Testing for unit roots in nonlinear dynamic heterogeneous panels

In this paper we present a unit root test against a nonlinear dynamic heterogenous panel with each cross section modelled as an LSTAR model. All parameters are viewed as cross section specific. We allow for serially correlated residuals over time and heterogenous variance among cross sections. The test is derived under three special cases: (i) the number of cross sections and observations over time are fixed, (ii) observations over time are fixed and the number of cross sections tend to infinity, and (iii) first letting the number of observations over time tend to infinity and thereafter the number of cross sections. Small sample properties of the test show modest size distortions and satisfactory power being superior to the Im, Pesaran, and Shin t-type of test. We also show clear improvements in power compared to a univariate unit root test allowing for nonlinearities under the alternative hypothesis.