-intensive industries to technology-intensive industries and knowledge intensive industries.
The investment has also shifted from one-way, onetime investment to continuous large-scale investment. The industrial operation style is changing from OEM (Original Equipment Manufacturer) to ODM (Original Design Manufacturer) and OBM (Own Branding & Manufacturing). According to Taiwan’s economic data, as of January 2011, there were 112 invested projects, with investment amounts reaching US$137 million. The spatial distribution of Taiwan firms expanded from the northern and western parts to the southern and eastern parts of Mainland China. However, the investment pattern over the last 20 years did not seem to change significantly before the assignment of Economic cooperation and Framework Agreement (ECFA). Conversely, the ECFA has promoted the close economic collaboration between Mainland China and Taiwan. Doses it will cause the change of factors impacting Taiwan enterprises’ investment in Mainland China? The paper will discuss that issue along with the comparison between factors of Taiwan enterprises’ investment before and after ECFA. The paper is organized as follows. Section 2 is literature review. Section 3 discusses the overall pattern of Taiwan investment in Mainland China. Section 4 uses Moran’ I index to evaluate the spatial agglomeration and then adopt panel data to assess the impacting factors influencing Taiwan firms’ location in the Mainland China. Section 5 is conclusions and policy implications.
Based on location theory such as Weber’s industrial location and Thünen’s agricultural theory, factors such as market capacity, labor cost, transportation cost, and the relative technology level proposed by Alibe, Hirsch and Vermon have empirically explored the cost-driven FDI mode. In recent years, factors such as tariffs, non-tariff barriers, tax policy, language, culture and environment have also bloomed in the mainstream of investment theory. Second, the perspective of agglomeration economies can be applied to study the motivation and purpose of FDI, and these discussions are mainly based on aspects such as the scale of the economy, specialization, supply chain integrity, and learning by doing. Third, Dunning’s eclectic theory has been used to describe the regional advantage. Dunning mainly emphasizes the balanced judgment of the host country, ownership advantages and internalization advantages as the real motivation for FDI. In recent years, the study of Taiwan investment in Mainland China reflects three features. First, macro retrospective analysis emphasizes the summarized experiences and goals of Taiwan investment over 20 years. Zou Xiao Juan (2005) systematically reviewed the characteristics of several phases of Taiwan investments in Mainland China and gave a predictive analysis of future investment [
Another concern is the location analysis of investment in the mainland. As Duan Xiao Mei (2006) pointed out, northern China is the new hot spot for Taiwan investment. She proposed a strategy to improve the investment environment in Mainland China with respect to aspects such as legal and financial considerations as well as intellectual property, infrastructure, and medical services [
The third focus is the study of the development of Taiwan investment in Mainland China by evaluating the differences between regions or industries that receive investment from Taiwan, such as the Yangtze River Delta [
Based on the above literature, it is clear that Taiwan investment in Mainland China cannot be described by a simple explanation of the investment situation. Instead, the development of Taiwan investment in Mainland China is continuously changing. This subject has prompted investigation into topics such as traceability of the theory, the development of research methods and public policy to carefully observe and explore the trajectory, and the motivation and future trends in the development of Taiwan investment in Mainland China. Simultaneously, scholars have considered the development of the new era of Taiwan investment in Mainland China from a global perspective, considering the high degree of technological innovation and the new regionalism, which will also require follow-up study.
According to the data from 2010 Environment and Risk Survey of Regional investment in Mainland China of Taiwan Electrical and Electronic Manufacturers Association report (hereinafter referred to as the TEEMA report), in 2010, Taiwan entrepreneurs evaluated the investment environment in Mainland China based on the certain major advantages in terms of infrastructure, local approval of Taiwan businesses, regional economic development and preferences of local governmental sections, as opposed to the disadvantages in terms of common market, R & D alliance, investigation and punishment of poor-quality goods, funding the upgrading of Taiwan businesses and gaining local human capital. Thus, the current Taiwan investment mainly focused on hardware infrastructure, and the investment environment of the software facilities remains unsatisfying. The fact also reflects the declining driving factors of the current Taiwan investment cost and the formation of an environment-oriented investment model. From the perspective of regional distribution, eastern and northern China remains dominant regions for investment, but central and southwestern China have recently become investment hot spots.
From an investment impact point of view, Taiwan preferences are well connected with the geographic environment, the available infrastructure, the economic environment and the business environment. Subsequently, these factors have become the dominant factors in Taiwan investment policy (see
In general, the main reasons for FDI enterprises, including market-driven foreign competitiveness, cost-driven production sharing factors, and the complete construction supply, are connected to the progressive formation of industry clusters. National and international academic studies have been conducted on the regional distribution decision factors of FDI from different viewpoints, including the market capacity, the degree of aggregation, labor costs, infrastructure, human capital, industrial structure, the level of openness, the level of market orientation, and the presence of incentives (Zhang Yan Bo, 2008).
In Mainland China, because of the significant intercountry differences in customs and language [
Studies of FDI in Mainland China have regarded the region as an independent factor without considering the effect of the spatial correlation between the FDI and its spatial distribution. Therefore, there are a few empirical analyses of the number of neighboring areas of FDI and the spatial effect on FDI production. Anselin, Rey and Montouri suggest that many regional economic phe-nomena are autocorrelated because of the effect of spatial continuity on regional distribution, which also applies to FDI. In recent years, the space dependence of regional economic growth has received increasing academic attention; the rapid development of spatial econometrics allows us to conduct empirical investigations of regional issues. Therefore, the use of spatial econometric models to examine the spatial distribution of non-equilibrium FDI will lead to more quantitative and accurate conclusions. This article will address questions regarding the spatial aspects of Taiwan investment in Mainland China and whether these aspects will affect the pattern of future Taiwan investment in Mainland China.
1) Moran’s I index In order to identify the impacting factors of Taiwan enterprises’ location strategy, we adopt spatial data model to examine the outcome. First, Moran’s I Index is applied so as to find whether spatial heterogeneity does exist. Moran’s I index has a significant advantage in the study of the relevance of spatial variables. It reflects the degree of similarity among the properties of the unit attribute values of adjacent regions to analyze the spatial autocorrelation coefficient between the variables. Moran’s I index is calculated as follows where represents the regional observations, is the total number of regions, is the weight matrix for the adjacent space representing any of these elements. In general, the selection criterion is I which is in the range −1 ≤ I ≤ 1. When the economic behavior of the two regions is positively correlated, I is positive. When the behavior is negatively correlated, I is negative. If I is 0, the economies of these two regions are independent. The use of global and local Moran’s I indices can reveal the spatial correlation of the overall region and the surrounding area.
The test of the regional spatial correlation is mainly based on testing of the maximum likelihood estimation hypothesis through Wald, LR and LM statistics, Moran’s I spatial correlation index, and Geary C. However, Moran’s I (Moran, 1948), LMerr (Burridge, 1980), LMsar, Lratios, Walds (Anselin, 1988), and other space-related tests proposed by single cross-section regression models cannot be directly used for a panel data model. In this article, we adopt Moran’s I to measure the spatial distribution of Taiwan investment in Mainland China in the last 20 years, and we used the block-diagonal matrix as a spatial weight matrix. These tests can be extended to panel data analysis.
2) The spatial panel data model The spatial panel data model is generally divided into a spatial lag model and a spatial error model. LMerr and LMsar and their steady-form spatial correlation test can provide insight into the model setting and identify the selectivity of the spatial lag model and spatial error model. If LMsar (LMerr) is more significant than LMerr (LMsar), then the spatial lag model (spatial error model) is appropriate. This study obtained LMerr (0.2144) > LMsar (0.0742) based on the outcome, thus, the spatial error model was more appropriate than the spatial lag model. Based on global spatial correlation, the spatial autoregressive model means that all explanatory variables of a region’s economic growth act on other areas through a spatial transmission mechanism. The spatial error model shows that the impact of regional spillovers is the result of random shocks. Spatial error model (SEM) is as follows:
Here, y is the dependent variable, X is the vector of the independent variables (including a constant term), is the variable coefficient, and are spatial autoregressive and spatial autocorrelation coefficients, respectively, and is the error component. In the one-dimensional error decomposition model, or ; in the two-dimensional error decomposition model, , , and... is the time dimension and cross-section dimension respectively. is the T-dimensional unit matrix, and is the spatial weight matrix n × n (n is the number of regions). The weight factor can be determined based on the actual situation.
This paper will use the spatial panel data to analyze Taiwan investment patterns in Mainland China. Therefore, we must select samples that contain measurement indicators. According to the geographical distribution of Taiwan investment in Mainland China published by the Investment Commission of the Taiwan Ministry of Economic Affairs, we selected 21 provinces in Mainland China that are autonomous regions as the statistical samples. They include Heilongjiang, Jilin, Liaoning, Hebei, Beijing, Shanxi, Tianjin, Shandong, Jiangsu, Anhui, Sichuan, Hubei, Chongqing, Shanghai, Zhejiang, Hunan, Jiangxi, Yunnan, Fujian, Guangdong, and Guangxi. The K-nearest neighbor (KNN) algorithm was used to estimate the Moran’s I index test results, which are shown in
According to the TEEMA report, the factors that affected Taiwan investment in Mainland China in 2008, 2009 and 2010 included twelve indices (see
Source: Based on the data gathered from the TEEMA report.
trepreneurs would like to recommend Mainland China to other Taiwan entrepreneurs as a investment regions. The data from the twelve indices in the samples in 2008, 2009 and 2010 included 2612 questionnaires, 2588 ones and 2618 ones, respectively. As the TEEMA report itself had corrected the statistical errors, we directly reported the TEEMA-published score values (between 0 and 5) as the independent variable data.
As the data presented in
3) Results analysis a) Spatial agglomeration
Totally, data from 20 years of Taiwan regional investment in Mainland China show the characteristics of a positive spatial correlation. The data for 1991, 2004, 2005, 2008, and 2009 were significant at the 5% level; 1993, 1994, 2002, 2006, and 2007 were significant at the 10% level. The positive correlation for the other years was not significant (see
We choose 1991, 2005, and 2009 as three years with significant features of spatial concentration as well as adopting the year 2010 (the year ECFA was signed) to give four space location figures.
The LISA spatial concentration chart measured using the local Moran’s I index shows (
The spatial correlation analysis shows that the regional distribution trajectories of Taiwan investment in Mainland China over the past 20 years have no significant continuous spatial correlation. The index of a year was selected to analyze the impact factors that affected Taiwan investment in Mainland China, regardless of the time factors. The key factors that truly affected Taiwan investment within a period of time could not be determined. Therefore, the longer the selected data sequence is, the more accurate the measurement results are. However, because the impact factors of Taiwan investment in Mainland China are complex, many indicators are difficult to quantify. Therefore, this paper selected the amount of Taiwan investment in Mainland China and the environment evaluation data describing Taiwan business investment in Mainland China released by the Taiwan Electrical and Electronic Manufacturers Association in 2008, 2009 and 2010, and we analyzed the impact factors based on data availability and index consistency. The three years were very meaningful. In 2008, when Taiwan’s leader, Ma Ying-Jeou, took office, the cross-strait economic and trade exchanges improved. However, the 2009 global economic crisis seriously affected the exportoriented economic model used in Taiwan and caused a certain change in the cross-strait economic trade. In 2010, the global economy rebounded, the mainland economy rapidly stabilized, and the cross-strait countries signed an economic cooperation framework agreement (ECFA), which enhanced the confidence index of Taiwan enterprises to invest in the mainland. Whether it was macroeconomic indices in Taiwan or the cross-strait trade exchanges, a quick recovery was achieved.
b) Impact factors of the special agglomeration of Taiwan investment First, the general spatial panel data were used for analysis. If the spatial correlation factors were not taken into account, the F test was significant (0.0758 > Fα (2,48) = 5.79), and mixed effects models were established. To improve the impact of collinearity, we applied a two-stage cross-sectional regression weights model; the obtained results are presented in
The mixed-effects model is appropriate for understanding the distribution of Taiwan investment in Mainland China because the model showed spatial characteristics as well as time features. Moreover, in 2008 and 2009, the global Moran’s I index revealed that 21 Taiwan-funded investment regions had certain spatial correlations. Therefore, a geo-spatial correlation model was established for the tests. The SEM estimation results are shown in
The results presented in
The results of the spatial error model (SEM) showed that the difference in inter-regional geographical and market environments and the gradual improvement in
Note: *p < 0.1; **p < 0.05; ***p < 0.01. Model 1: No fixed effects; Model 2: Spatial fixed effects; Model 3: Time-fixed effects; Model 4: Spatial and time-fixed effects.
each region of these two factors had a significant positive impact on Taiwan investment in Mainland China over the past three years. The difference in the inter-regional infrastructure and the legal environment significantly restricted the Taiwan regional investment strategy. The results of Model 4 indicate that business risks and the degree of Taiwan entrepreneurs’ appreciation more significantly (5% level) restricted the choice strategy of Taiwan investment in Mainland China. If regional differences were not taken into account, of the most important factors regarding Taiwan investment in Mainland China in the past three years, the social environment, the economic environment, the market environment, social risk and market risk had significantly positive impacts. The legal environment, the innovation environment, legal risk and the degree of Taiwan entrepreneurs’ appreciation had significant negative impacts on Taiwan investment.
This paper highlights the spatial agglomeration and impacting factors of Taiwan enterprises investment in Mainland China. When compared with the overall effect that did not take regional and time factors into accountregional differences and time factors had significant impacts on Taiwan investment in the mainland.
First, Taiwan investment in Mainland China and business risk are negatively correlated, indicating that Taiwan investment in Mainland China was affected by the overall cross-strait economic relations, especially in Taiwan in recent years. Taiwan’s major political events affected the desirability of Taiwan investment in the mainland, and the nationalist party returned to power after 5.20 of 2010, which gradually reduced the positive message that was conducive to the cross-strait economic interaction. It also affected and changed the confidence, behavioral strategies and regional distribution of Taiwan investment in Mainland China. Thus, the low-risk business region was more attractive to Taiwan companies.
Second, there was a significant positive correlation between the geographical environment and the market environment and Taiwan investment in Mainland China, indicating that the investment motivation of Taiwan businesses over the past 20 years has gradually changed from cost-driven to market-oriented and environmentoriented motives. Thus, when the market environment of a region was favorable and the geographical environment was also adapted to the development of Taiwan business, this area was likely to grow into a Taiwan investment gathering area. However, one drawback was that region’s infrastructure and legal environment show a significant negative correlation with Taiwan investment in the mainland. One possible explanation is that the scoring of these independent variables of Taiwan enterprise was based on the combined results of the last year’s overall evaluation and this year’s forecasts1. Thus, Taiwan investment in Mainland China was motivated by faith in the quality of the mainland regional infrastructure and legal environment in the last year and in its improvement in the coming year. In fact, Taiwan investment in Mainland China showed a negative correlation with the geographical environment and market environment.
Third, we would like to assume that Taiwan entrepreneurs’ appreciation and the amount of Taiwan investment would be positively correlated, but our results are the opposite. We believe that the main reason was that more information about Taiwan enterprises invested in the mainland in the past three years was provided from the rational investment decisions and market research, which had changed the original “rooted network” Taiwan-funded enterprises’ developmental pattern and had gradually changed to an open style of co-investment. Therefore, Taiwan investment increased in only a few situations through Taiwan entrepreneurs’ appreciation. At the same time, because the degree of Taiwan recommendation was the combined evaluation of the effect of Taiwan investment in Mainland China over the last year and the comprehensive evaluated value of this year’s forecasts, it is possible that the larger the Taiwan investment amount was, the lower the Taiwan recommendation because Taiwan entrepreneurs believed that investment in a Taiwan gathering area (investments in some areas were close to saturation) would offer less flexibility for business growth. The businesses have difficulty obtaining sufficient social capital for their development (such as reputation, government support, social network); there may even be rigorous competition. Therefore, the degree of Taiwan entrepreneurs’ appreciation and the amount of Taiwan investment are negatively correlated.
Overall, based on the performance of Taiwan investment in Mainland China in the last three years, Taiwan-funded enterprises paid more attention to the optimal environment to develop the best regional investment strategy. Taiwan’s investment motive was more comprehensive and objective, which challenges the mentality that Taiwan investment in Mainland China is based on the lowest operating cost 20 years ago, taking the ideas of “human networking” and “just having a try” to invest. The increased investment was the result of emotional judgments and subjective selection. Therefore, the overall pattern of Taiwan enterprise investment in Mainland China is changing.
From spatial correlation analysis and the spatial panel data analysis of Taiwan investment in Mainland China in the past three years, the pattern of the Taiwan-funded enterprises in the mainland area shows a certain spatial correlation. In three typical years, 2008, 2009 and 2010, time-fixed effects were evident in the pattern of Taiwan investment in Mainland China. The regional distribution of Taiwan investment in Mainland China shows the individual characteristics of specific years and regions. In this situation, the future investment distribution of Taiwan enterprise in the mainland could not be macroscopically based on the changes of the overall investment environment in the mainland. Instead, future Taiwan investment distribution should start with different regions and carry out regional distribution with focus.
1) The spatial heterogeneity of the regional distribution of the Taiwan investment in the mainland increased; thus, the combination of investment factors should be considered based on the characteristics of different regions.
The regional economic development potential and the regions with good basic conditions better attracted Taiwan investment, which was particularly evident in the eastern coastal areas (Jiangsu). Factors such as infrastructure, legal risk, and social risk decreased the willingness of Taiwan entrepreneurs to invest in the mainland. Therefore, each region can attract Taiwan investment from such angles as improving the region’s infrastructure environment and reducing the legal risk, market risk, and social risk.
At the same time, Taiwan investment in Mainland China over a longer period of time showed characteristics of gradual accumulation. The individual provinces indicated a clear spatial autocorrelation, but the provinces that received Taiwan investment did not show overall spatial dependence. The spatial distribution of Taiwan investment in the mainland showed an obvious polarization effect. In attracting Taiwan investors, interregional competition is stronger than inter-regional cooperation. The continuation of this situation is not conducive to a balanced pattern of Taiwan investment in Mainland China. If only from the overall study to affect the factors of Taiwan investment in the mainland, ignoring the factors of regional investment differences, this situation may expand the distribution characteristics. Therefore, future Taiwan-funded enterprises need to consider in detail the specific factors that affect the future development of Taiwan investment from the district level; various regions also need to develop differentiation strategies to attract foreign capital based on its own natural resources and storage and space development.
2) The time effect of the spatial distribution of Taiwan investment in Mainland China gradually increases; therefore, rational methods should be used to develop investment strategies.
Twenty years of investment data show that changes in cross-strait relations are an important factor in Taiwan investment in Mainland China. The pattern of Taiwan investment in Mainland China shows clear time-fixed effects. However, if the space-fixed effects are combined, we find that Taiwan investment in Mainland China is no longer solely the result of perceptual judgments of regional distribution; instead, rational, objective investment factors affect Taiwan. Recent years have been marked by improved openness of the global investment environment, increased popularity of the information network platform, expansion of communication channels, and changes in investment philosophy. A company’s outside investment is not only a purely economic behavior; it has slowly evolved into a fusion behavior of psychology, information, systems, and policies. Therefore, it is only dependent on the investment of “lowest cost” or “acquaintances”, and investment does not seem to have become less important. It is supplemented by more factors from the investor’s perspective and gradually affects the investor’s success. In this regard, the spatial distribution of Taiwan investment in Mainland China relies on long-term, cumulative study of the changes in the investment environment to scientifically and rationally develop investment strategies; it is not possible to immediately develop regional investment strategies.
3) In addition to the supply of the regional elements, Taiwan investment in Mainland China also needs to consider the industrial areas that match the regional elements.
Currently, Taiwan’s leading investment in the mainland is in the electronics and information industry; other high-end manufacturing and modern service industries are not significant. Therefore, in this study, some significant indices that affect Taiwan’s investment in the mainland are important for the manufacturing industry, especially the electronic information manufacturing industry. Nevertheless, these indicators may not affect the future of the factors of the Taiwan business pattern. If more Taiwan investment in Mainland China goes to the service industry or other areas of the manufacturing industry, perhaps the regions that do not have significant spatial characteristics will become evident.
After ECFA, an important aspect of the cross-strait industrial negotiation is the possible areas and mechanisms of industrial cooperation. In this context, Taiwan entrepreneurs can consider whether the supply of the environmental factors of an area meets the needs of the overall Taiwan business at this stage, convert the industry to its own operation, and choose to invest in the industrial areas that match the regional elements, or a business can invest in different areas. For example, in the absence of an environmental support platform, but with the advantage of the region of the upstream manufacturing industry, Taiwan investment may consider introduceing production services (such as R & D center, testing center, and financial platform) in the areas of developed service, rich capital. However, the technological innovation capacity needs to be improved, extending the manufacturing industry and requiring it to swim upstream to R & D and downstream to market development and brand development areas.
4) Regions must start by improving the conditions of employment, increasing the education level, strengthening the legal system, and enhancing regional industry chain integrity to start attracting Taiwan investment.
Currently, the regional governments of the mainland are concerned about the development of Taiwan investment. However, the unavoidable regional differences present a natural barrier, which leads to spatial heterogeneity of Taiwan businesses in Mainland China. In this situation, each region should start with its own specific characteristics to improve the investment environment conditions and accumulate the stock elements. However, as a whole, the employment environment, education level, legal conditions, and the integrity of the industry chain of any one region will have a long-term impact on FDI. These factors and conditions cannot be changed overnight. They must be addressed early and persistently. From an economic sense, the improvement of these elements is a process of increasing returns to scale. The returns may not be seen at the beginning of the investment, but with the increase in the stock investment, the effect will appear gradually.
In this regard, the construction of local job market through improvement of the educational institutions to enhance the level of human capital stock, increasing efforts to create a safe and legal environment for business, and assembly of a complete upstream and downstream industry chain, which will fundamentally change the embarrassing situation of “hard labor, lack of talent, fear for infringement, less foreign aid”, will increase the chances of success for the regions that attract Taiwan investors.
In cross-strait industrial cooperation, the pattern of Taiwan investment in Mainland China has shown a certain degree of heterogeneity and spatial dependence that is not strong, yet is doubly affected by time and spatial effects. The national “ten two five” plan aimed to “strengthen and improve inter-regional cooperation mechanism, remove market barriers to promote factors mobility, and guide the orderly transfer of industry”, which emphasizes regional cooperation. Furthermore, the concept of coordinative development has become the primary regional guiding strategy in the current economic map of Mainland China. Therefore, this strategic thinking should be integrated into the pattern of Taiwan investment in the mainland, and focus should be placed on improving the quality of Taiwan investment, balancing the proportion of Taiwan-funded regions and emphasizing cooperation to avoid the competition and rent-seeking of different areas that attract Taiwan investment. This information will be important after the subsequent implementation of the cross-strait economic cooperation framework agreement (ECFA), which is also an important mechanism for promoting cross-strait industrial cooperation.