Creative Destruction Cycles: Schumpeterian Growth in an Estimated DSGE Model
In this paper I incorporate a Schumpeterian mechanism of creative destruction in a standard DSGE framework. In the model, a sector of profit-maximizing innovators endogenously generates the process of technological advancement, ultimately determining the economy’s growth rate. I estimate the model using Bayesian methods on U.S. data (1984q2-2016q4), and show that the model is able to produce Total Factor Productivity (TFP) estimates comparable to those generated using a standard reduced form methodology. The endogenous TFP channel acts as a persistence mechanism in the model. Furthermore, the model provides an alternative solution to the Jaimovich and Rebelo (2009) puzzle, which does not rely on the use of non-standard preferences. The impulse responses analysis shows that while temporary TFP gains due to exogenous TFP shocks have a contractionary effect on hours worked, while shocks to the innovation step and to the marginal efficiency of R&D investment produce TFP gains associated with an increase of hours worked. The variance decomposition analysis shows that the shocks originated in the innovation sector are key drivers of the business cycle, while the effect of standard Smets-Wouters shocks (e.g. shocks to the risk premium, government spending, etc.) on TFP dynamics is limited. Jointly, shocks to R&D are shown to account for the 25% - 45% of the GDP volatility and for the 50% - 85% of the TFP volatility. These findings challenge the view that the TFP slowdown was triggered by a demand slump, and support the hypothesis of Fernald (2015), who associates the 2000s U.S. productivity slowdown to the end of a phase of intense and unsustainable development and adoption of IT technologies started in the late 1990s and culminated with the dot-com bubble.