Constructing a Poverty-Focused Social Accounting Matrix for Tunisia Under the New National Accounts System: Methodology, Data Integration, and Policy Applications

This paper constructs the first Social Accounting Matrix (SAM) for Tunisia calibrated to the new Base 2015 national accounts system (SCNT 2015), aligned with the United Nations System of National Accounts 2008. We develop a disaggregated 50-account SAM for 2022 incorporating 18 production sectors, two labor types (skilled and unskilled), five household categories differentiated by poverty status and location (urban poor, rural poor, farm households, rural nonpoor, and urban non-poor), and four institutional sectors. The household disaggregation is calibrated using product-level expenditure data from the 2021 National Survey on Household Budget, Consumption, and Standard of Living (EBCNV), which covers 1,085 consumption items across ten expenditure deciles. We employ Fofana's proportional allocation method for interinstitutional transfers, derive the intermediate consumption matrix from the official supply-use table (Tableau des Ressources-Emplois), allocate factor income to households using occupationbased population weights from the EBCNV, and balance the matrix using weighted crossentropy minimization following Robinson et al. (2001). The resulting SAM reveals that urban poor households allocate 39.9 percent of expenditure to food compared to 31.9 percent for urban non-poor households, an 8-percentage-point Engel gradient that constitutes the primary channel through which food price shocks transmit to poverty outcomes. The inter-industry structure reveals that agriculture's dominant forward linkage runs to food processing (11,640 million TND in intermediate sales), while the chemical industry depends critically on mining inputs (phosphates), establishing the sectoral transmission channels through which global commodity price shocks propagate domestically. To demonstrate the SAM's analytical utility, we implement a computable general equilibrium model with CES production technology using the R CGE package and simulate eight policy scenarios including food subsidy removal, targeted cash transfers, and trade liberalization. The pro-poor policy package combining transfers, agricultural productivity gains, and public service expansion reduces the inequality ratio by 14.5 percent while increasing poor household welfare by 18 percent. The SAM and all replication materials are publicly available.