我先将我的步骤大致说一下
首先检验收益率序列R的自相关性,发现与滞后5阶显著相关
然后对r(-5)回归
之后对残差做ARCH-LM检验
这是滞后9阶的结果
结果如下
Heteroskedasticity Test: ARCH
F-statistic 2.348236 Prob. F(9,724) 0.0130
Obs*R-squared 20.81833 Prob. Chi-Square(9) 0.0135
Test Equation:
Dependent Variable: RESID^2
Method: Least Squares
Date: 06/03/13 Time: 13:54
Sample (adjusted): 5/12/2010 5/23/2013
Included observations: 734 after adjustments
Variable Coefficient Std. Error t-Statistic Prob.
C 0.000132 2.38E-05 5.559009 0.0000
RESID^2(-1) 0.012711 0.036722 0.346147 0.7293
RESID^2(-2) 0.040154 0.036723 1.093422 0.2746
RESID^2(-3) -0.041296 0.036677 -1.125938 0.2606
RESID^2(-4) 0.024989 0.036416 0.686215 0.4928
RESID^2(-5) 0.081969 0.036306 2.257732 0.0243
RESID^2(-6) 0.066544 0.036436 1.826316 0.0682
RESID^2(-7) 0.062423 0.036504 1.710024 0.0877
RESID^2(-8) -0.002296 0.036546 -0.062836 0.9499
RESID^2(-9) 0.076758 0.036531 2.101169 0.0360
R-squared 0.028363 Mean dependent var 0.000197
Adjusted R-squared 0.016284 S.D. dependent var 0.000388
S.E. of regression 0.000384 Akaike info criterion -12.87609
Sum squared resid 0.000107 Schwarz criterion -12.81344
Log likelihood 4735.527 Hannan-Quinn criter. -12.85193
F-statistic 2.348236 Durbin-Watson stat 2.004669
Prob(F-statistic) 0.012956
可以说明存在ARCH效应吗
还有我的步骤正确吗?
希望能解释的详细一些!谢谢大家!


雷达卡



京公网安备 11010802022788号







