Predicting Revenue for Electricity Distribution Companies: a Multiple Linear Regression Approach

ABSTRACT
Electric power is an intrinsic ingredient to the survival and economic development of every country. This study focusses on the contribution of some distribution companies to the overall revenue generation of distribution companies. The electricity data used for this research is secondary data obtained from National Bureau of Statistics (NBS) websites. The data covers a period of 9 years from January 2015 to December 2023. The descriptive statistics of all variables (DISCOSR, AEDC, EKEDC, KEDC and PEDC) showed a normal distribution, suggesting that revenue trends for each distribution company have consistent central tendencies and moderate variability. Augmented Dickey-Fuller (ADF) test was used and it confirmed that all variables are stationary after first differencing as p-values for all the variables are less that the level of significance 0.05, making them suitable for time series modeling. The multiple regression model had R-squared value of approximately 79.96%, indicating that AEDC, EKEDC, KEDC, and PEDC together explain nearly 80% of the variance in DISCOS’s revenue. AEDC and EKEDC negatively impacted DISCOSR’s revenue while KEDC and PEDC positively impacted DISCOS’s revenue. The regression model for the study is DISCOSR =10781.43 -2.921311*AEDC-3.787607*EKEDC+23.59891*KEDC+ 5.644977*PEDC. The findings suggest that focusing on the factors influencing AEDC and EKEDC could help optimize DISCOS’s revenue, while efforts to enhance revenue trends in KEDC and PEDC can further strengthen their overall revenue generation and performance.