Quantile Regression Model Research Between GDP And Consumer Spending Constitute

As we all known, Gross Domestic Product(GDP) as an important index to measure theeconomic development of a country, which has been influenced by many factors. Amongthis factors, the influence of the consumer spending is worth paying more attention.This paper investigate the relationship between GDP and consumer spending constituteusing some countries data that coming from the statistical yearbook, most previous studiesare based on the Keynes consumption theory to research the relationship between them byCointegration Relationship. Recently, quantile regression is very concerned because of thecomputing convenience and the characteristics of the theory. This article adopt the quantileregression to study the relationship between GDP and consumer spending constitute, GDPis the dependent variable and the consumer spending constitute is the independent variable.The traditional least squares estimate suggests, the influence of the consumer variablesto GDP is not remarkable. On account of that the traditional least squares model mustbased on the assumption, which error is zero and variance is σ2, and neglect the situation ofleptokurtosis and fat-tail, but quantile regression is on the contrary.Considering the correlation between the variable, we will apply the cluster analysis andthe factor analysis to process the data of consumer spending constitute, this will increase thesignificant of parameter estimation through decrease the relevance of them. The processeddata was divided into two classes, one which called the personality consumption has astrong correlation with food and so on, the other which called generality consumption hasa strong relevant with the education and the medical treatment. At last, we conclude thatGDP has a strong negative correlation with the generality consumption and has a strongpositive correlation with the personality consumption.