tvtpnormalX: Dynamic Gaussian Copula Model with exogenous variable

In woraphonyamaka/TVTPcop: Dynamic Copula

Description

This function estimates the parameter(s) of a bivariate dynamic gaussian copula with exogenous variable using maximum likelihood estimation.

Usage

dynamicnormalX(data,z, plot=TRUE)

Arguments

data	The matrix T x 2 margins of the two random varibles
data	plot the time varying dependence parameter (plot=TRUE, default)
z	The vector T x exogenous variable

Details

This model can be used as an alternative way of connecting the marginal distributions to restore the joint distribution. The ARMA(1,10) process of Patton(2006) is extened to ARMAX(1,10), thus the additional exogenous variable is added in the time varying copula equaiton.

Value

result	Estimated parameter, Standard error, t-stat, p-value
AIC	Akaiki Information Criteria
BIC	Bayesain Information Criteria
Loglikelihood	Log Likeihood function'
tvtpdep	the vector of time varying dependence parameter

Author(s)

Woraphon Yamaka

References

Patton, A. J. (2006). Modelling asymmetric exchange rate dependence. International economic review, 47(2), 527-556.

Maneejuk, P., & Yamaka, W. (2019). Predicting Contagion from the US Financial Crisis to International Stock Markets Using Dynamic Copula with Google Trends. Mathematics, 7(11), 1032.

Examples

library(VineCopula)
library("tseries")
library("quantmod")
library("PerformanceAnalytics")

tckk <- c("^N225", "CL=F","^DJI") # ticker names defined
numtk <- length(tckk);
ustart <- "2010-12-30";
uend <- "2020-2-29" # start and end date
all_dat <- list(); # empty list to fill in the data
for(i in 1:numtk)
{
  all_dat[[i]] <- xxx <- get.hist.quote(instrument = tckk[i], start=ustart, end=uend, quote = c("Close"), provider = "yahoo", compression = "m")
}
OIL=all_dat[[2]]+0.000001
N225=all_dat[[1]]
DOWJ=all_dat[[3]]
rOIL=diff(log(OIL))     # exogenous variable
rN225=diff(log(N225))
rDOWJ=diff(log(DOWJ))
# normal margins
u=pnorm(rDOWJ/sd(rDOWJ))
v=pnorm(rN225/sd(rN225))

# Correlation rho
cor(u,v)
# maximum likelihood estimates for comparison
BiCopEst(u, v, family = 1, method = "mle")
# Dynamic Gaussian Copula
data=cbind(u,v)
model=dynamicnormalx(data, z=rOIL, plot=TRUE)

woraphonyamaka/TVTPcop documentation built on June 6, 2020, 9:47 a.m.