Wavelet Improvement of the Over-rejection of Unit root test under GARCH errors

In this paper, we use the wavelet technique to improve the over-rejection problem of the traditional Dickey-Fuller test for unit root when the data suffers from GARCH (1,1) effect. The logic is based on that the wavelet spectrum decomposition can separate out information of different frequencies in the data series. We prove that the asymptotic distribution of our test is similar to the traditional Dickey-Fuller(1979 and 1981) type of tests. The small sample distribution of the new test is assessed by means of Monte Carlo simulation. An empirical example with data on immigration to Sweden during the period 1950 to 2000 is used to illustrate the test. The results reveal that using the traditional Dickey-Fuller type of test, the unit root is rejected while our wavelet improved test shows the opposite result.