A Flexible Modeling Framework to Estimate Interregional Trade Patterns and Input-Output Accounts

Abstract:     
There are tremendous disparities in economic development across regions in large developing countries such as Brazil, China, India, and Indonesia. Globalization may have different impacts on urban and coastal areas compared with less developed rural and inland regions. A major obstacle in conducting policy analysis for regional economic development under globalization is the lack of consistent, reliable regional data, especially data on interregional trade and interindustry transactions.

This study implements and tests a mathematical programing model to estimate interregional, interindustry transaction flows in a national system of economic regions based on an interregional accounting framework and initial information on interregional shipments. A complete national input-output table plus regional sectoral data on gross output, value-added, exports, imports, and final demand are used as inputs to generate an interregional input-output system that reconciles regional market data and interregional transactions. The model is tested on a four-region 10-sector example against data aggregated from a multi-regional global input-output database, and test results from seven experiments are evaluated against eight mean absolute percentage error indexes. The model has capacity to discover the true interregional trade pattern from highly distorted initial estimates. The paper also discusses some general guidelines for implementing the model for a large-dimension multi-regional account based on real national and regional data.

This paper is a product of the Trade Team, Development Research Group.

Working Paper Series