VARsim: Simulate from given (pre-estimated) VAR process

In sstoeckl/pensionfinanceLi: Optimal Pension Decisions in Liechtenstein

Description

Simulate from given (pre-estimated) VAR process

Usage

VARsim(
  V,
  simN = 1000,
  age.max = 122,
  age = 0,
  frequency = 4,
  s = NULL,
  covres = NULL
)

Arguments

V	Output of a VAR Estimation (in package tsDyn)
simN	optional, number of simulated return series to generate, defaults to 1000
age.max	optional,maximum age, defaults to 122
age	optional,starting age, defaults to one, as we usually want to generate all lifetime returns and select from them accordingly
frequency	optional,defaults to 4 (quarterly), could also be 12 (monthly)
s	optional, (from VAR.sim): Starting values (matrix of dimension lag x k) for the VAR to simulate, in our case this defaults to the steady state calculated as $s=(diag(k)-B)^-1 * c$, where k is the dimension of the VAR, B the coefficients of parameters and equations and c the intercepts of the equations, all taken from V
covres	optional, Variance-covariance matrix for the innovations, defaults to the covariance matrix of the VAR given by V as $var(V$residuals)/(V$T-2-V$k)*(V$T-2)$

Value

An array of dimension (T,k,simN) containing simN simulated return series for T years and k variables

Examples

data(V)
retPA <- VARsim(V,simN=10)

sstoeckl/pensionfinanceLi documentation built on Dec. 2, 2020, 3:26 a.m.