Social accounting matrix: A new estimation methodology

In this article, we present a new methodology to develop bootstrap estimates of a social accounting matrix (SAM), by combining entropy minimization and Monte Carlo simulation techniques. An application is presented to the Italian economy, demonstrating how a set of policy measures can be evaluated by incorporating the prior degree of uncertainty on the model parameters and the historical volatility in the main variables. The results of this exercise show that the methodology proposed provides specific evaluations of the policy measures considered, as well as a rich informational structure on the extension and the limitations of the inference from the data and the economic model.