Examining the Supply Chain Integration Impact on Economy by Regression Model

Abstract

Abstract— In this study, we examine the current state of supply chain integration, estimate the economic impact of inadequate integration, and identify opportunities for governmental organizations to provide critical standards infrastructures that will improve the efficiency of supply chain communications. The development of methods to reduce the impact of multicollinearity in the construction of a linear regression model is an urgent task of applied econometrics. The article proposes a method for reducing multicollinearity in the construction of a linear regression model for evaluating the supply chain impact on economy. In the case of non-stationary multidimensional time series, it is assumed that all variables have a polynomial trend. Each predictor x_j(t) is decomposed into a trend and a remainder u_j(t), and then a regression y(t) is constructed for time t and the remainder u_j(t). In this case, the regression coefficients for u_j(t) are equal to the regression coefficients for x_j(t), but they are estimated using less correlated regressors SCM. The article gives a quantitative assessment of the increase in the accuracy of the forecast of the considered model in comparison with other models. In the case of spatial variables, the proposed approach is that some X_j regressors SCM correlated with others are replaced by the sum of two summands. One of them is the predicted value of X_j obtained from the regression equation X_j on the predictor correlated with it; the other is the remainder of this regression U_j. As a result, we get a new set of regressors SCM that are much less correlated with each other. The new regressors - the remnants of U_j - are susceptible to meaningful interpretation. However, the new regression equation changes the regression coefficients only for variables that act as dependent variables in auxiliary regressions. The application of the proposed method is illustrated by examples through the supply chain process. Calculations are performed in the R software environment.