variables facilitating the adjustment process through macroeconomic equilibrium.
In this paper we analyze whether the apparent comovement of the unemployment ratesfor the Central Eastern and European Countries (CEEC) can be explained by a common
force, possibly linked to the process of economic integration. To test for comovement we
need first to asses the order of integration of unemployment rates. If unemployment
rates are non-stationary, comovement can be explained in terms of cointegration. On the
contrary, if unemployment rates are stationary, something else is going on. We show that
the observed common behavior of the unemployment rates analyzed can be explained by
the existence of a common nonlinear component.
Unit root tests have been traditionally used in the empirical literature on unemploy-
ment to test for the natural rate hypothesis against either the hysteresis or the structural-
ist view. Hysteresis in unemployment states that unemployment shocks have permanent
effects over the long run path of the variable and therefore the variable will be well
characterized as a unit root process. There are a number of possible justifications for
explaining unemployment hysteresis. Examples include the existence of unions with high
negotiation power, soft protection schemes, too high real wages and the social stigma of
being long term unemployed. Also, it is possible to observe a slow speed
of adjustment towards the equilibrium (or even moving equilibrium) of unemployment
rates. This is the so-called "persistence" hypothesis which implies that the unemploy-
ment rate may be characterized as a near unit root or as a fractional integrated process.
On the other hand, the structuralist view im-
plies that most shocks cause temporary movements of unemployment around the natural
rate, but occasional shocks might cause permanent changes in the natural rate itself.
Unemployment fluctuations are therefore characterized by movements around a shifting
natural rate. The structuralist view implies that unemployment is stationary around a
process that is subject to structural breaks.
The contribution of this paper is twofold. First, we apply unit root tests that take into
account two sources of nonlinearities, i.e. in the deterministic components, Leybourne et
al. (1998), and in the autoregressive parameter, Kapetanios et al. (2003); and second,
we test whether there is a common nonlinear trend, Anderson and Vahid (1998) between
those stationary unemployment rates.