Event Details

Inference on Time Series with Changing Mean and Variance (Liudas Giraitis )

The paper develops point estimation and large sample statistical inference with respect to a semiparametric model for time series with moving mean and unconditional heteroscedasticity. These two features are modelled non parametrically, whereas autocorrelations are described by a short memory stationary parametric time series model. When the mean is correctly assumed to be constant, Whittle estimates that ignore the heteroscedasticity are found to be consistent for the dependence parameters, and asymptotically normal with parametric rate. Allowing a slowly time-varying mean we resort to trimming out of low frequencies to achieve the same outcome. Returning to finite order autoregression, nonparametric estimates of the varying mean and variance arengiven asymptotic justification, and forecasting formulae developed. Finite sample properties are studied by a small Monte Carlo simulations, and an empirical example is also included.