We now describe the method
by which we combine our 13 variables within the four
subindices. We collect our data from publicly available
data sources and use the companies’ systematic
disclosure of the data to ensure that the data are fully
transparent and reproducible. Our challenge here is
to objectively form an index to quantify the firm’s
CG. There are two similar ways to address problems
of this kind: PCA and factor analysis. We use PCA to
condense our variables into various CG internalmechanism
components7 because PCA includes all
possibilities in the variables, whereas factor analysis
considers only common variability. This method has
also been previously applied to a form index (e.g.
Linck et al., 2008). We generate index scores for each
year for each of the four mechanisms using the largest
eigenvalue vector. For comparative purposes, we also
use factor analysis and find empirical results similar
to those that PCA yields. Table 4 displays the first
two components of the factor loadings and the
eigenvalues from the PCA.