Computes simultaneous posterior bands for impulse responses

Transforms posterior draws of selected impulse responses into simultaneous credible bands using the empirical sup-norm around the posterior median response path. The empirical order statistic covers at least the requested fraction of retained draws over the complete selection. Ties can increase coverage. This is not a guarantee of exact population posterior coverage. Boundaries are rounded outward when necessary to retain the empirical guarantee in floating-point arithmetic.

Usage

simultaneous_irf(object, ...)

# Default S3 method
simultaneous_irf(object, ...)

# S3 method for class 'PosteriorIR'
simultaneous_irf(
  object,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

# S3 method for class 'PosteriorBSVAR'
simultaneous_irf(
  object,
  horizon = NULL,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

# S3 method for class 'PosteriorBSVARMIX'
simultaneous_irf(
  object,
  horizon = NULL,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

# S3 method for class 'PosteriorBSVARMSH'
simultaneous_irf(
  object,
  horizon = NULL,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

# S3 method for class 'PosteriorBSVARSV'
simultaneous_irf(
  object,
  horizon = NULL,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

# S3 method for class 'PosteriorBSVART'
simultaneous_irf(
  object,
  horizon = NULL,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

# S3 method for class 'PosteriorBSVARSIGN'
simultaneous_irf(
  object,
  horizon = NULL,
  probability = 0.9,
  variables = NULL,
  shocks = NULL,
  variable = NULL,
  shock = NULL,
  model = "model1",
  ...
)

Arguments

object: A posterior model object or a PosteriorIR object.
...: Additional arguments passed to computation methods.
probability: Posterior coverage probability of the simultaneous credible band.
variables: Response variables to include (character or integer vector).
shocks: Structural shocks to include (character or integer vector).
variable: Deprecated. Use variables instead.
shock: Deprecated. Use shocks instead.
model: Label identifying the model specification.
horizon: Maximum horizon used when object is a posterior model object.

Value

A bsvar_post_tbl with columns model, object_type, variable, shock, horizon, median, lower, upper, simultaneous_prob, achieved_coverage, and critical_value. Additional provenance columns are described in posterior_provenance.

Examples

data(us_fiscal_lsuw, package = "bsvars")
spec <- bsvars::specify_bsvar$new(us_fiscal_lsuw, p = 1)
#> The identification is set to the default option of lower-triangular structural matrix.
post <- bsvars::estimate(spec, S = 5, show_progress = FALSE)

sb <- simultaneous_irf(post, horizon = 3)
head(sb)
#> # A tibble: 6 × 21
#>   model  object_type      variable shock horizon  median   lower  upper
#>   <chr>  <chr>            <chr>    <chr>   <dbl>   <dbl>   <dbl>  <dbl>
#> 1 model1 simultaneous_irf ttr      ttr         0  0.0561 -0.0892 0.201 
#> 2 model1 simultaneous_irf gs       ttr         0 -0.0523 -0.198  0.0930
#> 3 model1 simultaneous_irf gdp      ttr         0  0.0841 -0.0612 0.229 
#> 4 model1 simultaneous_irf ttr      gs          0  0      -0.145  0.145 
#> 5 model1 simultaneous_irf gs       gs          0  0.0432 -0.102  0.188 
#> 6 model1 simultaneous_irf gdp      gs          0 -0.0402 -0.185  0.105 
#> # ℹ 13 more variables: simultaneous_prob <dbl>, achieved_coverage <dbl>,
#> #   critical_value <dbl>, analysis_type <chr>, interval_probability <dbl>,
#> #   interval_method <chr>, scale_by <chr>, scale_var <chr>, scale_factor <dbl>,
#> #   units <chr>, source_package <chr>, source_version <chr>, n_draws <int>