Hey everyone,
I am currently writing a master's thesis in strategic management and I am investigating the relationship of CSR, CSI, and firm performance through all six relations. I am quasi-replicating Kang et al. (2016), and extending their framework by incorporating two additional mechanisms (bad management: CSI --> FP) and (Performance Pressure: FP --> CSI). Furthermore, I will investigate the 2010-2019 period, using the updated KLD framework (indicators nearly doubled). I have collected all my data from Wharton Research Database (Compustat/CSRP for financial data, MSCI KLD Stats for CSR data). I have imported both data sets in Stata 18, but I am having quite some trouble merging my data sets. Using Wharton I was able to retrieve both the TICKER, and CUSIP codes for my data, however, KLD is displayed in CUSIP8, and I have found various issues where multiple companies have reported the same TICKER. In my attempt to still merge my datasets, I decided to trim CUSIP9 to CUSIP8 in my financial data set, and then create a Unique Identifier using cusip8 + year, and matching that with KLD data. This only leads to 20.432 matched observations, which makes me worry about there being sufficient data to run my reduced form PVAR model. Similarly, I also tried merging using the TICKER + year combination, which resulted in 21.904 matched observations, so slightly more, but I know there to be some problematic/false entries being there.
I chose this topic to push myself and see if I can get it done, but I must admit I am not the biggest expert in terms of using Stata, therefore I am wondering if anyone has advice for me given my initial setup and approach. Am I unnecessarily worrying about my observation count, or are there real problems here. I plan to use CUSIP if possible, given I know ticker to have false entries. Any thoughts, advice or general feedback is appreciated!
Thanks for taking the time,
Jesse
I am currently writing a master's thesis in strategic management and I am investigating the relationship of CSR, CSI, and firm performance through all six relations. I am quasi-replicating Kang et al. (2016), and extending their framework by incorporating two additional mechanisms (bad management: CSI --> FP) and (Performance Pressure: FP --> CSI). Furthermore, I will investigate the 2010-2019 period, using the updated KLD framework (indicators nearly doubled). I have collected all my data from Wharton Research Database (Compustat/CSRP for financial data, MSCI KLD Stats for CSR data). I have imported both data sets in Stata 18, but I am having quite some trouble merging my data sets. Using Wharton I was able to retrieve both the TICKER, and CUSIP codes for my data, however, KLD is displayed in CUSIP8, and I have found various issues where multiple companies have reported the same TICKER. In my attempt to still merge my datasets, I decided to trim CUSIP9 to CUSIP8 in my financial data set, and then create a Unique Identifier using cusip8 + year, and matching that with KLD data. This only leads to 20.432 matched observations, which makes me worry about there being sufficient data to run my reduced form PVAR model. Similarly, I also tried merging using the TICKER + year combination, which resulted in 21.904 matched observations, so slightly more, but I know there to be some problematic/false entries being there.
I chose this topic to push myself and see if I can get it done, but I must admit I am not the biggest expert in terms of using Stata, therefore I am wondering if anyone has advice for me given my initial setup and approach. Am I unnecessarily worrying about my observation count, or are there real problems here. I plan to use CUSIP if possible, given I know ticker to have false entries. Any thoughts, advice or general feedback is appreciated!
Thanks for taking the time,
Jesse
