Guido Ferrari¹, Tiziana Laureti², Luca Secondi¹¹Department of Statistics, University of Florence
²Department of Statistics and Mathematics for economic research, University of Naples “Parthenope”
Abstract
In the computation process of a General Equilibrium (GE) Model, the behavioural parameter values play a key role. Usually, these parameters, are estimated using external data such as time series or cross-section data, in addition to the original dataset, namely the Social Accounting Matrix (SAM). When this additional information are not available, they are taken from similar contexts in literature. However, if the data used do not reflect the same level of aggregation, comparability and classification problems arise. In order to overcome these problems, which may affect the model results, we suggest a new estimation approach based on the macro information contained in a Regional SAM (RSAM) only. On the other hand, since a RSAM does not contain enough information to obtain statistical satisfactory estimates and by using some a priori information coming from the economic theory, we will take advantage of the Generalized Cross Entropy (GCE) which provides a flexible estimation framework.
In this paper we will provide the estimates of the Constant Elasticity of Substitution (CES) production function for the Italian region Sardinia, based on the RSAM for the year 2001. According to the GCE philosophy we will introduce in the estimation process prior information about the unknown parameters both in terms of support space bounds and in terms of estimated prior probability distributions obtained as the results of a Maximum Entropy (ME) estimation as well.
Keywords: CGE models, RSAM, ill-posed, GCE,  CES production function