Mark

Mark

Research question#

  • Board composition on CEO Turnover

    • CEO Turnover is indicative of corporate governance

      • How well are the interests of shareholders represented
      • You have to explain why this is important, why does it matter how well the interests of shareholders are represented

    • When the company is underperforming relative to its benchmark

    • This seems superfluous

      • Why is it only interesting when the company is underperforming?
      • Do you expect there to be a different effect when the company is overperforming?
        • (Assuming some CEOs are also fired when the company is performing well)
      • Mediation Analysis
      • (Effect of Board composition on CEO Turnover is mediated by Underperformance/Overperformance)?

  • However, it’s within the realm of possibilities that a female from a non-white ethnicity influences the board in a different way than a malefrom a non-white ethnicity, yet at the same time the females from a non-white ethnicity might at the same time influence the board differently than a white female.

    You also have to expand upon why!

Variables & Data:

  • So which kind of CEO turnover do you want to focus on, internal or external?
    • Internal turnover
  • Stock performance
    • Why not a simple asset pricing model (and a corresponding negative $\alpha$?
    • CAPM, Fama French 3 factor
  • I think you also have to search for other measures of CEO entrenchment/managerial power (making it difficult to fire CEOs)
  • Are there different CEO-level variables other than ethnicity you can also extract from WRDS?

Methodology:

  • If turnover is a yes/no variable, you need a linear probability model, or logit
    • Justify why LPM and not logit
  • Fixed effects account for “corporate culture”, not for changes in corporate culture!
  • Selection: only certain companies may chose to hire CEOs of certain ethnicities/genders
    • Is the decision to hire female/ethnically diverse CEOs also worth investigating?

Conclusion:

  • I think the proposal is feasible. The advantage of choosing a variable like ethnicity/gender, is that it’s obviously exogenous (like a random treatment and not subject to change/influence of other factors), but the challenge consists of:
    • Controlling for other factors influencing CEO turnover
    • Controlling for the selection process
    • Correlations within-company over time
      • Find out how the literature deals with it