问题描述
早上好
我正在尝试根据GARCH（1,1）模型对货币的历史波动率模型进行建模。重要的是，我对诸如债券价格，Xrates和股票市场数据之类的其他回归变量的重要性感兴趣。
我得到的结果非常令人惊讶：平均方程结果（mxreg1-14）符合预期，但方差结果（vxreg1-14）困扰我，所有p值均接近1。重要的是，它不是系数的大小，而是我关注的值。我不认为p值应该接近1，因为之前有多项研究证实了这种关系。
任何帮助将不胜感激。我已经坚持了好几天，无法根据其他问题或答案来解释这些结果。
注意：外部回归变量的传递滞后1个周期
某些代码：
ug_spec_basic_ext_std <- ugarchspec(variance.model = list(model="sGARCH",garchOrder=c(1,1),external.regressors = matrix_garch_ext_reg_lagged),mean.model = (list(armaOrder = c(1,0),external.regressors = matrix_garch_ext_reg_lagged)),distribution.model = "std")
这是外部回归矩阵
head(matrix_garch_ext_reg_lagged)
Bond10Y_CNY Bond10Y_EUR Bond10Y_USD Bond10Y_JPY Bond10Y_KRW EURUSD_log_return CNYUSD_log_return JPYUSD_log_return KRWUSD_log_return
3 4.07 1.798 2.659 0.659 3.470 -9.464257e-05 0 4.711896e-03 0.000000000
4 4.07 1.807 2.622 0.646 3.425 -4.066948e-05 0 9.792881e-05 0.000000000
5 4.07 1.818 2.617 0.643 3.470 4.975261e-03 0 5.955595e-03 0.000000000
6 4.07 1.797 2.610 0.642 3.425 1.397992e-05 0 3.892944e-04 0.000000000
7 4.07 1.811 2.614 0.643 3.415 2.395036e-03 0 4.864050e-04 0.003213715
8 4.07 1.843 2.650 0.658 3.420 4.037028e-05 0 0.000000e+00 -0.001070091
Stock_CNY_log_return Stock_EUR_log_return Stock_USD_log_return Stock_JPY_log_return Stock_KRW_log_return
3 -0.01762379 -3.237398e-03 -1.830583e-03 0.0005349035 0.0065820220
4 0.01762379 -1.790529e-03 1.163694e-03 -0.0224675710 -0.0062818974
5 -0.01762379 5.378379e-04 -8.977563e-04 -0.0021461934 0.0062818974
6 0.01762379 -6.101015e-03 -8.122239e-03 0.0012114206 -0.0062818974
7 -0.01762379 4.823943e-05 7.743972e-05 -0.0090482109 0.0002250343
8 0.01762379 8.935721e-03 6.951047e-03 -0.0003863971 -0.0014711404
>
以下是完整结果：
*---------------------------------*
* GARCH Model Fit *
*---------------------------------*
Conditional Variance Dynamics
-----------------------------------
GARCH Model : sGARCH(1,1)
Mean Model : ARFIMA(1,0)
distribution : std
Optimal Parameters
------------------------------------
Estimate Std. Error t value Pr(>|t|)
mu 0.000866 0.000997 0.869016 0.384838
ar1 -0.001645 0.003966 -0.414777 0.678305
mxreg1 -0.000579 0.000499 -1.160978 0.245651
mxreg2 -0.000036 0.000624 -0.058076 0.953688
mxreg3 0.000177 0.000277 0.638928 0.522870
mxreg4 -0.000529 0.000837 -0.631988 0.527395
mxreg5 0.000443 0.000758 0.584109 0.559147
mxreg6 -0.054437 0.032554 -1.672219 0.094481
mxreg7 0.035150 0.031437 1.118115 0.263518
mxreg8 0.025732 0.027726 0.928062 0.353375
mxreg9 -0.074760 0.029803 -2.508457 0.012126
mxreg10 -0.000368 0.006587 -0.055924 0.955402
mxreg11 -0.007448 0.009493 -0.784552 0.432716
mxreg12 0.003114 0.009220 0.337755 0.735548
mxreg13 0.005641 0.010255 0.550027 0.582301
mxreg14 -0.005992 0.015217 -0.393750 0.693766
omega 0.000000 0.000002 0.000961 0.999233
alpha1 0.092502 0.014291 6.472855 0.000000
beta1 0.941712 0.006535 144.107614 0.000000
vxreg1 0.000001 0.000000 194.413956 0.000000
vxreg2 0.000000 0.000001 0.037197 0.970328
vxreg3 0.000000 0.000000 0.000146 0.999883
vxreg4 0.000000 0.000000 0.960760 0.336673
vxreg5 0.000000 0.000000 3.424647 0.000616
vxreg6 0.000000 0.001676 0.000006 0.999995
vxreg7 0.000000 0.001392 0.000007 0.999994
vxreg8 0.000000 0.001025 0.000010 0.999992
vxreg9 0.000000 0.000643 0.000015 0.999988
vxreg10 0.000000 0.000668 0.000015 0.999988
vxreg11 0.000000 0.001479 0.000007 0.999995
vxreg12 0.000000 0.000883 0.000011 0.999991
vxreg13 0.000000 0.001565 0.000006 0.999995
vxreg14 0.000000 0.001814 0.000006 0.999996
shape 2.100003 0.003168 662.952256 0.000000
Robust Standard Errors:
Estimate Std. Error t value Pr(>|t|)
mu 0.000866 0.001220 0.709915 0.477757
ar1 -0.001645 0.001274 -1.291507 0.196528
mxreg1 -0.000579 0.000709 -0.816031 0.414482
mxreg2 -0.000036 0.000461 -0.078623 0.937333
mxreg3 0.000177 0.000199 0.887280 0.374928
mxreg4 -0.000529 0.000560 -0.945175 0.344569
mxreg5 0.000443 0.000834 0.530824 0.595540
mxreg6 -0.054437 0.042201 -1.289941 0.197071
mxreg7 0.035150 0.032186 1.092104 0.274788
mxreg8 0.025732 0.035428 0.726308 0.467650
mxreg9 -0.074760 0.041088 -1.819505 0.068834
mxreg10 -0.000368 0.003911 -0.094189 0.924959
mxreg11 -0.007448 0.007105 -1.048185 0.294553
mxreg12 0.003114 0.009531 0.326751 0.743857
mxreg13 0.005641 0.010204 0.552774 0.580418
mxreg14 -0.005992 0.015624 -0.383508 0.701343
omega 0.000000 0.000008 0.000307 0.999755
alpha1 0.092502 0.039366 2.349771 0.018785
beta1 0.941712 0.018012 52.282533 0.000000
vxreg1 0.000001 0.000000 36.050806 0.000000
vxreg2 0.000000 0.000003 0.007509 0.994009
vxreg3 0.000000 0.000000 0.000007 0.999994
vxreg4 0.000000 0.000001 0.207830 0.835362
vxreg5 0.000000 0.000000 1.455378 0.145565
vxreg6 0.000000 0.003420 0.000003 0.999998
vxreg7 0.000000 0.003128 0.000003 0.999997
vxreg8 0.000000 0.002010 0.000005 0.999996
vxreg9 0.000000 0.002625 0.000004 0.999997
vxreg10 0.000000 0.001559 0.000006 0.999995
vxreg11 0.000000 0.003531 0.000003 0.999998
vxreg12 0.000000 0.002581 0.000004 0.999997
vxreg13 0.000000 0.004337 0.000002 0.999998
vxreg14 0.000000 0.003620 0.000003 0.999998
shape 2.100003 0.000940 2234.601332 0.000000
LogLikelihood : 12611.46
information Criteria
------------------------------------
Akaike -5.0030
Bayes -4.9589
Shibata -5.0031
Hannan-Quinn -4.9875
Weighted Ljung-Box Test on Standardized Residuals
------------------------------------
statistic p-value
Lag[1] 3.531 6.023e-02
Lag[2*(p+q)+(p+q)-1][2] 6.125 2.916e-05
Lag[4*(p+q)+(p+q)-1][5] 10.454 6.952e-04
d.o.f=1
H0 : No serial correlation
Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
statistic p-value
Lag[1] 1.705 1.917e-01
Lag[2*(p+q)+(p+q)-1][5] 34.761 1.426e-09
Lag[4*(p+q)+(p+q)-1][9] 44.700 3.377e-11
d.o.f=2
Weighted ARCH LM Tests
------------------------------------
Statistic Shape Scale P-Value
ARCH Lag[3] 2.507 0.500 2.000 0.113326
ARCH Lag[5] 7.559 1.440 1.667 0.025617
ARCH Lag[7] 11.649 2.315 1.543 0.007529
Nyblom stability test
------------------------------------
Joint Statistic: no.parameters>20 (not available)
Individual Statistics:
mu 0.21808
ar1 0.12373
mxreg1 0.26017
mxreg2 0.20581
mxreg3 0.18702
mxreg4 0.25610
mxreg5 0.25062
mxreg6 0.04071
mxreg7 0.02329
mxreg8 0.04260
mxreg9 0.11774
mxreg10 0.09217
mxreg11 0.15354
mxreg12 0.01729
mxreg13 0.01046
mxreg14 0.03266
omega 3.41688
alpha1 0.33271
beta1 0.26139
vxreg1 1.14788
vxreg2 0.79055
vxreg3 2.71528
vxreg4 0.02621
vxreg5 2.24046
vxreg6 0.05371
vxreg7 0.14095
vxreg8 0.08598
vxreg9 0.23041
vxreg10 0.40868
vxreg11 0.05558
vxreg12 0.06718
vxreg13 0.03007
vxreg14 0.02700
shape 2.05937
Asymptotic Critical Values (10% 5% 1%)
Individual Statistic: 0.35 0.47 0.75
Sign Bias Test
------------------------------------
Adjusted Pearson Goodness-of-Fit Test:
------------------------------------
group statistic p-value(g-1)
1 20 9887 0
2 30 13538 0
3 40 16256 0
4 50 17871 0
Elapsed time : 11.60321
有关数据的一些总体信息
Descriptive statistics
========================================================================
Statistic N Mean St. Dev. Min Pctl(25) Pctl(75) Max
------------------------------------------------------------------------
Bond10Y_CNY 5,029 3.48 0.51 2.49 3.08 3.72 5.00
Bond10Y_EUR 5,029 0.39 0.59 -0.86 0.03 0.62 1.96
Bond10Y_USD 5,029 2.19 0.59 0.51 1.86 2.61 3.24
Bond10Y_JPY 5,029 0.15 0.25 -0.29 -0.02 0.36 0.74
Bond10Y_KRW 5,029 2.28 0.62 1.18 1.81 2.61 3.79
EURUSD_log_return 5,029 -0.0000 0.01 -0.13 -0.0002 0.0002 0.13
CNYUSD_log_return 5,029 -0.0000 0.01 -0 -0.000 0.000 0
JPYUSD_log_return 5,029 -0.0000 0.01 -0.11 -0.0002 0.0002 0.11
KRWUSD_log_return 5,029 -0.0000 0.01 -0.10 -0.001 0.001 0.10
Stock_CNY_log_return 5,029 0.0001 0.02 -0.47 -0.004 0.004 0.46
Stock_EUR_log_return 5,029 0.0000 0.01 -0.21 -0.003 0.003 0.21
Stock_USD_log_return 5,029 0.0001 0.04 -0.69 -0.002 0.003 0.70
Stock_JPY_log_return 5,029 0.0001 0.03 -0.50 -0.004 0.005 0.49
Stock_KRW_log_return 5,029 0.0000 0.02 -0.30 -0.003 0.004 0.32
解决方法
暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！
如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。
小编邮箱:dio#foxmail.com (将#修改为@）