Marketing i Menedžment Innovacij (Dec 2014)
Menges model applied to the economy of the USA
Abstract
The aim of the article. This research is prolonged to investigate whether the particular economic model works in the certain economics. The results of the analysis. The Menges model was chosen – the model which shows the dependence of GNI on GNI of the previous period and on net investment. The second one is more interesting because it reflects one of the most fundamental laws of economics. There is no sense to deepen in the interrelationship between GNI and lagged GNI because of its evidence. The country in which the model is tested is USA – one of the most developed countries. There are too small amount of such kind of articles and Menges model research that is why this work can be interested in. To ensure the adequacy of the model the complex analysis was applied, the prolonged conclusion lighten the statements of the coefficients significance, the quality and randomnicity of R2, how the variance of repressors explains the variance of regressing, the homoscedasticity or heteroscedasticity of residuals and the presence or absence of correlation between them and the conditions of applying the OLS method in order to estimate parameters, the confidence interval and the adequacy of the model. For visual evidence of correlation the correlation analysis and scatter diagram are developed. Conclusions and directions of further researches. After all the research done the conclusion that the Menges model works in the economics of US. The model is adequate instead of some tests have not been passed. T- test haven’t been passed only for a0, so it means that only this free member is not significant, F-test have been passed, so it means that R2 is not random and quality of specification is high. R2-test showed that variation of GNI almost perfectly (99,9%) is described by the variation of repressors. The autocorrelation of residuals or its absence cannot be stated because by the test of Durbin-Watson, it came to be in the area of ambiguity. GQ-test was not passed, it means that the method of OLS cannot be used and that residuals are heteroscedastic, the valuations lose the feature of accuracy, the third condition of Gauss-Markov theorem is not performed. It can occur because of some important variable was not included in the model, or because of only one equation of the whole system of Menges equations was taken instead of all the system. Nevertheless, it can be treated if lagged variables will be included. And the statement that the Menges model is suitable for this particular country is the absolute truth because the predicted value satisfies the confidence interval even taken in account the difficult foreign economic situation dealed with.
