irf.ZerosignR.result: Calculate structural IRFs for zero and sign restrictions...

In roootra/ZerosignR: SVAR identification by zero-sign restrictions

Calculate structural IRFs for zero and sign restrictions estimation result

Description

This function calculates arbitrary-horizon structural IRFs conditinal on given set of matrices Q — orthonormal matrices that satisfy zero and sign restrictions.

Usage

## S3 method for class 'ZerosignR.result'
irf(zerosignr, horizon = 0, LR = FALSE, quantil = 0.05)

Arguments

zerosignr	ZerosignR.result object
horizon	Numeric. Horizon to calculate IRF for. 0 = only contemporaneous IRF, 1 = contemporaneous + next period after shock IRF, etc.
LR	Boolean. Whether to calculate long-run IRF additionally to finite-run ones.
quantil	Numeric. Number in (0,1), calculate IRF quantile band.

Details

Arias et al. (2018) show that structural shock IRF for period h can be calculated in matrix form as follows (see definitions for J and F in the original article):

Lh(A0, A+) = (A0 * J' * F^h * J)'.

The special case is for long-run IRF, which is given as:

Linf(A0, A+) = (A0' - ∑ A_l')^(-1).

In order to obtain structural IRF conditional on orthonormal matrix Q, that satisfies zero and sign restrictions, one has just multiply IRF matrix by Q:

Lh(A0, A+) * Q = Lh(A0 * Q, A+ * Q).

Value

ZerosignR.irf, a list containing:

• struc_irfs, 3-dim. array with IRFs. The 1-st dimension is for specific Q-matrix, the 2-nd one is nvars*(horizon+contemporaneous+LR) — row-stacked IRFs for different periods, starting from the earliest one, the 3-rd one is for shocks;

• median_irf, pointwise-median IRF, aggregated for Q-matrices;

• quantile_irf, IRF sample quantiles for band;

• ctm_irf, IRF from closest-to-median model, Euclidean distance;

• ..., auxillary constants.

Note

struc_irfs in returned value ZerosignR.irf is a row-stacked IRF with columns for shocks. Shocks' order is as in the restriction matrix from ZerosignR.result object.

First nvars rows are for variables' contemporaneous response (order is as in underlying model), Rows from nvars + 1 to 2*nvars are for variables' response in the period right after the shock, etc. If LR is TRUE, the last nvars rows are long-run IRF.

Author(s)

Artur Zmanovskii. [e-mail]

References

Arias, J.E. and Rubio-Ramirez, J. F. and Waggoner, D. F. (2018) Inference Based on Structural Vector Autoregressions Identifiied with Sign and Zero Restrictions: Theory and Applications. Econometrica, 86, 2, 685-720, https://doi.org/10.3982/ECTA14468.

See Also

zerosign_restr

roootra/ZerosignR documentation built on May 23, 2022, 2:06 p.m.