Cointegration in multivariate time series

Cointegration in multivariate time series'. The aim of this paper is to est ablish a forecast equation with three non-stationary time series variables. An attempt will be made to show the effect of the size of a community and the level of atmospheric pollution on the number of admissions with chest i llnesses to the emergency department of a hospital. Various regression syst ems are put to the test: regression with variables without transformation, with differentiated variables and through a Box-Jenkins transfer function, all of which show problem in fitting the data. It is found that the three v ariables are cointegrated and that a regression equation with an error corr ection mechanism which correctly adjusts the variables actually exists. The advantages of the cointegration and the error correction mechanism over ot her regression systems are shown.