Forecasting, Structural Time Series Models and the Kalman Filter

Inthisbook,AndrewHarveysetsouttoprovideaunifiedandcomprehensivetheoryofstructuraltimeseriesmodels.UnlikethetraditionalARIMAmodels,structuraltimeseriesmodelsconsistexplicitlyofunobservedcomponents,suchastrendsandseasonals,whichhaveadirectinterpretation.Asaresultthemodelselectionmethodologyassociatedwithstructuralmodelsismuchclosertoeconometricmethodology.Thelinkwitheconometricsismadeevencloserbythenaturalwayinwhichthemodelscanbeextendedtoincludeexplanatoryvariablesandtocopewithmultivariatetimeseries.Fromthetechnicalpointofview,statespacemodelsandtheKalmanfilterplayakeyroleinthestatisticaltreatmentofstructuraltimeseriesmodels.ThebookincludesadetailedtreatmentoftheKalmanfilter.Thistechniquewasoriginallydevelopedincontrolengineering,butisbecomingincreasinglyimportantinfieldssuchaseconomicsandoperationsresearch.Thisbookisconcernedprimarilywithmodellingeconomicandsocialtimeseries,andwithaddressingthespecialproblemswhichthetreatmentofsuchseriesposes.Thepropertiesofthemodelsandthemethodologicaltechniquesusedtoselectthemareillustratedwithvariousapplications.TheserangefromthemodelllingoftrendsandcyclesinUSmacroeconomictimeseriestotoanevaluationoftheeffectsofseatbeltlegislationintheUK.