[R-SIG-Finance] ugarchfit - Weighted Ljung-Box Test and ARCH LM Test

[Lukas Halbeisen]

Hi all,

I am currently working on my master thesis and have to evaluate some 
GARCH-Modells.

What I am wondering:

How are the lags picked up for the Weighted LB-Test as well as for the 
ARCH LM Test.

For instance I have fitted a GARCH(3,3)-Modell:

Weighted Ljung-Box Test on Standardized Squared Residuals
------------------------------------
                          statistic  p-value
Lag[1]                       9.052 0.002624
Lag[2*(p+q)+(p+q)-1][17]    14.593 0.070672
Lag[4*(p+q)+(p+q)-1][29]    20.134 0.129824
d.o.f=6


The lags are calculated by the formular 2*(p+q)+(p+q)-1. But what is the 
reason? Is there any reference where this kind of calculation has been 
described?


Weighted ARCH LM Tests

------------------------------------
              Statistic Shape Scale P-Value
ARCH Lag[7]      1.375 0.500 2.000  0.2409
ARCH Lag[9]      1.462 1.485 1.796  0.6476
ARCH Lag[11]     2.596 2.440 1.677  0.6695


Same for the ARCH-LM modell. I have 6 d.o.f. but why does rugarch 
consider the lags 7,9 and 11 as the most important ones?

Thank you and best regards,
Lukas