ESTIMATING THE PARAMETERS OF A REGRESSION MODEL USING COVARIATIONS OF VARIABLES AND INTERPRETING REGRESSION COEFFICIENTS

UDC 330.43 (075.8)
DOI: 10.36871/2618-9976.2023.10.001

Authors

Irina V. Orlova,
Candidate of Economic Sciences, Professor, Financial University under the Government of the Russian Federation, Moscow, Russia

Abstract

This article proposes a method for estimating linear regression coefficients other than least squares and a nonstandard approach to interpreting regression coefficients associated with this method. Estimates of the regression coefficients are selected in such a way that the partial derivatives with respect to the regressors of the average value of the endogenous variable Y calculated using the model are equal to the estimates of the partial derivatives of the average value of Y with respect to the regressors calculated without the model and, accordingly, independent of the model parameters. The values obtained during the implementation of the method make it possible to quantitatively assess the correctness of generally accepted assumptions when interpreting regression coefficients.

Keywords

Linear regression, Correlation, Covariance