There are three main methods for heteroscedasticity testing

1) Graphical test method: ①Correlation graph analysis. ②Residual graph analysis.

Since heteroskedasticity is usually considered to be due to the fact that the size of the residual changes with the size of the independent variable, whether there is heteroskedasticity can be simply judged by using a scatter plot. The specific method is to draw a scatter plot with the square 2ie of the regression residual as the ordinate and a certain explanatory variable ix in the regression formula as the abscissa. If the scatter plot shows a certain trend, it can be judged that heteroskedasticity exists.

2) Goldfeld-Quandt test (disadvantage, it can only handle single-up and single-down heteroscedasticity)

The Goldfeld-Quandt test is also called the sample segmentation method and the group method , proposed by Goldfeld and Quandt in 1965. The idea of ??this test is to divide the entire sample into two subsamples by removing the intermediate values ??in order of the size of the explanatory variables that cause heteroskedasticity.

3) White test and Glejser test.

The Goldfeld-Quandt test, also known as the sample segmentation method and the group method, was proposed by Goldfeld and Quandt in 1965. The idea of ??this test is to divide the entire sample into two subsamples by removing the intermediate values ??in order of the size of the explanatory variables that cause heteroskedasticity.

The most famous and commonly used is the third White test. The core principle is to judge the degree to which ui is explained by xi. The higher the degree, the more heteroskedasticity there is.

1. He Dexu and Wang Chaoyang; Time series econometrics: cointegration and conditional heteroscedastic autoregression [N]; Proceedings of the Chinese Academy of Social Sciences; 2003

2. Tang Shaoming; Futures position risk quantification method based on APARCH-GED model[N]; Futures Daily; 2008

3. Dr. Jiang Yingkun and Peng Yan of Guotai Junan Securities Ma Zhongqiang, head of futures research at Guotai Junan; futures index expiration A review of daily effect empirical results and classic empirical testing methods [N]; Futures Daily; 2007

4. Dai Weigao of GF Futures; Real Trading Competition and Sequence Analysis Technology [N]; Futures Daily; 2008 Year

5, Pei Yong; Application of Nobel Prize winner Engel's autoregressive conditional heteroskedasticity model in futures research (1)[N]; Futures Daily; 2004

6 , Research on the correlation between inflation and economic situation in recent years [N]; Shanghai Securities News; 2005

7 . Kong Xiangyi; Studying the complexity of finance from a macro perspective [N]; Chinese Journal of Social Sciences; 2009

8. Wang Longyun; The Nobel Prize in Economics values ??research on technical methods[N]; Economic Information Daily; 2003

9. Yu Qinyi and Zhang Hongjun; Applying econometric statistics to economic research[ N]; Journal of the Chinese Academy of Social Sciences; 2004

10. Reporter Wang Jieming and Wu Ping; Solving the problem of economic time series [N]; Xinhua Daily Telegraph; 2003