rcoint: Random generation of cointegrated sequences

In egcm: Engle-Granger Cointegration Models

Random generation of cointegrated sequences

Description

Generates a random pair of cointegrated sequences

Usage

rcoint(n, 
  alpha = runif(1, -10, 10), 
  beta = runif(1, -10, 10), 
  rho = runif(1, 0, 1), 
  sd_eps = 1, 
  sd_delta = 1,
  X0=0,
  Y0=0)

Arguments

n	number of observations in each sequence
alpha	constant term of linear relation
beta	slope term of linear relation
rho	coefficient of mean reversion
sd_eps	standard deviation of innovations in first sequence
sd_delta	standard deviation of innovations in residual sequence
X0	initial value of first sequence
Y0	initial value of second sequence

Details

Generates a random pair of cointegrated sequences. The sequences are constructed by first generating two random sequences that are independent and normally distributed. The elements of the first sequence, epsilon[i], have standard deviation sd_eps, while those of the second sequence, delta[i], have standard deviation sd_delta. Having generated these two sequences, the cointegrated sequences X[i] and Y[i] are generated according to the following relations:

X[i] = X[i-1] + epsilon[i]

R[i] = rho * R[i-1] + delta[i]

Y[i] = alpha + beta * X[i] + R[i]

Value

Returns a two-column data.frame containing the randomly generated cointegrated sequences.

Author(s)

Matthew Clegg [e-mail]

See Also

rar1 sim.egcm egcm

Examples

xy <- rcoint(1000, alpha = 1, beta = 2, rho = 0.8)
egcm(xy)

egcm documentation built on March 7, 2023, 6:31 p.m.