Multisectoral structures and policy design

Abstract

In this article, an attempt is made to identify a convenient structure of a policy variable, the final demand vector, through the use of a multisectoral model. The method used relies on a specific spectral decomposition that allows for the quantification of the scale-effect of each structure that the policy variable can assume on the structures of the objective variable, the total output vector. This quantification is of aggregate type, since each scalar obtained is valid for all the sectoral components of both the policy variable and the objective variable. But, more relevant, they are consistent with the multisectoral feature of the model, overcoming the objections put forward by the theory of aggregation. In fact, the aggregation theory states that if we aggregate sectors, we obtain a new model with different structural properties, while, in our case, the aggregate scalars that we obtain for each structure are perfectly consistent with the original model. We call these scalars macroeconomic multipliers since they say how many times the modulus of the multisectoral policy variable is multiplied to obtain the modulus of the multisectoral objective variable. Once the structures and the associated macro multipliers are identified, the policy maker can have a complete picture of the economic structure of the objective variable that can be attained and determine a convenient structure of the policy variable choosing either one structure or a combination of the structures identified.

Keywords:

• structural change
• policy control
• macro multipliers analysis
• I-O model

Acknowledgement

The authors gratefully thank the referees for their helpful comments.

Notes

¹ The non-square matrix case is easily developed along the same lines.

² This type of standardisation is particularly convenient since it is invariant with respect to an axes rotation, a characteristic that the sum of the vector's elements does not have. This feature allows for the consistent separation of the scale effects from the structure effects on the output vector variable. Employment can be immediately determined from total output, through the use of sectoral labour/output coefficients. On the other hand, in standard empirical IO tables the value-added components cannot be reduced to one, wages and salaries, since the IO accounting criteria usually imply the consideration, among value added, of at least seven components: taxes on products, other indirect taxes on products, subsidies on products, consumption of fixed capital, together with wages and salaries, employers' social contributions, other incomes.

Notes

1. See Lee (1986) and Polenske and Jordan (1988) for the all type of multipliers.

2. See Appendix A for a mathematical derivation.

3. See Table A2 for the classification of the branches.