Guido Ferrari¹, Tiziana Laureti², Luca Secondi¹¹Department of Statistics, University of Florence
² Department of Statistics and Mathematics for economic research, University of Naples “Parthenope”
Abstract
The computation process of a general equilibrium (GE) model, based on a Social Accounting Matrix (SAM), consisting of both calibration and estimation phases, requires parameters and elasticity values in order to entirely define the model equations. Concerning the behavioural functions of the consumption sphere, data contained in the SAM  do not represent sufficient information to allow a consistent estimation of a demand function system. In order to overcome this problem, cross-section or time-series data, available from household expenditure survey, are usually joined together with the national accounts information. However, lack of additional data, which may afflict both national or sub-national context, together with the related comparability and classification data problems represent a considerable hurdle for the model implementation. In this paper we suggest a new estimation strategy, based on the Generalized Maximum Entropy (GME) method by only using the expenditure information contained in a Regional SAM (RSAM). This procedure allow us to obtain consistent estimates of the demand functions unknown parameters.
Estimation of a Working Leser demand model parameters will be carried out by using the regional SAM of the Italian region Sardinia for the year 2001.The results, which are suitable for a RCGE context, will be obtained at a level of economic branches and income groups.
Keywords: demand functions, Working-Leser model, GME, CGE models, RSAM.