Computes posterior draws of regime probabilities

Each of the draws from the posterior estimation of a model is transformed into a draw from the posterior distribution of the regime probabilities. These represent either the realisations of the regime indicators, when type = "realized", filtered probabilities, when type = "filtered", forecasted regime probabilities, when type = "forecasted", or the smoothed probabilities, when type = "smoothed", .

Usage

compute_regime_probabilities(
  posterior,
  type = c("realized", "filtered", "forecasted", "smoothed")
)

Arguments

posterior: posterior estimation outcome of regime-dependent heteroskedastic models - an object of either of the classes: PosteriorBSVARMSH, or PosteriorBSVARMIX obtained by running the estimate function.
type: one of the values "realized", "filtered", "forecasted", or "smoothed" denoting the type of probabilities to be computed.

Value

An object of class PosteriorRegimePr, that is, an MxTxS array with attribute PosteriorRegimePr containing S draws of the regime probabilities.

References

Song, Y., and Woźniak, T., (2021) Markov Switching. Oxford Research Encyclopedia of Economics and Finance, Oxford University Press, doi:10.1093/acrefore/9780190625979.013.174 .

Author

Tomasz Woźniak wozniak.tom@pm.me

Examples

# upload data
data(us_fiscal_lsuw)

# specify the model and set seed
set.seed(123)
specification  = specify_bsvar_msh$new(us_fiscal_lsuw, p = 2, M = 2)
#> The identification is set to the default option of lower-triangular structural matrix.

# run the burn-in
burn_in        = estimate(specification, 10)
#> **************************************************|
#> bsvars: Bayesian Structural Vector Autoregressions|
#> **************************************************|
#>  Gibbs sampler for the SVAR-stationaryMSH model             |
#> **************************************************|
#>  Progress of the MCMC simulation for 10 draws
#>     Every 10th draw is saved via MCMC thinning
#>  Press Esc to interrupt the computations
#> **************************************************|

# estimate the model
posterior      = estimate(burn_in, 50)
#> **************************************************|
#> bsvars: Bayesian Structural Vector Autoregressions|
#> **************************************************|
#>  Gibbs sampler for the SVAR-stationaryMSH model             |
#> **************************************************|
#>  Progress of the MCMC simulation for 50 draws
#>     Every 10th draw is saved via MCMC thinning
#>  Press Esc to interrupt the computations
#> **************************************************|

# compute the posterior draws of realized regime indicators
regimes        = compute_regime_probabilities(posterior)

# compute the posterior draws of filtered probabilities
filtered       = compute_regime_probabilities(posterior, "filtered")

# workflow with the pipe |>
############################################################
set.seed(123)
us_fiscal_lsuw |>
  specify_bsvar_msh$new(p = 1, M = 2) |>
  estimate(S = 10) |> 
  estimate(S = 50) -> posterior
#> The identification is set to the default option of lower-triangular structural matrix.
#> **************************************************|
#> bsvars: Bayesian Structural Vector Autoregressions|
#> **************************************************|
#>  Gibbs sampler for the SVAR-stationaryMSH model             |
#> **************************************************|
#>  Progress of the MCMC simulation for 10 draws
#>     Every 10th draw is saved via MCMC thinning
#>  Press Esc to interrupt the computations
#> **************************************************|
#> **************************************************|
#> bsvars: Bayesian Structural Vector Autoregressions|
#> **************************************************|
#>  Gibbs sampler for the SVAR-stationaryMSH model             |
#> **************************************************|
#>  Progress of the MCMC simulation for 50 draws
#>     Every 10th draw is saved via MCMC thinning
#>  Press Esc to interrupt the computations
#> **************************************************|
regimes        = compute_regime_probabilities(posterior)
filtered       = compute_regime_probabilities(posterior, "filtered")