<< msvar_irf Markov-switching models MSVAR_stderr >>

Grocer >> Markov-switching models > msvar_irf_cb

msvar_irf_cb

msvar irf with confidence bands

CALLING SEQUENCE

resirfcb=msvar_irf_cb(res,hor,niter,siz,arg1,…,argn)

PARAMETERS

Input

• res = a tlist result from a ms var estimation

• hor = a scalar, the horizon of the irf (note that the shocks are supposed to happen at time 0; therefore (hor+1) impulses are calculated)

• niter = number of iterations

• siz = siz of confidence bands

• mres = optional argument that can be:

  - x = chol1 (cholesky decomposition)

  - x = chol2 (triangular factorisation)

  - x = original (original residuals)

  - x = soriginal (standardized (one standard deviation shock) uncorrelated residuals)

  - x = girf (Generalized Impulse Response based on Pesaran and Shin) (default = soriginal)

Output

• resirf = a result tlist with:

• resirf('meth') = 'msvar cb irf'

• resirf('msvar cb res') = the input ms var results tlist

• resirf('# irf') = the horizon of the irf calculations

• resirf('nb states') = the # of states

• resirf('nb irf') = the # of different AR parts

  * 1 if they do not switch

  * 1 if they do not switch

  * the # states if they switcch

  * the # states if they switcch

• resirf('sigma switch') = the # of switching wariances

  * 1 if they do not switch

  * 1 if they do not switch

  * the # states if they switch

  * the # states if they switch

• resirf('irf part state') = the partial impulse response

• resirf('irf full state') = the full impulse response

• resirf('irf part lower cb') = the lower part of the confidence band for the partial impulse response

• resirf('irf full lower cb') = the lower part of the confidence band for the full impulse response

• resirf('irf part upper cb') = the upper part of the confidence band for the partial impulse response

• resirf('irf full upper cb') = the upper part of the confidence band for the full imptulse response

• resirf('irf part median') = the median part of the confidence band for the partial impulse response

• resirf('irf full median') = the median part of the confidence band for the full impulse response

• resirf('irf part state # i') = partial impulse response for state i

• resirf('irf full state # i') = the full impulse response for state i

• resirf('irf part lower cb state # i') = lower part of the confidence band for partial impulse response of state i

• resirf('irf part upper cb state # i') = upper part of the confidence band for partial impulse response of state i

• resirf('irf part median state # i') = median partial impulse response of state i

• resirf('irf full lower cb state # i') = lower part of the confidence band for full impulse response of state i

• resirf('irf full upper cb state # i') = upper part of the confidence band for full impulse response of state i

• resirf('irf full median state # i') = median full impulse response of state i

DESCRIPTION

Calculates confidence bands for impulse response functions for a Markov-switching VAR.

EXAMPLE

global GROCERDIR ;
load(GROCERDIR+'\data\msirf_d.dat')
rms=ms_var('all',2,['100*delts(log(us_gdp))','us_un'],2,2,3)
rirf_cb=irf_cb(rms,20,99,0.1)
// example taken from function msvar_irf_d: provides irf confidence bands for the msvar
// estimate stored in tlist rms, at a 20 periods (quarters ) horizon, with 99 draws

AUTHOR

Stefan Fiesel 2015

Report an issue
<< msvar_irf Markov-switching models MSVAR_stderr >>