This thesis develops a new structure that explicitly combines two CGE models, a national and a regional, in an integrated structure that gives the thesis model the name IRES, in short for the Integrated Regional Equation System. The typical features of the integrated structure are the adding-up conditions and the two-way linkages between the national and the regional modules facilitated by the interface shifters. The adding-up conditions ensure the two modules produce consistent results and updated databases. The inclusion of the interface shifters on the one hand plays a role in ensuring compatibility of results of the two modules, i.e. no distortion occurs because technical or taste changes are transferred across modules. On the other hand, the interface shifters assist the operation of IRES in different modes: the model can be used as a top-down model, a bottom-up model or an integrated model where national and regional shocks can be introduced at the same time. Hence, IRES has more flexibility in its application than a regional model or a national model alone, as IRES can make use of availability of data at any levels in the economy. IRES has a new labour market in which regional migration is no longer the only factor that settles the labour market as in the original setting of the MMRF model. Regional unemployment and regional participation rates are modelled to response to changes in regional employment growth using elasticities estimated econometrically in this thesis. IRES implements historical patterns of regional migration so that results of regional migration are consistent with observed patterns. Altogether, regional migration, regional unemployment and participation rates determine the equilibrium of the labour market. IRES adopts new approaches to modelling margin demands and indirect taxes. These new approaches are very effective in reducing the size of IRES but they do not compromise the use of the model. These approaches are readily applicable to any other regional CGE models.