|
|
| |
| Generalized Two-Stage (OLS, LAD) Estimator |
Musaphau Animashahun, Prof. Kayode Ayinde
|
Abstract:
In any organization expenditure is generally believed to be related to income generated, among other factors. These factors need to be examined so as to know their effect and contribution on expenditure. Consequently, this work aims at modelling monthly expenditure and sales of a subsidiary of Nigerian National Petroleum Corporation’s products using linear regression model. The data used in this study were monthly expenditure and sales of petro-chemical products for a period of ten years. The Dickey-Fuller test statistic was used to test for the stationarity of the data.. Other statistics tests used include the Jaque-Bera, Glejser/LM and Durbin-Watson test statistic for testing normality, homoscedasticity and independence of error terms respectively. The Variance Inflation Factor (VIF) was also used to examine the presence of multicollinearity. Arising from various linear regression assumptions’ violation called for the use of Generalised Two-Stage(OLS, LAD) estimator. The results show that the original data is stationary. The OLS estimation results reveal the presence of multicollinearity, non-normality, heteroscedasticity and autocorellated error terms in the dataset. Attempting to improve the analysis using Generalised Two-Stage (OLS,LAD) shows that the results of the latter estimator are more efficient than the former, in terms of their significant variables. In conclusion, the results reveal that sales of Kerosene and Air Transport Kerosene(ATK) are positive and have significant contribution to NNPC monthly expenditure.
|
|
|
| | |