The Persistent Statistical Structure of the US Input-Output Coefficient Matrices: 1963-2007

01 April 2018

Luis Daniel Torres Gonzalez and Jangho Yang
View Journal Article / Working Paper

The paper finds evidence for the existence of a statistical structure in the US input-output (I-O) coefficient matrices A = {a i j } for 1963-2007. For various aspects of matrices A we find smooth and unimodal empirical frequency distributions (EFD) with a remarkable stability in their functional form for most of the samples. The EFD of all entries, diagonal entries, row sums, and the (left and right) Perron-Frobenius eigenvectors are well described by fat-tailed distributions while the EFD of column sums and eigenvalue moduli are well explained by the normal distribution and the Beta distribution, respectively. The paper provides several economic interpretations of these statistical results based on the recent developments in the I-O analysis and the price of production literature. Our findings question some probabilistic assumptions conventionally adopted in the stochastic I-O analysis literature and call for a statistical approach to the discussion of the structure of I-O matrices.