Distribution for Unit Root Statistics
Purpose
The way to generate distribution for bias and Pseudo t statistic
Bias simulation
> set.seed(1977)
> results <- data.frame()
> i <- 1
> for (i in 1:5000) {
+ T <- 1000
+ inc <- rnorm(T)
+ y <- cumsum(inc)
+ num <- T^-1 * sum(y[1:(T - 1)] * inc[2:T])
+ denom <- T^-2 * sum(y[1:(T - 1)]^2)
+ temp <- num/denom
+ results <- rbind(results, temp)
+ }
> quantile(results[, 1], probs = c(0.01, 0.05, 0.1))
1% 5% 10%
-14.371762 -8.183875 -5.866235 |
The above quantiles are same as the ones given in Fuller Table.
> par(mfrow = c(1, 2)) > hist(1000 * (results[, 1]), xlab = "", main = expression(paste(hat(delta), + " Density"))) > plot(ecdf(1000 * (results[, 1])), xlab = "", main = expression(paste(hat(delta), + " CDF"))) |

Pseudo t statistic Simulation
> set.seed(1977)
> results.t <- data.frame()
> i <- 1
> for (i in 1:5000) {
+ T <- 1000
+ inc <- rnorm(T)
+ y <- cumsum(inc)
+ num <- sum(y[1:(T - 1)] * inc[2:T])
+ denom <- sum(y[1:(T - 1)]^2)
+ temp <- num/denom
+ sigma.inc <- T^(-1) * (sum(inc^2))
+ sigma <- sigma.inc * sum(y[1:(T - 1)]^2)^(-1/2)
+ tau <- (temp)/sigma
+ results.t <- rbind(results.t, tau)
+ }
> quantile(results.t[, 1], probs = c(0.01, 0.05, 0.1))
1% 5% 10%
-2.640077 -1.979768 -1.659552 |
> par(mfrow = c(1, 2)) > hist((results.t[, 1]), main = expression(paste(hat(tau), " density"))) > plot(ecdf((results.t[, 1])), main = expression(paste(hat(tau), + " CDF"))) |

Comparison of cdf Standard Normal Vs Tauu
> par(mfrow = c(1, 1))
> plot(ecdf(rnorm(T)), main = expression(paste(hat(tau), " CDF")),
+ pch = 19, col.points = "red", xlim = c(-4, 4))
> par(new = T)
> plot(ecdf(results.t[, 1]), main = expression(paste(hat(tau),
+ " CDF")), pch = 19, col.points = "blue", xlim = c(-4, 4))
> legend("topleft", legend = (c("StdNormal", "Tau")), fill = c("red",
+ "blue")) |

Comparison of density Standard Normal Vs Tauu
> par(mfrow = c(1, 1))
> plot(density(rnorm(T)), main = "", col = "red", lwd = 4, xlim = c(-4,
+ 4), ylim = c(0, 0.5), xlab = "", ylab = "")
> par(new = T)
> plot(density(results.t[, 1]), main = "", col = "blue", lwd = 4,
+ xlim = c(-4, 4), ylim = c(0, 0.5), xlab = "", ylab = "")
> legend("topleft", legend = (c("StdNormal", "Tau")), fill = c("red",
+ "blue")) |
