3.1. Heterogeneity of Foreign Directed Investment (FDI) Dynamic Performance in China
Dynamic FDI performance in China is denoted as dynamic total factor productivity (dynamic Malmquist productivity index, DMPI) in Figure 2
. The diachronic results from 1997 to 2013 indicate that dynamic FDI performance exhibits a rising trend with fluctuations in China. The fluctuations subsequently became less pronounced after 2004. Compared with the benchmark of 1, DMPI was 1.08 in China during the whole period. This indicates a small progress of dynamic FDI performance in China. To obtain a diachronic perspective, the whole period is divided into two stages, with 2004 as a boundary. In the first stage, from 1997 to 2004, DMPI exhibits a “W”-shaped curve with up-and-down cycles for four years. The two slight rises indicate that DMPI in 2004 peaked for the whole period. Subsequently, the DMPI curve became relatively flat with a slight decline. Additionally, DMPI in 2012 was less than 1. Although DMPI increased in 2013, the trend remained declining overall. The trend of DMPI indicates dynamic FDI performance. This means the sustainability of evaluation variables does not correspond with the economic development. Therefore, it is necessary to analyze the bottlenecks of dynamic performance development in China.
As shown in Figure 2
, both DTEC and DTC are determinants of DMPI. Therefore, the attractive power is denoted by DTEC while the potential power is denoted by DTC. The two indices are critical to the total factor productivity of FDI in China. However, the influences of the indices are significantly different. In the first two years, DTEC substantially contributed to DMPI, whereas DTC became the major factor for improvement in DMPIs after 2000. This suggests that current changes in potential power, such as employee quality and advanced management, have been a big step forward for provincial FDI performance. It is dangerous to make effort to attract a single performance. This means all the resources are put into a single operation without considering the subsequent influence. This retrogression means the inter-temporal effect is gradually ignored. This ignored phenomenon leads provincial FDI to give up opportunities in the long run for the special benefit in the short run performance.
Thus, the dominant determinant of FDI performance is quality efficiency improvement with the promotion of technical development. The allocation efficiency and scale efficiency contributed to the stable basic in the first two years. However, DTEC was lower than DMPI after 2000. This implies that the allocation efficiency and scale efficiency change into a constraint to performance improvement. For example, DTEC in 2000 tended to decrease. Additionally, that in 2009 was less than 1, which means to retrogress. With the policy bonus of China, such as Western Development in 2000 and the Four Trillion Yuan Stimulus Plan in 2009, decision makers reinforced the government investment to attract FDI. However, they have focused on the quantity and ignored the quality without control of variable efficiency.
DTEC also constrains dynamic FDI performance, as shown in Figure 2
. DTEC can be further composed into DPTC (dynamic pure technology efficiency change) and DSEC (dynamic scale efficiency change) as shown in Figure 3
. This implies that the allocation efficiency of FDI variables is crucial for the attractive power. Based on the decomposition of the Malmquist model, the allocation efficiency (DTEC) can be further decomposed into technology efficiency (DPTC) and scale efficiency (DSEC). DPTC in 2000 initially tended to decline, combined with a continued declining tendency of DSEC since 2000. This result is consistent with those of previous studies on allocation efficiency in 2000, with the influence of Western Development policy in China.
The provincial diachronic result from 1997 to 2013 highlights the differences in dynamic FDI performance in China. As shown in Figure 4
, the deeper color province indicates the higher FDI performance. The different color provincial boundary shows the different regions in China. The white boundary is midland in China. The brighter boundary is the western province in China and the deeper one is the eastern province in China. Using a regional perspective, the advantaged provinces always belong to the western region. All the efficiency indices are greater than 1. This indicates that provincial FDI performance always progresses during the evaluation periods. The improvement of dynamic FDI performance in the western provinces was facilitated by the policy bonus of Western Development in China.
The first three advantaged provinces and the last three disadvantaged provinces are shown in Table 3
according to the performance indices. For the decomposed efficiency indices, DTC significantly enhances dynamic FDI performance. This result proves that an indirect constraint on FDI performance is the attractive power. Therefore, the effective use of evaluation variables is the solution to promoting performance. For example, Xinjiang and Guangxi in the western region are disadvantaged provinces of dynamic FDI performance. The decomposition is used in order to identify the critical restriction variables. First, from the allocation efficiency perspective, all the advantaged provinces are good at DPTCs in the western region, such as Inner Mongolia, Ningxia, and Guizhou. However, the disadvantaged provinces are in different regions, such as Fujian in the eastern region, Heilongjiang in the midland region, and Xinjiang in the western region. Thus, regional differences are not a critical constraint. This proves the proportionality of macro-control. Additionally, from a perspective of scale efficiency, the advantaged provinces of DSECs are Qinghai, Jiangxi, and Guizhou. While the disadvantaged provinces are Guangxi, Hainan, and Gansu.
In conclusion, the solution of dynamic FDI performance improvement is the combined promotion of multiple decomposed efficiency indices. Additionally, policy bonus is beneficial to FDI performance. However, the key is realizing the sustainability of the effects from the policy bonus.
3.2. Hierarchical Clustering of Provincial FDI Performance
To identify the interprovincial cooperative effect for dynamic FDI performance, all the provinces are classified using hierarchical clustering. For the previous calculation by the dynamic Malmquist model, the different performances of provincial FDI include the influence of the regional resource endowments. Therefore, three decomposed efficiency indices (DTC, DSEC, and DPTC) with the regional heterogeneity are the basics for the cluster analysis. Additionally, several other methods are tested including the Ward-method based on the theory of hierarchical cluster analysis. Only the furthest neighbor distance method can classify the discriminations of provincial FDI performance. For example, based on the Ward-method, cluster one includes 16 provinces which are more than one half of Chinese provinces. Both of cluster two and cluster four include three provinces. So, the furthest neighbor distance method is more suitable for this empirical analysis compared with other methods.
Based on the provincial regional resource endowment, all the provinces in China are classified into four clusters, as shown in Figure 5
. From a perspective of regional divide, the western provinces are classified into cluster two and cluster three. The midland provinces are classified into cluster three and cluster four, while the eastern provinces are classified into cluster one and cluster two. The regional influence is significant for the decomposed efficiency indices of FDI performance.
Based on cluster analysis, the average decomposed indices and the corresponding rankings are shown in Table 4
. The potential power effect is denoted by DTC. The allocation effect is denoted by DPTC. While, the scale effect is denoted by DSEC. Based on the thinking of the BCG Matrix (Boston consulting group matrix), the four clusters can be named as Dog group, Question Mark group, Cash Cow group, and Star group.
Cluster one has an absolute advantage for the first rank in DPTC. The allocation efficiency of FDI variables is the basis of the advantage. The dynamic FDI performance of cluster one is lower than that of the other clusters. Although cluster one is good at the basic efficiency index as a location of all the variables. The average of DTCs is less than 1. Therefore, it is a bottleneck (DTC = 0.977) for the sustainability development. The lowest rank of DTC indicates the absence of potential power control, such as sustainability of policy bonus and incorporation of advanced management. From a perspective of provinces in China, the municipalities, such as Beijing and Shanghai, belong to cluster one. Their decomposed efficiency indices exhibit a relatively flat trend around the benchmark of 1 during the whole period. The municipalities attract FDI at earlier years. They also had a higher degree of openness. Therefore, the attractive conditions of FDI are relatively mature. Nevertheless, the policy bonus is weaker than before. So, the performance of municipalities gradually regresses. Because of this bottleneck, FDI improvement in these regions is much more difficult than that of other regions. Inner Mongolia is a special province in cluster one for exhibiting an alternate trend. At the beginning of the 1990s, Inner Mongolia faced difficulties to attract FDI. The DTEC improved rapidly because of the lower basic status. Therefore, an absolute advantage in the allocation efficiency of evaluation variables makes Inner Mongolia a leader in the western region. Additionally, the regional advantage is beneficial to FDI because Inner Mongolia borders Central Asia with more policy bonus of One Belt One Road.
With a leading position in allocation efficiency, the main evaluation criterion presents profit growth. So, the provincial FDIs in cluster one are similar to the Dog group based on the thinking of the BCG Matrix. Based on the efficiency indices, the strategy of cluster one is stabilization development. The most efficient breakthrough for improving performance is keeping the advantage of pure economic growth in traditional evaluation. For more economic development, this strategy is to give full advantage of inter-temporal activity. This influence cannot be ignored.
Cluster two has an absolute advantage in DTC for the first rank in dynamic performance. These provinces take advantage of the sustainability for an advantage of potential power. Nevertheless, the average of DSECs (0.975) is a major concern. The retrogressive scale effect hinders the performance development. For example, the DSEC of Gansu (DSEC = 0.913) represents a typical province in cluster two. With the policy bonus of China Western Development, the DTPC of Gansu (DTPC = 1.9758) is much higher than other provinces in cluster two. Nevertheless, Gansu cannot use the extensive assistance provided which results in the scale effect problem.
With the obvious advantage of potential power, the strategy of cluster two is expansion. The most efficient breakthrough strategy for improving scale efficiency growth in a long run performance. Based on the thinking of the BCG Matrix, cluster two can improve the performance to strengthen a leading position as a typical Star group. Considering the sustainability, on one hand, cluster two should evaluate the inter-temporal efficiency changes of carry-over activity, such as physical capital stock in a long run development. On the other hand, cluster two should continue to consider attractive power strengthen to remedy the pure economic data without considering the completeness of outputs. In conclusion, the traditional advantages with a sustainability point of view can keep cluster two in a top position.
Cluster three has a disadvantage in DPTC which is the basic efficiency index for performance. This represents the allocation efficiency of FDI variables. The retrogressive DPTC influences other efficiency indices. Another threat is the lack of an absolute advantage according to the results of decomposed efficiency indices. For example, similar to Xinjiang, Heilongjiang has a regional advantage because it borders Central Asia. Heilongjiang has capitalized on the opportunities of One Belt and One Road. Additionally, the allocation of evaluation variables is remarkably inefficient. Cluster three is the largest cluster with generality concerns. Most of the provinces blindly pursue the quantity increase. This means, emphasis on quantity while neglecting quality constitute a bottleneck for FDI performance.
The strategy of cluster three should be retrenchment. This means they should gather strength to break through in order to avoid changing into a typical Dog group. The most efficient breakthrough for improving performance in this cluster is the control of carry-over activities. This means that strengthened control of carry-over activity, such as physical capital stock, can enhance dynamic FDI performance. The specific method is incorporating the inter-temporal influence into current incentive mechanisms. The embodiment is that the evaluation has not only focused on one period, but also on the inter-temporal period. The special dynamic variables are multiple, such as inter-temporal benefit increasing rate, average daily balance, and average daily contribution. This breakthrough can not only increase the advantage, but also match the strategy of sustainable development.
Cluster four has an absolute advantage in DSEC. Additionally, all the other efficiency indices exhibit an increasing trend. This implies that the indices are higher than 1. Because of the inadequate DTCs, the performance of cluster four is in the last ranking. This means the dynamic FDI performance of cluster four is the lowest. However, Yunnan is an exception to cluster four for an absolute advantage of DTEC. The investors in Yunnan are interested in the objective factors as labor-intensive and low-technology industries. For example, Yunnan has lower price levels, lower energy thresholds, and lower human capital stock. Policy bonus has resulted in considerable improvement on the dynamic FDI performance of Yunnan. However, from a diachronic perspective, the potential power of dynamic FDI performance in Yunnan is still lower than that of other provinces.
With an advantage position for the scale efficiency, cluster four sets a better example for other clusters. However, all the provinces in cluster four still have problems that need to be solved urgently with the traditional pure economic competition. The strategy of cluster four is reinforcement. The most efficient breakthrough strategy is modernizing the management to develop profit markets rapidly to catch up with other competitors. For example, using think-tank-management-mode transforms traditional pure economic benefit into traditional pure economic efficiency for a higher traditional index, such as TC (technical change) and PTC (pure technology change). If cluster four systematically selects strategies to pursue pure economic growth according to the strength and advantage, cluster four will change into the Star group.
3.3. Contribution Analysis of FDI Evvaluation Variables
Dynamic FDI performance in China is affected by the potential power, allocation, and scale of variables. This analysis focuses on the contribution of FDI evaluation variables. Additionally, the variables of Chinese provinces are evaluated using the non-radical and non-oriented slacks-based measure model.
The diachronic results are divided into two stages basing on the different trends in dynamic FDI performance. At the first stage, the average of DMPI exhibits a “W”-shaped curve with up-and-down cycles for four years. At the second stage, the average of DMPI flattens relatively with a slight decline. The average of DMPI declines to lower than 1. This indicates performance regression in 2012. Therefore, three periods of the first stage, the second stage, and 2012 are compared to reveal the different contributions of evaluation variables. The evaluation variables are human capital stock, energy consumption, the export rate, and physical capital stock. The contributions are relatively different during performance evaluation. From a regional perspective, the different contributions of Chinese provinces are shown as in Figure 6
, Figure 7
and Figure 8
. The cumulative percentages of contributions are plotted in a cumulative histogram with the ranking from left to right which indicates those of the first stage, the second stage, and 2012. The provinces are ranked according to the average contribution of evaluation variables.
The contribution scores of evaluation variables in the eastern region are the highest of three regions in China. As shown in Figure 6
, the contribution scores range from 0.43 to 1. The contributions of physical capital stock and export rate provide an absolute advantage in FDI performance, whereas energy consumption rate is relatively lower. Human capital stock has the lowest contribution scores. It is useful in few municipalities and provinces, such as Shanghai and Guangdong.
During the whole period, the contribution of physical capital stock is considerably weak, whereas that of the export rate strengthens. The contribution of human capital stock has an increasing trend. When dynamic FDI performance declined in 2012, the contribution of physical capital stock was gradually replaced by that of the export rate. The three municipalities, including Shanghai, Tianjin, and Beijing, have absolute advantages in FDI performance because of the higher contribution scores. From a contribution perspective, the contribution of the export rate has an absolutely increasing trend. While, that of physical capital stock has a significantly decreasing trend.
The contribution characteristics of Hainan are relatively similar to that of the three municipalities. For the economy of the Beijing-Tianjin-Hebei metropolitan region, the contribution of physical capital stock is weak. Even the region has a decreasing trend. Compared with the decreasing trend, the contribution scores of export rate increased substantially. The Yangtze River Delta, the Pearl River Delta, and the Pohai Economic Circle have developed with the new industrialization. Therefore, the contribution of physical capital stock is advantageous, though the increase trend is slow.
The contribution scores of evaluation variables in the midland are between the western and eastern in China. As shown in Figure 7
, the range of contribution scores was from 0.29 to 0.88. The contribution of the export rate provides an absolute advantage in FDI performance, whereas that of physical capital stock was relatively lower. Energy consumption and human capital stock have the lowest contribution scores. Additionally, the contribution of energy consumption and human capital stock in the midland were also weaker than that of the eastern region.
During the whole period, the contribution had a stable trend. However, the maximum fluctuations were caused by energy consumption, although the contribution of energy consumption increased from 2% to 3%. In contrast to the variance rate of evaluation variables, the contribution of human capital stock changes was much bigger, with a variance rate of 33%. The variance rate of the export rate and physical capital stock were 1% and 3%, respectively. When dynamic FDI performance declined in 2012, the contribution trend in the midland was different from that in the eastern region. This implies that the contribution of export rate was replaced by that of physical capital stock. From a provincial perspective, Jiangxi, Hubei, and Hunan, which border the western region, have an absolute advantage in FDI because of their higher contribution scores (0.9, 0.7, and 0.6, respectively). However, Anhui and Shanxi, which border the eastern region, have lower contribution scores (0.4 and 0.3, respectively). Thus, the midland does not receive a free ride from the eastern region for FDI performance improvement.
The contribution scores of evaluation variables in the western region were the lowest among the three regions in China. As shown in Figure 8
, the range of contribution scores was from 0.09 to 0.8. The contribution of export rate provides an absolute advantage in FDI performance. Whereas those of energy consumption and human capital stock were relatively lower.
During the whole period, the contribution trends of evaluation variables changed significantly. The contribution of physical capital stock was replaced by that of export rate in the western region, except Guangxi and Yunnan. When dynamic FDI performance declined in 2012, the contribution trend was relatively similar to that of the eastern region. This implies that the contribution of physical capital stock was replaced by that of export rate. From a provincial perspective, the contribution of export rate was particularly high in Inner Mongolia, Qinghai, Gansu, and Ningxia. However, the contribution scores of western provinces are differences in the distribution. For example, the contribution scores of Inner Mongolia, Chongqing, and Shaanxi were 0.8, 0.6, and 0.6, respectively. Whereas those of Ningxia and Guizhou were both 0.1. Thus, the control of evaluation variables in different provinces was relatively diverse. The western provinces have dual characteristics for FDI performance. Inner Mongolia, as a representative of the western region, has the advantages of a lower consumption level, more energy resources, and lower environmental thresholds, whereas non-negligible disadvantages are lower performance in FDI for a higher input demand.