Bayesian Macroeconometrics in R (BMR) is an R interface to BM++, a templated C++ library for estimating Bayesian Vector Autoregression (BVAR) and Dynamic Stochastic General Equilibrium (DSGE) models.
Features:
- BVAR models with flexible prior specification, including: the Minnesota prior; normal-inverse-Wishart prior; and Matias Villani's steady-state prior.
- BVARs with time-varying parameters.
- Solve DSGE models using the method of undetermined coefficients (Uhlig's method) and Sims' QZ method.
- Estimate DSGE and DSGE-VAR models.
The simplest installation method is via the devtools package.
install.packages("devtools")
library(devtools)
install_github("kthohr/BMR")
Note that BMR requires compilation, so approproate development tools are necessary to install the package.
See http://www.kthohr.com/bmr for documentation and replication files.
Estimate a BVAR model with Minnesota prior:
library("BMR")
data(BMRMCData)
# create bvarm object
bvar_obj <- new(bvarm)
# add data; add a constant; and set lags (p) = 2
bvar_obj$build(bvarMCdata,TRUE,2)
# prior of 0.5 on own-first-lags, and take default values for the hyperparameters
bvar_obj$prior(c(0.5,0.5))
# draw from the posterior distribution
bvar_obj$gibbs(10000)
# posterior mean and variance
bvar_obj$alpha_pt_mean
bvar_obj$alpha_pt_var
# plot IRFs
IRF(bvar_obj,20,save=FALSE)
Solve a three-equation New-Keynesian model:
#
# Solve the NK model with gensys
#
rm(list=ls())
library(BMR)
#
alpha <- 0.33
vartheta <- 6
beta <- 0.99
theta <- 0.6667
eta <- 1
phi <- 1
phi_pi <- 1.5
phi_y <- 0.5/4
rho_a <- 0.90
rho_v <- 0.5
BigTheta <- (1-alpha)/(1-alpha+alpha*vartheta)
kappa <- (((1-theta)*(1-beta*theta))/(theta))*BigTheta*((1/eta)+((phi+alpha)/(1-alpha)))
psi <- (eta*(1+phi))/(1-alpha+eta*(phi + alpha))
sigma_T <- 1
sigma_M <- 0.25
G0_47 <- (1/eta)*psi*(rho_a - 1)
#Order: yg, y, pi, rn, i, n, a, v, yg_t+1, pi_t+1
Gamma0 <- rbind(c( -1, 0, 0, eta, -eta/4, 0, 0, 0, 1, eta/4),
c( kappa, 0, -1/4, 0, 0, 0, 0, 0, 0, beta/4),
c( phi_y, 0,phi_pi/4, 0, -1/4, 0, 0, 1, 0, 0),
c( 0, 0, 0, -1, 0, 0, G0_47, 0, 0, 0),
c( 0, -1, 0, 0, 0, 1-alpha, 1, 0, 0, 0),
c( -1, 1, 0, 0, 0, 0, -psi, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 1, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 1, 0, 0),
c( 1, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 1, 0, 0, 0, 0, 0, 0, 0))
Gamma1 <- rbind(c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, rho_a, 0, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, rho_v, 0, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 1, 0),
c( 0, 0, 0, 0, 0, 0, 0, 0, 0, 1))
#
C <- matrix(0,10,1)
Psi <- matrix(0,10,2)
Psi[7,1] <- 1
Psi[8,2] <- 1
Pi <- matrix(0,10,2)
Pi[9,1] <- 1
Pi[10,2] <- 1
#
Sigma <- rbind(c(sigma_T^2, 0),
c( 0, sigma_M^2))
#
# build the model and solve it
dsge_obj <- new(gensys)
dsge_obj$build(Gamma0,Gamma1,C,Psi,Pi)
dsge_obj$solve()
# Y_t = G * Y_{t-1} + Impact * e_t
dsge_obj$G_sol
dsge_obj$impact_sol
# set shocks (Sigma); then plot IRFs and simulate some data
dsge_obj$shocks_cov = Sigma
var_names <- c("Output Gap","Output","Inflation","Natural Int",
"Nominal Int","Labour Supply","Technology","MonetaryPolicy")
dsge_irf <- IRF(dsge_obj,12,var_names=var_names)
dsge_sim <- dsge_obj$simulate(800,200)
Keith O'Hara
GPL (>= 2)