Claudia

Talking points#

  • Abstract: too long, the first half must be condensed in one or two sentences. Also, leave out the argumentation underlying your research questions, and save it for the introduction. Furthermore, you should explain the meaning of abbreviations, i.e., ESG and EU.

  • Introduction: some claims could be supplemented with references, e.g. to newspaper articles.

  • Enumarate the figures and provide a short note to make them self-sufficient.

  • Are you sure the returns in table 1 are cumulative returns? So it means that the buy-hold return starting from “Incubation” to 3/19/20 was negative?

The analyses aim to investigate the impact of Environmental, Social and Governance (ESG) scores on stock returns while controlling for firm- specific characteristics related to their size, book-to-market equity value, liquidity and leverage as well as country and industry interaction effects.

You should briefly explain in the introduction why we need an analysis of ESG scores on stock returns, by providing strong arguments!

  • 1st paragraph of chapter 2: try to explain the structure of chapter two briefly to the reader (still difficult to grasp for me)

The results suggest that the effect of ESG scores had almost the same magnitude as cash holdings and long- term debt.

  • What does this specifically mean?

Methodology#

  • Hypothesis 1:

To test hypotheses 1 and 2 we refer back to the methodology used by Garel and Petit-Romec (2021) who performed an empirical analysis on the cross-section of stock returns on the most recent firms’ ESG scores and on the selected control variables. Therefore, we first perform a cross-sectional regression by keeping country and industry fixed effects through the use of dummy variables.

  • This seems to be not Fama-MacBeth, but in the explanation afterwards, you seem to do Fama-MacBeth. Why, what is the reason?

General questions#

Does the process make sense? Currently, I am a bit confuse on how to integrate the industry fixed effects and I am looking for papers in which the process is clearly explained.

Industry fixed effects should be simple: just create a dummy var for each variable.

More specifically, am I following the right process by not running time-series regressions for each factor/control variable? If I understand correctly in my case the firm-specific data available in my dataset should be sufficient to run directly the cross-sectional regressions. Could you confirm that?

I don’t know, because it seems as if you are contradicting yourself in the methodology section. You have to show me, and we can discuss.

  • Correction after your meeting: Fama-Macbeth is the approach you are taking:

  • Step 1: for each company in your sample, do a regression on all the observations that you have:

$$ R_t = \alpha + \beta_1 \cdot R_{m,t} + \beta_2 \cdot SMB_t + \beta_3 \cdot HML_t + \epsilon_t $$

Then, take the beta’s $\beta_1, \beta_2, \beta_3$ for each company and run these regressions, one time for every date t:

$$ R_{i} = \alpha + \lambda_1 \cdot \beta_1 + \lambda_2 \cdot \beta_2 + \lambda_3 \cdot \beta_3 + \text{Whatever other variables you want to add} + \epsilon_i $$

Then, for each date, you get a vector of $\lambda$’s, and you can average them as you describe in the analysis.

General questions#

One doubt that I have is: is it correct to use the European risk- free rate (Euribor) to calculate the excess return (nominator) for all indexes? The alternative would be using three risk-free rates one for each region, but I think that would alter the comparison method.

Euribor is correct.