#' Solving an autonomous LRE model according to Blanchard and Kahn
#'
#' @param A Coefficient matrix for the step function
#' @param n Number of predetermined variables
#'
#' @return Two functions, decision rule g() and a motion function h()
#' @export
#'
#' @examples
lre_auto_bk <- function(A, n) {
sch <- QZ::qz(A)
sch <- QZ::qz.dtrsen(sch$T, sch$Q, abs(sch$W) <= 1)
Q1S <- sch$Q[1:n,1:n]
Q2S <- sch$Q[(n+1):(nrow(sch$Q)),1:n]
g <- function(x0) Q2S %*% solve(Q1S) %*% x0
h <- function(x0) A[1:n,1:n] %*% x0 + A[1:n,(n+1):ncol(A)] %*% Q2S %*% solve(Q1S) %*% x0
return(list(g,h))
}
## ---------------------------------------------------------------------------------------------------------
#' Solving an autonomous LRE model according to the Klein method with use of QZ decomposition
#'
#' @param A Coefficient matrix on the previous step of the difference equation
#' @param E Coefficient matrix of the current(next) step
#' @param n Number of predeterminate parameters
#'
#' @return Two functions g() and h() for a decision rule and law of motion respectively
#' @export
#'
#' @examples
lre_auto_klein <- function(A, E, n) {
sch <- QZ::ordqz(E,A, keyword = "udo")
Z21 <- sch$Z[n:(n+1), 1:n] ## not certain about this
Z11 <- sch$Z[1:n, 1:n]
S11 <- sch$S[1:n, 1:n]
T11 <- sch$T[1:n, 1:n]
ST <- solve(S11) %*% T11
g <- function(x0) Z21 %*% solve(Z11) %*% x0
h <- function(x0) Z11 %*% ST %*% solve(Z11) %*% x0
return(list(g,h))
}
## --------------------------------------------------------------------------------------------------------
#' A combination of the Klein and Blanchard, Kahn methods for the autonomous case.
#'
#' @param A Coefficient matrix on x(t)
#' @param E Coefficient matrix on x(t+1), if not given, the function evaluates the BK method
#' @param n number of predetermined variables
#'
#' @return Motion and decision rule functions according to either the Blanchard Kahn approach or through the Klein method
#' @export
#'
#' @examples
lre_auto <- function(A, E = NULL, n) {
if (is.null(E)) {
lre_auto_bk(A, n)
} else {
lre_auto_klein(A, E, n)
}
}
## ------------------------------------------------------------------------------------------------------
#' Solving LRE with exogenous shocks of the AR variety
#'
#' @param A Coefficient matrix on x(t)
#' @param E Coefficient matrix on x(t+1)
#' @param B Coefficient matrix on the error term u
#' @param Phi The AR process for the error term
#' @param n number of predetermined variables(excluding the exogenous shocks)
#'
#' @return Motion and decision rule functions for the LRE with AR exogenous shocks
#' @export
#'
#' @examples
lre_ar <- function(A, E, B, Phi, n) {
E1 <- dbind(diag(length(Phi)), E) # for new E
zerovec <- as.vector(matrix(0, nrow = dim(A)))
Ax <- rbind(zerovec, A)
Bx <- rbind(t(Phi), t(t(B)))
A1 <- cbind(Bx, Ax) # for new A
n1 <- n + length(Phi)
lre_auto_klein(A1, E1, n1)
}
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.