Using two waves of the Community Innovation Survey (CIS) for the Netherlands, this paper integrates recent lines of research to estimate the contribution of innovation to manufacturing multifactor productivity (MFP) growth. The model uses CIS data to control for the complementarity between internal and external knowledge bases, and investigates the importance of within-firm time interdependencies for inputs into innovation and innovation output. The results show the benefits of including more information on the technological environment of firms. Furthermore, the model shows that tracking the innovation performance of the same firms over time leads to a lower persistence of innovativeness when measured from the output side than when measured from the input side through use of R&D. Moreover, the contribution of innovation to MFP increases when estimating a static innovation model that uses the data obtained after pooling the two waves of CIS. The latter result reflects the ...