Fitting Bayesian VAR-SV

This function fits VAR-SV. It can have Minnesota, SSVS, and Horseshoe prior.

Usage

bvar_sv(
  y,
  p,
  num_chains = 1,
  num_iter = 1000,
  num_burn = floor(num_iter/2),
  thinning = 1,
  bayes_spec = set_bvar(),
  sv_spec = set_sv(),
  intercept = set_intercept(),
  include_mean = TRUE,
  minnesota = TRUE,
  save_init = FALSE,
  convergence = NULL,
  verbose = FALSE,
  num_thread = 1
)

# S3 method for class 'bvarsv'
print(x, digits = max(3L, getOption("digits") - 3L), ...)

# S3 method for class 'bvarsv'
knit_print(x, ...)

Arguments

y

Time series data of which columns indicate the variables

p

VAR lag

num_chains

Number of MCMC chains

num_iter

MCMC iteration number

num_burn

Number of burn-in (warm-up). Half of the iteration is the default choice.

thinning

Thinning every thinning-th iteration

bayes_spec

A BVAR model specification by set_bvar(), set_ssvs(), or set_horseshoe().

sv_spec

SV specification by set_sv().

intercept

Prior for the constant term by set_intercept().

include_mean

Add constant term (Default: TRUE) or not (FALSE)

minnesota

Apply cross-variable shrinkage structure (Minnesota-way). By default, TRUE.

save_init

Save every record starting from the initial values (TRUE). By default, exclude the initial values in the record (FALSE), even when num_burn = 0 and thinning = 1. If num_burn > 0 or thinning != 1, this option is ignored.

convergence

Convergence threshold for rhat < convergence. By default, NULL which means no warning.

verbose

Print the progress bar in the console. By default, FALSE.

num_thread

Number of threads

x

bvarsv object

digits

digit option to print

...

not used

Value

bvar_sv() returns an object named bvarsv class.

coefficients

Posterior mean of coefficients.

chol_posterior

Posterior mean of contemporaneous effects.

param

Every set of MCMC trace.

param_names

Name of every parameter.

group

Indicators for group.

num_group

Number of groups.

df

Numer of Coefficients: 3m + 1 or 3m

p

VAR lag

m

Dimension of the data

obs

Sample size used when training = totobs - p

totobs

Total number of the observation

call

Matched call

process

Description of the model, e.g. VHAR_SSVS_SV, VHAR_Horseshoe_SV, or VHAR_minnesota-part_SV

type

include constant term (const) or not (none)

spec

Coefficients prior specification

sv

log volatility prior specification

intercept

Intercept prior specification

init

Initial values

chain

The numer of chains

iter

Total iterations

burn

Burn-in

thin

Thinning

y0

\(Y_0\)

design

\(X_0\)

y

Raw input

If it is SSVS or Horseshoe:

pip

Posterior inclusion probabilities.

Details

Cholesky stochastic volatility modeling for VAR based on $$\Sigma_t^{-1} = L^T D_t^{-1} L$$, and implements corrected triangular algorithm for Gibbs sampler.

References

Carriero, A., Chan, J., Clark, T. E., & Marcellino, M. (2022). Corrigendum to “Large Bayesian vector autoregressions with stochastic volatility and non-conjugate priors” [J. Econometrics 212 (1)(2019) 137-154]. Journal of Econometrics, 227(2), 506-512.

Chan, J., Koop, G., Poirier, D., & Tobias, J. (2019). Bayesian Econometric Methods (2nd ed., Econometric Exercises). Cambridge: Cambridge University Press.

Cogley, T., & Sargent, T. J. (2005). Drifts and volatilities: monetary policies and outcomes in the post WWII US. Review of Economic Dynamics, 8(2), 262-302.

Gruber, L., & Kastner, G. (2022). Forecasting macroeconomic data with Bayesian VARs: Sparse or dense? It depends! arXiv.