ESG Score Effect on Stock Prices

[Derick Coleman]

Good afternoon,

I am trying to understand the effect of a companies ESG Score on their Stock Price. I have a large Panel Data Set for each quarter between 2015- 2022, I have the following data for each company:
Stock Price
Earnings Before Interest
Total Assets Percentage of Total Assets
Total Assets to Total Equity
Return on Equity
Return on Assets
Total Assets
ESG Score
Does anyone have any advice on how I would go about this, of which models/ regressions/ type of analysis I would use. I am quite new to Stata and quite overwhelmed in all honesty. Any feedback would be appreciated!

[Daniel Schaefer]

Hi Derick, welcome to the Stata forum. You don't really tell us much about your background, so I'm going to assume you are a complete beginner.

If you are feeling overwhelmed, consider taking a course in Stata or go through a detailed tutorial online. When you know how to use Stata generally and are ready to work with your data set, you should start by making sure your data is clean, that it looks the way you might expect, that you are correctly accounting for missing values, and so on. You can get summary statistics for quantitative variables using the -summary- command. On a new line in a do file, write summary followed by your variables of interest.

summary StockPrice EarningsBeforeInterest

You can also use the -tab1- command in the same way to get frequency tables for categorical variables. Actually cleaning your data, or deriving other variables of interest can be difficult. There are tutorials online that can help you with this, and googling and looking at other forum posts can be a great way to learn. Once you have a strong sense of the univariate statistics, you could start to calculate some bivariate statistics for your relationship of interest. You can get a correlation of ESG score on stock price with the -corr- command.

Eventually, you're probably going to want to run a panel style regression with the xt set of commands. You should look through the longitudinal analysis manual and make sure to pay particular attention to xtset and xtreg. You should also carefully consider which variables you might want to use as controls and include them in the model as well.

If you are absolutely new to statistics, you are going to want to take a course in the statistical methods that are typical of your field to get a baseline understanding of how things are done. You should also look for papers or studies that look at your relationship of interest to see how they do things, and to get a sense of the underlying theoretical justification for why the phenomenon is modeled the way that it is.

[Derick Coleman]

Hi Daniel,

Thank you for your response, I will spend some time looking into this and I will get back to you. Thank you for your help!