Is the supermultiplier stable?

Supermultiplier models, which show how autonomous demand can drive both business cycles and long-run GDP growth, are based on a stability assumption. In this paper I look at recent efforts to justify this assumption, and argue that they are not convincing. The supermultiplier literature generally assumes that business investment reacts very slowly to changes in the state of the economy, but faster adjustment speeds are consistent with US data and can generate instability.