Step 1:
Literature Review: review the relevant literature in the area that you are interested in. Determines the model that you are going to use. And then identify the variables that you are going to use. Relevant variables include:
Board Size
The number of directors serving on the company's board at the beginning of the financial year
Committee Size
The number of directors serving on the company's committee (audit, compensation, nomination, governance) at the beginning of the financial year
Board Independence
The percentage of independent directors on the company's board at the beginning of the financial year
Committee Independence
The percentage of independent directors on the company's committee (audit, compensation, nomination, governance) at the beginning of the financial year
Board Meetings
The number of board meetings held at the end of the financial year
Committee Meetings
The number of committee meetings held at the end of the financial year
CEO Duality
CEO of the company who also serves as the chairman of the board
Inside Director Ownership
Total percentage of shareholdings of inside and affiliate directors serving on the board
Outside Director Ownership
Total percentage of shareholdings of independent directors serving on the board
etc.
Step 2
Collecting data.
At this stage, use the existing data set or collect it yourself.
A) If you want to use the existing data set, you can use the data from the Corporate Library (available in the WRDS, U.S. companies).
Or
B) if you want to collect the new data set for your research, you can use the data collection system provided by Survey Partner (www.survey-partner.co.uk or www.survey-partner.com).
The first choice reduce your time in collecting and cleaning your data, but it might be not affordable for individual researchers.
The second choice would cost you some time but it is a cheaper choice. It may cost you 3-month-time to collect about 5-6 years data for 300 companies.
Step 3
Analyzing the model you specified in Step 1, and revise them if necessary. At this stage, you may need to control for the following sensitivities:
A) Endogeneity
B) Firm and time effects if you use panel data
C) Other sensitivities may incur in your model.
Step 4
Revise your hypotheses, discover the reasons for the unexpected outcomes and write up your reports.
Revising the hypotheses always happen since there would be unexpected outcomes after you have controlled for different factors. The results may not support your hypothesis and you have to discover the reason behind this from a lot of sources other than academic journals.
No comments:
Post a Comment