AN LM TEST FOR A UNIT-ROOT IN THE PRESENCE OF A STRUCTURAL-CHANGE

In this paper, we examine a suitably modified version of the unit root test proposed by Schmidt and Phillips (1992). A one-time structural break in the intercept does not affect its asymptotic distribution under the null hypothesis, and this is true whether the break is allowed for in the model or not. This implies that the asymptotic validity of this test statistic under the null is not affected by the incorrect placement of the structural break, by the allowance for a break when there is no break, or by no allowance for a break when there is a break.