Chapter 2 Does Financial Restructuring Change the Relationship between Corporate Governance
2. Methodology and Framework
2.2 Tobit Regression Model and Bootstrapping Method
Each term in the square brackets of (4) expresses the efficiency of term t measured by the relative slacks of inputs and links, and is equal to unity if all slacks are zero. It is units-invariant and its value is between 0 and 1.
),
Hence, (4) is the weighted average of term efficiencies over the whole term, which we call the input-oriented overall efficiency and it is also between 0 and 1.
1 *.
2.2 Tobit Regression Model and Bootstrapping Method 2.2.1 Model and Variable
This study selects static efficiency estimated by the SBM model as the explained variable.
The efficiency scores (as the explained variable) from DEA are limited to a value between 0 and 1. Because the explained variable in the regression equation cannot be expected to have a normal distribution, we cannot expect the regression error to also meet the assumption of normal distribution. The OLS method often leads to biased and inconsistent parameter estimates (Greene, 1981). We therefore used Tobit estimation (Garza-Garcia, 2012; Lin, 2012;
Cheng, 2012) in this study.
This study first applied DEA to estimate bank efficiency and then used the Tobit regression model to analyze the relationship between efficiency and corporate governance.
Such an approach, however, could be logically inconsistent because of different methodologies (nonparametric and parametric) in different stages of estimation. Tobit regression could also lead to biased and inconsistent results. We therefore applied the bootstrapping6 method (Efron, 1979) to alleviate this problem. Drawing on the replacement method, we tested whether the statistically significant relationship is the same between the Tobit regression model and the bootstrapping method.
6We first resampled 90% of the whole sample 15 times and estimated the coefficients with OLS. For each of the corporate governance variables, we obtained 15 regression coefficient estimates. We then used 5,000 bootstrap replications of the 15 coefficient estimates to produce the confidence interval.
Finally, we tested whether the regression coefficients significantly differ from zero.
‧
characterized by a high degree of ownership concentration, a predominance of family control, and an abundance of pyramidal groups and cross-holdings. Young, Peng, Ahlstrom, Bruton, and Jiang (2008) showed that controlling shareholders’ expropriation of minority shareholders’ wealth is an important concern in Taiwan. Whether to merge is an important decision for banks. To protect shareholder equity, the obligation of bank management and the board of directors in the merger process are important. Therefore, we selected cash flow right, deviation between voting right and cash flow right, and deviation between seat control right and voting right as the explanatory variables. We also chose shareholdings by managers, director and supervisor pledge ratio, and seat control right to form the ownership structure characteristics. We finally used bank size, establishment time, and merger experience as control variables to level their effects on corporate governance. Table 2-2 lists the definitions for these variables. Our estimated model is as follows.Yt = bXt +εt ， εt ~ N(0,σt 2 )
Y (static efficiency) = ao +b1×CF +b2×D_VC +b3×D_SV +b4×Pledge +b5×Manager +b6×Seat +b7×Size +b8×Merger +εt…... (8)
Table 2-2 Definition of explanatory variables Variable Code Variable Name Description
CF cash flow right The cash flow right is the fraction of the dividends paid by a bank that is (eventually) received by the ultimate owners.7 Cash flow right equals direct cash flow right plus indirect cash flow.
D_VC deviation between voting right and cash flow right
Voting right divided by cash flow right, where
voting right is the shareholding ratio controlled by the ultimate owners, that is, direct shareholding plus indirect shareholding.
D_SV seat control right and voting right
Proportion of board members controlled by the ultimate owners to the board size to voting right.
Pledge director and supervisor pledge ratio
Ratio of shares pledged to shares held by directors and supervisors.
Manager shareholdings by managers
Shareholdings by internal manager or group manager.
Seat seat control right Proportion of board members controlled by the
7 The ultimate owners are the people who have the greatest influence in company decision-making.
They are typically major shareholders, chairman, general manager, family members, and management team.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
ultimate owners to the board size.
Size bank size Natural logarithm of total assets.
Time establishment time Cumulative year of establishment time Exp merger experience Cumulative number of mergers per year.
For example, a bank merges five financial institutions in 2001 and we set the merger experience as 5. If the bank merges an additional three financial institutions the next year, we then set it as 8 in 2002.
2.2.2 Hypothesis
According to our model, we construct six hypotheses for the relationship between static efficiency and ownership structure, as shown in Table 2-3.
Table 2-3 Hypotheses
Hypothesis Description Related literature
H1 Cash flow right has no effect on bank efficiency.
Li(2012); Lin (2009) H2 Deviation between voting right and
cash flow right has no effect on bank efficiency.
Hung(2011); Huang(2012)
H3 Seat control right and voting right have no effect on bank efficiency.
Lin and Hsu (2008) H4 Director and supervisor pledge ratio
has no effect on bank efficiency.
Yang(2012); Lai(2010) H5 Manager shareholdings have no effect
on bank efficiency.
Ko(2009); Huang(2012) H6 Seat control right has no effect on
bank efficiency.
Chi(2012); Lin(2010)
Hypothesis 1: Cash flow right is the share of capital that the ultimate owner has invested and owns. Because controlling shareholders may have the same interest as the bank, reducing the incentive of using bank resources to enhance their utility, an increase in cash flow right can improve bank efficiency and reduce agency problems (La Porta, Lopez-de-Silanes, Shleifer, and Vishny, 2002). However, controlling shareholders may expropriate other investor wealth, but harm bank efficiency. Li (2012) found that there was no significant correlation between cash flow rights and corporate performance during financial tsunami. Lin (2009) showed that firm value increases with cash-flow right in the hands of controlling shareholders.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Hypothesis 2: If shareholders control a company through a pyramidal structure or cross-shareholdings, controlling voting rights are greater than cash flow rights. Hung (2011) and Huang (2012) revealed that the deviation between voting right and cash flow right is positively related to long term performance.
Hypothesis 3: When directors and supervisors gain more seats than control rights deserve, they may make unfavorable decisions that cost bank efficiency. The dominant controlling shareholder and his family members8 of a privately owned bank in Taiwan often hold positions in both management and the board. Lin and Hsu (2008) examined 14 Taiwanese business groups, along with their 63 affiliated companies, and found that family-controlled firms or firms with a higher deviation of control from cash flow rights underperformed in return on investment.
Hypothesis 4: The director and supervisor pledge is another approach to gain control rights. Shareholders can borrow from banks by collateralizing the shares they own. Yang (2012) showed that the pledged shares ratio of directors show a significant negative correlation with the ROA. Lai (2010) indicated that the higher percentage of directors and supervisors pledging with the worse corporate performance and the higher variability of corporate performance.
Hypothesis 5: Higher management shareholding leads to better interest alignment between shareholders and managers, and hence, improves corporate efficiency. It might decrease corporate performance, and managers become entrenched and pursue private benefits. Ko (2009) found that there is significantly positive relationship between managers’
ownership ratio and financial performance. Huang (2012) showed that ownership of managers is negatively linked to performance, especially significant in the second year after M&A.
Hypothesis 6: Seat-controlling shareholders hold decisive votes and may select their family members as board directors. If directors are not duly responsible, they may create agency problems at the expense of minority shareholders. Chi (2012) and Lin (2010) found that control of seats of the major shareholders and deviation of cash flow right are negatively related with firm performance.