Table 2, Panel A presents a summary of all
variables used in our model that is for constructing
CG indices using PCA. Table 2, Panel B displays an
alternative model, which consists of all binary variables
for the simple aggregated CG indices. The
nonbinary attributes are converted to binary variables.
The percentage of INEDs on board is
substituted with a dummy that equals 1 if the
percentage is larger than 50%, and otherwise 0, so
as to proxy whether the board majority is formed by
INEDs to demonstrate board independence. For
directors’ shareholding percentage and large shareholder’s
shareholding percentage, dummy variables
are equal to 1 if the shareholding is larger than 50%.
Over this percentage, the shareholder/directors would
have absolute control and significant interest in the
company. The number of substantial shareholders is
substituted by ‘if there is more than one substantial
shareholder’, which proxies for the market of corporate
control. The square of the largest shareholder’s
shareholding percentage is dropped because of the
difficulty associated with proxying nonlinearity with
binary-constructed attributes.