ARMA Sieve bootstrap unit root tests

Augmented Dickey-Fuller unit root tests may severely overreject when the DGP is a general linear process. The use of the AR sieve bootstrap, proposed by Park (2002) and Chang and Park (2003), may alleviate this problem. We propose sieve bootstraps based on MA and ARMA approximations. Invariance principles for the partial sum processes built from these sieve bootstrap DGPs are established and a proof of the asymptotic validity of the resulting ADF bootstrap tests is provided. Through Monte Carlo experiments, we find that the rejection probabilities of the MA and ARMA sieve bootstraps are often lower and more robust to the underlying DGP than that of the AR sieve bootstrap. In particular, the new sieve bootstraps perform much better than the AR sieve when a large MA root is present. We also find that the ARMA sieve bootstrap requires only a very parsimonious specification to achieve excellent results.