<< tvp_var_diagnos_hyperp VAR estimations var1 >>

Grocer >> VAR estimations > tvpvar_prior0

tvpvar_prior0

returns prior tlist needed by Hetero_TVP_VAR function

CALLING SEQUENCE

prior=tvpvar_prior0(nlag,k_B0,k_A0,k_W,k_Q,k_S,endo,arg1,...,argn)

PARAMETERS

Input

* nlag = a scalar, the # of lags in the VAR

* k_B0 = a scalar, the dilatation factor applied to the variance of the VAR coefficients over the training period to determine the variance of the prior of the VAR coefficients

* k_A0 = a scalar, the dilatation factor applied to the variance of the coefficients of the contemporaneous endogenous vraiables over the training period to determine the variance of the prior of the contemporaneous endogenous vraiables

* k_W = a scalar, the dilatation factor applied, squared, to (n+1) for the inverse Wishart prior of W

* k_Q = a scalar, the dilatation factor applied, squared, to (n+1) for the inverse Wishart prior of Q

* k_S = a scalar, the dilatation factor applied, squared, for the inverse Wishart prior of S

* arg1,...,argn = an optional argument, that can be:

  - 'exo=list(exo1,..., exon) if the user wants to add exogenous variables to the VAR, exoi then collects the exogenous variables of equation i

  - 'nocte' if the user does not want constants in the VAR

  - 'dropna' if the user wants to drop NA values from the data instead of generating an error

 

Output

* r = a prior tlist with:

  -r('B_0_prmean') = the prior mean of the B_0 parameters (time varying coefficients of the VAR)

  -r('B_0_prvar') = the prior variance of the B_0 parameters

  -r('A_0_prmean') = the prior mean of the A_0 parameters (time varying coefficients of the contemporaneous endogenous variables of the VAR equations)

  -r('A_0_prvar') = the prior variance of the A_0 parameters

  -r('Q_prmean') = the prior mean of the Q parameters

  -r('Q_prvar') = the prior variance of the Q parameters

  -r('W_prmean') = the prior mean of the W parameters

  -r('W_prvar') = the prior variance of the W parameters

  -r('S_prmean') = the prior mean of the W parameters

  -r('S_prvar') = the prior variance of the W parameters

DESCRIPTION

from broad parameters entered by the user, provide the prior tlist needed by fucntion Hetero_TVP_VAR function.

EXAMPLE

stacksize('max')
load(GROCERDIR+'\data\primiceri.dat')
mkdir(GROCERDIR+'\temp')
nlag=2
bounds('1953q3','1963q2')
prior=tvpvar_prior0(nlag,4,4,0.01,0.01,0.1,['Inflation';'Unemployment';'rate_3m'])
// returns the prior as defined by Primiceri (2005): dilatation factors applied to the ols estimates
// of A0 and B0 over the training sample '1953q3' to '1963q2' set to 4, the the dilatation factors
// applied, squared, to (n+1) for the inverse Wishart prior of W, Q and S set 0.01, 0.01 and 0.1
// respectively

AUTHOR

Éric Dubois 2015

Report an issue
<< tvp_var_diagnos_hyperp VAR estimations var1 >>