where:
Following is defined as the number of analysts following UK firms collected from FactSet dataset;CG is the aggregated CG ranking provided by CGQ; andSize is the natural logarithm of total assets.
CGQ also provides sub-scores for each firm in four particular governance areas; board, takeover defence, executive and director compensation and ownership and audit. These sub-scores are expressed in numbers from 1-5 (5 indicates firm in the top quintile in a governance area, while 1 indicates firm is in the bottom quintile in a governance area).
To understand which of these components better explains the number of analyst following, we run model (1) by replacing the overall ranking of CG with sub-scores pertaining to different aspects of CG. We expect a positive sign with each individual CG variable. That is, we expect b1 through b4 to be positive in the following model: Equation 2
where:
Following is defined as the number of analysts following the UK firms collected from FactSet dataset;Board is the board composition sub-score;Composition is the executive and director composition sub-score;Takeover is the anti-takeover provisions sub-score;Audit is the audit committee/audit fees/audit rotation/auditor ratification sub-score; and Size is the natural logarithm of total assets.
Moreover, we construct another aggregate measure of CG (NCG) from the addition of the four sub-scores (Board + Compensation + Takeover + Audit) and re-run regression (1) as follows: Equation 3
We control for firm size in the above three models using the natural logarithm of total assets (in £ millions). Based on prior literature (Lang et al., 2003, 2004), we expect a positive sign for Size.