國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
16
used before to compute the corporate governance score for every sample company called CG-Index here, hereby we examine Hypothesis 3 by adding CG-Index into the model, the results of this test would be sufficient to let us discuss whether a company’s corporate governance extent affect all its investors investment strategies and trading activities.
A. Descriptive Statistics
In order to examine the connections existed among corporate governance, investors’
behavior, and companies’ trading volume, we use the listed companies’ stocks from 1997 to 2011 in Taiwan as research sample, total is 274 firms.
Here we replace all of variables with clearer codes in our models and tables for the convenience of observing, the codes as showed below:
MGT: Management stockholdings (%)
MGT2: The square of Management stockholdings BLOCK: Blockholders’ stockholdings (%)
BOARD: Board size
BOARD2: The square of Board Size OUTDTR: Independent outside directors INV: Investments in asset (%)
ABC: Information transparency and disclosures level CEOCHR: CEO duality
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
17
PARTCHR: Inside dominance of chairman of the board PARTCEO: Inside dominance of CEO
MGTPARTCEO: The cross item of management stockholdings and inside dominance of CEO
CG-Index: Corporate governance scores SIZE: Ln of asset
LVG: Leverage ratio
RTN (t-x)i : The Return of (t-x) period, where x=1, 2 INDIV: Individuals Stockholdings
DOINDIV: Domestic Individuals Stockholdings FOINDIV: Foreign Individuals Stockholdings INSTI: Institutions Stockholdings
FOINSTI: Foreign Institutions Stockholdings LnValue: Ln of Trading Value
LnSDValue: Ln of Standard Deviation of Trading Value Ratio: Turnover Ratio
SDRatio: Standard Deviation of Turnover Ratio Rtn: Returns
SDRtn: Standard Deviation of Returns
‧
Table 2. Descriptive Statistics- Results for 15 years, yearly frequency
Variable Mean SD Min Max
CG-Index denotes the governance index, we obtain CG-Index for each firm by awarding one point for each governance standard that is met, the standard would be explained later. The dependent variable evaluated is separated into two parts, which are investors’ shares holding ratio and companies’ trading volume, besides, we measure the variability of trading volume for 15 years, yearly frequency.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
19
The results of descriptive statistics for 15 years data are showed in Table 2 and Table 3. The extreme values of blockholders’ stockholdings (BLOCK) are very large (73.40, 77.16), and the mean of independent outside directors (OUTDTR) are below 1 (0.616, 0.616), somehow it indicate corporations centralizing the ownership rights at a few shareholders, and this reflect the enterprise culture in Taiwan even other Asia countries, like Claessens and Fan (2002) address, the lack of protection of minority rights has been the major corporate governance issue in Asia. The extreme values of investment in asset (INV) are quite large (95.06, 95.42), indicating some firms may need to manage their firm structures as well to avoid occurring financial crisis.
CG-Index denotes the governance index, we obtain CG-Index for each firm by awarding one point for each governance standard that is met, the standard would be explained later.
The individuals Stockholdings is almost concentrated at domestic investors, however, there is about twenty percent shares holding by foreign institutional investors of all the institutional stockholdings. It shows the trading volume related variables here, we use trading value, trading turnover ratio, and stocks return as factors to measure the impact of corporate governance extents to firms’ trading activity, we adopt yearly frequency data to test firms’ trading persistence especially.
Two clear firm characterized variables - firm size and leverage level are included to our models, also for monthly frequency data, we add past one and two month returns to the models in order to increasing the credibility of the results.
‧
Table 3. Descriptive Statistics- Results for 15 years, monthly frequency
Variable Mean SD Min Max
CG-Index denotes the governance index, we obtain CG-Index for each firm by awarding one point for each governance standard that is met, the standard would be explained later. The dependent variable evaluated is separated into two parts, which are investors’ shares holding ratio and companies’ trading volume. The controlling variables include past returns for 15 years, monthly frequency.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
21
To reinforce the research results, we test the corporate governance impact for shorter sample period from 2007 to 2011, as we know, the regime of information transparency and disclosure ranking (ABC) is carried out for only nine years from 2003, so we add this variable in shorter period model. The information transparency and disclosure ranking is classified into seven levels, but the mean of it is below 4 (3.556) in Table 4, indicating many firms got the level below “A-”. CG-Index included the new variable (ABC) sum up the grades of nine corporate governance variables resulting in higher statistic values in mean, standard deviation, and maximum value.
The t-statistics showed in the second column of Table 4 denote that the mean value of the variable for 5 years data is significantly different from the corresponding value for 15 years data in Table 3 except for two incomparable variables - ABC and CG-Index. We find several values existing differences between two sample periods, the values indicate block shareholders’ stockholdings and investments in asset increased, board size down, the phenomenon of CEO duality grew up and part management decreased recent years. There’s a tendency of declining individual stockholdings and raising institutional stockholdings close years, furthermore the trading value has been risen up, and so as the firm size, but leverage level cut down.
‧
Table 4. Descriptive Statistics- Results for 5 years, monthly frequency
Variable Mean SD Min Max
CG-Index denotes the governance index, we obtain CG-Index for each firm by awarding one point for each governance standard that is met, the standard would be explained later. The dependent variable to be evaluated is separated into two parts, which are investors’ shares holding ratio and companies’ trading volume. The controlling variables include past returns for 5 years, monthly frequency. *** denotes that the mean value of the variable for 5 years data is significantly (at the 1% level) different from the corresponding value for 15 years data.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
23