Empirical results

This study employs the mathematical programming software LINGO 11.0 to calculate the optimal values of intermediate measures and value of technology efficiency (TE) with different areas. Due to the different degrees of regional development in China, we separate a country in three areas: Eastern, middle and western. Besides, the effect of financial crisis extended around the world in 2008. Hence, we compare the relationship between the before and after financial crisis.

We find that eastern is better than western and middle in economic TE before and after financial crisis on the means of TE in Table 3. In environmental TE, eastern is better than middle and eastern is better than western before financial crisis. In social TE, eastern is better than middle and western is better than middle. The highest value in three dimensions are in eastern regions, and the lowest value are in middle area.

Table 3 Mean of technology efficiency (TE) with different areas Area

Mean Eastern Middle Western Total

TE1 2006-07 0.80642 0.30015 0.49593 0.59132

(Economic) 2008-09 0.76428 0.29659 0.50342 0.57509

TE2 2006-07 0.80669 0.58075 0.65442 0.70567

(Environmental) 2008-09 0.77589 0.60014 0.66451 0.69990

TE3 2006-07 0.94638 0.74939 0.80357 0.85462

(Social) 2008-09 0.90094 0.72050 0.76566 0.81525

Average Efficiency 2006-07 0.85316 0.54343 0.65131 0.71720 2008-09 0.81370 0.53907 0.64453 0.69675

The output variables separate into economic, environmental and social dimensions.

Hence, efficiency in one region can be measured into these three dimensions. The empirical result in Table 4, at 10% level of significance, all dimensions of technology efficiency (TE)

25

among three areas are significantly different except environmental TE in 2008-09. It indicates that the environmental TE among three areas are the same. We further use the Mann Whitney U test to investigate whether the TE for three dimensions between any two of areas are the same. (Wackerly et al., 2002) The result in table 4 indicates that eastern with middle and eastern with western are significant in TE for three dimensions before financial crisis while middle with western is significant in economic and social TE. However, after financial crisis, environmental TE is not significant in three areas.

Table 4 Nonparametric Test of technology efficiency (TE) for different areas

Kruskal Wallis Test 2006-07 2008-09

Technology Efficiency (TE) Chi-Square df P-Value Chi-Square df P-Value TE1(Economic) 25.909^＊＊＊ 2 <0.001 22.424^＊＊＊ 2 <0.001

TE2(Environment) 8.192^＊＊＊ 2 0.017 4.090 2 0.129

TE3(Society) 10.338^＊＊＊ 2 0.006 8.277^＊＊＊ 2 0.016

Average Efficiency (AveE) 21.981^＊＊＊ 2 <0.001 17.159^＊＊＊ 2 <0.001

Mann-Whitney U Test 2006-07 2008-09

Z P-Value Z P-Value

TE1 E vs. M ^a -4.537^＊＊＊ <0.001 -4.244^＊＊＊ <0.001

(Economic) E vs. W ^a -3.529^＊＊＊ <0.001 -3.143^＊＊＊ 0.002

M vs. W ^a -2.595^＊＊＊ 0.009 -2.451^＊＊ 0.014

TE2 E vs. M ^a -2.723^＊＊＊ 0.006 -2.051 0.040

(Environmental) E vs. W ^a -2.097^＊＊ 0.036 -1.286 0.198

M vs. W ^a -0.578 0.563 -0.509 0.611

TE3 E vs. M ^a -3.355^＊＊＊ 0.001 -2.579^＊＊ 0.010

(Social) E vs. W ^a -1.918^＊ 0.055 -2.205^＊＊ 0.027

M vs. W ^a -0.958 0.338 -0.727 0.467

AveE E vs. M ^a -4.059^＊＊＊ <0.001 -3.613^＊＊＊ <0.001

E vs. W ^a -3.437^＊＊＊ 0.001 -2.664^＊＊＊ 0.008

M vs. W ^a -2.162^＊＊ 0.031 -2.487^＊＊ 0.013

Note: (1) ^a E: Eastern, M: Middle and W: Western.

(2) ^＊,^＊＊, and ^＊＊＊represent the 10%, 5%, and 1% levels of significance, respectively.

26

We find that eastern is higher than western and middle higher than western on the means of optimal values in Table 5. Moreover, we find all optimal values have increased after financial crisis except labors in eastern. It indicates the regional development can go toward more capital intensity.

Table 5 Mean of optimal value with different areas Area

Optimal value Year Eastern Middle Western Total

Labor 2006-07 504.4943 469.9397 240.1388 400.6530 2008-09 501.0712 494.1548 243.0013 405.0623 Energy 2006-07 10,610.6086 9,489.1526 5,846.4162 8,639.4468 2008-09 11,129.2505 9,950.8114 6,340.7490 9,137.7788 Capital 2006-07 25,939.0545 22,091.3891 11,181.0264 19,758.2444 2008-09 34,530.8578 27,961.2297 14,704.3836 25,947.2250 Funds from 2006-07 1,736.9962 1,176.1267 500.2681 1,171.3553 HMT 2008-09 1,795.4678 1,423.8260 569.7545 1,271.7112 Foreign 2006-07 2,725.9184 2,285.7091 963.4675 1,991.6445 Funded 2008-09 3,634.1349 2,552.9975 1,218.7948 2,532.2827

This study wants to find the optimal value of intermediate measures to achieve efficiency in one region. Table 6 shows that the optimal values of origins of FDI in different areas are significantly different by Mann Whitney U test. It indicates that the optimal value of funds from HMT is different from foreign funded in three regions separately.

27

Table 6 Mann-Whitney U Test of optimal value of origins of FDI

Mann-Whitney U Test 2006-07 2008-09

Area Z P-Value Z P-Value

Eastern -2.855^＊＊＊ 0.004 -4.278^＊＊＊ <0.001

Middle -2.944^＊＊＊ 0.003 -2.309^＊＊ 0.021

Western -2.418^＊＊ 0.016 -2.723^＊＊＊ 0.006

Total -3.842^＊＊＊ <0.001 -4.278^＊＊＊ <0.001

Note:^＊,^＊＊, and ^＊＊＊represent the 10%, 5%, and 1% levels of significance, respectively.

The result in table 7, we find that optimal values of funds from HMT and foreign funded are all significant at 1% level of significance. It means that there are different optimal values of intermediate measures in different areas, then we can support hypothesis 1 the optimal demands for origins of FDI in different regions are various. In energy

consumptions, it support hypothesis 2 the optimal energy consumption volumes in different regions are various. We test any two of areas by Mann Whitney U test in table 7. We find

that eastern with middle is not significant in all optimal values before and after financial crisis. It means that there does not exist the difference of optimal values in eastern and middle. Nevertheless, eastern with western and middle with western are significant in all optimal values before and after financial crisis. It indicates that there exists the difference of optimal values in eastern with western and middle with western.

28

Table 7 Nonparametric Test of optimal value for different areas

Kruskal Wallis Test 2006-07 2008-09

Optimal Value Chi-Square df P-Value Chi-Square df P-Value

Labor 20.775^＊＊＊ 2 <0.001 21.386^＊＊＊ 2 <0.001

Energy 14.246^＊＊＊ 2 0.001 11.811^＊＊＊ 2 0.003

Capital 21.191^＊＊＊ 2 <0.001 21.634^＊＊＊ 2 <0.001

Funds from HMT 19.257^＊＊＊ 2 <0.001 21.901^＊＊＊ 2 <0.001 Foreign funded 19.927^＊＊＊ 2 <0.001 22.208^＊＊＊ 2 <0.001

Mann-Whitney U Test 2006-07 2008-09

Optimal Value Z P-Value Z P-Value

Labor E vs. M ^a -0.597 0.551 -0.157 0.875

E vs. W ^a -3.952^＊＊＊ <0.001 -3.890^＊＊＊ <0.001 M vs. W ^a -3.820^＊＊＊ <0.001 -4.072^＊＊＊ <0.001

Energy E vs. M ^a -0.471 0.638 -0.220 0.826

E vs. W ^a -3.249^＊＊＊ 0.001 -2.959^＊＊＊ 0.003

M vs. W ^a -3.207^＊＊＊ 0.001 -2.919^＊＊＊ 0.004

Capital E vs. M ^a -0.628 0.530 -0.722 0.470

E vs. W ^a -3.952^＊＊＊ <0.001 -3.993^＊＊＊ <0.001 M vs. W ^a -3.928^＊＊＊ <0.001 -3.964^＊＊＊ <0.001

Funds from E vs. M ^a -0.722 0.470 -0.188 0.851

HMT E vs. W ^a -3.952^＊＊＊ <0.001 -4.304^＊＊＊ <0.001 M vs. W ^a -3.388^＊＊＊ 0.001 -3.496^＊＊＊ <0.001

Foreign E vs. M ^a -0.220 0.826 -0.911 0.362

funded E vs. W ^a -3.952^＊＊＊ <0.001 -4.407^＊＊＊ <0.001 M vs. W ^a -3.604^＊＊＊ <0.001 -3.279^＊＊＊ 0.001 Note: (1) ^a E: Eastern, M: Middle and W: Western.

(2) ^＊,^＊＊, and ^＊＊＊represent the 10%, 5%, and 1% levels of significance, respectively.

In this study, we employed two-stage DEA approach to find the optimal value of intermediate measures. We find that the mean of difference of origins of funds in middle is higher than eastern and western in table 8. Additionally, we find that the difference in

29

energy consumptions is more serious in three areas. Although the optimal of energy consumptions is increased, the volume of real use is more than optimal values. Hence, the difference is larger. Funds from HMT and foreign funded have to invest more in middle and western, and funds from HMT have to decrease in eastern in 2008-09.

Table 8 Mean of optimal value minus observations with different areas Area

Difference Year Eastern Middle Western Total

Labor 2006-07 -0.5326 31.5980 -1.5612 5.5164

2008-09 -30.0750 45.3548 -8.2941 -7.0027 Energy 2006-07 -1,745.8844 -1,126.5050 -724.2719 -1,247.4172 2008-09 -2,405.0351 -2,206.5532 -1,402.4951 -1,997.7408 Capital 2006-07 -92.1423 4,193.7757 1,238.3056 1,252.8722 2008-09 -18.5381 1,788.3932 321.0125 467.3500 Funds from HMT 2006-07 128.8827 695.8213 329.7275 315.9135 2008-09 -311.7412 744.5846 301.3701 124.3314 Foreign Funded 2006-07 327.4543 1,738.8292 750.3150 764.7782 2008-09 629.2893 1,820.6963 911.6569 971.1055

We further test the means of optimal values minus observations. The difference between them in table 9 indicates that labor and energy consumptions are not significant in three areas before and after financial crisis. It means that the improved values of labor and energy consumptions are no difference in three areas. The difference of capital is significant different before financial crisis, and after financial crisis, it is not significant. It indicates that the improved value in three region tend to close. The difference of funds from HMT and foreign funded are significant in three areas before and after financial crisis. It indicates that the improved value of funds from HMT and foreign funded are different in three areas. We further use the Mann Whitney U test to investigate whether the difference between any two of areas are the same in table 9. The result shows that any two of the areas are significantly different for different origins of funds.

30

Table 9 Nonparametric Test for means of optimal values minus observations for different areas

Kruskal Wallis Test 2006-07 2008-09

Difference Chi-Square df P-Value Chi-Square df P-Value

Labor 2.225 2 0.329 2.854 2 0.240

Energy 0.186 2 0.911 0.152 2 0.927

Capital 11.302^＊＊＊ 2 0.004 2.155 2 0.340

Funds from HMT 17.394^＊＊＊ 2 <0.001 11.467^＊＊＊ 2 0.003 Foreign funded 25.219^＊＊＊ 2 <0.001 14.859^＊＊＊ 2 0.001

Mann-Whitney U Test 2006-07 2008-09

Difference Z P-Value Z P-Value

Labor E vs. M ^b -1.346 0.178 -1.612 0.107

E vs. W ^b -0.420 0.674 -0.177 0.860

M vs. W ^b -1.297 0.194 -1.441 0.149

Energy E vs. M ^b -0.604 0.546 -0.190 0.849

E vs. W ^b -0.104 0.917 -0.146 0.884

M vs. W ^b -0.036 0.971 -0.541 0.589

Capital E vs. M ^b -3.414^＊＊＊ 0.001 -1.363 0.173

E vs. W ^b -1.173 0.241 -0.333 0.739

M vs. W ^b -2.234^＊＊ 0.025 -1.261 0.207

Funds from E vs. M ^b -3.878^＊＊＊ <0.001 -2.750^＊＊＊ 0.006

HMT E vs. W ^b -1.944^＊ 0.052 -2.255^＊＊ 0.024

M vs. W ^b -2.955^＊＊＊ 0.003 -2.342^＊＊ 0.019 Foreign E vs. M ^b -4.355^＊＊＊ <0.001 -3.423^＊＊＊ 0.001

funded E vs. W ^b -2.968^＊＊＊ 0.003 -2.040^＊＊ 0.041

M vs. W ^b -3.460^＊＊＊ 0.001 -2.775^＊＊＊ 0.006 Note:(1) ^b E: Eastern, M: Middle and W: Western.

(2) ^＊,^＊＊, and ^＊＊＊represent the 10%, 5%, and 1% levels of significance, respectively.