ANALYZING SPATIAL AND TEMPORAL DYNAMICS OF CSR IN INDIA AND ODISHA THROUGH PREDICTIVE MODELS

Abstract

India was also among the rare countries to require corporate social responsibility (CSR) expenditure with the introduction of the Section 135 of the Companies Act, 2013. In the last ten years, this legislation has been spearheading massive capital inflows into the social development sectors. This paper compares CSR expenditure in India and Odisha through two regression equations; linear regression and Polynomial Regression to a ten-year period of analysis (FY 2014-15 to 2023-24). The comparison of the two unique spatial and temporal growth rates draws a picture of the linear growth and nonlinear and fluctuating growth of India and Odisha respectively. The comparative analytical evaluation is also done with the help of R², RMSE, MAE, and MAPE measures of predictive accuracy and model suitability. It was found that India CSR expenditure has risen by 3.47x in FY 2023-24 to ₹34,908.75 crore as compared to FY 2014-15, and Odisha CSR expenditure has risen by 5.51x to ₹1,389.39 crore based on the dataset. The Linear model for India achieved an R² of 0.98, confirming strong linear consistency, while Odisha’s 2^nd-degree polynomial model achieved an R² of 0.97, effectively capturing volatility. Comparative evaluation confirmed that India’s CSR is predictable with an MAPE below 6%, whereas Odisha’s higher MAPE (≈12%) indicates greater unpredictability. These findings highlight spatial disparity in CSR flows and reinforce the importance of adopting region-specific forecasting strategies for balanced social development.