Smooth transition VAR with Impulse response

[ABDHUT123]

hi, everyone, I am looking code for Smooth transition VAR model with Generalised impulse response....can anyone help me...

[TomDoan]

https://estima.com/forum/viewtopic.php?f=8&t=2126 estimates (or at least tries to estimate) an STVAR. Until you actually have a working model, don't worry about the non-linear IRF.

From the RATS Structural Breaks and Switching Models course:

6.4 More Complicated Models
An alphabet soup of smooth transition models such as STVAR (vector autoregression),
STVECM (vector error correction model), STGARCH (GARCH) and combinations
such as STAR-STGARCH (smooth transition mean model with smooth
transition GARCH variance model) have been proposed in the literature. As
a general rule, these have not been particularly successful, and most can be
probably best be described as an estimator in search of an application. If you
have an interest in these, you might want to take a look at Chan and McAleer
(2003), for which we have a replication, which is an (unfortunately) rare paper
which shows how complicated “trendy” models can fail, and fail badly. A major
problem with this whole literature is that most of these models have multiple
transition functions, for instance, the standard description of the STVAR
has a separate smooth transition for each equation (different threshold values,
sometimes even different threshold variables). However, there’s no reason to
suspect that each equation will actually need that, or, if it has a transition,
that it should be smooth. Either failure on any single equation can make it
impossible to estimate the model.