BVAR (version 0.1.5)

bv_irf: Impulse response settings

Description

Provide settings for the computation of impulse responses to bvar. Allows setting the horizon for which impulse responses should be computed, whether or not forecast error variance decompositions (FEVDs) should be computed and, if and what kind of identification should be used.

Usage

bv_irf(horizon = 12, fevd = FALSE, identification = TRUE,
  sign_restr = NULL)

Arguments

horizon

Integer scalar. The horizon for which impulse responses (and FEVDs) should be computed.

fevd

Logical scalar. Whether or not forecast error variance decompositions should be calculated.

identification

Logical scalar. Whether or not the shocks used for calculating impulses should be identified. Defaults to TRUE, i.e. identification via Cholesky decomposition unless sign_restr is provided.

sign_restr

Numeric matrix. Sign restrictions for identification. Elements should be set to \(1\) (\(-1\)) to restrict for positive (negative) impacts. If no presumption about the impact can be made the corresponding elements can be set to \(0\). The default value is NULL, meaning identification would be performed via Cholesky decomposition.

Value

Returns a named list of class bv_irf with options for bvar.

Details

Identification can be performed via Cholesky decomposition and sign restrictions. The algorithm for generating suitable sign restrictions follows Rubio-Ramirez et al. (2010).

References

Rubio-Ramirez, J. F., Waggoner, D. F., & Zha, T. (2010). Structural Vector Autoregressions: Theory of Identification and Algorithms for Inference. The Review of Economic Studies, 77, 665-696. https://doi.org/10.1111/j.1467-937X.2009.00578.x

Examples

# NOT RUN {
# Set impulse responses to a horizon of 20 time periods and enable FEVD
# (Identification is performed via Cholesky decomposition)
bv_irf(horizon = 20, fevd = TRUE)

# Identify impulse responses using sign restrictions
data("fred_qd")
data <- fred_qd[, c("CPIAUCSL", "UNRATE", "FEDFUNDS")]
data[5:nrow(data), 1] <- diff(log(data[, 1]), lag = 4) * 100
data <- data[5:nrow(data), ]

# Signs should be based on economic theory
signs <- matrix(c(1, 1, -1, -1, 1, -1, -1, 1, 1), nrow = 3)

irf_signs <- bv_irf(sign_restr = signs)
# }
# NOT RUN {
bvar(data, lags = 5, irf = irf_signs)
# }