Technological Change, Projection of the Technology Matrix and the Hypothesis of Negative Coefficient Changes: Parametric and Non-parametric Tests with Swedish Input–Output Data

Abstract

Input–output coefficients at constant prices have a role in economic analysis when we want to distinguish between real effects and pure price effects. The speed and extent of technological change in modern economies is one of the main reasons why such coefficients change over time and the empirical evidence is substantial. Two models suggested in the literature for bringing about changes in the technology matrix are examined. We find support in historical comparisons for use of models with negative coefficient changes over time when projecting the technology matrix. Whatever the source of changes in technical coeficients, the null hypothesis of random changes could be rejected in favour of negative coefficient changes. Strongest support is given to this alternative hypothesis when relative coefficient changes are considered.