Predicting the regional impact of a full employment policy in Canada: a Box-Jenkins approach

Recently developed time series techniques are applied to the problem of predicting the possible regional impacts of a full employment policy in Canada. A brief critique of existing spatial labor market methods is presented. The focus of this critique is concerned with issues of serial autocor-relation and the high risks of finding spurious relationships associated with Brechling-type models. It is argued, on the basis of the derived results, that a full employment policy would substantially benefit regions with already high levels of unemployment. However, it also appears that many depressed regions would not attain national levels of full employment.