Abstract:
Empirical evidence on the cyclical behavior of technology shocks, or the relative importance of technology shocks versus other structural shocks as sources of fluctuations, hinges crucially on the identification of technological changes. In this paper, we study different measures of technology in order to find out (i) to what extent they capture the same underlying phenomenon and (ii) whether the implications for macroeconomic theory vary between approaches. Several variations of the production function approach and structural VAR models are investigated: the classic Solow residual, the refined Solow residuals of Burnside et al (1995) and Basu and Kimball (1997), large cointegrated VAR models as in King et al (1991) and a small VAR in first differences à la Galí (1999). It turns out that the different measures of technological change are reasonably coherent when applied to US data. However, they are often insignificantly related in the case of Sweden. Furthermore, our results do not support the hypothesis that business cycle fluctuations are primarily drive by changes in technology.