Analysis of Corporate Social Performance on Financial Performance

This project investigates the impact of corporate social performance (CSP) on financial performance in the US market, using secondary data from S&P 500 companies. Financial performance is measured by Return on Equity (ROE) and Return on Assets (ROA), while CSP is assessed using KLD scores, considering environmental, employee, community, customer, and governance factors. The research employs data cleaning, descriptive statistics, and inferential analysis, including correlation and regression analysis (fixed and random effects), to determine the relationship between CSP and financial outcomes. Control variables such as sales margin and current assets are also considered. The analysis utilizes SPSS statistical software to examine the data, including descriptive statistics (mean, median, skewness, kurtosis, etc.) and tests for multicollinearity, normality and heteroskedasticity. The findings are then compared with previous research to draw conclusions about the impact of CSP on financial performance.

Impact of the corporate social performance on the financial performance in US
market
Data analysis
The current research is aimed to investigate the impact of the corporate social
performance on the financial performance taking into account the US market. For
the analysis purpose the secondary data has been collected for s&P 500 companies.
The financial performance of the companies has been measured in terms of Return
on equity(ROE)and returns on assets (ROA). Furthermore the corporate social
performance of the companies has been measured by Kinder, Kydenberg and Domini
& Co. Inc which provides transparency for not only the past but also about the future
environmental performance of the companies. The KLD score reflects the ethical
commitment of the firms which helps the investors to analyze about the
sustainability of the firms. The KLD score is calculated taking into consideration five
factors which includes environment, employee and supply chain, community and
society, customers and governance and ethics. The KLD scores in the independent
variables for the current research. Furthermore some of the variables such as sales
margin, current assets have been taken as the control variables in the study.
For the data analysis purpose, researcher have used different data analytic
techniques. However before performing the data analysis the data cleaning process
was conducted wheres the missing values were identified and imputed through the
mean values. Also the outliers and other anomolies in the data were treated so that
the results are robust. Once the data cleaning process was completed the data was
imported to the statistical software SPSS for further analysis.
In the first section the descriptive statistics of dependent , independent and the
control variables has been discussed. For the descriptive statistics mean, median,
kurtosis, skewness, minimum value, maximum value and standard deviation has
been taken into consideration. The descriptive statistics helps to provide the
overview of the data for the researcher and also the reader of the research.
In the second section the findings from the inferential analysis has been discussed.
This includes the normality test of the dependent variable and also the test of
heteroskedasticity has also been conducted to test if there is any multicollinearity
among the independent variables. After the testing of the assumptions, the
inferential analysis has been performed which included the correlation analysis and
the regression analysis. The correlation analysis is used to test whether the two
variables are related to each other or not. If the correlation coefficient is greater than
0.6 it can be concluded that the variables are strongly correlated. The direction of the
correlation is measured by the sign of the correlation coefficient. Positive correlation
coefficient infers positive correlation whereas the negative sign indicates the
negative correlation. After the correlation analysis, the regression analysis has been
performed. Since the data is panel data ( which includes both the time series and the
cross sectional) both the fixed and the random effect regression has been performed.
Then the hauzman test has been performed to examine whether the random effect is

more appropriate or the fixed effect. In the last section the results has been
discussed along with the findings from the previous research.