Dear all,
I'm currently working on my master's thesis that looks at firm performance given certain CEO characteristics in the United States, using a panel dataset of all firm years from 2000-2020 extracted from Compustat.
I've ran some regressions for ln(Tobin's Q), ROE and ROA.
My ceo characteristic dummy variable is significant for log of Tobin's Q (0.021), however it's nothing close to any significance for ROE (0.862) and ROA (0.934)
I got my data from Compustat and I used the following definitions for my independent variables:
Log of Tobin's Q = ln((Market Value + Total Debt + Preferred stock liq. value)/Total Assets)
ROA = Net Income/Total Assets
ROE = Net Income/Stockholder's Equity
Also, as a side note, I noticed that the ROA and ROE values calculated seem very strange, please see attached pictures of summary tables.
I have 2 questions.
1) Is there something wrong with the ROA/ROE calculation? If so, what might it be? Again, the calculation was simply done by dividing the raw data of Net Income and Total Assets from Compustat database... I also checked SEC 10-k to confirm some weird companies, and in fact, some actually do have either 0 assets or like $80 in total assets which is baffling. see:
https://sec.report/Document/0001144204-12-062282/
https://sec.report/Document/0001262463-16-000974/
2) In case ROA/ROE cannot be fixed, what other performance measures are popular to be used on US firms for the purpose of robustness checks? Suggested papers would be extremely appreciated!
Thank you very much!
Hello! From your comments, it seems to me that there are problems with your data. If there are measurements that seem strange to you, it is a symptom that something is wrong. If you have a large sample, you could debug your data to make sure the structure is correct. For example, you could check for missing values or outliers. Also, that everything is correctly typed in the database, sometimes there are errors such as entering the percentage in decimals in some cases and the percentage in others; these typing errors are quite common.
I think it is very interesting that there are differences in the significance of the performance measures because Tobin's Q measures a different thing than ROE/ROA.
Regards!
- Badges
- Science topic
- Similar topics
- Thesis Research