7. Conclusion
This paper has provided a dynamic panel data analysis of patent filings and business cycles. Thus far, limited empirical studies have been conducted on the relationship between patenting and business fluctuations. Moreover, prior work has not applied methods for distinguishing between market trends and cycles, such as filtering methods that extract cyclical shocks from movements in a time-series, but has relied instead on ad-hoc, informal measures of shocks. By applying more rigorous methods to a broad panel dataset of source countries of the EPO, this paper finds that the response of patent filings to GDP shocks is quite elastic and pro-cyclical, lending support to the resource effect of business cycles on innovation. The findings are also robust to alternative measures of business cycles, and thus do not depend on the HP-filtering method we adopted. The sensitivity of patent filings to business cycles is also no different during the global financial crisis of 2008–2009 than during the rest of the sample period. Lastly, some out-of-sample forecasting experiments show that incorporating business cycles can help improve the accuracy of predicting patenting behavior. They show, however, that while GDP shocks can result in significant perturbations in patenting, the effects of cyclical shocks on patenting are short-lived.
