This paper investigates the permanent effect on total factor productivity (TFP) of temporary shocks. We estimate a structural vector autoregression to test the predictions of endogenous growth models over the business cycle. According to theory, the stock of technological knowledge promotes its flow as researchers “stand on the shoulders of giants.” Therefore, if R&D investment is pro-cyclical—as data show and theory predicts—a recession leads to a temporary deviation of the R&D level from its trend, thus reducing new knowledge creation. The lost technological advancements cause the economy to follow a parallel but permanently lower growth path. Our findings align with the primary theoretical prediction. Quantitatively, the US economy forgoes approximately 1.3% in TFP following an increase in cyclical unemployment that peaks at 1 percentage point above the mean. The historical variance decomposition shows a strong positive effect during the boom of the late 1960s and strong negative effects around the Volcker disinflation period and the Great Recession. Finally, we estimate the effects on R&D of a TFP shock to differentiate between different explanations on how the R&D pro-cyclicality arises. Our results align with models where financial frictions or nominal rigidities drive it.