Broken mean stationarity and the validity of the Dickey-Fuller test: the case of controlled inflation

Authors

  • Daniel Ventosa-Santaulária Departamento de Economía y Finanzas, DCEA-Campus Marfil Fracc
  • Manuel Gómez-Zaldívar Departamento de Economía y Finanzas Universidad de Guanajuato

DOI:

https://doi.org/10.12660/bre.v29n12009.2693

Keywords:

Dickey-Fuller test, Mean Stationary Process, Structural Breaks

Abstract

This article proves the asymptotic efficiency of the Dickey Fuller (DF) test when the data generating process of the variable under consideration is in fact mean stationary with breaks. Monte Carlo simulations show that asymptotic properties remain valid for sample sizes of practical interest. Our results complement those already available in the literature, which can be summarized as follows: DF-type tests (i) over-accept the null hypothesis of unit root when there is a break in the trend-stationary process, and; (ii) over-reject the null hypothesis when there is a level/trend break in the unit root process. We illustrate the performance of the DF-t test by studying inflation rate series, a variable that should be stationary if the monetary authority follows an effective monetary policy: shocks are short-lived, therefore, inflation fluctuates randomly around specific goals

Downloads

Published

2009-05-01

Issue

Section

Articles