This paper analyses the persistence of research intensity in the OECD over the period 1870-2018. The goal is to test if the conclusion of the study conducted by Ang and Madsen (Ang, J. B., & Madsen, J. B. (2011). Can second-generation endogenous growth models explain the productivity trends and knowledge production in the Asian miracle economies? The Review of Economics and Statistics 2011, 93(4), 1360-1373), namely that the Schumpeterian growth models predict that research intensity is stationary, is correct. Using fractional integration methods on annual research intensity from 16 OECD countries, we observe that the series are very persistent. The order of integration is observed to be statistically higher than 1 in all the countries except Spain, rejecting thus the hypothesis of stationarity. When the likelihood of non-linear trends is considered in the analysis, the results are not materially different. An implication of the results is that policies aimed at boosting research activities will have a long-term impact on research intensity.