5.5 Empirical analyses
5.5.2 Ownership influences
dependent variable, Eff(re_level)it, is applied to both Model S1 and Model J1.
Model S1 and Model J1 only measure the influence of bank ownership (STATE or JOINT) with year dummies. Policy banks are neither attached to the category of state-owned banks nor the group of joint-stock banks, because of its disputed role and outstanding cost efficiency discussed in previous chapters. That is to say, Model 1s only discusses a single influence factor of bank ownership, wherein the author treats the policy banks and joint-equity banks as a control group in Model S1, and treats the policy banks and the Big Five state-owned banks as a control group in Models J1. The state-owned majority ownership shows a positive significance to cost efficiency. Moreover, we have not measured the ownership influence of policy banks with other variables or separately because of multi-collinearity. The influences of majority ownerships to cost efficiency are estimated. And the empirical results are shown in Table 5.15.
.
Table 5.15: Cost Efficiency Influence Factor of Majority Ownership (2002-2009)
Independent Variables
Dependent Variables Eff(re_level)it
Model S1 (Fixed Effect)
Model J1 (Fixed Effect)
STATE 0.015***
(3.000) Omitted
JOINT Omitted -0.015***
(0.005) YEAR 2002 -0.012***
(-4.64)
-0.012***
(0.003) YEAR 2003 -0.010***
(-3.94)
-0.010***
(0.003) YEAR 2004 -0.022***
(-9.34)
-0.022***
(0.002) YEAR 2005 -0.017***
(-7.27)
-0.017***
(0.002) YEAR 2006 -0.014***
(-6.03)
-0.014***
(0.002) YEAR 2007 -0.009***
(-3.91)
-0.009***
(0.002) YEAR 2008 -0.005**
(-2.10)
-0.005**
(0.002)
Constant 0.890***
(423.79)
0.902***
(0.003)
N 134 134
Clusters 20 20
F-statistics 470.31*** 779.33***
R-square 0.184 0.033
Corr(u_i, Xb) -0.524 -0.305
Hausman test Non-positive 30.14***
Note: This table shows the panel regression results of cost efficiency by Stata 11.0SE, with the standard error of the coefficients shown in the parentheses. ***, **, and * are significant at 1, 5, and 10 percent significance levels, respectively. The null hypothesis of the Hausman specification test in this table is that the systematic difference of individual-level effects measured by a fixed-effects model or random model is insignificantly different. And when the Hausman test is non-positive, the fixed-effect model will be selected.
The coefficients of all the independent variables are significant in Model 1s, which infers that the ownerships of both state and joint stock significantly affect the banking cost efficiency. However, the signs of the ownership variable coefficients are different. The coefficients of STATE are significantly
positive, while the coefficients of JOINT are significantly negative to the same relative level-dependent value, Eff(re_level)it, at a similar 1% significance level.
It seems opposite to the results of Table 5.13 and Table 5.14, but coincides with Table 5.1. Combing the results of Table 5.13, Table 5.14, and Table 5.15, an inference of majority ownership of state-owned and non-state-owned is deduced. The state ownership is propitious to cost-efficiency ‘improvement’ of the whole industry (including policy banks, the Big Five banks, and the joint-stock banks in this thesis), but the state-owned banks have the lowest efficiency rank in the industry. This ‘improvement’ is probably caused by the big asset share of the Big Five banks. However, Table 5.1 and Table 5.4 weaken this cause, for the higher-than-average industrial profitability of the Big Five banks. And whether this ‘improvement’ is real progress or operated by government intervention will be discussed in the following government intervention section. Regardless of the possible ‘improvement’ method, the state-owned majority ownership significantly promotes the cost efficiency of the banking industry.
Since the restriction relaxation agreement with WTO, debates about influences of foreign banks or foreign shareholders of local banks have prevailed. But until writing this thesis, the foreign capital in the banking industry, either invested in the foreign banking institutions or as a minority ownership of the domestic banks, is still very small. In the former cost efficiency rank section, we have already established that the three banks of lowest cost efficiency have foreign participators. Whether the foreign investment is insignificant to cost efficiency is the study goal of this section.
Owing to most banks lacking foreign shareholders, pooled data of individual banks is employed for foreign minority ownership measurement.
The influence of foreign shareholders is firstly applied to a full sample scope (see Table 5.16, Model F_p and Model F_s), then to the state-owned banks (see Table 5.16, Model S_p and Model S_s), and the joint-stock banks (see Table 5.16, Model J_p and Model J_s). No foreign shareholder information is reported within policy banks by the former referred data source. Therefore, the
author only separately investigates the influence of foreign shareholders within the state-owned banks and joint-equity banks in the measurement of both percentage value and dummy. The dependent variables of all the modes in this section are similarly Eff(re_level)it.
Table 5.16: Cost Efficiency Influence Factors of Ownerships (2006-2009)
Independent Variables
Dependent Variables Eff(re_level)it Model
F_p
Model F_s
Model S_p
Model S_s
Model J_p
Model J_s Sample
Scope Full sample Big Five banks Joint-stock banks STATE -0.097***
(-5.72)
-0.163***
(-10.70) JOINT Omitted -0.058***
(-4.28) FOREIGN_P -0.052
(-0.49)
-0.127*
(-1.91)
-0.042 (-0.29)
FOREIGN_S 0.034***
(2.89)
0.048***
(4.58)
0.031*
(1.77) YEAR 2006 0.007
(0.31)
-0.012 (-0.81)
-0.017 (-1.56)
-0.024*
(-1.72)
0.022 (0.62)
-0.011 (-0.48) YEAR 2007 0.001
(0.03)
-0.011 (-0.72)
-0.008 (-0.74)
-0.017 (-1.19)
0.006 (0.18)
-0.013 (-0.54) YEAR 2008 -0.003
(-0.15)
-0.012 (-0.84)
-0.001 (-0.09)
-0.009 (-0.66)
-0.003 (-0.10)
-0.019 (-0.80) Constant 0.922***
(45.56)
0.950***
(81.18)
0.845***
(72.79)
0.782***
(127.52)
0.917***
(32.38)
0.894***
(84.60)
N 50 134 16 34 34 76
F-statistics 6.71*** 22.82*** 1.97 5.94*** 0.15 0.043
R-square 0.433 0.519 0.417 0.450 0.020 0.79
Note: This table shows the pooled regression results of cost efficiency by Stata 11.0SE, with the standard error of the coefficients shown in the parentheses. ***, **, and * are significant at 1, 5, and 10 percent significance levels, respectively.
The evidence of Model F_p and Model F_s shows that during the sample period of 2006 to 2009, the majority ownership of both STATE and JOINT display a significant effect on cost efficiency, while the foreign minority ownership of FOREIGN_S is significantly positive to the cost efficiency. But
the foreign minority ownership of FOREIGN_P is insignificant. In the same period, the foreign minority ownership, measured by dummy (FOREIGN_P), is significantly positive within both state-owned banks as well as the joint-stock banks. The values of F statistics of Model S_p and Model J_p are insignificant, and the observer number of Model S_p is only 16. For this reason the results of Model S_s and Model J_s are more appropriate for ‘foreign shares within bank majority ownership’ discussion. From Model S_s and Model J_s, the foreign minority ownership of FOREIGN_S is also significantly positive to the cost efficiency both within the state-owned banks and the joint-stock banks.
In other words, opposite to the majority ownership, foreign share is significantly positive to cost efficiency improvement of the whole banking industry as well as within the Big Five banks and within the joint-stock banks.
But only the effect of foreign shares reflects in ‘participation’, not in the percentage (for the percentage of foreign shares is merely small). And the significant levels of the whole industry and of within the state-owned banks (1%) are lower than those of within the joint-stock banks (10%). The influence of foreign shares impacts more effectively on the cost efficiency of the Big Five banks than of the joint-stock banks. And because the state-owned banks take a larger pie of the whole banking industry, the evidence of the state-owned banks certainly impacts more on the whole banking industry than others.