Error when getting variance decomp from SVAR in statsmodels

Code Libraries
1

Hi
I have set up an SVAR model. When I attempt to get the variance decomposition, the following error raised:

TypeError Traceback (most recent call last)
~\AppData\Local\Temp/ipykernel_10304/4106933268.py in
----> 1 fevd = results.fevd()
2 print(fevd.summary())

~\anaconda3-new\lib\site-packages\statsmodels\tsa\vector_ar\var_model.py in fevd(self, periods, var_decomp)
1881 fevd : FEVD instance
1882 “”"
→ 1883 return FEVD(self, P=var_decomp, periods=periods)
1884
1885 def reorder(self, order):

~\anaconda3-new\lib\site-packages\statsmodels\tsa\vector_ar\var_model.py in init(self, model, P, periods)
2382 self.names = model.model.endog_names
2383
→ 2384 self.irfobj = model.irf(var_decomp=P, periods=periods)
2385 self.orth_irfs = self.irfobj.orth_irfs
2386

TypeError: irf() got an unexpected keyword argument ‘var_decomp’

How should i do?

2

Hi Mahdi, I am new around here and was not able to replicate this error.

I think the error may be occurring because var_decomp is not an argument in the irf method of the class which you are calling.

It looks like a known bug as per this link, am not sure if it has been patched. website link

Hope it helps!

3

Hi Sa1tyBiscuit,

Thank you for your help.
The problem is with the sourcce code, I didn’t find the answers in that forum helpful. I’m really confused.