Modelling the Sectoral Structure of the Final Output

This paper examines the modelling complications that appear when some macroeconomic behavioral relationships interact with structural variables, even under a given A matrix. The main problem is related to the situation when: a) the final consumption, gross fixed capital formation, inventory changes, export, import (all of them at market prices), and gross value added (at production prices) are estimated as macro-indicators; and b) the output (at production prices) is determined at a disaggregated level. The so-called demand-side or supply-side approaches are possible; here, the supply-side approach is in focus. With such a goal, the regression and linear weighted average (in the Fisher version) techniques are discussed as the main tools for estimating sectoral weights of the final output. For the linear weighted average method, the paper sketches – as a discussion proposal – a methodology for the optimal selection of the length (number of terms) of the moving average. As a primary database, the Romanian input-output tables for 1989–2009, aggregated into 10 sectors, were used