We consider a recently proposed class of nonlinear time series models and focus mainly onmisspecification testing for models of such type. Following the modeling cycle for nonlineartime series models of specification, estimation and evaluation we first treat how to choosean adequate transition...

Determining good parameter estimates in ESTAR models is known to be difficult. We showthat the phenomena of getting strongly biased estimators is a consequence of the so-calledidentication problem, the problem of properly distinguishing the transition function in relationto extreme parameter...

While it is widely agreed that Purchasing Power Parity (PPP) holds as a long-runconcept the specific dynamic driving the process is largely build upon a priori economicbelief rather than a thorough statistical modeling procedure. The two prevailing timeseries models, i.e. the exponential smooth...

We study the empirical behaviour of semi-parametric log-periodogram estimation forlong memory models when the true process exhibits a change in persistence. Simulationresults confirm theoretical arguments which suggest that evidence for long memory islikely to be found...

We show that tests for a break in the persistence of a time series in the classicalI(0) - I(1) framework have serious size distortions when the actual data generatingprocess exhibits long-range dependencies. We prove that the limiting distributionof a CUSUM of squares based test depends on the...

We examine the asymptotic behavior of unit root tests against nonlinear alternativesof the exponential smooth transition type if the data is erroneously nonlinearly transformed...

We analyze the cross-national distribution of GDP per capita and its evolutionfrom 1970 to 2003. We argue that peaks are not a suitable measure for distinctgrowth regimes, because the number of peaks is not invariant under strictlymonotonic transformations of the data (e.g. original vs. log...

We show that the power of the KPSS-test against inte-gration, as measured by divergence rates of the test statistic underthe alternative, remains the same when residuals from an OLS-regression rather than true observations are used.[...]

Atheoretical regression trees (ART) are applied to detect changes in the mean of a stationarylong memory time series when location and number are unknown. It is shownthat the BIC, which is almost always used as a pruning method, does not operate well inthe long memory framework.[...]