An Inventory of Datasets for the Compilation of Regional Social Accounting Matrices for the EU

Due to the ever-increasing demand for model-based analyses of regional development policies in a multi-sector context, the Institute for Prospective Technological Studies (IPTS) launched in 2009 a project on the compilation of regional Social Accounting Matrices (SAM) for the NUTS2 regions of the EU (IOTNUTS2). The SAMs cover the time span between 2000 and 2005. This database shall permit general equilibrium analyses of policies like reforestation programmes, the promotion of investment in agro-tourism or environmental services, and the support for the production of renewable energy by farming enterprises, and more generally to evaluate the rural development pillar of the European Common Agricultural Policy. Such measures target primarily the agricultural sector, but are likely to have an impact on other economic sectors and aggregate regional income, depending on the regional economic structure and the dominance of agriculture. Addressing regional heterogeneity requires multisector data on sub-national scale. Such datasets are usually not sufficiently detailed, if available at all, which gave raise to numerous non-survey methods to generate regional InputOutput tables based on combinations of available regional indicators and national datasets (e.g. Location Quotients, GRIT methods). One particular challenge encountered during the IOTNUTS2 project was the high level of sectoral aggregation in regional branch accounts provided by ESTAT, where e.g. agriculture, forestry, and fisheries are merged. Given the interest in spill over effects of dominantly agricultural policies, more detailed information was required. Therefore, statistical organisations of the 27 EU Member States were contacted and the results of previous projects on regional databases were screened. This paper gives an overview on the compiled inventory on regional datasets for EU27, starting with the target structure of the database and available national and regional datasets from ESTAT. Based on this, we discuss the datasets obtained from national statistical departments (NSO) and from previous projects with comparable aims. In general, we achieved a relevant informational gain over the exclusive use of ESTAT datasets for several member states while for some of them (i.e. Bulgaria) not as large as initially expected. Furthermore, we used the obtained NSO data to test the reliability of non-survey methods for the combination of national and regional datasets. It appeared that particularly forestry, mining/quarrying, and fuel industries displayed substantial deviations between derived indicators and those obtained from NSO, namely intermediate demand and gross output. For other branches, information could either be obtained (e.g. agriculture) or derived indicators proved to be close to the NSO values (most service sectors). In general, we conclude that for the majority of considered economic sectors, non-survey methods can generate reliable substitutes for otherwise collected indicators, but not for some critical branches which are usually concentrated in some regions and may 2 dominate the regional economic structure (forestry, mining, fisheries). This result can be helpful for future projects with comparable aims as we suggest that instead of attempting to sample economy-wide datasets, a focus on the mentioned critical sectors would provide higher marginal informational gains. The data collected from all the different sources are firstly utilized to populate national Input-Output tables for the EU 27 Member States. These matrixes are then balances following standard cross-entropy methods. These tables, with the suitable level of disaggregation, could be utilized as starting point to update the EU InputOutput tables that IPTS provided to the GTAP Consortium.