Mark
Proposal comments#
In my view, the proposal is a pass. The introduction starts off nicely, attempting to integrate the theory into the debate, but needs improvement in structure, as several seemingly unrelated questions are introduced at once in a very ad hoc way.
First paragraph: you talk about: it is difficult to measure performance of individuals in firms. However, the CEO is a special individual. Second paragraph: “Morck, Shleifer and Vishny (1988) argue that CEO turnover is a way to determine the quality of corporate controls, their paper found that if the firm does badly but the firm is well-governed, the CEO is more likely to get fired.”
But: you want to measure the quality (performance) of a CEO, not of corporate controls!
“Jenter and Kanaan (2015) specifically focused on forced turnovers”
But: you should explain the link between measuring CEO performance and forced turnovers first.
“Kanter (1997) argues that under tokenism, once the threshold of at least 15% of a certain minority is reached, this minority tends to be treated more equally.”
What is tokenism? What does it have to do with the study of CEO performance?
My suggestion would be to start off with anecdotes that spark interest in the question, then pose it, then go to the theoretical perspective, then go to the empirical literature. Sometimes the introduction also feels too much like a list of items, at the expense of a good flow.
“Thus, the research question that will be answered is whether board characteristics play a mediating factor on the relationship between stock returns and CEO turnover in the S&P1500.”
In my view, you should pose the research question after you’ve illustrated (by means of examples or previous studies) why that question and similar questions are interesting and relevant.
The literature features many relevant studies, but the structure could also be improved upon: it is not clear why you need section 2.2, as in section 2.1 you treat the X-Y relationship, and section 2.3 the possible moderating factors. The literature is skewed towards empirical studies rather than theoretical mechanisms. I like the hypothesis section, in which you should try to elaborate a little bit more what the main arguments underlying the hypotheses are.
Section 4, the methodology, contains the basic approach that you will take. That approach seems uncontroversial, but try to elaborate on the assumptions underlying your approach, the extent to which your estimates can be interpreted as causal effects, and risks such as endogeneity, selection bias or attrition (and many more). If you can, try to anticipate on these concerns, and design and conduct additional analyses that can counter these criticisms.
“If the CEO on the time of departure was in poor health, death or accepted a position elsewhere, the turnover will be classified as not forced. Moreover, if the CEO is close to retirement age, meaning above the age of 60, the departure will also be classified as not forced, assuming an early retirement.”
- How do you find out whether the first is the case?
- How do you know the second assumption is warranted?
Questions#
I have some ideas, but I am still a little uncertain on how to address the feedback of my co-reader. Moreover, I remember you suggesting to do a probit regression as well as an OLS regression in one of our earlier meetings, however, I still don’t fully understand the advantage of doing both, what would the advantage of doing the probit as well?
Probit and OLS rely on slightly different econometric assumptions - you want to make sure that your analyses are robust to the usage of different methods.
- measurement of ethnicity in board of directors
- Data on WRDS