[R-SIG-Finance] Interpretation of sign bias test in rugarch output?

[Jen Bohold]

I fitted a standard GARCH(1,1) model to my data using r and the rugarch package. In the model checking, I looked at the sign bias test output:

Sign Bias Test
------------------------------------
                            t-value      prob sig
Sign Bias               2.020     4.344e-02  **
Negative Sign Bias  1.014     3.105e-01    
Positive Sign Bias   1.579     1.145e-01    
Joint Effect             21.632    7.779e-05 ***


I am not sure about the interpretation of this?

I said the following:

The test for sign bias tests if positive and negative shocks have different impacts on the volatility. 
The positive size bias test tests, if the size of positive shocks affects the volatility differently. So if large and small shocks have different impacts on the volatility, which are not predicted by the volatility model (the negative size bias test is for negative shocks).
Since I used a standard GARCH model, I already assumed, that larger shocks have a larger influcence on the volatility (since the epsilon term is squared in the volatility equation). The output shows, that the main problem here is the sign bias. So my sGARCH models the impact of the size of the shocks on the volatility correctly, but positive and negative shocks have different impacts on the volatility and not a symmetric impact.

Is this correct, can anyone give further hings/comments?

Thanks a lot for your wisdom!
Jen