Stochastic Technical Change and Procyclical TFP The Italian Agriculture Case

TFP comparisons across nations, sectors and time have been frequently used to represent the relative technological levels of the production systems. In this paper the use of the TFP as a proxy of the technological level in Italian agriculture is questioned. The cyclical behaviour seems to be the most important shortcoming of the TFP measure. It clearly reveals a short-run component due to shocks on the supply side. To separate these business cycle effects from real long-run technological change, we use a stochastic process for the generation of technical change induced by R&D and Extension expenditure and a Kalman filter procedure is applied. Doing this, it is possible to separate the short-run and long-run components of the traditional TFP measure. The empirical evidence clearly reveals that supply-side shocks enter the traditional measure of productivity therefore, it represents an erroneous indicator of technological progress. A better indicator can be obtained estimating the longrun component of productivity that seems to be close to a linear trend and significantly induced by R&D and Extension public expenditure.