This paper uses Australian data to test whether an additive growth model explains productivity growth better than the exponential growth model typically used. It finds additive growth model forecasts are more accurate. The additive model also appears to resolve the puzzle that total factor productivity growth has slowed down in advanced economies, something that has defied definitive explanation to date.
The paper also explores some of the implications of the additive growth model. These include that productivity growth should be expected to slow over time, absent new general-purpose technologies. Compared to exponential growth models, additive growth models lead us to expect smaller intertemporal spillovers from R&D activity.