The RATS Software Forum

Hi,

I hope this finds you fine,

I would like to implement a TVP-VAR with Heteroskedastic errors, like that of Primiceri (2005). I would like to quantify the sources of Business Cycles.


My VAR Model includes External and domestics variables. In addition to exogeniety restriction implied by the small open economy assumption, I would like to impose short run restrictions for External variables, and long run ones to the domestic variables.

Is there a code that handles this kind of models? Or could you kindly guide me how to modify an existing code to implement the model.

Thank you in advance,

Ahmed Sahloul

AhmedSahlool wrote:Hi,

I hope this finds you fine,

I would like to implement a TVP-VAR with Heteroskedastic errors, like that of Primiceri (2005). I would like to quantify the sources of Business Cycles.


My VAR Model includes External and domestics variables. In addition to exogeniety restriction implied by the small open economy assumption, I would like to impose short run restrictions for External variables, and long run ones to the domestic variables.

Is there a code that handles this kind of models? Or could you kindly guide me how to modify an existing code to implement the model.

Thank you in advance,

Ahmed Sahloul

Have you found a paper which does that type of model?

Dear Tom,

Thank you for your reply.

I thought that using long and short run identification restrictions is an option that we have, as in ordinary SVAR model where we can choose between: short run, long run restrictions, and /or both. I was wondering if this is applicable with Bayesian inference and while allowing for time varying coefficients and stochastic volatility.

I read a paper " Can External Shocks Explain the Asian Side of Global Imbalances? Lessons from a Structural VAR Model with Block Exogeneity" which use Bayesian inference and apply Long and short run restriction. However, the authors don't allow for time varying coefficients and stochastic volatility.

Therefore, do you think this is applicable, or not?

Thank you and best regards

AhmedSahlool wrote:I read a paper " Can External Shocks Explain the Asian Side of Global Imbalances? Lessons from a Structural VAR Model with Block Exogeneity" which use Bayesian inference and apply Long and short run restriction. However, the authors don't allow for time varying coefficients and stochastic volatility.

Therefore, do you think this is applicable, or not?

Thank you and best regards

Probably not. A long run restriction says the effect of a shock converges to zero. If the coefficients are time-varying, how do you enforce that?