Decomposition of electricity consumption in China by primary component analysis.

The empirical relationship between electricity consumption and gross domestic product, population, the product of primary industry, second industry, and tertiary industry are investigated. The strong multicollinearity among EC's affecting factors does not meet the criteria of the ordinary least square regression (OLS) regression model. Principle component analysis is used to eliminate multicollinearity. Three principle components with no multicollinearity can explain 99.34 % of affecting factors' variance. The three principle components seemed as independent, and EC seemed as dependent variables when OLS regression is employed. The results show that: gross domestic product, primary industrial production value, second industrial production value, and tertiary industrial production value codetermined the trend of electricity consumption, while the proportion of primary industrial production value, second industrial production value, and tertiary industrial production value and population codetermined the starting point and fluctuation of electricity consumption; the economic scale is the mainly affecting factors on electricity consumption; as some parts of electricity consumed by primary industry are not included in the state grid, there is an illusion that the primary industry can produce electricity. [ABSTRACT FROM AUTHOR]