Econometric inference with transitional macroeconomic time series data : Some theoretical properties and practical implications (CROSBI ID 752605)
Druge vrste radova | elaborat/studija
Podaci o odgovornosti
Cziraky, Dario ; Tišma, Sanja ; Pisarović, Anamarija
engleski
Econometric inference with transitional macroeconomic time series data : Some theoretical properties and practical implications
Transitional macroeconomic time series data are short and generated by DGPs likely to be variant over time with a known "beginning of time". Several assumption needed for standard econometric inference are thus violated. It is shown that problems caused by data properties cannot be solved by imposing distributional assumptions. Subsequently, quarterly data sets still cannot be used for reliable inference with such data. An example is shown where an apparently small change in the length of the estimation sample can cause drastically differing results. However, it is shown that monthly data, if available, might provide solid basis for reliable inference and model building even if the available series are disaggregated proxies. Furthermore, it is argued that due to the nature of transitional DGPs and lack of firm theory of transitional economies, the general-to-specific approach should be preferred even in light of some theoretically appealing results that apparently prefer simple models. In this regard dynamic mis-specification is discussed and illustrated in a simple simulated example.
Data generating process; statistical assumptions; data properties; inference; small samples; asymptotic distributions; ADL models
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o izdanju
Econometric inference with transitional macroeconomic time series data: Some theoretical properties and practical implications
2001.
nije evidentirano
objavljeno