Welfare implications of increasing spatial resolution in national CGE models for climate change impact assessment

Computable General Equilibrium (CGE) models are a widespread tool to assess the economic impacts of climate change because they allow to capture the inter-sectoral and inter-regional interactions of the environmental and climatic shocks. The geographical resolution adopted in CGE models is usually the country level, however, climate impacts can be heterogenous even within countries as regions located at different latitudes or altitudes respond differently to changes in temperature and precipitations. Our aim in this work is to evaluate the welfare implications of increasing the spatial resolution of a CGE model when combined with sub-national level impacts of climatic change on sectoral Gross Value Added (GVA) derived through spatial econometrics. We compare a standard version of a national CGE model calibrated for Italy with its sub-national version (twenty Italian NUTS-2 regions). The objective is not only to show that the aggregated results obtained can differ in terms of national GDP loss but also to provide an economic explanation for these differences. Specifically, we will explain the importance of intra-national trade (flows of imports and exports between sub-national units) and capital/labour re-location across regions. These economic channels are absent in the national version of the CGE.